0413
This assumption is supported by the transcription pattern, as STM activity is restricted to the SAM but down-regulated in founder cells (P0) of newly arising lateral organ primordia (5).
0107
A functional interaction between plant MEINOX-HD and BLH genes is substantiated by the synergism observed when pennywise a null-allele of BLH9 (18), allelic to bellringer, replumless, vaamana (19,22,23) is combined with the brevipedicellus1 (bp1) mutant (24).

0114
Guided by an evolutionarily conserved structure and heterodimerization of two TALE-HD super gene family members in plants and animals, it has been proposed that KNOX/BLH as MEIS/PBC genes share a common ancestry (25). However, amino acid sequence conservation outside the homeodomain gene is low, although members of the plant BLH or animal PBC gene families each share a typical and lineage-specific domain, the BELL domain and the bipartite PBC-A/B domain, respectively (11,17,25,26).
0030
This dominant-negative function upon expression of the MEINOXSTM domain is strengthened in combination with the Drosophila engrailed repressor domain and best explained by squelching: the abundant MEINOXSTM domain depleting the native STM protein from essential interaction partners.
0018
The meristem-enriched cDNA library in the vector pACT2 (Clonetech) was prepared from immature inflorescence shoots (Lumbrineris erecta, 4–5 mm size) containing inflorescence and multiple floral meristems (FMs). cDNA synthesis exactly followed the manual for the cDNA synthesis kit (Stratagene). Directional cloning into pACT2 was achieved via an XhoI site addition to the oligo-dT primer (3′) and an EcoRI adapter at the cDNA 5′-terminus. The cDNA library (1.25 × 106 clones) was colony amplified in Escherichia coli, plasmid DNA purified by CsCl2-gradient centrifugation and used for yeast transformation (strain AH109). Screening of the library followed the Matchmaker manual applying quadruple selection (HIS3, ADE2, LacZ and MEL1).

0018
Positive clones (223) were isolated and cDNA inserts in pACT2 were subjected to N-terminal sequence analysis. The specificity of each protein interaction was confirmed by retransformation of sequence-verified cDNA clones (ATH1, BHL3 or BHL9) into yeast strains providing the MEINOXSTM or MEINOXBP1 bait constructs in pGDKT7.
0809
Verified cDNA inserts were either directly transferred into appropriate secondary vectors for bimolecular fluorescence complementation (BiFC), co-immunoprecipitation or transgenic experiments and served as templates in second-round PCR-amplification to remove the stop codon or to adjust open reading frames (ORFs) for translational fusions.

0019
Verified cDNA inserts were either directly transferred into appropriate secondary vectors for bimolecular fluorescence complementation (BiFC), co-immunoprecipitation or transgenic experiments and served as templates in second-round PCR-amplification to remove the stop codon or to adjust open reading frames (ORFs) for translational fusions.
0030
Non-radioactive in situ hybridization experiments were performed essentially as described in Bradley et al.

0889
Non-radioactive in situ hybridization experiments were performed essentially as described in Bradley et al.
0018
Epitope-tagged proteins were prepared in the EasyXpress protein system (Qiagen), via an in vitro transcription/translation system based on the T7 promoter. T7 promoter transcription templates of STM partner proteins were obtained via nested PCRs on the ATH1, BHL3 or BHL9 constructs in pUC-SPYCE which encodes the HA epitope prior the YFP C-terminal domain.
0018
The MACS epitope tagged protein isolation kit (Miltenyi Biotec) was used to precipitate candidate partner proteins via the HA epitope. Eluted protein samples were split (1/3 and 2/3), size-fractionated in parallel on two polyacrylamide SDS-gels (concentration adjusted to protein size), electroblotted onto Immobilon-P membrane (Millipore) and subjected to antibody detection.

0096
The MACS epitope tagged protein isolation kit (Miltenyi Biotec) was used to precipitate candidate partner proteins via the HA epitope.

0007
HA-tagged (1/3 aliquot) partner proteins were identified using an anti-HA horseradish peroxidase (HRP)-conjugated antibody (clone 3F10; Roche 2012819). The co-precipitated 6× his-tagged STM protein was visualized by a primary penta-His mouse antibody [α-(H)5; Qiagen 34660] and secondary HRP-coupled goat anti-mouse IGG (Dianova 115-035-062).

0004
Epidermal proteins after bombardement of leek epidermal cells were isolated in 50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 1% Triton and size-fractionated by SDS–PAGE.

0007
GFP fusion proteins were detected after transfer to Immobilon-P membrane (Millipore) by a mouse monoclonal anti-GFP HRP-conjugated antibody (IgG1; Milteny Biotec 130-091-833).
0055
The method of choice to prove a physical interaction between candidate partner proteins in plant cells is BiFC (31,36). As a prerequisite to substantiate interactions between potential partner proteins identified in a yeast two-hybrid screen, C- and N-terminal fusions between the ORFs of STM and the GFP were constructed and expressed in leek or onion epidermal cells after particle bombardment. As a potential transcription factor, STM was expected to direct GFP fluorescence into the plant cell nucleus.

0107
The method of choice to prove a physical interaction between candidate partner proteins in plant cells is BiFC (31,36).

0663
As a potential transcription factor, STM was expected to direct GFP fluorescence into the plant cell nucleus.

0096
However, both fusion proteins, GFP-STM and STM-GFP remained in the cytoplasm (Figure 1B and C).

0413
However, in contrast to ectopic STM activity affecting the SAM or causing meristems on leaves (37), the expression of a nuclear targeted NLS-STM version from the CaMV 35S promoter in transgenic Arabidopsis plant predominantly resulted in a lobed leaf phenotype (Figure 1E), which is reminiscent of KNAT1 overexpression (38) rather than STM over expression (37).
0006
Protein extracts prepard from leek epidermal cells after particle bombardement showed no degradation of the GFP-STM fusion protein with an antibody directed against the GFP (Figure 1I).

0007
Protein extracts prepard from leek epidermal cells after particle bombardement showed no degradation of the GFP-STM fusion protein with an antibody directed against the GFP (Figure 1I).

0055
Protein extracts prepard from leek epidermal cells after particle bombardement showed no degradation of the GFP-STM fusion protein with an antibody directed against the GFP (Figure 1I).

0055
Additional support for nuclear exclusion was obtained by several STM deletion constructs, none of which targeted GFP fluorescence to the plant cell nucleus (data not shown). This nuclear import deficiency is noteworthy since the ELK domain conserved between the MEINOX domain and the homeodomain in plant KNOX proteins (Figure 1A) has been assumed to provide a functional NLS (39).

0663
Additional support for nuclear exclusion was obtained by several STM deletion constructs, none of which targeted GFP fluorescence to the plant cell nucleus (data not shown).
0018
Besides obvious controls in yeast such as retransformation of bait or prey plasmids and the combination with the empty bait vector, we included the MEINOXBp1 domain of the BP1 gene as bait as it encodes the closest relative to the MEINOXSTM domain in the Arabidopsis genome.

0114
All three BLH proteins interact with both the MEINOXSTM and the MEINOXBp1 domain and thus recognize sequence or structural features common to the MEINOX domains of BP1 or STM (Figure 2A).

0019
In addition to the interactions in yeast, we performed co-immunoprecipitation experiments with epitope-tagged full-length proteins.

0018
In addition to the interactions in yeast, we performed co-immunoprecipitation experiments with epitope-tagged full-length proteins.

0096
The HA-tagged BLH proteins shown in Figure 2B were used to co-precipitate epitope-tagged STM-His (Figure 2C).

0018
The detection of the STM-His protein was strictly dependent on the presence of BLH partner proteins. The co-immunoprecipitation experiments therefore confirm that the full-length STM protein interacts with full-length ATH1, BLH3 and BLH9 proteins in vitro and substantiate the affinity of BLH/STM interactions.

0019
The co-immunoprecipitation experiments therefore confirm that the full-length STM protein interacts with full-length ATH1, BLH3 and BLH9 proteins in vitro and substantiate the affinity of BLH/STM interactions.
0018
To verify STM interaction partners we constructed translational fusions between the N- or the C-terminal YFP sub-domains and the full-length STM protein or individual full-length BLH partner proteins.

0096
To verify STM interaction partners we constructed translational fusions between the N- or the C-terminal YFP sub-domains and the full-length STM protein or individual full-length BLH partner proteins.

0107
To verify STM interaction partners we constructed translational fusions between the N- or the C-terminal YFP sub-domains and the full-length STM protein or individual full-length BLH partner proteins. Optimal combinations of chimeric gene constructs are indicated in Figure 4 above each corresponding experiment.

0055
Any detectable fluorescence was dependent on the combination of the C-terminal with the N-terminal YFP sub-domain fused to STM or BLH partner proteins. In contrast, the fusion of the C- or the N-terminal YFP sub-domain to the STM protein and their co-bombardment in plant cells gave no fluorescence (data not shown).

0416
Any detectable fluorescence was dependent on the combination of the C-terminal with the N-terminal YFP sub-domain fused to STM or BLH partner proteins. In contrast, the fusion of the C- or the N-terminal YFP sub-domain to the STM protein and their co-bombardment in plant cells gave no fluorescence (data not shown).
0055
The BiFC results therefore, confirm that full-length STM and BLH proteins interact in plant cells, but more notably, that BLH/STM heterodimers are efficiently incorporated into the nuclear compartment.

0096
Expectedly, the GFP-ATH1, GFP-BLH3 and GFP-BLH9 fusion proteins exerted a preference for the nuclear compartment (Figure 4D–F).

0030
However, without analysing the cellular localization of the STM protein, Bhatt et al.

0889
However, without analysing the cellular localization of the STM protein, Bhatt et al.

0055
This may indicate a weak interaction between STM and BLH9 as suggested by the yeast two-hybrid results, that chimeric BLH9-NYFP protein levels remain low in leak epidermal cells, or that steric constraints interfere with a reconstitution of YFP fluorescence.

0107
This may indicate a weak interaction between STM and BLH9 as suggested by the yeast two-hybrid results, that chimeric BLH9-NYFP protein levels remain low in leak epidermal cells, or that steric constraints interfere with a reconstitution of YFP fluorescence.

0416
This may indicate a weak interaction between STM and BLH9 as suggested by the yeast two-hybrid results, that chimeric BLH9-NYFP protein levels remain low in leak epidermal cells, or that steric constraints interfere with a reconstitution of YFP fluorescence.
0055
However, we have been unable to reconstitute YFP fluorescence in BiFC experiments with the isolated MEINOXSTM and BELLBLH3 domains although several compatible N- or C-terminal combinations with each YFP sub-domain were tested. One explanation for this may be that the relative arrangement of YFP sub-domains sterically interferes with their functional association.
0809
According to our BiFC results, however, STM is targeted into the nucleus as a heterodimer with ATH1, BLH3 and BLH9.
0018
BLH proteins comprise one of two TALE-HD protein subfamilies in plants (10), STM is the most prominent member of the MEINOX protein sub-branch in Arabidopsis. Our findings that STM resides in the cytoplasm but is nuclear in heterodimers with BLH partners is reminiscent of the situation in animals, where nuclear import of MEIS proteins is directed by heterodimerization with PBC gene products.

0030
In Drosophila leg imaginal discs, the EXD protein is nuclear localized in proximal regions, which correspond to the expression domain of the HTH protein.

0107
The accepted scheme of nuclear import control of MEINOX TALE-HD proteins through interaction with the second class of TALE-HD proteins, BLH or PBC in plants and animals, respectively, suggests an ancient and conserved mechanism. In animals, the interaction between PBC and MEINOX proteins requires integrity of the bipartite PBC-A and B domains in the PBC and the MEINOX domain in MEIS proteins (13).
0018
Based on protein–protein interaction data, the animal PBC-A/B domain finds its counterpart in the BELL domain of plant BLH proteins although conservation at the amino acid level is weak.

0096
Based on protein–protein interaction data, the animal PBC-A/B domain finds its counterpart in the BELL domain of plant BLH proteins although conservation at the amino acid level is weak.

0104
Based on protein–protein interaction data, the animal PBC-A/B domain finds its counterpart in the BELL domain of plant BLH proteins although conservation at the amino acid level is weak.
0081
The PBC-B domain contains several conserved phosphorylation sites for Ser/Thr kinases and PBX1 sub-cellular localization correlates with the phosphorylation state of these residues whose dephosphorylation induces nuclear export (49).
0107
The evolutionary conservation of a functional interaction between members of two families of TALE-HD gene products, MEIS/PBX and KNOX/BLH proteins in animals or plants, respectively, suggests an ancient and possibly extant function.
0018
STM/BLH interactions in yeast and co-immunoprecipitations. (A) Galactosidase activity mediated by the interaction of the MEINOX domain cloned into pGBKT7 (bait) with BLH partner proteins expressed in pACT2. No signal is observed with the empty bait vector.

0107
The positive controls upper panel to the left shows the interaction between p53 and the SV40 T-antigen (pGBKT7-53 and pADT7-T, respectively, provided in the MATCHMAKER Biosensor kit).

0007
(B) In vitro translated full-length BLH proteins tagged with the hemaglutine (HA) epitope which were used in co-precipitation experiments with His-tagged STM protein and visualized by an anti-HA HRP-conjugated antibody. (C) Co-precipitated STM-His protein detected by a penta-His antibody.

0096
(B) In vitro translated full-length BLH proteins tagged with the hemaglutine (HA) epitope which were used in co-precipitation experiments with His-tagged STM protein and visualized by an anti-HA HRP-conjugated antibody.

0018
(B) In vitro translated full-length BLH proteins tagged with the hemaglutine (HA) epitope which were used in co-precipitation experiments with His-tagged STM protein and visualized by an anti-HA HRP-conjugated antibody. (C) Co-precipitated STM-His protein detected by a penta-His antibody.
0809
(A–C) BiFC staining of the nucleus obtained after coexpression of STM/BLH constructs as indicated above each photograph in leek epidermal cells.

0107
(H) Interaction between BLH3ΔC-NYFP and STM-CYFP.
0107
Functional interaction between STM-GR and partner proteins in transgenic Arabidopsis plants.
0416
Herpes simplex virus type-1 expresses a heterodimeric Fc receptor, gE-gI, on the surfaces of virions and infected cells that binds the Fc region of host immunoglobulin G and is implicated in the cell-to-cell spread of virus.

0114
Here we identify the C-terminal domain of the gE ectodomain (CgE) as the minimal Fc-binding domain and present a 1.78-Å CgE structure. A 5-Å gE-gI/Fc crystal structure, which was independently verified by a theoretical prediction method, reveals that CgE binds Fc at the CH2-CH3 interface, the binding site for several mammalian and bacterial Fc-binding proteins. The structure identifies interface histidines that may confer pH-dependent binding and regions of CgE implicated in cell-to-cell spread of virus.

0096
A 5-Å gE-gI/Fc crystal structure, which was independently verified by a theoretical prediction method, reveals that CgE binds Fc at the CH2-CH3 interface, the binding site for several mammalian and bacterial Fc-binding proteins. The structure identifies interface histidines that may confer pH-dependent binding and regions of CgE implicated in cell-to-cell spread of virus.
0411
Herpes simplex virus type-1 (HSV-1) has evolved several strategies to escape detection by the host's immune system, including the expression of an Fc receptor (FcR) called gE-gI that is found on the surface of virions and infected cells [1–3].

0416
Herpes simplex virus type-1 (HSV-1) has evolved several strategies to escape detection by the host's immune system, including the expression of an Fc receptor (FcR) called gE-gI that is found on the surface of virions and infected cells [1–3].

0006
Previous studies suggested that anti-HSV IgG antibodies participate in antibody bipolar bridging, whereby an antibody molecule simultaneously binds to gE-gI with its Fc region and to a specific HSV-antigen (e.g., gC or gD) with its Fab arms [5–8]. Antibody bipolar bridging has been shown to protect the virus and infected cells from IgG-mediated immune responses, such as antibody- and complement-dependent neutralization [6], antibody-dependent cell-mediated cytotoxicity [5], and granulocyte attachment [8]. Experiments in HSV-1–infected mice comparing the effectiveness of human anti-HSV IgG versus nonimmune IgG or murine anti-HSV IgG (which does not bind gE-gI) have provided support for the importance of antibody bipolar bridging mediated by gE-gI [7].

0416
Antibody bipolar bridging has been shown to protect the virus and infected cells from IgG-mediated immune responses, such as antibody- and complement-dependent neutralization [6], antibody-dependent cell-mediated cytotoxicity [5], and granulocyte attachment [8]. Experiments in HSV-1–infected mice comparing the effectiveness of human anti-HSV IgG versus nonimmune IgG or murine anti-HSV IgG (which does not bind gE-gI) have provided support for the importance of antibody bipolar bridging mediated by gE-gI [7].
0676
The Fc binding region on gE has been localized to the C-terminal domain of the gE ectodomain (CgE) [14,15], whereas the N-terminal domain of gE (NgE) associates with gI, forming a complex that does not bind to IgG [16].
0096
The gE-gI-binding site on IgG has been localized to the Fc CH2-CH3 interdomain hinge, a “hot spot” that serves as the binding site for several other mammalian and bacterial Fc-binding proteins [19,20].

0107
Residues critical to the interaction were identified in binding studies comparing the affinities of human IgG subtypes for gE-gI, which found that gE-gI binds to IgG1, IgG2, and IgG4 with similar affinities (equilibrium dissociation constant [KD] ˜40–400 nM), but does not bind several IgG3 allotypes or an IgG4 mutant in which His435 was changed to an arginine [21,22].

0104
Residues critical to the interaction were identified in binding studies comparing the affinities of human IgG subtypes for gE-gI, which found that gE-gI binds to IgG1, IgG2, and IgG4 with similar affinities (equilibrium dissociation constant [KD] ˜40–400 nM), but does not bind several IgG3 allotypes or an IgG4 mutant in which His435 was changed to an arginine [21,22]. These results implicate a role for Fc His435, which is an arginine in most human IgG3 allotypes, and are consistent with a previous suggestion that gE-gI and protein A have overlapping binding sites at the CH2-CH3 domain interface of Fc [23].

0114
Additional studies using a nonbinding Fc mutant (nbFc), which contains six point mutations in the CH2-CH3 hinge, confirmed the gE-gI binding site on Fc and established the stoichiometry as two gE-gI heterodimers per one Fc dimer (2:1) [12], analogous to the binding stoichiometries for all known FcR/Fcγ interactions that involve the Fc CH2-CH3 domain interface [19,24–28].
0107
In order to study the interaction between gE-gI and Fc in more detail, we initiated structural studies of gE, gE-gI, and a gE-gI/Fc complex.

0104
The gE-gI/Fc model resulting from prediction and crystallographic methods is consistent with biochemical data characterizing the interaction, provides insight into the molecular basis for the observed pH dependence of the gE-gI/Fc interaction, and allows mapping of CgE residues important for IgG binding and cell-to-cell spread.
0096
To determine whether isolated CgE (gE residues 213–390, where residue 1 is the first residue of the hydrophobic leader peptide in the immature gE and residue 420 is the first residue of the predicted transmembrane region) binds Fc and to measure the affinity of the interaction, we performed a surface plasmon resonance assay using two forms of recombinant Fc derived from human IgG1: wild-type Fc (wtFc) and nbFc, which contains six point mutations in the CH2-CH3 linker that abrogate binding to gE-gI (Met252 to Gly, Ile253 to Gly, His310 to Glu, His433 to Glu, His435 to Glu, and Tyr436 to Ala) [12]. The Fc proteins were immobilized on the surface of a biosensor chip, and the binding of CgE was assayed at pH 8 and pH 6.

0676
To determine whether isolated CgE (gE residues 213–390, where residue 1 is the first residue of the hydrophobic leader peptide in the immature gE and residue 420 is the first residue of the predicted transmembrane region) binds Fc and to measure the affinity of the interaction, we performed a surface plasmon resonance assay using two forms of recombinant Fc derived from human IgG1: wild-type Fc (wtFc) and nbFc, which contains six point mutations in the CH2-CH3 linker that abrogate binding to gE-gI (Met252 to Gly, Ile253 to Gly, His310 to Glu, His433 to Glu, His435 to Glu, and Tyr436 to Ala) [12].

0081
To determine whether isolated CgE (gE residues 213–390, where residue 1 is the first residue of the hydrophobic leader peptide in the immature gE and residue 420 is the first residue of the predicted transmembrane region) binds Fc and to measure the affinity of the interaction, we performed a surface plasmon resonance assay using two forms of recombinant Fc derived from human IgG1: wild-type Fc (wtFc) and nbFc, which contains six point mutations in the CH2-CH3 linker that abrogate binding to gE-gI (Met252 to Gly, Ile253 to Gly, His310 to Glu, His433 to Glu, His435 to Glu, and Tyr436 to Ala) [12].

0104
Significant binding of CgE to nbFc was not observed under any conditions (unpublished data), but CgE binds wtFc with a KD of ˜7 μM at pH 8 (Figure 1A and1C).

0019
The relatively low affinity of the CgE/Fc interaction likely explains why a CgE/Fc binding interaction was not detectable by coimmunoprecipitation analyses [16].

0104
The affinity of CgE for Fc is similar to that determined for the interaction of gE with Fc at pH 8 (KD ˜30 μM), but ˜50-fold weaker than the affinity of the gE-gI/Fc interaction, which has KD values of 340 nM and 930 nM for the first and second binding sites on wtFc, respectively [12].

0107
The affinity of CgE for Fc is similar to that determined for the interaction of gE with Fc at pH 8 (KD ˜30 μM), but ˜50-fold weaker than the affinity of the gE-gI/Fc interaction, which has KD values of 340 nM and 930 nM for the first and second binding sites on wtFc, respectively [12].

0096
As previously observed for the interaction of gE-gI with Fc [12], the binding of CgE to Fc is much weaker at pH 6 than at pH 8 (Figure 1A and1B). In combination with previous data, these results suggest that the gI chain contributes to the gE-gI interaction with Fc either directly or through stabilizing the binding conformation of gE, that NgE does not contribute to Fc binding, and that at least some of the pH dependence of the gE-gI interaction with Fc can be attributed to the CgE/Fc interface.
0114
A 3-D search of the protein database using the DALI server [29] identified the coxsackie virus and adenovirus receptor domain 1 (pdb entry 1eaj) and a mouse Fv (pdb entry 1mfa) as two of its closest structural neighbors, with root mean square deviations of 2.6 Å over 108 Cα atoms and 2.8 Å over 106 Cα atoms, respectively (Figure S1). The structural similarity between CgE and its structural homologs does not include the third β-sheet found in CgE.

0114
Further structural differences include the extension of strands C, C′, and C′′ as well as the A′–B and C′–C′′ loops, and the shortening of strands D and E as compared to a representative Ig V domain (Figure S1A), which leads to a more open organization of the first two β-sheets (sheets DEBB′A and A′GFCC′C′′).
0114
The only other FcR of known structure to adopt an Ig V domain fold is the polymeric Ig receptor (pIgR), which contains an N-terminal V-like domain that binds to polymeric IgA and IgM [31]. A comparison of the CgE and pIgR domain 1 structures does not reveal significant similarity beyond a region of common folding topology (unpublished data).
0081
An initial analysis of a KIrCl3 derivative identified five iridium sites located near the Fc molecule as positioned in the molecular replacement solution, four of which are related by the Fc dimer symmetry and two of which make a chemically-plausible interaction with a methionine (Met358) on each half of the Fc dimer (Figure 3B). Phases were calculated after locating heavy atom sites for the other derivatives (PIP, EMP, and Na2WO4) and used to find the locations of 13 of 18 potential selenium-substituted methionines in a crystal of SeMet-substituted gE-gI complexed with Fc. A comparison of the identified selenium sites with the molecular replacement model showed that eight of the 13 sites mapped to the eight methionine residues in CgE (Figure 3B). The clustering of all of the heavy atom sites around the Fc molecule provides further support for only one copy of the complex in the asymmetric unit.

0676
Thus, experimental phasing methods using a combination of heavy-atom derivatives and SeMet-substituted crystals provides further confirmation of the molecular replacement solution and validates its use as a model for the CgE/Fc interaction in the context of the gE-gI/Fc complex.
0096
In the first step, a global search through possible rigid-body orientations of the CgE and Fc proteins was carried out with the constraint that the CgE/Fc interface contains at least one of the six residues in the consensus binding site on Fc (Met252, Ile253, Ser254, Asn434, His435, and Tyr436) [19].

0107
The magnitude of the energy gap between the correct models and other models is correlated with the binding affinity of the interaction [33].

0096
Perhaps because of the relatively low binding affinity of the CgE/Fc interaction (Figure 1C), we did not observe a significant energy gap in this case, and hence could not confidently identify a single model as correct based on the global docking calculations alone. Instead, we proceeded by clustering the lowest-energy models based on structural similarity and further pruning high-ranked (large-sized) clusters by imposing a second constraint that CgE interacts with both the CH2 and CH3 domains of Fc, which has been observed for other interactions with the consensus Fc binding site [19].
0114
The complex structure shows that CgE interacts with Fc using regions of the first and second β-sheets, which correspond to the two β-sheets found in Ig V domains: strands A and B′ and the B′–C and D–E loops in the DEBB′A sheet and strands C, C′, and C′′ and the C′–C′′ loop in the A′GFCC′C′′ sheet (Figures 2C and4).

0676
The complex structure shows that CgE interacts with Fc using regions of the first and second β-sheets, which correspond to the two β-sheets found in Ig V domains: strands A and B′ and the B′–C and D–E loops in the DEBB′A sheet and strands C, C′, and C′′ and the C′–C′′ loop in the A′GFCC′C′′ sheet (Figures 2C and4).
0018
The CH2-CH3 linker region on Fc has been identified as a “hot spot” for protein–protein interactions [19,20]. Crystal structures are available for Fc complexed with four other proteins, domain B1 of protein A [25], domain C2 of protein G [28], rheumatoid factor [24], and FcRn [27], and with one peptide that bind to this region [19].

0114
Crystal structures are available for Fc complexed with four other proteins, domain B1 of protein A [25], domain C2 of protein G [28], rheumatoid factor [24], and FcRn [27], and with one peptide that bind to this region [19].

0096
Each of the Fc-binding proteins is different in sequence and structure; however, their binding interactions with Fc are similar, sharing a set of six contact residues on Fc that have been referred to as the consensus region [19,27].

0081
Interface interactions involving side-chain atoms of Fc residues Met252, Ile253, Ser254, Asn434, His435, and Tyr436 and the adjacent peptide backbone are found in each of the five FcR/Fc interfaces [19,27], and are predicted to occur in the CgE/Fc interface (Figure 4A).

0114
Likewise, the gE-gI/Fc structure suggests that the interface with CgE contains at least two conserved interactions with Fc that are found in three or more of the five FcR/Fc interfaces [19,27]: hydrophobic packing onto Fc His435 by gE Pro319 and gE Pro321, and hydrophobic packing onto Fc Met252 and Fc Tyr436 by gE His247 (Figure 4A).

0096
Similar to the binding interfaces for the other five proteins that bind the CH2-CH3 linker region of Fc [19,20,27], the CgE/Fc interface is primarily nonpolar with a total of ˜1800 Å2 of buried surface area.
0413
Previous studies demonstrated a sharply pH-dependent interaction for HSV-1–infected cells binding to rabbit IgG [37] and for soluble gE-gI or gE monomers binding to human Fc [12], such that binding occurs at neutral or slightly basic pH but not at acidic pH.

0416
Previous studies demonstrated a sharply pH-dependent interaction for HSV-1–infected cells binding to rabbit IgG [37] and for soluble gE-gI or gE monomers binding to human Fc [12], such that binding occurs at neutral or slightly basic pH but not at acidic pH.

0676
An analysis of the change in the affinity of the gE-gI/Fc complex versus pH suggested that two ionizable residues, likely histidines because they have a pKa near neutral pH, participate in the pH-dependent affinity transition [12].

0096
The pH-dependent binding of Fc to CgE (Figure 1A and1B) suggests a direct role for CgE in modulating the gE-gI/Fc interaction.
0114
Although we cannot rule out local conformational variability that occurs as a function of pH, it seems unlikely that CgE undergoes large changes in conformation due to its relative compactness as a single domain, and arguments against pH-induced conformational changes in Fc have been made based on the similarity of Fc crystal structures at pH values between 4.1 and 8.0 [27].

0419
Although we cannot rule out local conformational variability that occurs as a function of pH, it seems unlikely that CgE undergoes large changes in conformation due to its relative compactness as a single domain, and arguments against pH-induced conformational changes in Fc have been made based on the similarity of Fc crystal structures at pH values between 4.1 and 8.0 [27].
0114
Two studies used insertion mutagenesis to localize the IgG-binding site on gE [14,17]. Five of 15 insertions that fall within the CgE domain were introduced into surface-exposed loops and are therefore unlikely to have global effects on the gE structure (Figure 4B and4C).
0416
Insertion mutants in gE were also tested for their effects on viral spread in cultured epithelial cells and HSV-1 infected mice [7,17,18,38,39]. Although the mechanism of viral spread is unknown, it has been hypothesized that the extracellular domains of gE-gI bind to receptors at cell junctions, allowing entry into adjacent cells [17,40].
0676
The CgE/Fc model, taken together with the locations of introduced heavy atoms and SeMet residues, provides insight into the organization of the gE-gI/Fc ternary complex and its orientation on a cell membrane. The position of the C-terminus of CgE in the gE-gI/Fc complex structure, which is ˜30 amino acids before the predicted transmembrane domain, suggests that IgG bound by gE-gI is in an upright orientation with respect to the membrane (Figure 5), similar to FcαRI-bound IgA [41], but opposite to the orientation predicted for antibodies bound by other cellular Fc receptors such as FcγRIII and FcɛRI [42–44].

0114
The position of the C-terminus of CgE in the gE-gI/Fc complex structure, which is ˜30 amino acids before the predicted transmembrane domain, suggests that IgG bound by gE-gI is in an upright orientation with respect to the membrane (Figure 5), similar to FcαRI-bound IgA [41], but opposite to the orientation predicted for antibodies bound by other cellular Fc receptors such as FcγRIII and FcɛRI [42–44].
0081
The locations of NgE and gI in the gE-gI/Fc complex can be approximated by an analysis of heavy-atom and SeMet sites that are not accounted for by the CgE and Fc models. The extra heavy-atom sites are located in a space that is bounded by the CH3 domains of Fc on the top, the presumed membrane location on the bottom, and symmetry-related CgE molecules on the sides (Figure 3B). Six of the sites that are not accounted for by CgE or Fc, and thus attributed to NgE or gI, are related by the Fc dyad axis (i.e., two sets of three heavy-atom sites), consistent with the two-fold symmetry of the complex that results from the inherent symmetry of the Fc dimer.

0676
The proposed positions of NgE and gI, together with the upright orientation of Fc on a cell membrane, would allow the Fab arms of an intact anti-HSV IgG to interact with HSV antigens on the surface of the same membrane in a bipolar bridging complex (Figure 5).
0114
Because we were able to solve the crystal structure of the CgE subunit of the gE-gI ectodomain at high resolution, the identification of CgE as a minimal Fc-binding domain (Figure 1) made its structure useful for interpreting the low-resolution gE-gI/Fc complex structure.

0676
An independent prediction of the complex structure using the coordinates of the unbound CgE and Fc molecules revealed a close agreement with the experimentally determined gE-gI/Fc structure (Figure 3C), demonstrating the power of new protein docking methods [33,34], and their potential for facilitating difficult crystallographic problems.

0030
An independent prediction of the complex structure using the coordinates of the unbound CgE and Fc molecules revealed a close agreement with the experimentally determined gE-gI/Fc structure (Figure 3C), demonstrating the power of new protein docking methods [33,34], and their potential for facilitating difficult crystallographic problems.
0081
The structure of the CgE domain of gE-gI represents a new variant of the Ig superfamily that is distinct from the structures of host FcRs and other known Fc-binding proteins (Figure 2C and2D). CgE contains two β-sheets with a Greek-key folding topology similar to the topology of Ig V domains, but includes an extra β-sheet that packs against the Ig V–like portion of CgE (Figure 2A and2B).

0114
The structure of the CgE domain of gE-gI represents a new variant of the Ig superfamily that is distinct from the structures of host FcRs and other known Fc-binding proteins (Figure 2C and2D).
0114
As revealed in the low resolution crystal structure of a gE-gI/Fc complex, the CgE portion of gE-gI binds to Fc at the CH2-CH3 interdomain junction (Figure 3A), consistent with previous studies of the gE-gI/Fc interaction [12,21,23] and allowing a comparison of the Fc-binding properties of CgE with other proteins that bind to the same site on Fc [19,20,27].

0081
As revealed in the low resolution crystal structure of a gE-gI/Fc complex, the CgE portion of gE-gI binds to Fc at the CH2-CH3 interdomain junction (Figure 3A), consistent with previous studies of the gE-gI/Fc interaction [12,21,23] and allowing a comparison of the Fc-binding properties of CgE with other proteins that bind to the same site on Fc [19,20,27]. An analysis of the CgE/Fc interface predicts that the six residues that are the hallmark of the consensus binding site on Fc [19,27] are also central to Fc binding by CgE (Figure 4A).

0096
As revealed in the low resolution crystal structure of a gE-gI/Fc complex, the CgE portion of gE-gI binds to Fc at the CH2-CH3 interdomain junction (Figure 3A), consistent with previous studies of the gE-gI/Fc interaction [12,21,23] and allowing a comparison of the Fc-binding properties of CgE with other proteins that bind to the same site on Fc [19,20,27].
0096
Our analysis of gE-gI/Fc affinity as a function of pH suggested that the pH-dependent binding was due to protonation of two residues, likely histidines, in Fc and/or gE [12]. As CgE also shows pH-dependent binding to Fc (Figure 1), the CgE/Fc model derived from the gE-gI/Fc structure can be used to identify candidate residues involved in the pH-dependent interaction, which include four histidines in the CgE/Fc interface (His247 on gE and His310, His433, and His435 on Fc) (Figure 4A).

0081
As CgE also shows pH-dependent binding to Fc (Figure 1), the CgE/Fc model derived from the gE-gI/Fc structure can be used to identify candidate residues involved in the pH-dependent interaction, which include four histidines in the CgE/Fc interface (His247 on gE and His310, His433, and His435 on Fc) (Figure 4A).
0676
The model for CgE binding to Fc derived from the crystal structures of CgE and a gE-gI/Fc complex can be used to gain insight into the regions of CgE that are implicated in cell-to-cell spread of HSV. Mapping of insertion mutations in gE that affect viral spread [17,18,39] on the CgE and gE-gI/Fc complex structures identifies a region of CgE that could interact with receptors for gE-gI (Figure 4B).

0096
Although the structure-based interpretation of the insertion mutagenesis studies on CgE suggests that the surface of the protein involved in cell-to-cell spread of the virus is distant from the Fc-binding interface (Figure 4B), it is currently unclear whether the cell-to-cell spread and IgG binding functions of gE-gI are correlated. Viruses with insertions in gE after positions 333 and 339, which are at or near the crystallographically observed CgE/Fc interface, are disrupted for IgG binding but spread normally, suggesting that IgG binding is not required for cell-to-cell spread [7,17,38].
0081
The crystallographic data can be used to deduce a model for how IgG binds gE-gI on the surface of virions or infected cells, whereby CgE binds the consensus Fc-binding site on each half of the Fc dimer, and gI and the NgE are located underneath the CgE domain between the CH3 domains of Fc and the membrane (Figure 5).

0416
The crystallographic data can be used to deduce a model for how IgG binds gE-gI on the surface of virions or infected cells, whereby CgE binds the consensus Fc-binding site on each half of the Fc dimer, and gI and the NgE are located underneath the CgE domain between the CH3 domains of Fc and the membrane (Figure 5).

0426
The crystallographic data can be used to deduce a model for how IgG binds gE-gI on the surface of virions or infected cells, whereby CgE binds the consensus Fc-binding site on each half of the Fc dimer, and gI and the NgE are located underneath the CgE domain between the CH3 domains of Fc and the membrane (Figure 5).

0676
The upright orientation of the Fc and the positioning of the bulk of gE-gI to the side and underneath the Fc leave sufficient space for the Fab arms of an anti-HSV IgG bound to gE-gI to interact with antigens using antibody bipolar bridging (Figure 5).
0114
The gE-gI binding site on Fc does not directly overlap with the binding sites for the FcγRs or the C1q component of complement, which bind on or near the CH2 domain [43,44,51], thus the structure of the gE-gI/Fc complex does not directly suggest how gE-gI binding to the Fc region of IgG leads to evasion of FcγR- and complement-mediated immune responses.

0081
The gE-gI binding site on Fc does not directly overlap with the binding sites for the FcγRs or the C1q component of complement, which bind on or near the CH2 domain [43,44,51], thus the structure of the gE-gI/Fc complex does not directly suggest how gE-gI binding to the Fc region of IgG leads to evasion of FcγR- and complement-mediated immune responses.

0676
The gE-gI binding site on Fc does not directly overlap with the binding sites for the FcγRs or the C1q component of complement, which bind on or near the CH2 domain [43,44,51], thus the structure of the gE-gI/Fc complex does not directly suggest how gE-gI binding to the Fc region of IgG leads to evasion of FcγR- and complement-mediated immune responses.

0007
However, an anti-HSV antibody that is bound to both gE-gI and an HSV antigen on the surface of an infected cell could be sterically hindered from also binding to host FcγRs or C1q due to the close proximity of the proteins in the antibody bipolar bridging complex.

0096
However, an anti-HSV antibody that is bound to both gE-gI and an HSV antigen on the surface of an infected cell could be sterically hindered from also binding to host FcγRs or C1q due to the close proximity of the proteins in the antibody bipolar bridging complex.

0006
The demonstration that the gE-gI/Fc structure is compatible with antibody bipolar bridging (Figure 5) raises the possibility that anti-HSV IgG/HSV antigen complexes interacting with gE-gI on the surface of infected cells are endocytosed by gE-gI and degraded in the lysosomes after dissociation at acidic pH, resulting in destruction of antiviral antibodies and removal of viral antigens from the cell surface.

0416
The demonstration that the gE-gI/Fc structure is compatible with antibody bipolar bridging (Figure 5) raises the possibility that anti-HSV IgG/HSV antigen complexes interacting with gE-gI on the surface of infected cells are endocytosed by gE-gI and degraded in the lysosomes after dissociation at acidic pH, resulting in destruction of antiviral antibodies and removal of viral antigens from the cell surface.
0104
N-terminal sequencing and mass spectroscopy analyses of degradation products present in a preparation of purified gE-gI suggested that gE residues 210–395 comprise a stable domain (unpublished data) in agreement with previous proteolytic analyses of soluble gE-gI that showed that gE was cleaved into N- and C-terminal fragments with a domain boundary in the vicinity of gE residue 208 [16].

0081
Standard PCR-based subcloning techniques were used to insert a fragment of the gE gene from HSV strain KOS, encoding residues 213–390, downstream of the p10 promoter in the pAcUW51 baculovirus transfer vector (PharMingen, San Diego, California, United States). The expression vector also encoded the gE hydrophobic leader peptide N-terminal to the inserted fragment and a C-terminal Factor Xa cleavage site and 6x-His tag.

0411
Recombinant baculovirus stocks were generated by cationic liposome cotransfection of the expression plasmid with linear wild-type baculovirus DNA in High 5 insect cells (Invitrogen, Carlsbad, California, United States) as described by the manufacturer.

0004
Insect cell supernatants containing secreted CgE were buffer-exchanged into nickel-binding buffer (40 mM Tris [pH 8], 300 mM NaCl, 10 mM imidazole) and passed over a Ni-NTA agarose column (Qiagen, Valencia, California, United States). Bound protein was eluted in buffer containing 40 mM Tris (pH 8), 300 mM NaCl, and 250 mM imidazole, and further purified by size-exclusion chromatography on a Superdex 75 HiLoad 16/60 column (GE Healthcare, Piscataway, New Jersey, United States) that was equilibrated in 10 mM HEPES (pH 7.6) and 150 mM NaCl. Peak fractions were concentrated and buffer exchanged into 10 mM HEPES (pH 7.5).

0071
Insect cell supernatants containing secreted CgE were buffer-exchanged into nickel-binding buffer (40 mM Tris [pH 8], 300 mM NaCl, 10 mM imidazole) and passed over a Ni-NTA agarose column (Qiagen, Valencia, California, United States). Bound protein was eluted in buffer containing 40 mM Tris (pH 8), 300 mM NaCl, and 250 mM imidazole, and further purified by size-exclusion chromatography on a Superdex 75 HiLoad 16/60 column (GE Healthcare, Piscataway, New Jersey, United States) that was equilibrated in 10 mM HEPES (pH 7.6) and 150 mM NaCl. Peak fractions were concentrated and buffer exchanged into 10 mM HEPES (pH 7.5).

0416
Briefly, insect cells that were infected with the CgE-recombinant virus were grown in Ex-Cell 400 media (JRH Biosciences, Lenexa, Kansas, United States) for 16 hours, pelleted, and then resuspended in Sf-900 II SFM media (Invitrogen) supplemented with 30 mg/l of L-cysteine (Sigma, St.
0004
Equilibrium binding data for CgE binding to wtFc (coupling density, 440 resonance units), nbFc (coupling density, 405 resonance units), and a mock-coupled flow cell were collected for a CgE concentration series (three-fold dilutions of CgE from 30 μM to 5 nM) at pH 8 (50 mM HEPES [pH 8.0], 150 mM NaCl, 3 mM EDTA, 0.005% [vol/vol] P-20 surfactant) and pH 6 (50 mM sodium phosphate [pH 6.0], 150 mM NaCl, 3 mM EDTA, 0.005% [vol/vol] P-20 surfactant). After each injection of CgE, a 30-s injection of 250 mM di-ammonium hydrogen citrate (pH 5.0) was used to disrupt the interaction and restore the surface.

0104
Equilibrium binding data for CgE binding to wtFc (coupling density, 440 resonance units), nbFc (coupling density, 405 resonance units), and a mock-coupled flow cell were collected for a CgE concentration series (three-fold dilutions of CgE from 30 μM to 5 nM) at pH 8 (50 mM HEPES [pH 8.0], 150 mM NaCl, 3 mM EDTA, 0.005% [vol/vol] P-20 surfactant) and pH 6 (50 mM sodium phosphate [pH 6.0], 150 mM NaCl, 3 mM EDTA, 0.005% [vol/vol] P-20 surfactant).
0104
SeMet-substituted CgE crystals were grown in conditions similar to the native crystals with the addition of 10 mM MgCl2 in the well solution, which decreased the disorder that was often observed in the diffraction of the native crystals. All crystals were cryoprotected in well solution with an additional 15% PEG 4000, added in 5% increments, and stored in liquid nitrogen prior to data collection at −180 °C. Native and SeMet data were collected to 1.78 Å and 2.0 Å, respectively (Table S1).
0081
Phases were derived from a three-wavelength multiple anomalous dispersion experiment using SeMet-substituted CgE crystals. Solve [56] was used for local scaling of the data, location and refinement of Se positions, and phasing (figure of merit of 0.36 from 30–2 Å) followed by Resolve [57,58] for solvent flattening and automated model building (60% of the asymmetric unit was built by Resolve [57,58]).
0676
Various forms of gE and the gE-gI heterodimer were subcloned, expressed, and purified from baculovirus-infected insect cell supernatants by nickel affinity and/or IgG affinity and gel-filtration chromatography as described previously [12].

0096
Two recombinant forms of the Fc fragment of IgG1, wtFc and heterodimeric Fc, which contain two and one gE-gI binding sites, respectively, were also produced in CHO cells and purified as described previously [12].

0413
Two recombinant forms of the Fc fragment of IgG1, wtFc and heterodimeric Fc, which contain two and one gE-gI binding sites, respectively, were also produced in CHO cells and purified as described previously [12].
0114
The search models were the CgE structure (reported in this paper) truncated after residue 390 (the last residue in gE before the Factor Xa cleavage site) and the human IgG Fc structure (pdb entry 1dn2), which together account for 45% of the total molecular mass of the gE-gI/Fc complex.
0096
In the next stage, a second constraint was imposed requiring that CgE make contact with both the CH2 and CH3 domains of the Fc protein, as observed for other proteins binding to the Fc “hot spot,” [19] which resulted in pruning some of the top-ranking clusters. The centers of each of the top five remaining clusters were then subjected to more extensive local rigid-body and side-chain refinement to produce five final predicted structures, of which one matches the crystallographic model of the CgE/Fc interaction (Figure 3C) (model “d,” described below).

0018
In the next stage, a second constraint was imposed requiring that CgE make contact with both the CH2 and CH3 domains of the Fc protein, as observed for other proteins binding to the Fc “hot spot,” [19] which resulted in pruning some of the top-ranking clusters.

0104
In the five RosettaDock-predicted CgE/Fc complexes, different regions of CgE are used to bind the consensus region of Fc (model “b,” CgE residues in strand A, A-A′ linker, strand A′, strand B′, CI–CI′ linker, strand G, G–I′′ linker, and strand I′′; model “c,” CgE residues in strand A, A–A′ linker, B–B′ linker, strand B′, CI–CI′ linker, C′–C′′ linker, strand G, and strand I′′; model “d,” which resembles the molecular replacement solution, CgE residues in strand A, B–B′ linker, strand B′, strand C, strand C′, C′–C′′ linker, and D–E linker; model “e,” CgE residues in CI′–C′ linker, strand C′′, C′′–D linker, strand D, and E–F linker; and model “f,” CgE residues in strand C, C–CI linker, CI–CI′ linker, strand C′, and strand C′′, C′′–D linker, E–F linker) (unpublished data).
0676
(C) Stereo superposition of the crystallographically determined CgE/Fc complex and the complex predicted with RosettaDock [34] using the structures of CgE and Fc.

0114
A half complex (one CgE and one chain of the Fc dimer) is shown with Fc in green, the CgE as positioned in the crystal structure in blue, and the CgE as positioned by the docking prediction in pink.
0114
A ribbon diagram of the CgE/Fc portion of the gE-gI/Fc structure is shown with the approximate positions of the gI chain (red open oval) and NgE domain (blue open oval) shown schematically.
0018
These steps are controlled by a complex protein machinery that contains components that have homologs in most types of intracellular membrane fusion such as the SNARE proteins syntaxin, synaptobrevin, and SNAP-25, the Sec1/Munc18 (SM) homolog Munc18–1, and the Rab3 small GTPases, as well as specialized proteins such as synaptotagmin 1, complexins, Munc13s, and α-RIMs (reviewed in [1]). Although the mechanism of release is still unclear, clues to this mechanism have emerged from the three-dimensional structures of several complexes of these proteins (reviewed in [2]).

0096
These steps are controlled by a complex protein machinery that contains components that have homologs in most types of intracellular membrane fusion such as the SNARE proteins syntaxin, synaptobrevin, and SNAP-25, the Sec1/Munc18 (SM) homolog Munc18–1, and the Rab3 small GTPases, as well as specialized proteins such as synaptotagmin 1, complexins, Munc13s, and α-RIMs (reviewed in [1]). Although the mechanism of release is still unclear, clues to this mechanism have emerged from the three-dimensional structures of several complexes of these proteins (reviewed in [2]).
0413
Whereas it was initially thought that these and other ubiquitous modules (e.g., PDZ, SH2, and SH3 domains) each perform a particular type of activity, it is now clear that many of these modules can participate in diverse interactions. Thus, C2 domains perform a variety of functions that often depend on their most common activity, Ca2+-dependent phospholipid binding, but they are also believed to act as protein–protein interaction modules (reviewed in [21]). Although the mechanisms of Ca2+/phospholipid binding to C2 domains have been extensively characterized [22–24], no high-resolution structures of protein complexes involving C2 domains have been described.

0096
Thus, C2 domains perform a variety of functions that often depend on their most common activity, Ca2+-dependent phospholipid binding, but they are also believed to act as protein–protein interaction modules (reviewed in [21]). Although the mechanisms of Ca2+/phospholipid binding to C2 domains have been extensively characterized [22–24], no high-resolution structures of protein complexes involving C2 domains have been described.

0107
Hence, understanding the interaction between the Munc13–1 C2A domain and the α-RIM ZF domain at atomic detail can provide key insights into the functional diversification of C2 domains and ZF domains in general.
0114
NMR spectroscopy and X-ray crystallography provide complementary tools to study protein structure. X-ray crystallography is better suited to elucidate the structures of large proteins and accurately define interfaces of protein complexes, but requires crystallization, which can yield artifacts due to crystal packing.

0077
NMR spectroscopy and X-ray crystallography provide complementary tools to study protein structure.

0018
X-ray crystallography is better suited to elucidate the structures of large proteins and accurately define interfaces of protein complexes, but requires crystallization, which can yield artifacts due to crystal packing.

0077
NMR spectroscopy can be performed in solution, and low-resolution information on the conformational and aggregation states of proteins can be quickly obtained even for large species using heteronuclear NMR experiments such as1H-15N heteronuclear single quantum coherence (HSQC).

0007
NMR spectroscopy can be performed in solution, and low-resolution information on the conformational and aggregation states of proteins can be quickly obtained even for large species using heteronuclear NMR experiments such as1H-15N heteronuclear single quantum coherence (HSQC).

0018
These spectra contain one cross-peak for each non-proline residue of a15N-labeled protein and exhibit well-dispersed cross-peak patterns for well-folded proteins, whereas unfolding or misfolding leads to poor cross-peak dispersion.

0114
In this study, we took advantage of the strengths of both techniques, using NMR spectroscopy to optimize protein complexes for crystallization and to obtain structural information in solution that was later employed to interpret the high-resolution structures of these complexes elucidated by X-ray crystallography.

0077
In this study, we took advantage of the strengths of both techniques, using NMR spectroscopy to optimize protein complexes for crystallization and to obtain structural information in solution that was later employed to interpret the high-resolution structures of these complexes elucidated by X-ray crystallography.
0104
Interestingly, the apparent molecular weights observed for isolated Munc13–13–132, Munc13–13–150, and Munc13–13–209 in gel filtration were significantly higher than their monomeric molecular weights (Figure 1B and unpublished data), suggesting that they form stable dimers.
0104
We next used analytical ultracentrifugation to determine the oligomerization state of Munc13–13–209, Munc13–13–150, and Munc13–13–132, and of a shorter fragment that we prepared during optimization for our crystallographic studies (Munc13–13–128; see below). Fitting the data to single ideal species yielded molecular weights of 49,100 Da, 35,200 Da, 30,100 Da, and 26,700 Da, respectively, which are approximately twice the predicted molecular weights of these fragments (24683.4 Da, 17821.3 Da, 16168.4 Da, and 14856.9 Da, respectively). Correspondingly, the best correlations were obtained when the data were fit to an equation describing a monomer/dimer equilibria, yielding dimer dissociation constants of 1.9 nM, 5.2 nM, 80 nM, and 310 nM, respectively (illustrated for Munc13–13–128 inFigure 1E). Note that dissociation constants below 50–100 nM are not quite accurate under the conditions of our experiments.
0114
The Munc13–1 C2A domain monomer exhibits a β-sandwich structure formed by two four-stranded β-sheets that is characteristic of C2 domains (Figure 2B and2C), but the aforementioned β-hairpin formed by a long loop in monomer C (strands 8 and 9) has not been previously observed in other C2 domains. A structural comparison using DALI [26] showed that, among C2 domains deposited in the Protein Data Bank (http://www.rcsb.org/pdb, the PLC-δ1 C2-domain shares the highest structural similarity with the Munc13–1 C2A domain (1.65 Å rms deviation for 106 equivalent Cα carbons). The superposition of the PLC-δ1 C2 domain and the Munc13–1 C2A domain (monomer C) shown inFigure 2D illustrates that the β-sandwich cores of both C2 domains are very similar, and that substantial divergence exists in the loops connecting the β-strands, with the most prominent difference being the unusual β-hairpin of the Munc13–1 C2A domain.

0104
A structural comparison using DALI [26] showed that, among C2 domains deposited in the Protein Data Bank (http://www.rcsb.org/pdb, the PLC-δ1 C2-domain shares the highest structural similarity with the Munc13–1 C2A domain (1.65 Å rms deviation for 106 equivalent Cα carbons).
0114
Most of the side chains that form the interface between the two monomers are highly conserved through evolution (Figure 3), suggesting that the ability of the Munc13–1 C2A domain to homodimerize is shared in a wide variety of species.
0114
However, this β-hairpin is well packed against strand 6 of monomer C itself and appears to be an intrinsic feature of the Munc13–1 C2A domain, because the two Munc13–1 monomers present in the crystal structure of the Munc13–1/RIM2 heterodimer also contain this β-hairpin (see below).
0114
Hence, we designed two charge-reversal mutations, K32E and E63K, to disrupt Munc13–1 homodimerization based on the crystal structure of the C2A-domain homodimer.
0104
Gel filtration showed that both point mutations disrupt dimerization of Munc13–13–128, Munc13–13–150, and Munc13–13–209, but preserve heterodimerization with RIM2α82–142 (Figure 4A and unpublished data).

0676
We attribute this increased apparent affinity compared to the wild-type complex to the lack of competition with homodimerization.

0018
Importantly, the Munc13–13–150(K32E)/RIM2α82–142 complex readily yielded crystals in more than one third of the conditions of a basic crystallization screen (Index screen, Hampton Research) (Figure S3), whereas crystallization trials with the Munc13–13–209(K32E)/RIM2α82–142 complex failed, as expected from the lack of a defined structure in most of the 151–209 sequence (see above).
0676
Condition optimization allowed us to determine the structure of the Munc13–13–150(K32E)/RIM2α82–142 complex at 1.78 Å resolution using molecular replacement (Figures 5 and6).

0104
Thus, the binding mode between monomer A and RIM2α82–142 observed in the crystals is consistent with all available data obtained in solution, whereas that involving monomer B is not.
0081
Superposition of monomer C of the Munc13–1 C2A-domain homodimer with monomer A of Munc13–13–150(K32E) from the heterodimer (Figure 5D) yielded an rms deviation of 0.52 Å for 127 Cα carbons, showing that binding to RIM2α82–142 does not involve large conformational changes in the C2A domain.

0114
The structure of the RIM2α ZF domain in the heterodimer with Munc13–13–150(K32E) (Figure 5E, blue) is similar to the solution structure of the isolated RIM2α ZF domain (Figure 5E, red) [20], exhibiting two zinc-binding sites and a central β-hairpin that is flanked on one side by N-terminal loops and on the other side by another β-hairpin and a short C-terminal α-helix (helix a2). Superposition of the two structures yielded an rms deviation of 1.9 Å for 54 Cα carbons, revealing that heterodimer formation induces significant conformational changes in the ZF domain that suggest a substantial malleability and that are consistent with the widespread changes in its1H-15N HSQC spectrum caused by Munc13–1 binding [20].

0081
The structure of the RIM2α ZF domain in the heterodimer with Munc13–13–150(K32E) (Figure 5E, blue) is similar to the solution structure of the isolated RIM2α ZF domain (Figure 5E, red) [20], exhibiting two zinc-binding sites and a central β-hairpin that is flanked on one side by N-terminal loops and on the other side by another β-hairpin and a short C-terminal α-helix (helix a2).
0107
As shown inFigure 6D, the interaction of the C-terminal α-helix of Munc13–13–150(K32E) with the crevice between the C-terminal β-hairpin and helix a2 of the ZF domain involves a combination of hydrophobic and polar interactions.
0081
Correspondingly, the surfaces of the C2A domain involved in homo- and heterodimerization are distinct but contiguous, exhibiting only a small degree of overlap that is sufficient to make both interactions incompatible (Figure 7A). Particularly noteworthy in this regard are the hydrogen bonds formed by the ZF domain K99 side chain with the S33 hydroxyl group and the backbone carbonyl group of residue 32 of Munc13–1 (Figure 6E), since these two groups are involved in intermolecular hydrogen bonds in the C2A-domain homodimer (Figure 2F and2G).
0114
Our Munc13–13–150(K32E)/RIM2α82–142 structure, together with the crystal structure of a rabphilin/Rab3A complex [27], and the sequence homology between rabphilin and RIM2α, also provide an explanation for our previous observations that Munc13–1, RIM2α, and Rab3A form a tripartite complex and that Munc13–1 alters the RIM2α/Rab3A binding mode [20].
0107
Disrupting the interaction between the N-terminal region of Munc13–1 and α-RIMs causes a severe impairment in priming [10,20] that is comparable to that observed in RIM1α knockout mice [16]. The marked decrease in Munc13–1 levels observed in these mice further supports the physiological importance of the Munc13–1/α-RIM interaction [14].

0081
In any case, the observation that only a small overlap exists between the homodimerization and heterodimerization surfaces of the Munc13–1 C2A domain (Figure 7A) suggests that the homodimer may not need to be fully disrupted for α-RIM binding to be initiated.
0114
The structure of the Munc13–1/RIM2α ZF domain complex described here shows that Munc13–1 binding to the α-RIM/Rab3A complex should lead to a partial steric clash with Rab3A (Figure 7B), explaining our previous observation that the interaction of the α-RIM SGAWFY motif with Rab3A is released upon Munc13–1 binding [20].

0096
The structure of the Munc13–1/RIM2α ZF domain complex described here shows that Munc13–1 binding to the α-RIM/Rab3A complex should lead to a partial steric clash with Rab3A (Figure 7B), explaining our previous observation that the interaction of the α-RIM SGAWFY motif with Rab3A is released upon Munc13–1 binding [20]. The Munc13–1/α-RIM interaction still allows formation of the tripartite complex because helix a1 of α-RIMs is sufficient for Rab3A binding, but is expected to induce a significant reorientation of the N-terminal structural elements of α-RIMs (Figure 7C). Alternatively, it can be envisaged that Rab3A binding may also cause a rearrangement in α-RIM molecules that were initially bound to Munc13–1. Independently of which event occurs first, the finding that contiguous but partially overlapping surfaces are used by α-RIMs for Munc13–1 and Rab3A binding, and by the Munc13–1 C2A domain for homodimerization and heterodimerization, suggests that a synchronized cascade of protein–protein interactions involving Munc13–1, α-RIMs, and Rab3s may control vesicle priming and facilitate fast refilling of the ready-releasable pool of vesicles during repetitive stimulation.
0081
It is worth noting that phosphorylation of RIM1α by PKA at an N-terminal site between the ZF and PDZ domains is key for mossy-fiber LTP [17], and that the surfaces of Munc13–1 involved in homodimerization and heterodimerization (Figures 2G,6D, and6E) contain several predicted phosphorylation sites (Y23, T25, and Y140). Hence, phosphorylation may be crucial for regulation of inter- and intramolecular interactions of Munc13–1 and α-RIMs during plasticity processes.
0096
In particular, Ca2+-dependent phospholipid binding is known to be a common activity of many C2 domains that is mediated by loops at one tip of the β-sandwich [21], but it is now clear that many C2 domains do not bind Ca2+ and likely act as protein–protein interaction modules.

0413
In particular, Ca2+-dependent phospholipid binding is known to be a common activity of many C2 domains that is mediated by loops at one tip of the β-sandwich [21], but it is now clear that many C2 domains do not bind Ca2+ and likely act as protein–protein interaction modules.

0114
Indeed, homodimerization of the piccolo C2A domain depends on sequences homologous to strand 3 of the Munc13–1 C2A domain [29,30], suggesting that the dimerization mode may be similar. Our demonstration that a C2 domain can participate in two types of protein–protein interactions through distinct surfaces, which in turn differ from the usual Ca2+/phospholipid-binding site of C2 domains, emphasizes the diversity of interactions that can be mediated by a given protein module. It will thus not be surprising if such versatility is eventually found also in other domain families.

0081
Our demonstration that a C2 domain can participate in two types of protein–protein interactions through distinct surfaces, which in turn differ from the usual Ca2+/phospholipid-binding site of C2 domains, emphasizes the diversity of interactions that can be mediated by a given protein module.
0114
In the past, NMR spectroscopy and X-ray crystallography were largely viewed as alternative methods for structure determination of biomolecules, but the usefulness of combining the strengths of both techniques is increasingly being recognized [32,33]. On the other hand, two structural genomics studies recently indicated that there is no overt correlation between1H-15N HSQC spectral quality and successful crystallization of protein targets [34,35]. However, it is unclear to what extent the fragment length of each target was optimized in these studies, and a separate structural genomics effort suggested that NMR spectra can be used to identify promising targets for structure determination by X-ray crystallography [36]. The data presented here, together with our previous NMR analysis of Munc13–1/α-RIM/Rab3 interactions [20], provide a particularly compelling illustration of how NMR spectroscopy can assist in X-ray diffraction studies of protein complexes that present particularly challenging problems for crystallization, and at the same time can provide complementary information.

0077
In the past, NMR spectroscopy and X-ray crystallography were largely viewed as alternative methods for structure determination of biomolecules, but the usefulness of combining the strengths of both techniques is increasingly being recognized [32,33].

0077
However, it is unclear to what extent the fragment length of each target was optimized in these studies, and a separate structural genomics effort suggested that NMR spectra can be used to identify promising targets for structure determination by X-ray crystallography [36]. The data presented here, together with our previous NMR analysis of Munc13–1/α-RIM/Rab3 interactions [20], provide a particularly compelling illustration of how NMR spectroscopy can assist in X-ray diffraction studies of protein complexes that present particularly challenging problems for crystallization, and at the same time can provide complementary information.

0114
Altogether, these observations suggest that combining the strengths of NMR spectroscopy in fragment optimization and analysis of protein–protein interactions in solution with the high accuracy of structure determination by X-ray crystallography for biomolecules of any size will be particularly useful to study complex protein networks.

0077
Altogether, these observations suggest that combining the strengths of NMR spectroscopy in fragment optimization and analysis of protein–protein interactions in solution with the high accuracy of structure determination by X-ray crystallography for biomolecules of any size will be particularly useful to study complex protein networks.
0004
The construct to express Munc13–13–128 was generated by PCR and subcloned into a modified pGEX-KG vector [37] including a TEV protease cleavage site.

0096
Unlabeled and isotopically labeled proteins were expressed in bacteria as GST fusions as described [20]. The fusion proteins were isolated on glutathione-Sepharose beads (Amersham Biosciences, Little Chalfont, United Kingdom), cleaved from the GST moiety, and further purified by size-exclusion or ion-exchange chromatography.

0071
The fusion proteins were isolated on glutathione-Sepharose beads (Amersham Biosciences, Little Chalfont, United Kingdom), cleaved from the GST moiety, and further purified by size-exclusion or ion-exchange chromatography. Gel-filtration binding experiments were performed on Superdex S75 or S200 columns (Amersham) in 30 mM Tris-HCl buffer containing 150 mM NaCl, and 1 mM Tris(2-carboxyethyl)-phosphine (TCEP) at pH 7.4.

0004
Gel-filtration binding experiments were performed on Superdex S75 or S200 columns (Amersham) in 30 mM Tris-HCl buffer containing 150 mM NaCl, and 1 mM Tris(2-carboxyethyl)-phosphine (TCEP) at pH 7.4.
0004
ITC experiments were performed using a VP-ITC system (MicroCal, Northampton, Massachusetts, United States) at 20 °C in a buffer composed of 30 mM Tris (pH 7.4), 150 mM NaCl, and 1 mM TCEP.

0065
ITC experiments were performed using a VP-ITC system (MicroCal, Northampton, Massachusetts, United States) at 20 °C in a buffer composed of 30 mM Tris (pH 7.4), 150 mM NaCl, and 1 mM TCEP.

0065
Data were fit with a non-linear least-squares routine using a single-site binding model with Origin for ITC v.5.0 (Microcal), varying the stoichiometry(n), the enthalpy of the reaction (ΔH) and the association constant (Ka).
0004
The purified Munc13–13–128 fragment and the Munc13–13–150(K32E)/Rim2α82–142 complex were concentrated to 12 and 10 mg/ml, respectively, in buffer containing 30 mM Tris (pH 7.4), 150 mM NaCl, and 1 mM TCEP. Munc13–13–128 was crystallized in 0.4 M magnesium formate, 0.1 M sodium acetate (pH 4.5) at 20 °C using the hanging-drop vapor-diffusion method.

0071
The purified Munc13–13–128 fragment and the Munc13–13–150(K32E)/Rim2α82–142 complex were concentrated to 12 and 10 mg/ml, respectively, in buffer containing 30 mM Tris (pH 7.4), 150 mM NaCl, and 1 mM TCEP.

0676
The Munc13–13–150(K32E)/Rim2α82–142 complex was crystallized in 1.3 M ammonium tartrate (pH 7.0) at 20 °C using the hanging-drop vapor-diffusion method.

0104
Prior to data collection, crystals were transferred into a solution of 1.4 M ammonium tartrate (pH 7.0) and 20% (v/v) ethylene glycol, and then flash-cooled in liquid propane. Diffraction data were collected at the Structural Biology Center beamline 19ID of the Advanced Photon Source at 100 K to a Bragg spacing (dmin) of 1.78 Å.
0004
The loading concentration of Munc13–13–128, Munc13–13–132, Munc13–13–150, and Munc13–13–209 were 13.52 μM, 13.65 μM, 12.46 μM, and 12.55 μM in 30 mM Tris (pH 8.0), 150 mM NaCl, 1 mM TCEP.
0114
(A) Domain structure of Munc13–1 and RIM2α. The calmodulin-binding sequence (CaMb) of Munc13–1 and the helices that flank the RIM2α ZF domain (labeled a1 and a2) are indicated below the diagrams, and residue numbers are indicated above them.
0065
(B) ITC analysis of the binding of Munc13–13–150(K32E) to RIM2α82–142.
0107
The model represents how formation of the tripartite Munc13–1/α-RIM/Rab3 complex involves dissociation of the Munc13–1 homodimer and release of the interaction between the SGAWFY motif and Rab3A.
0081
It requires the coordination of adhesion receptor–matrix interactions on the cell surface, trafficking of the receptors to and from the sites of adhesion (adhesion site turnover), and cytoskeletal reorganization inside the cell.
0030
Integrins, like several other proteins lacking the AP-2 localization signal are internalized from the cell membrane via nonclathrin-derived structures, and some integrins have been shown to subsequently fuse with compartments containing clathrin-derived cargo proteins (Ng et al., 1999; Naslavsky et al., 2003; Upla et al., 2004; Weigert et al., 2004).

0889
Integrins, like several other proteins lacking the AP-2 localization signal are internalized from the cell membrane via nonclathrin-derived structures, and some integrins have been shown to subsequently fuse with compartments containing clathrin-derived cargo proteins (Ng et al., 1999; Naslavsky et al., 2003; Upla et al., 2004; Weigert et al., 2004).

0663
Depending on the cell type and the stimulus, integrins have been shown to be transported through caveolin-1–positive structures or early endosomes, either directly or via the perinuclear recycling compartment, back to the plasma membrane (Caswell and Norman, 2006).
0030
Integrin traffic is regulated by several kinases (Ng et al., 1999; Roberts et al., 2001, 2003, 2004; Ivaska et al., 2002; Woods et al., 2004) and motor proteins (Zhang et al., 2004).

0889
Integrin traffic is regulated by several kinases (Ng et al., 1999; Roberts et al., 2001, 2003, 2004; Ivaska et al., 2002; Woods et al., 2004) and motor proteins (Zhang et al., 2004).

0107
Although Rab proteins are known to bind a multiplicity of diverse effectors (Zerial and McBride, 2001), only a few examples demonstrate an interaction between a Rab GTPase and a cargo molecule (Seachrist et al., 2002; van IJzendoorn et al., 2002).

0030
Although Rab proteins are known to bind a multiplicity of diverse effectors (Zerial and McBride, 2001), only a few examples demonstrate an interaction between a Rab GTPase and a cargo molecule (Seachrist et al., 2002; van IJzendoorn et al., 2002).

0889
Although Rab proteins are known to bind a multiplicity of diverse effectors (Zerial and McBride, 2001), only a few examples demonstrate an interaction between a Rab GTPase and a cargo molecule (Seachrist et al., 2002; van IJzendoorn et al., 2002).

0096
We identify Rab21 and Rab5 as integrin-associated proteins and positive regulators of integrin traffic.
0030
Rab21 is a ubiquitously expressed and poorly characterized member of the Rab family that has recently been shown to function on the endocytic pathway (Simpson et al., 2004).

0889
Rab21 is a ubiquitously expressed and poorly characterized member of the Rab family that has recently been shown to function on the endocytic pathway (Simpson et al., 2004).
0413
The ability of Rab21 to associate with integrins was further confirmed in human cells. GFP-tagged Rab21 coprecipitated with α2β1-integrin, a collagen binding molecule (Takada and Hemler, 1989), in cells plated either on collagen or on plastic (Fig.

0107
1 A), indicating a constant, rather than a matrix adhesion–inducible, association between the two proteins.

0096
Finally, endogenous β1-integrins readily associated with endogenous Rab21 and to some extent Rab5 proteins in vivo.

0006
These associations were specific, as Rab7 and Rab11 failed to coprecipitate β1-integrin from these cells, even though separately all proteins were efficiently immunoprecipitated (Fig.
0413
(B and C) Full-length Rab21WT tagged with Rluc (Rluc-Rab21), mutants (Rluc-Rab21GTP, -Rab21GDP, or -Rab21C-del), or Rluc alone were expressed in Saos-2 cells stably expressing chimeric α-integrin subunits (α2/α1) or (α2/α5) (B) or HT1080 cells (C), and immunoprecipitation was performed with the indicated antibodies. The coprecipitated luciferase activity is presented relative to the basal nonspecific activity detected in the relevant control immunoprecipitations (anti-EGFR).

0107
Interaction between p53 and large T antigen was used as a positive control.

0006
(E) CHO cells were cotransfected with Rluc-Rab21 and GFP-α2 variants. Immunoprecipitation was performed with anti-GFP antibody, and the coprecipitated luciferase activity is presented relative to the activity detected in immunoprecipitations from α2WT-expressing cells.

0413
(E) CHO cells were cotransfected with Rluc-Rab21 and GFP-α2 variants. Immunoprecipitation was performed with anti-GFP antibody, and the coprecipitated luciferase activity is presented relative to the activity detected in immunoprecipitations from α2WT-expressing cells.

0007
Immunoprecipitation was performed with anti-GFP antibody, and the coprecipitated luciferase activity is presented relative to the activity detected in immunoprecipitations from α2WT-expressing cells.

0006
(F) Endogenous proteins were immunoprecipitated from MDA-MB-231 cells using the indicated antibodies and blotted with anti-β1 or anti-Rab antibodies.

0413
(F) Endogenous proteins were immunoprecipitated from MDA-MB-231 cells using the indicated antibodies and blotted with anti-β1 or anti-Rab antibodies.

0007
(F) Endogenous proteins were immunoprecipitated from MDA-MB-231 cells using the indicated antibodies and blotted with anti-β1 or anti-Rab antibodies.

0416
(G) Nontransfected MDA-MB-231 cells were analyzed with three-color immunofluorescence staining (Rab5 or Rab21, green; EEA1, red; and β1-integrin, blue).
0030
A Rab21 mutant (Simpson et al., 2004) that has defects in GTP hydrolysis (Rab21 Q76L, designated Rab21GTP) showed a 10–73% increase in its ability to associate with integrins in different cell lines when compared with Rab21 wild type (WT).

0889
A Rab21 mutant (Simpson et al., 2004) that has defects in GTP hydrolysis (Rab21 Q76L, designated Rab21GTP) showed a 10–73% increase in its ability to associate with integrins in different cell lines when compared with Rab21 wild type (WT).
0413
The cellular localization of endogenous β1-integrins, Rab5A, and Rab21 was studied in MDA-MB-231 cells. β1-Integrin was detected by the NH2-terminally binding antibody in the lumen of endogenous Rab5- and Rab21-positive large vesicles (Fig.

0096
β1-Integrin was detected by the NH2-terminally binding antibody in the lumen of endogenous Rab5- and Rab21-positive large vesicles (Fig.

0413
In addition, these cells also harbor large endogenous EEA1-positive endosomes that stained positive for Rab5A but showed very limited overlap with Rab21 (Fig.

0030
α/β1- Integrin heterodimers can be expressed in active (extracellular domain detected by a monoclonal antibody HUTS21; active-β1 antibody) and inactive forms on the cell surface (extracellular domain of both forms detected by P5D2; pan-β1 antibody; Lenter et al., 1993; Luque et al., 1996).

0889
α/β1- Integrin heterodimers can be expressed in active (extracellular domain detected by a monoclonal antibody HUTS21; active-β1 antibody) and inactive forms on the cell surface (extracellular domain of both forms detected by P5D2; pan-β1 antibody; Lenter et al., 1993; Luque et al., 1996).

0416
Expression of Rab21 in MDA-MB-231 breast cancer cells altered the subcellular localization of the total pool of β1-integrins. Cells expressing GFP-Rab21 contained numerous β1-integrin−positive vesicles (compare the nontransfected cell [Fig.

0411
Expression of Rab21 in MDA-MB-231 breast cancer cells altered the subcellular localization of the total pool of β1-integrins. Cells expressing GFP-Rab21 contained numerous β1-integrin−positive vesicles (compare the nontransfected cell [Fig.

0030
The recruitment of integrins to Rab21-positive intracellular structures was confirmed using two additional antibodies recognizing active-β1-integrin (12G10 [Mould et al., 1995] and P4G11 [Wayner et al., 1993]; Fig.

0889
The recruitment of integrins to Rab21-positive intracellular structures was confirmed using two additional antibodies recognizing active-β1-integrin (12G10 [Mould et al., 1995] and P4G11 [Wayner et al., 1993]; Fig.

0030
Similar to previous observations on breast cancer cells (Ng et al., 1999), the ECM proteins collagen and fibronectin were also detected in the endocytic structures (Fig.

0411
Similar to previous observations on breast cancer cells (Ng et al., 1999), the ECM proteins collagen and fibronectin were also detected in the endocytic structures (Fig.

0889
Similar to previous observations on breast cancer cells (Ng et al., 1999), the ECM proteins collagen and fibronectin were also detected in the endocytic structures (Fig.
0413
(M) Immunoprecipitation (IP) was performed with anti–β1-integrin mAb or control mouse IgG from GFP-Rab21WT– or -Rab21CCSS–transfected MDA-MB-231 cells followed by Western blotting with anti-β1 or anti-GFP.
0030
Mutagenesis of the putative COOH-terminal prenylation motif (CCXXX; Opdam et al., 2000) in Rab21 resulted in a complete loss of vesicular localization of Rab21 (GFP-Rab21CCSS; Fig.

0889
Mutagenesis of the putative COOH-terminal prenylation motif (CCXXX; Opdam et al., 2000) in Rab21 resulted in a complete loss of vesicular localization of Rab21 (GFP-Rab21CCSS; Fig.

0006
Cells transfected with GFP-Rab21CCSS showed prominent focal adhesions (active-β1 antibody; Fig. 2 H, arrows) when compared with the vesicular pattern observed in the nontransfected cells (Fig.
0096
Interestingly, Rab5 and Rab21 antibodies very weakly coimmunoprecipitated biotinylated ∼160-kD proteins located on the plasma membrane (Fig.
0007
(A) Nontransfected MDA-MB-231 cells were surface labeled with cleavable biotin and lysed immediately or allowed to internalize cell surface proteins for 15 min. Immunoprecipitations (IPs) were performed as indicated, and the Rab coprecipitating proteins were detected first with anti-biotin antibody followed by stripping and reprobing with anti–β1-integrin antibody.

0055
(B and C) MDA-MB-231 cells transiently transfected with GFP or GFP-Rabs (B) or stably expressing Scr- or Rab21-shRNA (C) plated on collagen (1 h) and were surface labeled with cleavable biotin.

0055
Inset (C) shows the extent of Rab21 down-regulation in Rab21-shRNA–expressing cells. (D) GFP-, GFP-Rab21–, GFP-Rab21GTP–, or GFP-Rab21GDP–transfected HeLa cells (48 h) were lysed, and postnuclear supernatant was fractionated on a sucrose gradient and subjected to Western blot analysis with anti–α2-integrin and anti-GFP antibodies.
0030
We are aware that we have not excluded the effects of Rab21 expression for all types of unspecific internalization (e.g., macropinocytosis), but at least its closest homologues (Rab5 and Rab22) have not been shown to affect macropinocytosis (Rosenfeld et al., 2001; Kauppi et al., 2002).

0889
We are aware that we have not excluded the effects of Rab21 expression for all types of unspecific internalization (e.g., macropinocytosis), but at least its closest homologues (Rab5 and Rab22) have not been shown to affect macropinocytosis (Rosenfeld et al., 2001; Kauppi et al., 2002).

0055
Interestingly, the internalized receptor recycled rapidly back to the cell surface in Rab5- and Rab21-expressing cells with >50% of the labeled pool being recycled during a 15-min chase (detected by the reduction in the amount of biotinylated receptor in these cells where cell surface biotin is cleaved). Integrin internalization was reduced by expression of GFP-Rab21GDP, whereas GFP-Rab21GTP induced a steady accumulation of the internalized receptor when compared with GFP-, Rab5-, Rab11-, and Rab21GDP-expressing cells (Fig.

0411
Interestingly, the internalized receptor recycled rapidly back to the cell surface in Rab5- and Rab21-expressing cells with >50% of the labeled pool being recycled during a 15-min chase (detected by the reduction in the amount of biotinylated receptor in these cells where cell surface biotin is cleaved). Integrin internalization was reduced by expression of GFP-Rab21GDP, whereas GFP-Rab21GTP induced a steady accumulation of the internalized receptor when compared with GFP-, Rab5-, Rab11-, and Rab21GDP-expressing cells (Fig.
0030
In density gradient fractionations, expression of GFP-Rab21 shifted the integrins toward the denser Rab-positive fractions (Hughes et al., 2002), and GFP-Rab21 cofractionated with α2-integrin in fractions 3–9 (Fig.

0889
In density gradient fractionations, expression of GFP-Rab21 shifted the integrins toward the denser Rab-positive fractions (Hughes et al., 2002), and GFP-Rab21 cofractionated with α2-integrin in fractions 3–9 (Fig.

0027
In the lighter fractions (3–5), GFP-Rab21 and integrin were found to cosediment with the Golgi-marker GM130, whereas in the denser fractions, cosedimentation was observed with the ER marker P115 (fractions 6–8) and EEA1 (fractions 7–9; Fig.
0040
Further characterization of the large intracellular structures induced by GFP-Rab21 overexpression was performed by electron microscopy and immunogold labeling of GFP (Fig.
0416
Combined total internal reflection fluorescence microscopy (TIRFM; pseudocolored green) and conventional widefield epifluorescence analysis (pseudocolored red) showed GFP-Rab21 vesicles emanating from the membrane into the cell (changing from green to red/yellow) and back or vice versa (although formally this technique does not exclude the possibility that some vesicles will by chance come closer and further away from the plasma membrane).
0411
We also addressed whether Rab21 expression alters the motility of integrins in live cells.

0055
Rab21 expression induced motile GFP–α2-integrin–labeled vesicles (Fig. 5 B and Video 4, available at http://www.jcb.org/cgi/content/full/jcb.200509019/DC1), whereas cells expressing GFP–α2-integrin alone or with Rluc-Rab21GDP, showed a vesicular-tubular staining pattern with no obvious vesicles and that partially overlaps with ER tracker stain in live cells (Fig.
0055
(B and C) MDA-MB-231 cells transfected with GFP–α2-integrin and Rluc-Rab21 (B) or with GFP–α2-integrin alone (C) were plated on collagen for 1 h, and time-lapse series were acquired with widefield epifluorescence microscopy. Cotransfected cells (B; see Video 4, available at http://www.jcb.org/cgi/content/full/jcb.200509019/DC1) show green fluorescent vesicles moving in the cytosol. No vesicles are visible in cells transfected with GFP–α2-integrin alone (C; see Video 5).

0416
(B and C) MDA-MB-231 cells transfected with GFP–α2-integrin and Rluc-Rab21 (B) or with GFP–α2-integrin alone (C) were plated on collagen for 1 h, and time-lapse series were acquired with widefield epifluorescence microscopy. Cotransfected cells (B; see Video 4, available at http://www.jcb.org/cgi/content/full/jcb.200509019/DC1) show green fluorescent vesicles moving in the cytosol.

0663
(B and C) MDA-MB-231 cells transfected with GFP–α2-integrin and Rluc-Rab21 (B) or with GFP–α2-integrin alone (C) were plated on collagen for 1 h, and time-lapse series were acquired with widefield epifluorescence microscopy.
0411
To investigate the importance of Rab21-mediated integrin traffic in β1-integrin–dependent cell adhesion, we studied MDA-MB-231 breast and PC3 prostate cancer cells having relatively high and low endogenous Rab21 expression, respectively (unpublished data). Of all the Rabs tested, overexpressed Rab21 was most efficient in increasing cell adhesion to collagen (86 ± 8% in PC3 cells and 22% in MDA-MB-231 cells; Fig.

0413
To investigate the importance of Rab21-mediated integrin traffic in β1-integrin–dependent cell adhesion, we studied MDA-MB-231 breast and PC3 prostate cancer cells having relatively high and low endogenous Rab21 expression, respectively (unpublished data). Of all the Rabs tested, overexpressed Rab21 was most efficient in increasing cell adhesion to collagen (86 ± 8% in PC3 cells and 22% in MDA-MB-231 cells; Fig.

0411
Rab21 expression did not enhance the initial adhesion step of cells plated on fibronectin or vitronectin (Fig.

0411
Expression of Rab5 increased adhesion by 48 ± 13% in PC3 cells.

0411
6 C) and PC3 cells (not depicted), suggesting that efficient transport of endocytosed integrins to newly formed sites of adhesion, rather than a change in the steady-state expression of integrins on the cell surface, is the basis of Rab21-induced cell adhesion. Interestingly, GFP-Rab21GDP caused a modest increase in β1-integrin surface expression in MDA-MB-231 cells (Fig.

0413
CHO cells lack endogenous collagen receptors.

0411
Transient expression of α2WT, α2AA (deficient Rab21 association), or α2A (unaltered Rab21 association), together with Rluc, enabled equivalent adhesion of these cells to collagen (Fig.
0054
(A and B) Adhesion of transiently transfected PC3 or MDA-MB-231 cells was studied on collagen (0.25 μg/ml CI, 30 min; means ± SD; n = 12; **, P < 0.02; ***, P < 0.003). (C) Cell surface expression of β1-integrin was analyzed with dual-color FACS from transiently transfected MDA-MB-231 cells.

0411
(A and B) Adhesion of transiently transfected PC3 or MDA-MB-231 cells was studied on collagen (0.25 μg/ml CI, 30 min; means ± SD; n = 12; **, P < 0.02; ***, P < 0.003). (C) Cell surface expression of β1-integrin was analyzed with dual-color FACS from transiently transfected MDA-MB-231 cells.

0416
Mean fluorescence intensities (633-nm laser; β1-integrins) were detected from GFP-positive and -negative cells and are presented as a ratio (relative means ± SEM; n = 4). (D) MDA-MB-231 cells were transfected with two siRNAs specific for Rab21, scrambled control (Scr), or oligofectamine alone (no oligo).

0416
(F) Subcellular localization of GFP-integrin in the transfected collagen adhering cells was studied by widefield fluorescence microscopy.

0663
(F) Subcellular localization of GFP-integrin in the transfected collagen adhering cells was studied by widefield fluorescence microscopy.
0055
We also found that cells expressing GFP-Rab21WT migrated on plastic in a scratch wound assay to close the wound almost completely, whereas GFP-, GFP-Rab21GDP–, and GFP-Rab21GTP–expressing cells migrated less efficiently, covering only ∼70–75% of the wound area (Fig.

0411
Conversely, stable Rab21-shRNA (short hairpin RNA)–transfected cells with reduced Rab21 expression (Fig. 3 C) migrated poorly, covering only sim;59% of the wound area, whereas Scr-shRNA–transfected control cells closed the wound almost completely (Fig.

0411
During this time, there were no obvious differences in the proliferation of the stable transfected cells (not depicted). These data suggest that expression of Rab21 regulates migration in these cells.
0055
Number of vesicles per area near the edges (5 μm; mean ± SD; *, P < 0.05) were quantified at two time points after plating (1–2 h, 9 cells; overnight, 6 cells). (B and C) Migration of MDA-MB-231 cells stably expressing GFP, GFP-Rab21, GFP-Rab21GTP, GFP-Rab21GDP, Rab21-shRNA, or Scr-shRNA was studied using the scratch wound assay (migration 24 h in 10% FBS). Analyses of wound areas covered by the cells are shown (means ± SD; n = 4–7; *, P < 0.01; **, P < 0.001).
0413
Together, our data show that Rab21 activity regulates integrin-dependent adhesion to collagen and human bone as well as the motility of cancer cells on cell-secreted and serum-derived matrix components other than collagen.
0411
Furthermore, Rab21 expression is shown to regulate integrin-containing focal adhesions and adhesion and migration of breast and prostate cancer cells. Live-cell imaging revealed that Rab21-positive vesicles move between the plasma membrane and the cell body and that expression of Rab21 modulates vesicular motility of integrins in vivo.
0107
The identification of integrin association with the putative early endosomal Rab21 led us to identify a functional relationship between integrin association with Rab21 and Rab5 and the regulation of cell adhesion.

0030
Upon growth factor stimulus, internalized integrins either recycle very rapidly back to the plasma membrane from the perinuclear compartment (β1-integrins; Powelka et al., 2004) or are completely rerouted to a short-loop trafficking pathway directly back to the membrane (αvβ3-integrin; Roberts et al., 2004).

0889
Upon growth factor stimulus, internalized integrins either recycle very rapidly back to the plasma membrane from the perinuclear compartment (β1-integrins; Powelka et al., 2004) or are completely rerouted to a short-loop trafficking pathway directly back to the membrane (αvβ3-integrin; Roberts et al., 2004).

0030
This is in agreement with the recent finding that integrins in newly forming protrusions travel on actin cables associated to motor protein Myosin X and that normal cell adhesion and spreading during the initial stages of adhesion seem to require the efficient motility of integrins close to the plasma membrane (Zhang et al., 2004).

0889
This is in agreement with the recent finding that integrins in newly forming protrusions travel on actin cables associated to motor protein Myosin X and that normal cell adhesion and spreading during the initial stages of adhesion seem to require the efficient motility of integrins close to the plasma membrane (Zhang et al., 2004).
0030
One study demonstrates the internalization of GFP-tagged integrin in live cells (Laukaitis et al., 2001).

0889
One study demonstrates the internalization of GFP-tagged integrin in live cells (Laukaitis et al., 2001).

0889
Although the role of Rab proteins was not addressed in the study of Laukaitis et al.

0030
Although the role of Rab proteins was not addressed in the study of Laukaitis et al.

0096
However, we cannot rule out the possibility that as-yet-unknown proteins that are capable of binding to the integrin may serve as linkers between Rab21 and integrins.

0030
Evidence from detailed biochemical and nuclear magnetic resonance studies (Stefansson et al., 2004; Vinogradova et al., 2004) indicates that upon integrin activation the membrane-proximal regions of the integrin cytoplasmic tails move out of the membrane into the cytoplasm, revealing the highly conserved residues to the cytoplasmic face and involving substantial structural changes in the cytoplasmic tails.

0077
Evidence from detailed biochemical and nuclear magnetic resonance studies (Stefansson et al., 2004; Vinogradova et al., 2004) indicates that upon integrin activation the membrane-proximal regions of the integrin cytoplasmic tails move out of the membrane into the cytoplasm, revealing the highly conserved residues to the cytoplasmic face and involving substantial structural changes in the cytoplasmic tails.

0889
Evidence from detailed biochemical and nuclear magnetic resonance studies (Stefansson et al., 2004; Vinogradova et al., 2004) indicates that upon integrin activation the membrane-proximal regions of the integrin cytoplasmic tails move out of the membrane into the cytoplasm, revealing the highly conserved residues to the cytoplasmic face and involving substantial structural changes in the cytoplasmic tails.
0030
Prenylation-dependent membrane targeting of Rabs is crucial for Rab function as regulators of vesicle fusion in intracellular protein trafficking (Desnoyers et al., 1996).

0889
Prenylation-dependent membrane targeting of Rabs is crucial for Rab function as regulators of vesicle fusion in intracellular protein trafficking (Desnoyers et al., 1996).
0030
Integrins have been shown to internalize in a dynamin and PKCα-dependent manner after ligand binding (Ng et al., 1999). Internalized β1-integrins are then targeted to caveosomes (Upla et al., 2004) or transferrin-positive endosomes (Laukaitis et al., 2001) depending on the heterodimer, the cell type, and the stimulus used. Recent data showing that the caveolar and endosomal pathways intersect (Pelkmans et al., 2004) and our data showing caveolin-1–positive domains on Rab5- and Rab21-positive integrin–containing vesicles provide a possible explanation for these variable integrin trafficking routes.

0889
Integrins have been shown to internalize in a dynamin and PKCα-dependent manner after ligand binding (Ng et al., 1999). Internalized β1-integrins are then targeted to caveosomes (Upla et al., 2004) or transferrin-positive endosomes (Laukaitis et al., 2001) depending on the heterodimer, the cell type, and the stimulus used. Recent data showing that the caveolar and endosomal pathways intersect (Pelkmans et al., 2004) and our data showing caveolin-1–positive domains on Rab5- and Rab21-positive integrin–containing vesicles provide a possible explanation for these variable integrin trafficking routes.
0030
This correlates well with the previous work showing that α5-integrin engagement initiates the newly forming adhesions and serves to organize proteins like paxillin in these sites (Laukaitis et al., 2001).

0889
This correlates well with the previous work showing that α5-integrin engagement initiates the newly forming adhesions and serves to organize proteins like paxillin in these sites (Laukaitis et al., 2001).
0030
Interestingly, emerging evidence implicates alterations in the Rab small GTPases and their associated regulatory proteins and effectors in multiple human diseases, including cancer (Cheng et al., 2004).

0889
Interestingly, emerging evidence implicates alterations in the Rab small GTPases and their associated regulatory proteins and effectors in multiple human diseases, including cancer (Cheng et al., 2004).
0030
Antibodies against the following antigens were used: EEA1, Rab5A, Rab7, Rab11, caveolin-1 (all from Santa Cruz Biotechnology, Inc.), Rab21 (Opdam et al., 2000), β1-integrin (P5D2, P4G11, and AIIB2), α5-integrin (BIIG2), EGFR (151-IgG; all from the Drosophila Studies Hybridoma Bank), α2 (mAb MCA2025; Serotec), pAb AB1934 (Chemicon), α1 (MAB1973; Chemicon), α6 (MAB699; Chemicon), β1 (HUTS-21 [BD Biosciences] and MAB2252 [Chemicon]), collagen type 1 (RAHC11; Imtek), GFP polyclonal antibody, fluorescently conjugated secondary antibodies, Cell Tracker dyes, and labeled transferrin (all from Invitrogen).

0889
Antibodies against the following antigens were used: EEA1, Rab5A, Rab7, Rab11, caveolin-1 (all from Santa Cruz Biotechnology, Inc.), Rab21 (Opdam et al., 2000), β1-integrin (P5D2, P4G11, and AIIB2), α5-integrin (BIIG2), EGFR (151-IgG; all from the Drosophila Studies Hybridoma Bank), α2 (mAb MCA2025; Serotec), pAb AB1934 (Chemicon), α1 (MAB1973; Chemicon), α6 (MAB699; Chemicon), β1 (HUTS-21 [BD Biosciences] and MAB2252 [Chemicon]), collagen type 1 (RAHC11; Imtek), GFP polyclonal antibody, fluorescently conjugated secondary antibodies, Cell Tracker dyes, and labeled transferrin (all from Invitrogen).
0030
Plasmids encoding GFP-Rab5a, GFP-Rab7, YFP-Rab9, and GFP-Rab11 have been described (Wilcke et al., 2000; Barbero et al., 2002; Lebrand et al., 2002; Gomes et al., 2003), and YFP-mouse talin was provided by D.

0889
Plasmids encoding GFP-Rab5a, GFP-Rab7, YFP-Rab9, and GFP-Rab11 have been described (Wilcke et al., 2000; Barbero et al., 2002; Lebrand et al., 2002; Gomes et al., 2003), and YFP-mouse talin was provided by D.

0030
α2-Integrin was subcloned from the α2 cDNA in pawneo2 vector (Ivaska et al., 1999) into pEGFP-C2 vector.

0889
α2-Integrin was subcloned from the α2 cDNA in pawneo2 vector (Ivaska et al., 1999) into pEGFP-C2 vector.
0413
Rluc-tagged Rab21 constructs alone or with GFP-α2 variants (CHO cells that lack endogenous collagen binding integrins) were transfected into 95% confluent cells using Lipofectamine 2000 and incubated for 18 h. For immunoprecipitations with endogenous proteins, confluent MDA-MB-231 cells (20 × 106) were collected from plastic plates with cold PBS. For analysis of association with cell surface–labeled integrin, MDA-MB-231 cells were plated on collagen-coated dishes for 1 h and surface biotinylated with cleavable biotin (0.5 mg/ml EZ-sulfo-NHS-SS-biotin in HANKS buffer) for 30 min on ice.

0030
Reimmunoprecipitations were performed as described earlier (Mattila et al., 2005).

0889
Reimmunoprecipitations were performed as described earlier (Mattila et al., 2005).
0030
These were performed as described previously (Roberts et al., 2001; Ivaska et al., 2002) with some modifications.

0889
These were performed as described previously (Roberts et al., 2001; Ivaska et al., 2002) with some modifications.

0007
Biotin was removed from cell surface proteins by MesNa reduction and iodoacetamide quenching on ice. The cells were lysed (200 mM NaCl, 75 mM Tris, 15 mM NaF, 1.5 mM Na3VO4, 7.5 mM EDTA, 7.5 mM EGTA, 1.5% Triton-X-100, and Complete), and the amount of biotinylated integrin was assayed using the anti–β1-integrin antibody AIIB2 to capture the integrins and HRP anti-biotin antibody for ELISA detection.

0411
The cells were lysed (200 mM NaCl, 75 mM Tris, 15 mM NaF, 1.5 mM Na3VO4, 7.5 mM EDTA, 7.5 mM EGTA, 1.5% Triton-X-100, and Complete), and the amount of biotinylated integrin was assayed using the anti–β1-integrin antibody AIIB2 to capture the integrins and HRP anti-biotin antibody for ELISA detection. As control, the cells were lysed after the labeling to determine the amount of total biotinylated integrin.

0006
The cells were lysed (200 mM NaCl, 75 mM Tris, 15 mM NaF, 1.5 mM Na3VO4, 7.5 mM EDTA, 7.5 mM EGTA, 1.5% Triton-X-100, and Complete), and the amount of biotinylated integrin was assayed using the anti–β1-integrin antibody AIIB2 to capture the integrins and HRP anti-biotin antibody for ELISA detection. As control, the cells were lysed after the labeling to determine the amount of total biotinylated integrin.
0055
HeLa cells were transiently transfected with GFP, GFP-Rab21, or GFP-Rab21GTP using Lipofectamine 2000 as described in the Immunoprecipitations section.

0030
48 h after transfection, the cells were harvested and fractionated on a sucrose density gradient and analyzed by Western blotting as described previously (Hughes et al., 2002).

0889
48 h after transfection, the cells were harvested and fractionated on a sucrose density gradient and analyzed by Western blotting as described previously (Hughes et al., 2002).
0018
The α2-integrin COOH-terminal tail (28 residues) Gal4 DNA binding domain fusion (pGBKT7 vector) was used to screen a mouse E17 Matchmaker cDNA library (CLONTECH Laboratories, Inc.) as described previously (Mattila et al., 2005). In yeast mating tests, pGADT7-Rab21 (95–222) prey was transformed in Y187 host strain and cytoplasmic tails of α2- and α11-integrin (pGBKT7-α2 and -α11) and their variants in AH109 host strain.

0030
The α2-integrin COOH-terminal tail (28 residues) Gal4 DNA binding domain fusion (pGBKT7 vector) was used to screen a mouse E17 Matchmaker cDNA library (CLONTECH Laboratories, Inc.) as described previously (Mattila et al., 2005).

0096
The α2-integrin COOH-terminal tail (28 residues) Gal4 DNA binding domain fusion (pGBKT7 vector) was used to screen a mouse E17 Matchmaker cDNA library (CLONTECH Laboratories, Inc.) as described previously (Mattila et al., 2005).

0889
The α2-integrin COOH-terminal tail (28 residues) Gal4 DNA binding domain fusion (pGBKT7 vector) was used to screen a mouse E17 Matchmaker cDNA library (CLONTECH Laboratories, Inc.) as described previously (Mattila et al., 2005).
0413
CHO cells (American Type Culture Collection) were grown in MEM Alpha Medium + 5% FBS. Saos-2 cells express no endogenous α2 (Ivaska et al., 1999).

0030
Saos-2 cells express no endogenous α2 (Ivaska et al., 1999). Stable Saos-2 cells expressing equal levels of chimeric integrins (extracellular domain of α2 fused with α1 or α5 cytoplasmic tails; Ivaska et al., 1999) have been described (Mattila et al., 2005).

0889
Saos-2 cells express no endogenous α2 (Ivaska et al., 1999). Stable Saos-2 cells expressing equal levels of chimeric integrins (extracellular domain of α2 fused with α1 or α5 cytoplasmic tails; Ivaska et al., 1999) have been described (Mattila et al., 2005).
0055
After one washing with PBS, cells were fixed (4% PFA, 10 min). Adhesion was measured by counting the number of green fluorescent cells using Acumen Assay Explorer 488 nm. The total number of GFP-positive cells was assayed after adhesion to collagen for 4 h in the presence of 10% FBS.
0055
For the scratch wound assay, stable MDA-MB-231 cells expressing GFP, GFP-Rab21, GFP-Rab21GTP, or GFP-Rab21GDP were generated. The cells were seeded onto collagen-coated 96-well plates at 35,000 cells/well and allowed to adhere overnight in the presence of 10% FBS.
0040
After permeabilization (PBS/0.02% saponin/10% FBS, 15 min), cells were stained with primary antibodies (in the same buffer) for 1 h at RT.

0663
Slides were examined using an inverted fluorescence microscope (Carl Zeiss MicroImaging, Inc.) or a confocal laser-scanning microscope (Axioplan 2 with LSM 510; Carl Zeiss MicroImaging, Inc.) equipped with 100×/1.4 Plan-Apochromat oil-immersion objectives. Confocal images represent a single z section of ∼1.0 μm.

0030
β1-Integrin and transferrin internalization were studied as described previously (Powelka et al., 2004).

0889
β1-Integrin and transferrin internalization were studied as described previously (Powelka et al., 2004).
0416
TIRFM was combined with conventional widefield epifluorescence microscopy and time-lapse series (frame rate ∼2/s) Widefield images were pseudocolored red and TIRFM images green. Transiently transfected GFP-Rab21 cells were plated on acid-washed glass-bottomed dishes (MatTek Corporation) coated with 10 μg/ml collagen type I and allowed to adhere for 1 h before microscopy.
0006
S1 shows the cellular localization of endogenous β1-integrin, caveolin-1, Rab21, and EEA1 in MDA-MB-231 cells expressing GFP-Rab5 or -Rab21 and internalization of β1-integrin antibody and labeled transferrin in GFP-Rab5– and GFP-Rab21–expressing cells.

0413
S1 shows the cellular localization of endogenous β1-integrin, caveolin-1, Rab21, and EEA1 in MDA-MB-231 cells expressing GFP-Rab5 or -Rab21 and internalization of β1-integrin antibody and labeled transferrin in GFP-Rab5– and GFP-Rab21–expressing cells.

0055
Table S1 demonstrates the association of Rab21WT and its variants with α/β1-integrin heterodimers in HT1080 cells. Video 1 shows MDA-MB-231 cells expressing GFP-Rab21, adhering to collagen recorded on GFP channel. Video 2 shows a combined widefield epifluorescence and TIRFM analysis of MDA-MB-231 cells expressing GFP-Rab21, adhering to collagen. Video 3 shows a combined widefield epifluorescence and TIRFM analysis of MDA-MB-231 cells expressing GFP-Rab21GDP mutant, adhering to collagen. Video 4 shows MDA-MB-231 cells cotransfected with GFP–α2-integrin and Rluc-Rab21WT, adhering to collagen recorded on GFP channel. Video 5 shows MDA-MB-231 cells transfected with GFP–α2-integrin alone adhering to collagen recorded on GFP channel.
0018
In the case of miRNAs, they are assembled into miRNA-induced silencing complexes (miRISC) that contain Dicer, argonaute proteins, transactivation-responsive RNA-binding protein [19–22], and other cellular factors [4].
0413
In human cells, the mechanism by which the endogenous miRNA and siRNA pathways are distinguished is not clearly understood.
0018
One P-body protein, RCK/p54, the human homolog of yeast Dhh1p, is a member of the ATP-dependent DEAD box helicase family and was originally identified as a proto-oncogene [42].
0413
Here, we show that RCK/p54 interacts with argonaute proteins, Ago1 and Ago2, in affinity-purified active RISC assemblies from human cells programmed with siRNA or endogenous miRNA; directly interacts with Ago1 and Ago2 in vivo, facilitates formation of cytoplasmic P-bodies, and acts as a general repressor of translation.
0018
To investigate the mechanism of miRNA-mediated repression of mRNA translation and to determine the interactions of P-body components with the RNAi machinery, we constructed expression vectors for the yellow fluorescent protein (YFP)-tagged P-body proteins, Lsm1, RCK/p54, Dcp2, and eIF4E.

0055
These vectors were co-expressed in HeLa cells with Myc-tagged Ago2 and immunopurified using anti-Myc antibodies. The protein composition of isolated complexes was analyzed by immunoblot using antibodies against green fluorescent protein (GFP) or Myc.

0096
When total cell extracts (TCE) were analyzed to determine the protein expression efficiencies of the vectors used in these experiments (Figure 1A, TCE lane), all YFP- and Myc-tagged proteins were expressed.

0018
When total cell extracts (TCE) were analyzed to determine the protein expression efficiencies of the vectors used in these experiments (Figure 1A, TCE lane), all YFP- and Myc-tagged proteins were expressed.
0018
Since P-bodies contain RNA and proteins, many protein components of P-bodies are likely to be assembled on a common RNA scaffold without forming functional protein–protein interactions.

0007
To address this possibility, HeLa cells were transfected with vectors to co-express Myc-Ago2 and the YFP-tagged P-body proteins, Lsm1, RCK/p54, Dcp2, and eIF4E, subjected to RNase A digestion, and immunopurified.
0055
To confirm whether these structures also contain Ago2 as recently reported [28, 31], we transfected HeLa cells with expression vectors containing YFP-Ago1, CFP (cyan fluorescent protein)-Ago1, YFP-Ago2, and CFP-Ago2. Transiently expressed YFP- and CFP-tagged Ago1 and Ago2 co-localized at specific foci in cytoplasm (unpublished data).

0411
To confirm whether these structures also contain Ago2 as recently reported [28, 31], we transfected HeLa cells with expression vectors containing YFP-Ago1, CFP (cyan fluorescent protein)-Ago1, YFP-Ago2, and CFP-Ago2.

0411
To examine the contents of these cytoplasmic foci, HeLa cells were transfected with expression vectors for YFP-Lsm1 and CFP-Ago2, or YFP-RCK/p54 and CFP-Ago2, and visualized 24 h later by confocal microscopy.

0663
To examine the contents of these cytoplasmic foci, HeLa cells were transfected with expression vectors for YFP-Lsm1 and CFP-Ago2, or YFP-RCK/p54 and CFP-Ago2, and visualized 24 h later by confocal microscopy.

0416
To examine the contents of these cytoplasmic foci, HeLa cells were transfected with expression vectors for YFP-Lsm1 and CFP-Ago2, or YFP-RCK/p54 and CFP-Ago2, and visualized 24 h later by confocal microscopy.

0413
To confirm that the localization of Ago2 and Lsm1 to P-bodies was not the result of over-expression in transiently transfected cells, we immunostained cells with antibodies to endogenous Ago2 and Lsm1 and found that the localization of Ago2 and Lsm1 to P-bodies (Figure S1) was similar to that in Figure 1B.
0019
In cells expressing CFP-Ago2 and YFP-Lsm1, FRET efficiency was not significant (1.62% ± 1.11%), corroborating our immunoprecipitation results (Figure 1A).

0107
Similar to the YFP-RCK/p54 and CFP-Ago2 pair, YFP-Ago1 and CFP-Ago2 showed an efficient FRET (19.61% ± 4.51%), indicating a direct interaction between Ago1 and Ago2 in vivo [28].
0018
To determine whether RCK/p54 is recruited into a functional RISC complex containing argonaute proteins or its association with Ago1/Ago2 is merely due to their co-localization in P-bodies, we affinity-purified active RISCs, analyzed their protein composition, and assayed for RISC function (Figure 2A).

0096
To determine whether RCK/p54 is recruited into a functional RISC complex containing argonaute proteins or its association with Ago1/Ago2 is merely due to their co-localization in P-bodies, we affinity-purified active RISCs, analyzed their protein composition, and assayed for RISC function (Figure 2A).
0007
To probe the involvement of P-body proteins in this purified active RISC, its protein composition was analyzed by immunoblot using antibodies against Flag tag or endogenous Ago2, RCK/p54, eIF4E, and Lsm1.

0018
To probe the involvement of P-body proteins in this purified active RISC, its protein composition was analyzed by immunoblot using antibodies against Flag tag or endogenous Ago2, RCK/p54, eIF4E, and Lsm1.

0104
The data also show that Lsm1 was not a RISC component, consistent with findings shown in Figure 1A.
0676
To determine the functional interactions of RCK/p54 with miRISC, we employed affinity purification of RISC and target mRNA cleavage capabilities of miRISC when the target has perfectly complementary sequences to the miRNAs.
0413
Incubation with let-7 inhibitors (with or without 3′-biotin) blocked the cleavage activity of let-7 miRISC in supernatant or beads (Figure 3B, lanes 2–3, and lanes 5–6). These results are consistent with previous reports [54,55] that adding 2′-O-Me oligonucleotides complementary to let-7 abolishes target cleavage activity by let-7 in cell extracts, indicating complete hybridization of the 2′-O-Me probe.
0007
To understand the function and role of RCK/p54 in the RNAi pathway, RCK/p54 was depleted in P-bodies of HeLa cells by siRNA-mediated RNAi. 24 h after transfecting cells with siRNA, real-time quantitative PCR showed that mRNA levels decreased by more than 90% and immunoblot analysis showed that RCK/p54 protein levels decreased significantly without affecting the levels of other P-body proteins including Lsm1 and Ago2 (Figure S2).

0018
24 h after transfecting cells with siRNA, real-time quantitative PCR showed that mRNA levels decreased by more than 90% and immunoblot analysis showed that RCK/p54 protein levels decreased significantly without affecting the levels of other P-body proteins including Lsm1 and Ago2 (Figure S2).

0096
24 h after transfecting cells with siRNA, real-time quantitative PCR showed that mRNA levels decreased by more than 90% and immunoblot analysis showed that RCK/p54 protein levels decreased significantly without affecting the levels of other P-body proteins including Lsm1 and Ago2 (Figure S2).
0411
HeLa cells were transfected with siRNA against Lsm1, harvested at 24, 48, and 72 h post-transfection, and TCEs were analyzed by immunoblot.

0676
We next visualized Lsm1 and Ago2 by immunofluorescence using antibodies against Lsm1 and Myc tag and found that Lsm1 and Ago2 co-localized in P-bodies (Figure 5B).
0006
HeLa cells were transfected for 48 h with Myc-Ago2 and control siRNA or siRNA against Lsm1, TCEs were prepared, and Myc-Ago2 was immunoprecipitated from an aliquot of TCE.

0423
Immunoprecipitation using Myc antibodies showed that RCK/p54 interacted with Ago2 and this interaction was not significantly changed by depleting Lsm1 (Figure 5C).
0411
P-bodies were disrupted in HeLa cells by transfecting them with siRNA against Lsm1, and cytoplasmic extracts were prepared 48 h post-transfection.

0019
After immunoprecipitation, RISC activities were analyzed by incubating the supernatant (S) or bead (B) phases with 182-nt 32P-cap-labeled let-7 substrate mRNAs having a perfectly complementary or mismatched sequence to the let-7 miRNA.

0676
Control experiments using IgG (Figure 5D, lanes 1 and 2) did not show any miRISC activity purified by magnetic beads, supporting specific capture by Ago2 and RCK/p54 antibodies (Figure 5D, lanes 4 and 6).

0006
Despite the significant loss of P-body structures in cells treated with Lsm1 siRNA, the efficiency of let-7 target mRNA cleavage by miRISC purified by anti-RCK/p54 antibody was not significantly affected (compare Figure 3A, lane 9 with Figure 5D, lane 6).
0007
To examine whether the P-body protein RCK/p54, also a component of RISC, is involved in siRNA-mediated gene silencing, the siRNA dose dependence of RNAi-mediated gene silencing was quantified in RCK/p54-depleted HeLa cells using a dual fluorescence reporter assay.

0055
To examine whether the P-body protein RCK/p54, also a component of RISC, is involved in siRNA-mediated gene silencing, the siRNA dose dependence of RNAi-mediated gene silencing was quantified in RCK/p54-depleted HeLa cells using a dual fluorescence reporter assay. Briefly, GFP and red fluorescent protein (RFP) were constitutively expressed in cells transfected with reporter plasmids for enhanced GFP (EGFP) and RFP, respectively. GFP expression was silenced by treating cells with a 21-nt siRNA targeting nt 238–258 of the EGFP mRNA. The fluorescence intensity ratio of target (GFP) to control (RFP) fluorophore was determined in the presence of siRNA duplexes and normalized to that observed in control-treated cells [52].

0416
To examine whether the P-body protein RCK/p54, also a component of RISC, is involved in siRNA-mediated gene silencing, the siRNA dose dependence of RNAi-mediated gene silencing was quantified in RCK/p54-depleted HeLa cells using a dual fluorescence reporter assay.

0411
Briefly, GFP and red fluorescent protein (RFP) were constitutively expressed in cells transfected with reporter plasmids for enhanced GFP (EGFP) and RFP, respectively. GFP expression was silenced by treating cells with a 21-nt siRNA targeting nt 238–258 of the EGFP mRNA.

0663
Briefly, GFP and red fluorescent protein (RFP) were constitutively expressed in cells transfected with reporter plasmids for enhanced GFP (EGFP) and RFP, respectively.

0663
The fluorescence intensity ratio of target (GFP) to control (RFP) fluorophore was determined in the presence of siRNA duplexes and normalized to that observed in control-treated cells [52].

0416
The fluorescence intensity ratio of target (GFP) to control (RFP) fluorophore was determined in the presence of siRNA duplexes and normalized to that observed in control-treated cells [52].
0411
When HeLa cells were transfected with siRNA mismatched for CDK9 (control), the second transfection with GFP siRNA silenced GFP expression in a dose-dependent manner, i.e., GFP/RFP ratios decreased with increasing concentrations of siRNA (Figure 6A, gray bars).

0055
When HeLa cells were transfected with siRNA mismatched for CDK9 (control), the second transfection with GFP siRNA silenced GFP expression in a dose-dependent manner, i.e., GFP/RFP ratios decreased with increasing concentrations of siRNA (Figure 6A, gray bars). Depleting RCK/p54 before quantifying GFP silencing resulted in moderate to no significant loss of RNAi activity, i.e., GFP/RFP again decreased with increasing concentrations of siRNA (Figure 6A, green bars). Similarly, depleting Lsm1 (Figure 6A, red bars) and the decapping enzyme, Dcp2 (unpublished data), before quantifying GFP silencing in HeLa cells did not significantly decrease RNAi activity.

0413
Similarly, depleting Lsm1 (Figure 6A, red bars) and the decapping enzyme, Dcp2 (unpublished data), before quantifying GFP silencing in HeLa cells did not significantly decrease RNAi activity. In contrast, prior depletion of Ago2 completely abolished RNAi activity (Figure 6A, blue bars), as expected and consistent with previous reports [14–16].
0413
To examine the role of RCK/p54 in RISC catalysis of mRNA processing, in vitro mRNA cleavage activity was assayed in extracts from RCK/p54-depleted HeLa cells.

0055
Cells were transfected with siRNAs against CDK9 mismatch (control), RCK/p54, Lsm1, or Ago2 and programmed 24 h later with GFP siRNA.

0413
Varying amounts of cytoplasmic extract (20–100 μg) were assayed in vitro for target mRNA cleavage activity. As shown in Figure 6B and 6C, cleavage activity increased with increasing protein concentration of extracts from both control and RCK/p54-depleted cells, reaching a robust cleavage (37%–41%) of target mRNA at the highest protein concentration (100 μg).

0081
As shown in Figure 6B and 6C, cleavage activity increased with increasing protein concentration of extracts from both control and RCK/p54-depleted cells, reaching a robust cleavage (37%–41%) of target mRNA at the highest protein concentration (100 μg). In consistent with in vivo results shown in Figure 6A, Ago2 depletion abolished the target mRNA cleavage activities of RISC.
0411
To test this hypothesis, we first determined whether RCK/p54 was a general translation repressor in human cells. To this end, RCK/p54 expression was silenced in HeLa cells and general translational activity was analyzed by [35S]methionine incorporation (Figure 7A).

0413
To test this hypothesis, we first determined whether RCK/p54 was a general translation repressor in human cells. To this end, RCK/p54 expression was silenced in HeLa cells and general translational activity was analyzed by [35S]methionine incorporation (Figure 7A). The translational rates measured by [35S]-labeling were significantly greater in RCK/p54-depleted cells than in control siRNA-treated cells (Figure 7A), indicating that RCK/p54 was a translational repressor in human cells, consistent with results in yeast and reticulocyte extracts [45].

0081
The translational rates measured by [35S]-labeling were significantly greater in RCK/p54-depleted cells than in control siRNA-treated cells (Figure 7A), indicating that RCK/p54 was a translational repressor in human cells, consistent with results in yeast and reticulocyte extracts [45].
0007
HeLa cells were co-transfected with siRNAs directed against P-body proteins (RCK/p54, GW182, Lsm1, and Ago2) and with siRNA or miRNA reporters in the absence or presence of 25 nM CXCR4 siRNA.

0411
At 24 h post-transfection, cells were harvested and RL activities were analyzed. RL signals were normalized to Photinus pyralis luciferase (FL) signals from cells co-transfected with pGL3 plasmid as control.
0411
We next hypothesized that the expression of a specific cellular protein, known to be controlled by miRNAs, might be up-regulated in RCK/p54-depleted cells.

0007
Total cellular protein extracts from equal numbers of cells were prepared and analyzed by immunoblot using monoclonal anti-RAS antibodies directly conjugated with HRP (see Materials and Methods). Actin protein levels were measured as a control using HRP-conjugated anti-actin primary antibody. When cells were treated with let-7 inhibitor, RAS protein levels increased, consistent with previous findings [13].

0411
Strikingly, RAS protein expression in RCK/p54-depleted cells also showed a robust increase (Figure 7C) while RAS levels were not significantly affected when cells were treated with control siRNA or 2′-O-Me oligonucleotides. RAS mRNA levels did not differ drastically in HeLa cells after treatment with let-7 inhibitor, RCK/p54 silencing, or mock treatment (Figure S3).
0413
To further probe our findings of endogenous RAS regulation by RCK/p54, we co-transfected HeLa cells with RAS 3′ UTR reporter constructs and let-7 2′ -O-Me inhibitor or siRNA against RCK/p54. For control experiments, we treated cells with a CDK9 mm siRNA and a 2′ -O-Me oligonucleotide complementary to HIV-1 TAR RNA sequence [53].

0007
To further probe our findings of endogenous RAS regulation by RCK/p54, we co-transfected HeLa cells with RAS 3′ UTR reporter constructs and let-7 2′ -O-Me inhibitor or siRNA against RCK/p54.

0411
Cells transfected with a control siRNA and let-7 inhibitor induced more firefly luciferase expression when reporter plasmids contained 3′ UTR sequences for NRAS and KRAS than for the control siRNA and a 2′-O-Me oligo control (Figure 7D), consistent with a recent report [13].
0018
These studies also demonstrated that human Ago1 and Ago2 co-localize in P-bodies with other cellular proteins, such as Dcp1a, Dcp2, GW182, Lsm1, and Xrn1 [28–31]. A homolog of the P-body protein GW182 in Caenorhabditis elegans is the developmental timing regulator AIN-1, which also interacts with miRISCs and may target argonaute proteins to P-bodies [57]. To dissect and understand the relationship between RNAi function and P-bodies, we affinity-purified RISC using Myc-Ago2 and expression vectors of the YFP-tagged P-body proteins, Lsm1, RCK/p54, Dcp2, and eIF4E.
0107
Where in the cell does this physical interaction between RISC and RCK/p54 take place and does it require P-body structures? To address these questions, we disrupted P-bodies in cells by depleting Lsm1 and immunopurified endogenous miRISC, and analyzed its ability to cleave a target mRNA with perfect complementarity to let-7 miRNA (Figure 5).

0413
Where in the cell does this physical interaction between RISC and RCK/p54 take place and does it require P-body structures? To address these questions, we disrupted P-bodies in cells by depleting Lsm1 and immunopurified endogenous miRISC, and analyzed its ability to cleave a target mRNA with perfect complementarity to let-7 miRNA (Figure 5). We observed no significant difference in miRISC activities purified from Lsm1-depleted and non-depleted cells (Figures 3 and 5) even though the P-body structures were disrupted and the total number of P-bodies per cell had significantly decreased.
0413
To test our hypothesis that RCK/p54 mediates the translational repression of endogenous miRNA targets in P-bodies, we examined RAS protein levels in RCK/p54-depleted cells. We chose RAS because it is an endogenous target of let-7, and 3′ UTRs of human RAS genes contain multiple complementary sites for let-7 to bind and regulate RAS expression levels [13].

0411
Furthermore, let-7 inhibitors are known to enhance RAS protein expression in HeLa cells [13].

0081
Depleting RCK/p54 in HeLa cells up-regulated RAS protein, and this increase in RAS levels was higher than that in general translation of control actin (Figure 7C), suggesting that multiple sites of let-7 miRISC binding to target 3′ UTR dictate the potency and specificity of translation suppression.

0413
Depleting RCK/p54 in HeLa cells up-regulated RAS protein, and this increase in RAS levels was higher than that in general translation of control actin (Figure 7C), suggesting that multiple sites of let-7 miRISC binding to target 3′ UTR dictate the potency and specificity of translation suppression.

0096
Depleting RCK/p54 in HeLa cells up-regulated RAS protein, and this increase in RAS levels was higher than that in general translation of control actin (Figure 7C), suggesting that multiple sites of let-7 miRISC binding to target 3′ UTR dictate the potency and specificity of translation suppression.
0114
Translation repression by miRISC and P-body localization require the 5′-cap structure in the target mRNA [29,58], which provides a unique and elegant control mechanism for translation and its regulation (reviewed in [59,60]).

0030
Recently, Petersen et al.

0889
Recently, Petersen et al.
0096
Once in P-bodies, translationally repressed mRNA could stay in oligomeric structures for storage or it could form a complex with decapping enzymes and cap binding proteins that would lead to the mRNA decay pathway.
0004
TCEs were prepared by incubating cells in lysis buffer (20 mM HEPES [pH 7.9], 10 mM NaCl, 1 mM MgCl2, 0.2 mM EDTA, 0.35% [v/v] Triton X-100, 1/100 [v/v] dilution in protease inhibitor cocktail) and centrifuging at 14,000 rpm for 10 min at 4 °C. Protein concentration was determined by Dc protein assay (Bio-Rad, Hercules, California, United States).

0019
To examine the RNA dependence of protein–protein interactions, TCEs (250 μg) were treated before immunoprecipitation with 0.2 μg/ul of RNase A for 20 min at room temperature.
0055
HeLa cells were cultured in 35 mm dishes with glass coverslip bottoms (MatTek Corporation, Ashland, Massachusetts, United States). Expression vectors for CFP- or YFP-tagged proteins were transfected into cells using Lipofectamine as described above. 24 h later, the live cells were monitored for CFP and YFP signals of the transiently expressed proteins.

0411
HeLa cells were cultured in 35 mm dishes with glass coverslip bottoms (MatTek Corporation, Ashland, Massachusetts, United States). Expression vectors for CFP- or YFP-tagged proteins were transfected into cells using Lipofectamine as described above.

0007
HeLa cells were cultured in 35 mm dishes with glass coverslip bottoms (MatTek Corporation, Ashland, Massachusetts, United States). Expression vectors for CFP- or YFP-tagged proteins were transfected into cells using Lipofectamine as described above.

0663
The signals were detected by a Leica (Wetzlar, Germany) confocal imaging spectrophotometer system (TCS-SP2) attached to a Leica DMIRE inverted fluorescence microscope equipped with an argon laser, two HeNe lasers, an acousto-optic tunable filter (AOTF) to attenuate individual visible laser lines, and a tunable acousto-optical beam splitter (AOBS).

0055
For FRET studies, HeLa cells co-expressing CFP- and YFP-tagged proteins were fixed with 4% paraformaldehyde in PBS at room temperature for 20 min and washed three times with PBS. After the final wash, cells were visualized with a Leica confocal imaging system as described above. FRET experiments were performed by an acceptor photobleaching method as described [48,49,68]. FRET efficiencies were measured and images were analyzed using Leica confocal software.

0007
For FRET studies, HeLa cells co-expressing CFP- and YFP-tagged proteins were fixed with 4% paraformaldehyde in PBS at room temperature for 20 min and washed three times with PBS.

0416
For immunofluorescence studies, cells transfected with Myc-Ago2 were fixed with 4% paraformaldehyde in PBS at room temperature for 20 min and permeabilized with 0.25% (v/v) Triton X-100 for 5 min.

0040
Samples were washed three times with PBST (0.1% [v/v] Triton X-100 in PBS) and blocked for 30 min in PBST containing 2% (w/v) BSA. Primary and secondary antibodies were diluted in blocking solution during incubation.
0663
A dual fluorescence assay was used to quantify the RNAi activity of siRNAs against GFP.

0006
To quantify RNAi effects, cell lysates were prepared from siRNA-treated cells 24 h post-transfection.

0411
To quantify RNAi effects, cell lysates were prepared from siRNA-treated cells 24 h post-transfection.

0663
GFP fluorescence was detected in cell lysates by exciting at 488 nm and recording emissions from 504–514 nm. The spectrum peak at 509 nm represents the fluorescence intensity of GFP. RFP fluorescence was detected in the same cell lysates by exciting at 558 nm and recording emissions from 578–588 nm. The spectrum peak at 583 nm represents the fluorescence intensity of RFP. The fluorescence intensity ratio of target (GFP) to control (RFP) fluorophores was determined in the presence of siRNA duplexes and normalized to the emissions measured in mock-treated cells.

0055
The fluorescence intensity ratio of target (GFP) to control (RFP) fluorophores was determined in the presence of siRNA duplexes and normalized to the emissions measured in mock-treated cells.

0416
The fluorescence intensity ratio of target (GFP) to control (RFP) fluorophores was determined in the presence of siRNA duplexes and normalized to the emissions measured in mock-treated cells.
0007
HeLa cells cultured in 6-well plates were co-transfected with 0.8 μg/well Pp-luc-expressing vectors (pGL3-control, pGL3-NRAS, or pGL3-KRAS) and with 100 nM let-7-2′-O-Me inhibitor or 50 nM siRNA against RCK/p54. In all experiments, transfection efficiencies were normalized to those of cells co-transfected with the Rr-luc-expressing vector (pRL-TK; 0.1 μg/well).
0007
HeLa cells transfected with siRNAs directed against P-body proteins (RCK/p54, GW182, Lsm1, Ago2) were co-transfected again with siRNA or miRNA reporters in the absence or presence of 25 nM CXCR4 siRNA.

0411
At 24 h post-transfection, cells were harvested and RL activities were analyzed. RL signals were normalized to the FL signals from cells co-transfected with pGL3 plasmid as control.
0055
TCEs from HeLa cells co-expressing Myc-Ago2 and YFP-Ago1, YFP-Dcp2, YFP-RCK/p54, YFP-eIF4E, YFP-Lsm1, or YFP were treated with +/− RNase A followed by Myc-Ago2 immunoprecipitation.
0411
HeLa cells expressing YFP-RCK/p54 and CFP-Ago2 were fixed at 24 h post-transfection.
0055
The guide strands of siRNA complexes targeting GFP (si-GFP) were conjugated with 3′ biotin (si-GFP-Bi; blue strands) and transfected into HeLa cells.
0007
(C) Biotin-captured RISC contains proteins associated with mRNA processing. Active human RISC from HeLa cells expressing Flag-Ago1 was captured by biotin-siRNA and its protein composition was analyzed by immunoblot using anti-Flag, anti-Ago2, anti-RCK/p54, anti-Lsm1, and anti-eIF4E antibodies.
0019
After immunoprecipitation, RISC activities were analyzed by incubating the supernatant (S) or bead (B) phases with 182-nt 32P-cap-labeled let-7 substrate mRNAs having a perfectly complementary or mismatched sequence to the let-7 miRNA.
0416
HeLa cells were transfected with siRNA against Lsm1. At 48 h post-transfection, cells were analyzed by immunofluorescence using antibodies against Lsm1 and Myc tag for Ago2. Cells were stained with Hoechst33258 to visualize nuclei, and images were digitally merged.

0411
HeLa cells were transfected with siRNA against Lsm1. At 48 h post-transfection, cells were analyzed by immunofluorescence using antibodies against Lsm1 and Myc tag for Ago2. Cells were stained with Hoechst33258 to visualize nuclei, and images were digitally merged.
0006
(C) RCK/p54 interacts with Myc-Ago2 in Lsm1-depleted cells. HeLa cells were transfected for 48 h with Myc-Ago2 and control siRNA or siRNA against Lsm1, TCEs were prepared, and Myc-Ago2 was immunoprecipitated from an aliquot of TCE.
0019
After immunoprecipitation, RISC activities were analyzed by incubating the supernatant (S) or bead (B) phases with 182-nt 32P-cap-labeled let-7 substrate mRNAs having a perfectly matched or a mismatched sequence to the let-7 miRNA.
0411
HeLa cells were transfected with siRNAs against CDK9 mismatch (control), RCK/p54, Lsm1, or Ago2. 24 h later cells were transfected again with EGFP and RFP reporter plasmids and varying amounts (2, 10, and 50 nM) of siRNA against EGFP.

0663
24 h later cells were transfected again with EGFP and RFP reporter plasmids and varying amounts (2, 10, and 50 nM) of siRNA against EGFP.

0663
To quantify the effect of depleting RCK/p54 and Ago2 on RNAi, the ratio of GFP/RFP signals was normalized to that observed in the absence of GFP siRNA (0 nM).
0413
(C) Quantification of siRISC cleavage activity in vitro after depletion of RCK/p54, Lsm1, or Ago2. Cleavage activity of siRISC targeting EGFP mRNA was quantified as a function of protein content in extracts of HeLa cells depleted of RCK/p54, Lsm1, or Ago2.
0426
Incorporation of [35S]methionine into HeLa cells was used to measure general translational activity. Cells were transfected with 50 nM siRNAs targeting mismatched CDK9 (control) or RCK/p54.

0413
Incorporation of [35S]methionine into HeLa cells was used to measure general translational activity. Cells were transfected with 50 nM siRNAs targeting mismatched CDK9 (control) or RCK/p54.

0411
Mock control cells were treated with the transfection reagent only. At 24 h post-transfection, cells were incubated for 1 h in medium lacking Met and Cys, and metabolically labeled with [35S]methionine (see Materials and Methods). As a control for passive uptake of [35S], mock cells were treated with 40 μg/ml of the translation inhibitor, cycloheximide.
0081
24 h later, cells were co-transfected with siRNA (1 × perfectly matched [PM] site) or miRNA (4 × bulged sites) luciferase reporters in the presence of CXCR4 siRNA.
0411
Sorting of ephrin-B1–positive and –negative cells following X-inactivation has been observed in ephrin-B1+/− mice; however, the mechanisms by which mosaic ephrin-B1 expression leads to cell sorting and phenotypic defects remain unknown. Here we show that ephrin-B1+/− mice exhibit calvarial defects, a phenotype autonomous to neural crest cells that correlates with cell sorting.
0411
As a result of random X-inactivation, X-linked ephrin-B1 expression is mosaic in ephrin-B1+/− mice and ephrin-B1–positive and ephrin-B1–negative cells segregate from one another.
0411
Indeed, studies in zebrafish have shown that expression of Eph receptors and ephrins in animal cap cells was sufficient to block GJC at the boundary between both cell populations.
0018
We chose to focus on ephrin-B2 since loss of this gene has been shown previously to affect migration of a sub-population of cranial NCCs [19] and ephrin-B2 is expressed in the craniofacial area (Figure 2) We generated a mouse line harboring a null mutation in the ephrin-B2 gene by placing a cDNA coding for a fusion protein between histone 2B (H2B) and the green fluorescent protein (GFP) under the control of the ephrin-B2 promoter (A.
0107
It has been reported previously that patchy expression of ephrin-B1 in ephrin-B1+/− limb buds reflects sorting between ephrin-B1–positive and –negative cell populations that are generated in the ephrin-B1 heterozygous females via random X-inactivation and that this abnormal expression of ephrin-B1 in the limb bud correlates with a polydactyly phenotype that is observed in ephrin-B1+/− females [7,8].

0411
Our data demonstrate that the calvarial phenotypes observed in ephrin-B1 heterozygous females correlate with an abnormal expression of ephrin-B1 and EphB2 in the presumptive frontal bone, likely due to cell sorting between ephrin-B1–positive and ephrin-B1–negative cells in the craniofacial mesenchyme.
0413
To confirm the differentiation defects observed in ephrin-B1 heterozygotes, we isolated presumptive osteogenic mesenchymal cells from E14.5 wild-type and ephrin-B1+/− embryos, and evaluated their ability to differentiate in vitro. Using AP activity as a marker for osteogenic differentiation, we observed that cells isolated from ephrin-B1+/− embryos were consistently less prone to differentiate in vitro (Figure 3C), indicating that these defects are autonomous to the osteoprogenitor cells. In these primary cultures, expression of ephrin-B1 was detected in a punctate pattern in both AP-positive as well as AP-negative cells (Figure S3A), indicating that expression of ephrin-B1 and AP do not strictly correlate, and suggesting that ephrin-B1 does not regulate AP activity directly.

0411
In these primary cultures, expression of ephrin-B1 was detected in a punctate pattern in both AP-positive as well as AP-negative cells (Figure S3A), indicating that expression of ephrin-B1 and AP do not strictly correlate, and suggesting that ephrin-B1 does not regulate AP activity directly.
0040
Paraffin sections of X-gal–stained E11.5 chimeric embryos were processed for immunofluorescence using the Cx43 antibody.
0030
Compagni et al.

0889
Compagni et al.

0107
We found that interaction between ephrin-B1 and Eph-B2 resulted in inhibition of GJC in vitro (Figure 5).

0107
These results demonstrate that interaction between EphB2 and ephrin-B1 impairs establishment of GJC.
0411
Expression of ephrin-B1 and Cx43 was partially overlapping in untreated NIH 3T3 cells, especially at interfaces between cells expressing ephrin-B1, which exhibited strong Cx43 staining (Figure 6Ba–c).

0413
To test whether ephrin-B1 and Cx43 physically interact, we performed a pull-down assay in NIH 3T3 cells expressing ephrin-B1.
0096
We generated a mutant form of ephrin-B1 that shows reduced binding to PDZ-containing proteins (ephrin-B1ΔPDZ [8]).

0081
Western-blot analysis of whole cell lysates indicated that both proteins were expressed, albeit at different levels, and that expression of ephrin-B1 did not influence the phosphorylation status of Cx43 (detected by differences in mobility on a SDS-PAGE), which is known to regulate GJC (Figure 7A).

0006
Cx43 could be detected in the pull downs from cells expressing either ephrin-B1 wild type or ephrin-B1ΔPDZ, however, the relative abundance of phosphorylated versus unphosphorylated band was changed.

0412
Cx43 could be detected in the pull downs from cells expressing either ephrin-B1 wild type or ephrin-B1ΔPDZ, however, the relative abundance of phosphorylated versus unphosphorylated band was changed.

0081
More phosphorylated Cx43 (slower mobility) was observed in the ephrin-B1 wild-type pull down, whereas more unphosphorylated Cx43 was detected in the ephrin-B1ΔPDZ pull down (Figure 7A).
0413
However, subtle but consistent defects in sternum development, a phenotype that is associated with complete loss of ephrin-B1 [7,8], were observed in almost all of the chimeras (Figure S4), indicating that the lack of calvarial and polydactyly phenotypes in the chimeric embryos is not due to an ineffective contribution of ephrin-B1ΔPDZ ES cells to bone. To test the effect of the ephrin-B1ΔPDZ mutation on the distribution of Cx43 in vivo, we generated chimeric embryos by injecting mutant ES cells carrying the ephrin-B1ΔPDZ allele (or ephrin-B1 null ES cells as a control) in wild-type ROSA 26 blastocysts. Unexpectedly, paraffin sections of X-gal–stained E11.5 chimeric embryos revealed that unlike ephrin-B1 null cells, cells expressing ephrin-B1ΔPDZ do not sort-out from wild-type cells (Figure 7C).
0413
The inability of ephrin-B1ΔPDZ to co-localize with Cx43 upon engagement and to drive cell sorting suggested that regulation of GJC itself may play a role in the sorting-out process between ephrin-B1–positive and ephrin-B1–negative cells. To test this hypothesis, we transiently transfected HEK293T cells (that have very low levels of endogenous Cx43) with Cx43 and allowed them to sort out following trypsinization.

0081
Unlike control transfected cells, Cx43-overexpressing cells segregated from untransfected cells and were consistently found in clusters (Figure 7Da), indicating that the establishment of GJC does indeed promote cell sorting.
0107
Because the lack of cell sorting in the experiments described above precluded the establishment of a functional link between the calvarial phenotype and the regulation of GJC in ephrin-B1 heterozygote embryos, we performed a genetic rescue experiment.
0030
Moreover, heterozygosity for the transcription factor Twist can lead to craniosynostosis, as well as to foramina, and both transcription factors regulate differentiation of frontal bone osteoprogenitors (Ishii et al., 2003).

0889
Moreover, heterozygosity for the transcription factor Twist can lead to craniosynostosis, as well as to foramina, and both transcription factors regulate differentiation of frontal bone osteoprogenitors (Ishii et al., 2003).

0107
The genetic interaction suggests, however, that the discrepancy in coronal suture phenotype between ephrin-B1 heterozygous mice and humans could be due to genetic modifiers.
0096
Our data using chimeric embryos demonstrate that a mutation in the PDZ binding domain of ephrin-B1, which is known to phenocopy some of the phenotypes observed in ephrin-B1 null embryos including cleft palate [8], does not induce calvarial defects or polydactyly. These results indicate that the mutations found in CFNS patients might impinge on protein stability, protein localization, or binding of effector molecules independent of the PDZ domain.
0107
In our pull-down assay, wild-type ephrin-B1 interacted preferentially with phosphorylated Cx43 whereas ephrin-B1ΔPDZ interacted preferentially with unphosphorylated Cx43, suggesting that the interaction between ephrin-B1 and Cx43 might not be direct, and that these proteins might interact differently when at the cell surface or in the cytoplasm.

0423
In our pull-down assay, wild-type ephrin-B1 interacted preferentially with phosphorylated Cx43 whereas ephrin-B1ΔPDZ interacted preferentially with unphosphorylated Cx43, suggesting that the interaction between ephrin-B1 and Cx43 might not be direct, and that these proteins might interact differently when at the cell surface or in the cytoplasm.
0052
Trypsin was inactivated by addition of complete medium (DMEM containing 15% FCS) and cells were dissociated by trituration with a glass Pasteur pipette.

0413
After 3 d, cells were fixed in 2% PFA and rinsed three times in NTMT. AP activity was detected by incubating fixed cells with NBT/BCIP (Roche, Basel, Switzerland).
0004
Cells were scraped in 1% NP40 lysis buffer (50 mM Hepes [pH 7.5], 150 mM NaCl, 10% glycerol, 1.5 mM MgCl2, 1mM EGTA, 100 mM NaF), except for the experiment presented in Figure 7A (100 mM Tris [pH 7.4], 150 mM NaCl, 1mM EDTA).

0006
Protein lysates were incubated with either 5-μg EphB2-Fc or 4-μg Cx43 monoclonal antibody (generous gift from P.

0676
Affinity complexes were analyzed by SDS-PAGE using the following antibodies: ephrin-B1 (A20 or C18, Santa Cruz Biotechnology, Santa Cruz, California, United States), Cx43 (Sigma, St.

0096
We have noted that the binding of Cx43 to ephrin-B1 is sensitive to the lysis buffer used: the interaction is lost in RIPA buffer.
0411
NIH 3T3 cells were plated on glass coverslips and transiently transfected with an expression vector for ephrin-B1. Forty hours after transfection, cells were either fixed in 2% PFA or incubated with 4-μg/ml EphB2-Fc for 30 min at 37 °C and then fixed in PFA.

0006
Cells were permeabilized with 0.1% Tx-100 for 3 min, rinsed in PBS, and incubated with a mix of EphB2-Fc (4 μg/ml) and either Cx43 monoclonal antibody or N-cadherin monoclonal antibody (Zymed, Carlsbad, California, United States). In primary cells, ephrin-B1 was detected using the 25H11 rat monoclonal antibody [47].

0007
In primary cells, ephrin-B1 was detected using the 25H11 rat monoclonal antibody [47].

0055
For the sorting experiments, HEK293T cells were transiently transfected with either DsRed or a Cx43 expression construct. Twenty-four hours after transfection, cells were trypsinized into a single-cell suspension and replated onto glass coverslips.

0411
For the sorting experiments, HEK293T cells were transiently transfected with either DsRed or a Cx43 expression construct. Twenty-four hours after transfection, cells were trypsinized into a single-cell suspension and replated onto glass coverslips.

0007
Immunofluorescence was performed at 48 h post-transfection using a Cx43 antibody (Sigma).
0040
For immunofluorescence on sections, paraffin sections were rehydrated and subjected to a citrate boil to reveal antigens. Tissue sections were blocked in Blocking solution (PBS/5% horse serum) 1 h at room temperature and incubated overnight at 4 °C in Cx43 antibody (1/100 in Blocking solution [Sigma]). Incubation with the Cy3-conjugated secondary antibody was for 1 h at room temperature (1/250 in Blocking solution).
0413
(A) Primary mesenchymal cells isolated from presumptive calvaria of wild-type (a) and (b) or ephrin-B1+/− embryos (c) and (d) were stained for AP activity (a) and (c) and subsequently processed for Cx43 immunofluorescence (b) and (d).

0416
(A) Primary mesenchymal cells isolated from presumptive calvaria of wild-type (a) and (b) or ephrin-B1+/− embryos (c) and (d) were stained for AP activity (a) and (c) and subsequently processed for Cx43 immunofluorescence (b) and (d). Junctional Cx43 evidenced by bright dots is readily detected in cultures from wild-type embryos (WT), but not in cultures from ephrin-B1+/− embryos.
0416
(B) Detection of Cx43 by immunofluorescence in limb bud sections of an X-gal–stained chimeric embryo obtained by injecting ephrin-B1 null cells into a ROSA26-βgal blastocyst.
0411
(B) Primary NCCs were loaded with Calcein-AM and dropped onto NIH 3T3 cells that were transfected either with a control plasmid (pcDNA3 [a]), or an expression construct for ephrin-B1 (b) or Eph-B2 (c). In the control situation, the majority of cells transferred the dye (arrowheads).
0676
(C) Ephrin-B1 was affinity precipitated from NIH 3T3 cells using EphB2-Fc. The presence of Cx43 in the affinity complex was assessed by Western-blot (a).

0006
Protein lysates from NIH 3T3 cells expressing ephrin-B1 were incubated with ephrinB1-Fc, and the presence of Cx43 in the affinity complexes was assessed by Western blot (b) Cx43 was immunoprecipitated from NIH 3T3 cells expressing ephrin-B1 and the presence of ephrin-B1 in the immunocomplexes was detected by Western blot (right panel).
0416
(B) NIH 3T3 cells were transiently transfected with ephrin-B1ΔPDZ and treated with EphB2-Fc. Subcellular localization of Cx43 (a) and ephrin-B1ΔPDZ (b) was analyzed by immunofluorescence.
0107
At a developmental boundary and at ectopic ephrin boundaries in ephrin-B1+/− embryos, Eph/ephrin interaction leads to inhibition of GJC, possibly through endocytosis of Cx43, concomitant with a loss of stable cell–cell interactions between the two cell types. Sorting between these cells and inhibition of GJC concur to establish distinct developmental compartments.
0018
Two proteins that are believed to be required for assembly of pre-RCs in fission yeast are Cdc18 (Cdc6) and Cdt1 (Cdt1) [14-18].
0423
Also, biochemical analysis of purified Cdc23 (Mcm10) protein demonstrates that it is required for efficient phosphorylation of the Mcm2-7 complex by Dfp1-Hsk1 (Dbf4-Cdc7) kinase in vitro, and that Cdc23 (Mcm10) can directly interact with Dfp1-Hsk1 (Dbf4-Cdc7) [37].
0018
Ars binding protein 1 (Abp1) was first identified in a search for fission yeast proteins that could retard an ARS (Autonomously Replicating Sequence)-containing DNA fragment in a gel-shift mobility assay [38]. Although Abp1 was shown to be non-essential [39], the protein could bind very tightly to ARS elements in vitro.

0096
Ars binding protein 1 (Abp1) was first identified in a search for fission yeast proteins that could retard an ARS (Autonomously Replicating Sequence)-containing DNA fragment in a gel-shift mobility assay [38].
0107
The two-hybrid interaction between Snf1 and Snf4 is shown as a positive control (Figure 1A, row 4).
0104
Consistent with the lacZ data, Cdc23 (Mcm10) fused to the DNA binding domain of Gal4 activated HIS3 under the control of the Gal4 promoter when co-expressed with Abp1 fused to the Gal4 activation domain (Figure 1B, row 1).

0081
Consistent with the lacZ data, Cdc23 (Mcm10) fused to the DNA binding domain of Gal4 activated HIS3 under the control of the Gal4 promoter when co-expressed with Abp1 fused to the Gal4 activation domain (Figure 1B, row 1).
0104
We found that the double mutant cdc23-M36 Δabp1 is less viable then either cdc23-M36 or Δabp1 when grown at 30°C (Figure 3A, lower panel, row 4), consistent with our two-hybrid data suggesting that Cdc23 (Mcm10) interacts directly with Abp1.

0107
Therefore the genetic interaction observed between cdc23-M36 and Δabp1 appears to be Abp1-specific.
0054
Cell cycle progression profiles of Δabp1, cdc23-M36 and cdc23-M36 Δabp1 at 30°C (from 1–6 hrs) and 25°C at time zero, by flow cytometry anaylsis.
0054
We analyzed DNA content by flow cytometry to determine the precise arrest points for the different strains (Figure 3B). As expected, wildtype or mutant cells grown at the permissive temperature of 25°C display a 2C DNA content, indicating that most cells when growing exponentially are in the G2 phase of the cell cycle.
0018
Yeast genetics has provided a powerful tool to identify some of the key proteins that are required for assembly of these complexes. One of these proteins, called Cdc23 (Mcm10), is conserved from yeast to man, and is required for the early events of DNA replication initiation.

0007
One of the proteins identified using this screen is Abp1, a protein that had previously been shown to bind to both ARS elements and centromeric DNA sequences.
0018
First, Abp1 not only interacts with Cdc23 (Mcm10) protein in the two-hybrid analysis, but deletion of Abp1 lowers the restrictive temperature for cdc23-M36, consistent with its proposed role in DNA replication. Although two other proteins show significant homology to Abp1, deletion of either of these does not alter the phenotype of the cdc23 temperature-sensitive mutant, suggesting that Cdc23 (Mcm10) specifically interacts with Abp1.
0054
Our flow cytometry analysis suggests that cdc23+ (MCM10) is required for DNA replication initiation. Previous studies also indicate that following a shift to the restrictive temperature of 36°C, cdc23 temperature-sensitive mutant cells arrest in early-S phase [32].
0054
In murine thymocytes and tumour cells expressing an oncogenic tyrosine kinase, this DNA damage–induced cascade is blocked. Enforced intracellular alkalinisation mimics the effects of DNA damage in murine tumour cells and human B-lineage chronic lymphocytic leukaemia cells, thereby causing Bcl-xL deamidation and increased apoptosis. Our results define a signalling pathway leading from DNA damage to up-regulation of the NHE-1 antiport, to intracellular alkalanisation to Bcl-xL deamidation, to apoptosis, representing the first example, to our knowledge, of how deamidation of internal asparagine residues can be regulated in a protein in vivo.

0426
In murine thymocytes and tumour cells expressing an oncogenic tyrosine kinase, this DNA damage–induced cascade is blocked. Enforced intracellular alkalinisation mimics the effects of DNA damage in murine tumour cells and human B-lineage chronic lymphocytic leukaemia cells, thereby causing Bcl-xL deamidation and increased apoptosis. Our results define a signalling pathway leading from DNA damage to up-regulation of the NHE-1 antiport, to intracellular alkalanisation to Bcl-xL deamidation, to apoptosis, representing the first example, to our knowledge, of how deamidation of internal asparagine residues can be regulated in a protein in vivo.
0018
The deamidation of internal asparaginyl and glutaminyl protein residues has attracted increasing attention over the past decade as a modification leading to significant changes in protein function [1,2]. The protein deamidation rates of more than 18,000 proteins have been computed, containing 230,000 individual asaparaginyl residues, generating Asn half-lives of less than 1 d to 50 y or more [3,4]. Protein deamidation has broad biological implications, ranging from changes in the specificity of antigen presentation [5], to modifications in eye lens proteins [6], to the activation of RhoA by cytotoxic necrotizing factor [7], to aging [1], to name but a few examples.
0018
Until recently, it was assumed that Asn protein deamidation rates in vivo were set up by a “fixed clock” that was defined only by the primary, secondary, and tertiary structures of proteins that specified the half-life of the particular Asn residue in question. However, this view has been radically changed by the recent observation that DNA damage induces the relatively rapid deamidation of the pro-survival protein Bcl-xL in an osteosarcoma cell line system [9], indicating that the deamidation “clock”, far from being fixed, is a dynamic process that can be regulated in vivo by biologically critical events.

0018
Initial work from the Weintraub laboratory suggested that when Asn52 and Asn66 are both mutated to Asp, then Bcl-xL loses its ability to bind to the BH3-only pro-apoptotic protein Bim, thereby providing a putative linkage between DNA damage and apoptosis [9].
0411
Surprisingly, DNA damage–triggered deamidation in primary wild-type cells is mediated not enzymatically, but by intracellular alkalinisation caused by increased expression of the NHE-1 Na+/H+ exchanger (antiport), events blocked by expression of the oncogenic tyrosine kinase (OTK).

0006
In the case of either murine or human cancer cells, enforced alkalinisation triggers Bcl-xL deamidation, crippling its ability to provide protection from the pro-apoptotic consequences of DNA damage, thereby indicating possible novel approaches to cancer therapy.
0081
We therefore used short hairpin RNA (shRNA) to deplete Bax and Bak from CD4−CD8− (double-negative, DN) thymocytes, confirmed that depletion was sufficient to block caspase 9 cleavage (Figure S1A), and showed that DNA damage–induced Bcl-xL deamidation proceeded normally in the absence of Bax and Bak (Figure 1C). We also showed that Bcl-xL deamidation was clearly detectable within 3–6 h after the instigation of DNA damage, and proceeded in parallel with increased apoptosis (Figure S1B and S1C).
0054
Because the BH3-only protein Puma, not Bim, plays a major role in DNA-damage triggered apoptosis [19,20], we also showed that both Puma and Bim are found in Bcl-xL immunoprecipitates from etoposide treated CD45−/−LckF505 thymocytes, whereas sequestration is ablated in wild-type cells, correlating with Bcl-xL deamidation (Figure 2B).

0423
Because the BH3-only protein Puma, not Bim, plays a major role in DNA-damage triggered apoptosis [19,20], we also showed that both Puma and Bim are found in Bcl-xL immunoprecipitates from etoposide treated CD45−/−LckF505 thymocytes, whereas sequestration is ablated in wild-type cells, correlating with Bcl-xL deamidation (Figure 2B).

0426
Because the BH3-only protein Puma, not Bim, plays a major role in DNA-damage triggered apoptosis [19,20], we also showed that both Puma and Bim are found in Bcl-xL immunoprecipitates from etoposide treated CD45−/−LckF505 thymocytes, whereas sequestration is ablated in wild-type cells, correlating with Bcl-xL deamidation (Figure 2B).
0096
Recombinant purified His-tagged Bcl-xL was exposed to alkaline conditions to cause partial deamidation and separated by anion-exchange chromatography into three peaks (Figure 2C, peaks A, B and C).
0081
Peak B represents rBcl-xL deamidated at either Asn52 or Asn66, whereas peak C is deamidated at both sites (Figure 2C and 2D).
0004
Until now, the in vivo mechanism for the deamidation of internal protein Asn residues has not been described for any protein. Because protein Asn deamidation is accelerated by increased pH in vitro, we investigated intracellular pH change (pHi) as a possible regulatory mechanism in thymocytes.

0054
To investigate Bcl-xL deamidation, pHi, and apoptosis in parallel, we manipulated pHi values artificially by incubating cells at varying pHe values in the absence of monensin.

0426
To investigate Bcl-xL deamidation, pHi, and apoptosis in parallel, we manipulated pHi values artificially by incubating cells at varying pHe values in the absence of monensin.
0054
Interestingly, in the cells expressing these mutant forms of Bcl-xL, the apoptosis induced by enforced alkalinisation was reduced 4-fold compared to cells transduced with empty vector, or more than 2-fold in comparison with the wild-type protein (Figure 3G, right panel), which of course undergoes deamidation in response to alkali treatment.

0426
Interestingly, in the cells expressing these mutant forms of Bcl-xL, the apoptosis induced by enforced alkalinisation was reduced 4-fold compared to cells transduced with empty vector, or more than 2-fold in comparison with the wild-type protein (Figure 3G, right panel), which of course undergoes deamidation in response to alkali treatment.

0054
Nevertheless, protection was not absolute, suggesting that Bcl-xL may not be the only mechanism protecting cells from apoptosis triggered by alkalinisation.

0426
Nevertheless, protection was not absolute, suggesting that Bcl-xL may not be the only mechanism protecting cells from apoptosis triggered by alkalinisation.

0081
As a further control, we have confirmed that Bcl-xL isolated from wild-type thymocytes exposed to a high pH buffer can no longer sequester Bim (Figure S2B), thereby mimicking the effects of DNA damage (Figure 2A).
0054
Taken overall, these results demonstrate that intracellular alkalinisation following DNA damage is both necessary and sufficient for nonenzymatic Bcl-xL deamidation, that the oncogenic suppression of Bcl-xL deamidation in pretumourigenic thymocytes is caused by inhibition of alkalinisation, and that versions of Bcl-xL competent for BH3-only protein sequestration are sufficient per se to protect cells from apoptosis at alkaline pHi.

0426
Taken overall, these results demonstrate that intracellular alkalinisation following DNA damage is both necessary and sufficient for nonenzymatic Bcl-xL deamidation, that the oncogenic suppression of Bcl-xL deamidation in pretumourigenic thymocytes is caused by inhibition of alkalinisation, and that versions of Bcl-xL competent for BH3-only protein sequestration are sufficient per se to protect cells from apoptosis at alkaline pHi.
0411
Because the NHE-1 Na/H antiport is a well-established regulator of pHi [22] and has previously been implicated in the regulation of thymic apoptosis [23], we measured its expression in wild-type thymocytes after DNA damage and found that the NHE-1 level increased 2.5-fold within 5 h, whereas this increase was completely suppressed in pretumourigenic thymocytes (Figure 4B). No inhibition of increased NHE-1 expression in wild-type thymocytes was observed following addition of the Z-VAD-fmk caspase inhibitor (Figure S4A) nor following depletion of Bax and Bak from the cells (Figure S4B). We therefore carried out a further series of experiments to demonstrate that there was a direct causal linkage between the regulation of NHE-1 expression, pHi, Bcl-xL deamidation, and apoptosis. Given that the OTK blocks DNA-damage induced NHE-1 expression in pretumourigenic thymocytes, this provides a powerful system for examining the consequences of experimentally enforcing NHE-1 expression in these cells by retroviral transduction. As Figure 4C illustrates (upper panel), an enforced 2-fold–3-fold increase in NHE-1 expression in pretumourigenic thymocytes, without DNA damage, restored Bcl-xL deamidation to a level comparable to that observed in a retrovirally transduced wild-type control in five separate experiments, thereby bypassing the OTK-mediated inhibition in deamidation.
0054
We measured apoptosis by two different methods to ensure that DNA damage–induced cell death following retroviral transduction was by apoptosis and not by necrosis. Figure 5C and Figure S6B illustrate that double staining for Annexin V and propidium iodide (PI) followed by FACS analysis revealed a major increase in Annexin V+ PI− (apoptotic) cells following transduction with the negative control shRNA followed by either γ irradiation or treatment with etoposide, whereas there was no increase in apoptotic cells above baseline in the cells depleted of NHE-1: DNA damage–induced apoptosis was blocked 100%. Comparable results were obtained by measuring the sub-G1 peak by FACS (unpublished data) and NHE-1 depletion also correlated with increased survival (Figure S5B).

0426
We measured apoptosis by two different methods to ensure that DNA damage–induced cell death following retroviral transduction was by apoptosis and not by necrosis. Figure 5C and Figure S6B illustrate that double staining for Annexin V and propidium iodide (PI) followed by FACS analysis revealed a major increase in Annexin V+ PI− (apoptotic) cells following transduction with the negative control shRNA followed by either γ irradiation or treatment with etoposide, whereas there was no increase in apoptotic cells above baseline in the cells depleted of NHE-1: DNA damage–induced apoptosis was blocked 100%.

0416
Figure 5C and Figure S6B illustrate that double staining for Annexin V and propidium iodide (PI) followed by FACS analysis revealed a major increase in Annexin V+ PI− (apoptotic) cells following transduction with the negative control shRNA followed by either γ irradiation or treatment with etoposide, whereas there was no increase in apoptotic cells above baseline in the cells depleted of NHE-1: DNA damage–induced apoptosis was blocked 100%.
0081
For example, a number of serine kinases have been shown to regulate NHE-1 phosphorylation and activity [24,25], so we investigated the pSer and pThr levels in NHE-1 immunoprecipitates from irradiated wild-type and pretumourigenic thymocytes, but the basal level of phosphorylation did not change after DNA damage and was comparable between the two cell types (Figure S6C). Nevertheless, we cannot formally exclude the possibility that not all pSer/pThr sites were recognised by the cocktail of monocolonal antibodies (mAbs) used.
0054
Figure S7 shows that this was indeed the case: murine tumour cells resistant to genotoxic insult at physiological pHi values can be sensitised to die by enforced alkalinisation leading to Bcl-xL deamidation. Furthermore, a modest rise in pHi following incubation in a mildly alkaline buffer produces levels of Bcl-xL deamidation and apoptosis in murine tumour cells comparable to those observed by adding a DNA damaging reagent to wild-type thymocytes incubated at physiological pH.

0426
Figure S7 shows that this was indeed the case: murine tumour cells resistant to genotoxic insult at physiological pHi values can be sensitised to die by enforced alkalinisation leading to Bcl-xL deamidation. Furthermore, a modest rise in pHi following incubation in a mildly alkaline buffer produces levels of Bcl-xL deamidation and apoptosis in murine tumour cells comparable to those observed by adding a DNA damaging reagent to wild-type thymocytes incubated at physiological pH.
0054
We therefore determined whether genotoxic treatment in vitro of primary human B lineage CLL (B-CLL) cells might cause increased NHE-1, alkalinisation, Bcl-xL deamidation, and apoptosis, as in primary murine thymocytes (Figures 3–5), or whether this might be inhibited, as with the murine cancer cells (Figure S7). In addition, we examined the consequences for these parameters of incubating cancer cells in alkaline pH buffers.

0426
We therefore determined whether genotoxic treatment in vitro of primary human B lineage CLL (B-CLL) cells might cause increased NHE-1, alkalinisation, Bcl-xL deamidation, and apoptosis, as in primary murine thymocytes (Figures 3–5), or whether this might be inhibited, as with the murine cancer cells (Figure S7). In addition, we examined the consequences for these parameters of incubating cancer cells in alkaline pH buffers. To perform these investigations, we divided each sample of patient cancer cells into nine aliquots that were either untreated, subjected to γ irradiation, or exposed to etoposide, followed by incubation at pH 7.2, pH 8.0, or pH 8.5 for 24 h. Each aliquot was then further subdivided into three samples to measure pHi, Bcl-xL deamidation, and apoptosis. As expected, exposure of cells to mildly alkaline buffers generated pHi values that displayed some variation between samples from different patients within a narrow range.

0426
Interestingly, unlike the murine tumour cells expressing an OTK, the B-CLL cells behaved somewhat more like wild-type thymocytes in that DNA damage at physiological pHe caused a mean increase of pHi of 0.22 units, an 8% increase in Bcl-xL deamidation, and an 18% increase in the number of cells undergoing apoptosis (Figure 6A and Figure S8A), compared to the higher thymocyte values of 0.45 pHi units, 40% increase, and 37% increase, respectively (Figure 3). The human cancer cell values for these parameters were greatly increased at alkaline pHe, generating tight correlations between increasing pHi, Bcl-xL deamidation, and apoptosis (r values shown in Figure 6A).

0054
Interestingly, unlike the murine tumour cells expressing an OTK, the B-CLL cells behaved somewhat more like wild-type thymocytes in that DNA damage at physiological pHe caused a mean increase of pHi of 0.22 units, an 8% increase in Bcl-xL deamidation, and an 18% increase in the number of cells undergoing apoptosis (Figure 6A and Figure S8A), compared to the higher thymocyte values of 0.45 pHi units, 40% increase, and 37% increase, respectively (Figure 3). The human cancer cell values for these parameters were greatly increased at alkaline pHe, generating tight correlations between increasing pHi, Bcl-xL deamidation, and apoptosis (r values shown in Figure 6A).

0054
This point is further illustrated by the gray shaded area shown in Figure 6A, which encompasses the overlap in sub-G1 (apoptosis) values that were obtained either by DNA damage at physiological pH or by enforced alkalinisation without DNA damage. Conversely, incubation of B-CLL cells at lower pH inhibited DNA damage–induced Bcl-xL deamidation and apoptosis (Figure 6B). Therefore with respect to enforced changes in pHi, the B-CLL cells behaved in a comparable way to both murine thymocytes and tumour cells.

0426
This point is further illustrated by the gray shaded area shown in Figure 6A, which encompasses the overlap in sub-G1 (apoptosis) values that were obtained either by DNA damage at physiological pH or by enforced alkalinisation without DNA damage. Conversely, incubation of B-CLL cells at lower pH inhibited DNA damage–induced Bcl-xL deamidation and apoptosis (Figure 6B). Therefore with respect to enforced changes in pHi, the B-CLL cells behaved in a comparable way to both murine thymocytes and tumour cells.
0054
These increases correlate with the observed increases in Bcl-xL deamidation and apoptosis in patients' cells (Figure 6A) and at the 2.6-fold level, at least, are comparable with the increases observed in wild-type thymocytes (Figure 4B).

0426
These increases correlate with the observed increases in Bcl-xL deamidation and apoptosis in patients' cells (Figure 6A) and at the 2.6-fold level, at least, are comparable with the increases observed in wild-type thymocytes (Figure 4B). Furthermore, DNA damage–induced Bcl-xL deamidation in B-CLL cells was prevented by addition of either cycloheximide (CHX) (Figure S8C) or DMA (Figure S8D), establishing a possible linkage between DNA damage, NHE-1 function, and Bcl-xL deamidation in human cancer cells.
0114
The structural importance of protein iso-Asp residues is likewise underlined by the expression of the putative repair enzyme L-isoaspartate O-methyltransferase which converts iso-Asp to Asp residues: its deletion has striking effects on protein functions [28–30]. Furthermore, comparison of the crystal structures of native rat Bcl-xL with its deamidated version has revealed significant differences [10]; the structural implications of introducing iso-Asp residues into the disordered loop environment of Asn52/Asn66 merits further work.
0081
We have identified critical elements in the signalling pathway leading from DNA damage to Bcl-xL deamidation in thymocytes and have shown, as Figure 7A illustrates, that deamidation is induced upon DNA damage by up-regulation of the NHE-1 antiport and consequent intracellular alkalinisation (Figures 3–5).

0096
The regulation of NHE-1 antiport function is complex, involving modulation of its expression, phosphorylation, and binding of regulatory proteins [24,25,31].

0104
Our data are consistent with a model in which DNA damage causes alkalinisation by a direct 2–3-fold increase in NHE-1 expression (Figures 4B and 6C), although we cannot exclude the possibility that undetected changes in phosphorylation might shift the pH dependence of the antiport to a more alkaline range as described for myocardial tissue [32].

0007
Furthermore, the calcineurin B homologous protein 1 (CHP-1) has been characterised as an essential cofactor for NHE-1 in normal tissues [33], whereas its CHP-2 homologue is up-regulated in transformed cells [34], so regulation of these proteins might also be involved in activation of the antiport.

0018
Furthermore, the calcineurin B homologous protein 1 (CHP-1) has been characterised as an essential cofactor for NHE-1 in normal tissues [33], whereas its CHP-2 homologue is up-regulated in transformed cells [34], so regulation of these proteins might also be involved in activation of the antiport.
0054
The resistance to genotoxic attack by CD45−/−lckF505 murine tumour cells correlates, as in their pretumourigenic counterparts, with the inhibition of DNA damage–induced NHE-1 antiport expression, alkalinisation, Bcl-xL deamidation, and apoptosis (Figure S7), which is an apparent example of ”oncogene addiction”, whereby oncogene expression continues to be important for survival [36]. By contrast, DNA damage of human B-CLL cells, which should not express OTKs, triggered increased NHE-1 expression and apoptosis, achieving levels comparable with wild-type thymocytes (Figure 6C). However, enforced alkalinisation of either the murine (Figure S7) or human (Figure 6) cancer cells triggered significant increases in Bcl-xL deamidation and apoptosis, even in the absence of genotoxic attack (Figure 7C). In the case of the B-CLL cells, we cannot yet exclude the possibility that the tight correlation observed between these events does not reflect causal efficacy, and further work will be necessary to elucidate this point.

0411
The resistance to genotoxic attack by CD45−/−lckF505 murine tumour cells correlates, as in their pretumourigenic counterparts, with the inhibition of DNA damage–induced NHE-1 antiport expression, alkalinisation, Bcl-xL deamidation, and apoptosis (Figure S7), which is an apparent example of ”oncogene addiction”, whereby oncogene expression continues to be important for survival [36]. By contrast, DNA damage of human B-CLL cells, which should not express OTKs, triggered increased NHE-1 expression and apoptosis, achieving levels comparable with wild-type thymocytes (Figure 6C).

0426
The resistance to genotoxic attack by CD45−/−lckF505 murine tumour cells correlates, as in their pretumourigenic counterparts, with the inhibition of DNA damage–induced NHE-1 antiport expression, alkalinisation, Bcl-xL deamidation, and apoptosis (Figure S7), which is an apparent example of ”oncogene addiction”, whereby oncogene expression continues to be important for survival [36]. By contrast, DNA damage of human B-CLL cells, which should not express OTKs, triggered increased NHE-1 expression and apoptosis, achieving levels comparable with wild-type thymocytes (Figure 6C). However, enforced alkalinisation of either the murine (Figure S7) or human (Figure 6) cancer cells triggered significant increases in Bcl-xL deamidation and apoptosis, even in the absence of genotoxic attack (Figure 7C). In the case of the B-CLL cells, we cannot yet exclude the possibility that the tight correlation observed between these events does not reflect causal efficacy, and further work will be necessary to elucidate this point.

0054
For example, the down-regulation of Bcl-xL promoted the apoptosis of KARPAS-299 cells derived from a patient with anaplastic large cell lymphoma [38], and down-regulation of Bcl-xL suppresses the tumourigenic potential of the causative NPM-ALK oncogenic fusion protein in vivo [39]. Knockdown of Bcl-xL also significantly reduces the viability of pancreatic cancer cells to tumour necrosis factor α (TNF-α)– and TNF-α - related apoptosis-inducing ligand (TRAIL)-mediated apoptosis by antitumour drugs [40].

0426
For example, the down-regulation of Bcl-xL promoted the apoptosis of KARPAS-299 cells derived from a patient with anaplastic large cell lymphoma [38], and down-regulation of Bcl-xL suppresses the tumourigenic potential of the causative NPM-ALK oncogenic fusion protein in vivo [39]. Knockdown of Bcl-xL also significantly reduces the viability of pancreatic cancer cells to tumour necrosis factor α (TNF-α)– and TNF-α - related apoptosis-inducing ligand (TRAIL)-mediated apoptosis by antitumour drugs [40].

0412
Knockdown of Bcl-xL also significantly reduces the viability of pancreatic cancer cells to tumour necrosis factor α (TNF-α)– and TNF-α - related apoptosis-inducing ligand (TRAIL)-mediated apoptosis by antitumour drugs [40].
0081
These digestion conditions were chosen after careful optimisation to give good and consistent yields of the peptides SDVEENRTEAPEGTESEMETPSAINGNPSW (peptide 1) and HLADSPAVNGATGHSSSL (peptide 2), containing the putative deamidation sites N52 and N66, respectively, but without inducing further deamidation.

0004
Peptides were eluted from the column (0.075 mm × 100 mm, Vydac C18) with a gradient of 5%–35% acetonitrile (containing 10 mM ammonium acetate pH 5.3) over 30 min at a flow rate of 250 nl/min.

0081
During the development phase of the methodology, the mass spectrometer was operated in MS/MS mode to conclusively identify the peptide digestion products and to confirm the sites of deamidation as N52 and N66. Once the identities of the peptides had been established, the mass spectrometer was operated in MS mode for subsequent analyses.
0004
Cells were lysed in 50 mM HEPES (pH 7.2), 150 mM NaCl, 1mM EDTA, 0.2% NP-40, and complete protease inhibitors. Cell lysates were resolved by standard Laemmli's SDS-PAGE (pH 8.8) unless otherwise stated.
0054
Intracellular pH was measured using a standard ratiometric method with a pH-sensitive fluorophore SNARF-1 by flow cytometry [44]. Briefly, cells in phosphate-buffered saline (PBS) were loaded with 10 μM SNARF-1 for 40 min at 37 °C, followed by washing and incubation in PBS at room temperature for 30 min prior to measurement of pHi.

0054
FACS data were analysed using Flowjo software to obtain the ratio based on the Fl3/Fl2 channels.
0055
(C) Plasmids of shRNA Bax (GFP) and shRNA Bak (DsRed) were cotransfected into purified DN thymocytes using an Amaxa nucleofactor kit. 48 h later, GFP+ DsRed+ cells were purified by flow cytometry and treated with etoposide (Etop, 25 μM) for 30 h or exposed to irradiation (IR, 5 Gy) followed by 30 h in culture.

0663
(C) Plasmids of shRNA Bax (GFP) and shRNA Bak (DsRed) were cotransfected into purified DN thymocytes using an Amaxa nucleofactor kit. 48 h later, GFP+ DsRed+ cells were purified by flow cytometry and treated with etoposide (Etop, 25 μM) for 30 h or exposed to irradiation (IR, 5 Gy) followed by 30 h in culture.

0054
48 h later, GFP+ DsRed+ cells were purified by flow cytometry and treated with etoposide (Etop, 25 μM) for 30 h or exposed to irradiation (IR, 5 Gy) followed by 30 h in culture.

0006
Cells were then processed for immunoblotting with Bcl-xL antibody.
0006
Wild-type (C57BL/6) thymocytes (1.5 × 107) were exposed to 5 Gy irradiation (IR) and then maintained in culture for the times shown, after which cells were lysed and either separated as whole cell lysates (WCL) or as Bim immunoprecipitates, followed by immunoblotting for either Bcl-xL or for Bim.
0006
(B) Bcl-xL was immunoprecipitated from lysates derived from purified DN thymocytes treated with/without etoposide (ut/E), followed by immunoblotting for Bim or Puma.
0006
Bcl-xL was immunoprecipitated from lysates derived from 1.5 × 106 sorted GFP-positive cells per lane, followed by immunoblotting for Bim or Puma.

0413
Note that in the vector lane, at this exposure endogenous Bcl-xL is not visible because of the small number of cells used.

0426
Note that in the vector lane, at this exposure endogenous Bcl-xL is not visible because of the small number of cells used.
0081
(F) Peptides SDVEENRTEAPEGTESEMETPSAINGNPSW (peptide 1) and HLADSPAVNGATGHSSSL (peptide 2), and the corresponding deamidated forms, containing the putative deamidation sites N52 and N66, respectively, were generated by digestion of rBcl-xL with chymotrypsin. The chromatographic conditions used for the separation of the peptides in the LC-MS analyses were optimised so as to resolve the Asn, Asp, and iso-Asp forms of peptides 1 and 2. The Asp and iso-Asp forms of the two peptides were identified by spiking an aliquot of a digestion mixture with Asp- or iso-Asp–containing synthetic peptides prior to LC-MS (Figure S3).
0004
Wild-type thymocytes were maintained in RPMI-1640/10% bovine fetal calf serum buffered at the indicated pH with Tris-HCl for 20 h in the presence of 20 μM monensin prior to lysis and immunoblotting for Bcl-xL. To minimize any deamidation produced during the gel-running process, the resolving gel buffer was adjusted to pH 8.0 in this experiment.
0416
72 h after the first round of infection, cells were immunoblotted for NHE-1 and Bcl-xL.

0054
The lower left FACS histogram shows the infection efficiency for nontransfected (non), empty-vector transfected (vector), or NHE-1 transfected (NHE-1) cells as percentage GFP-positive cells. The lower right histograms show the mean pHi and apoptosis (sub-G1) values ± SD (n = 5) analysed on GFP-negative and positive cells.

0416
The lower left FACS histogram shows the infection efficiency for nontransfected (non), empty-vector transfected (vector), or NHE-1 transfected (NHE-1) cells as percentage GFP-positive cells.

0426
The lower right histograms show the mean pHi and apoptosis (sub-G1) values ± SD (n = 5) analysed on GFP-negative and positive cells.
0054
pHi was measured by FACS on live CD4−CD8− cells, and the sub-G1 peak was analysed by FACS on CD4−CD8− cells to assess apoptosis.

0426
pHi was measured by FACS on live CD4−CD8− cells, and the sub-G1 peak was analysed by FACS on CD4−CD8− cells to assess apoptosis.
0054
Patients' cells (PBMC, in the range 85%–95% CD19+B220+) were incubated at pHe values of 7.2, 8.0, or 8.5, and the pHi values were monitored by SNARF-1 staining using flow cytometry. Apoptosis was evaluated by measurement of sub-G1 peaks using flow cytometry.
0052
The same cell aliquots cultured in RPMI/10% FCS for 24 h or 48 h were analysed for apoptosis by sub-G1 staining (right panel).
0054
(C) Enforced alkalinisation of murine tumour cells, or human B-CLL cells, causes Bcl-xL deamidation and subsequent apoptosis, even in the absence of external genotoxic attack.

0426
(C) Enforced alkalinisation of murine tumour cells, or human B-CLL cells, causes Bcl-xL deamidation and subsequent apoptosis, even in the absence of external genotoxic attack.
0030
Even though these highly homologous receptors both activate the G protein stimulatory for adenylyl cyclase (Gs), signaling through β1AR and β2AR produces clearly distinguishable biological effects (Xiang and Kobilka, 2003; Xiao et al, 2004).
0018
Biochemical, electrophysiological, and in vivo imaging studies are consolidating the idea that occupancy of different receptors generates a nonuniform pattern of activation of cAMP effector proteins such as PKA (cAMP-dependent protein kinase).
0006
To probe for a possible signaling complex including the β1AR and a PDE, mouse neonatal cardiomyocytes were infected with an adenovirus encoding a Flag-tagged β1AR, and the receptor was subsequently immunoprecipitated using an antibody against the tag.

0676
To probe for a possible signaling complex including the β1AR and a PDE, mouse neonatal cardiomyocytes were infected with an adenovirus encoding a Flag-tagged β1AR, and the receptor was subsequently immunoprecipitated using an antibody against the tag.

0019
A significant amount of endogenous PDE activity was recovered in the β1AR immunoprecipitation (IP) pellet (Figure 1A).

0413
A significant amount of endogenous PDE activity was recovered in the β1AR immunoprecipitation (IP) pellet (Figure 1A). The PDE activity associated with the β1AR was inhibited by the PDE4-selective inhibitor, Rolipram, identifying this activity as PDE4.

0019
Whereas ablation of PDE4A or PDE4B had no effect, inactivation of the PDE4D gene prevented co-IP of PDE activity with the β1AR (Figure 1B).
0030
Long forms contain a conserved UCR1/UCR2 (upstream conserved regions 1 and 2) motif, whereas short forms lack UCR1 and part of UCR2 (Conti et al, 2003; Houslay and Adams, 2003).

0411
The co-IP of these PDE4D splice variants expressed exogenously in HEK293 cells identified PDE4D8 as the variant that most efficiently interacts with β1AR.
0007
To further characterize the interaction, we performed IPs using PDE4D and βARs that were purified from a baculovirus expression system to >90% purity (see Supplementary Figure 1 for the characterization of the purified proteins).

0096
To further characterize the interaction, we performed IPs using PDE4D and βARs that were purified from a baculovirus expression system to >90% purity (see Supplementary Figure 1 for the characterization of the purified proteins).

0676
To further characterize the interaction, we performed IPs using PDE4D and βARs that were purified from a baculovirus expression system to >90% purity (see Supplementary Figure 1 for the characterization of the purified proteins).
0411
To determine whether receptor occupancy affects the β1AR/PDE4D complex, HEK293 cells expressing exogenous β1AR and PDE4D8 were incubated with different ligands.
0081
All PDE4 long forms are activated by phosphorylation at a conserved PKA consensus site in UCR1 (see Figure 2D); this mechanism provides a ubiquitous negative-feedback loop critical for cAMP signaling (Conti et al, 2003). Accordingly, stimulation of cultured neonatal cardiac myocytes with β-adrenergic agonists leads to a rapid PKA-mediated activation of PDE4D (Supplementary Figure 2A–C).

0412
If complexes composed of βARs and PDEs are present in these cells, phosphorylation should be biased toward the PDEs present in the vicinity of the occupied receptors.
0413
The presence of a PDE4D in the vicinity of the β1AR should affect the activity of PKA localized with the receptor as well as the PKA-phosphorylation state of the receptor itself. This possibility was tested by blocking PDE activity with selective PDE4 inhibitors in cardiomyocytes (Figure 6A and B), by using MEFs deficient in PDE4D (Figure 6C and D), or by overexpressing a catalytically inactive PDE4D in cardiomyocytes, which acts as a dominant-negative construct (Perry et al, 2002; Baillie et al, 2003) by displacing endogenous PDE4D from the β1AR complex (Figure 6E and F; Supplementary Figure 5). In all instances, blockage of PDE4 activity or, more specifically, ablation or displacement of PDE4D caused a significant increase in the phosphorylation of the transfected β1AR in the absence of β-adrenergic agonists.

0676
This possibility was tested by blocking PDE activity with selective PDE4 inhibitors in cardiomyocytes (Figure 6A and B), by using MEFs deficient in PDE4D (Figure 6C and D), or by overexpressing a catalytically inactive PDE4D in cardiomyocytes, which acts as a dominant-negative construct (Perry et al, 2002; Baillie et al, 2003) by displacing endogenous PDE4D from the β1AR complex (Figure 6E and F; Supplementary Figure 5).
0081
The rate of return to basal heart rate after the initial response to ISO was slightly faster in PDE4DKO mice compared with wild-type controls (Figure 7A); however, this effect was greatly magnified by pretreatment of mice with GLP1 (Figure 7B; P<0.0001). The faster decrease in heart rate is in agreement with our stated hypothesis that elevated levels of cAMP/PKA activity in the vicinity of the β1AR, due to absence of PDE4D in this compartment, causes an increased phosphorylation and heterologous desensitization of β1AR (see Figure 6).
0413
β1AR preferentially associates with PDE4D8 in cardiomyocytes as shown by co-IP of endogenous PDE with the β1AR (Figure 1C), as well as the selective activation of PDE4D8 in intact cells (Figure 5A).

0676
Conversely, PDE4D5 is the variant tethered to the β2AR/β-arrestin complex (Baillie et al, 2003) concurring with the preferential activation of PDE4D5 upon β2AR signaling (Figure 5B). In pull-down experiments using purified proteins (Figure 3D and E), β1AR efficiently interacts with PDE4D, whereas β2AR has negligible affinity for PDE4D, underscoring the direct mode of PDE4D–β1AR interaction versus the indirect, β-arrestin-dependent mode of PDE4D–β2AR interaction.

0019
In pull-down experiments using purified proteins (Figure 3D and E), β1AR efficiently interacts with PDE4D, whereas β2AR has negligible affinity for PDE4D, underscoring the direct mode of PDE4D–β1AR interaction versus the indirect, β-arrestin-dependent mode of PDE4D–β2AR interaction.

0413
The β1AR/PDE4D complex is present in the absence of agonist and dissociates after receptor occupancy, whereas agonist binding to the β2AR is a prerequisite for the recruitment of the β-arrestin/PDE4D complex to the receptor. Thus, under basal conditions, PDE4D is poised to control local cAMP concentration and PKA activity in the vicinity of the β1AR (see Figure 6), whereas it affects β2AR signaling only after ligand binding and β-arrestin recruitment.
0081
This, in turn, protects the β1AR from PKA-mediated phosphorylation and desensitization (Rapacciuolo et al, 2003; Gardner et al, 2006) and may control PKA-mediated phosphorylation of other localized substrates. Indeed, when PDE4 activity is inhibited in cardiomyocytes (Figure 6A and B), is absent, as in the PDE4D-deficient MEFs (Figure 6C and D), or is displaced, as in PDE4D-DN infected cardiomyocytes (Figure 6E and F), a substantial increase in basal β1AR receptor phosphorylation is observed.
0096
The latter may include the competition of PDE4D with other proteins for binding to the receptor or, in an opposite fashion, PDE4D may act as a scaffold by tethering additional proteins to the receptor complex such as the exchange protein activated by cAMP or PKA-anchoring proteins (Dodge-Kafka et al, 2005).

0018
The latter may include the competition of PDE4D with other proteins for binding to the receptor or, in an opposite fashion, PDE4D may act as a scaffold by tethering additional proteins to the receptor complex such as the exchange protein activated by cAMP or PKA-anchoring proteins (Dodge-Kafka et al, 2005).

0030
The latter may include the competition of PDE4D with other proteins for binding to the receptor or, in an opposite fashion, PDE4D may act as a scaffold by tethering additional proteins to the receptor complex such as the exchange protein activated by cAMP or PKA-anchoring proteins (Dodge-Kafka et al, 2005).
0081
The observed increased receptor phosphorylation that follows PDE4D deficiency (Figure 6) promotes progressive desensitization (Figure 7) and downregulation of this receptor, a hallmark in heart failure.
0081
PAN-selective antibodies against PDE4A (AC55), PDE4B (K118), and PDE4D (M3S1), as well as splice variant-selective antibodies against PDE4D3, 4, 5, 8, and 9 (Richter et al, 2005), were used in IPs to determine the expression of the respective PDE4 subtype and splice variant in cardiac myocytes (Figure 2), as well as their PKA-dependent activation after β-adrenergic stimulation (Figure 5; Supplementary Figure 2). A PKA-site-specific antibody from Cell Signaling (Danvers, MA) was used to measure PKA phosphorylation of the β1AR (Figure 6).

0426
PAN-selective antibodies against PDE4A (AC55), PDE4B (K118), and PDE4D (M3S1), as well as splice variant-selective antibodies against PDE4D3, 4, 5, 8, and 9 (Richter et al, 2005), were used in IPs to determine the expression of the respective PDE4 subtype and splice variant in cardiac myocytes (Figure 2), as well as their PKA-dependent activation after β-adrenergic stimulation (Figure 5; Supplementary Figure 2). A PKA-site-specific antibody from Cell Signaling (Danvers, MA) was used to measure PKA phosphorylation of the β1AR (Figure 6).
0411
HEK293 and MEF cells were cultured in DMEM supplemented with 10% FBS, 1 mM glutamine, 30 μg/ml penicillin, and 100 μg/ml streptomycin. All cells were cultured at 37°C and under a 5% CO2 atmosphere. For expression of exogenous βARs and/or PDE4D constructs, cells were infected with adenoviruses 40 h before experimentation at an MOI of 2–6 (HEK293), 20–40 (cardiac myocytes), and 100 (MEFs), respectively.

0416
For expression of exogenous βARs and/or PDE4D constructs, cells were infected with adenoviruses 40 h before experimentation at an MOI of 2–6 (HEK293), 20–40 (cardiac myocytes), and 100 (MEFs), respectively. As ‘mock' controls, cells were infected with comparable titers of an adenovirus encoding green fluorescent protein (GFP).

0055
As ‘mock' controls, cells were infected with comparable titers of an adenovirus encoding green fluorescent protein (GFP).
0004
After the respective cell treatment, cells were rinsed once with ice-cold PBS and then lysed in 500 μl of 20 mM HEPES, 150 mM NaCl, 2 mM EDTA, 10% glycerol, 1% N-dodecyl-β-D-maltopyranoside (DDM, Anatrace), 1 μM microcystin-LR (Calbiochem), and Complete protease inhibitor cocktail (Roche).

0007
Flag-tagged receptors were then immunoprecipitated using M1-affinity resin (α-Flag antibody resin; Sigma Aldrich). After incubation for 4 h at 4°C, the resin was washed three times and proteins were eluted in 40 μl of elution buffer (200 μg/ml Flag peptide, 20 mM HEPES, 50 mM NaCl, 0.1% cholesterol, and 8 mM EDTA).

0006
Flag-tagged receptors were then immunoprecipitated using M1-affinity resin (α-Flag antibody resin; Sigma Aldrich).
0006
Rat PDE4D3, expressed in Sf9 insect cells using a recombinant baculovirus, was purified to >90% purity using an anti-PDE4D antibody (M3S1) covalently coupled to ProteinG Sepharose as described previously (Salanova et al, 1998). Flag-tagged β1AR and β2AR were also expressed in Sf9 cells and subsequently purified in a two-step procedure consisting of an initial affinity chromatography using M1-resin (immobilized anti-Flag antibody; Sigma Aldrich), followed by an alprenolol-sepharose affinity column.

0007
Rat PDE4D3, expressed in Sf9 insect cells using a recombinant baculovirus, was purified to >90% purity using an anti-PDE4D antibody (M3S1) covalently coupled to ProteinG Sepharose as described previously (Salanova et al, 1998). Flag-tagged β1AR and β2AR were also expressed in Sf9 cells and subsequently purified in a two-step procedure consisting of an initial affinity chromatography using M1-resin (immobilized anti-Flag antibody; Sigma Aldrich), followed by an alprenolol-sepharose affinity column.

0019
Rat PDE4D3, expressed in Sf9 insect cells using a recombinant baculovirus, was purified to >90% purity using an anti-PDE4D antibody (M3S1) covalently coupled to ProteinG Sepharose as described previously (Salanova et al, 1998).

0676
Flag-tagged β1AR and β2AR were also expressed in Sf9 cells and subsequently purified in a two-step procedure consisting of an initial affinity chromatography using M1-resin (immobilized anti-Flag antibody; Sigma Aldrich), followed by an alprenolol-sepharose affinity column.

0004
For βAR/PDE4D IP, equal amounts of purified β1AR and β2AR (1 μg) were coupled to M1 resin and then incubated in 500 μl of 20 mM HEPES (pH 7.5), 100 mM NaCl, 0.1% DDM, 4 mM CaCl2, and 0.01% cholesterol hemisuccinate with 0.5 μg of purified PDE4D under continuous rotation for 4 h at 4°C.
0004
After 4 days of culture and the respective cell treatment, neonatal cardiac myocytes were harvested in buffer containing 50 mM Tris–HCl (pH 7.4), 1 mM EDTA, 0.2 mM EGTA, 150 mM NaCl, 5 mM β-mercaptoethanol, 10% glycerol, 1 μM microcystin-LR, Complete protease inhibitor cocktail (Roche Diagnostics), and 1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride (AEBSF; Roche Diagnostics).

0040
for 30 min), and soluble extracts were immunoprecipitated using 30 μl ProteinG Sepharose and the respective PDE4 subtype, or PDE4D splice variant antibodies, as well as IgG as a control.
0019
(B) Co-IP of β1AR and PDE activity from cardiomyocytes deficient in PDE4A, PDE4B, or PDE4D, and wild-type controls.
0413
(A) Total, PDE4, and non-PDE4 activity in detergent extracts of cultured neonatal cardiac myocytes. PDE3 is the major non-PDE4 subtype expressed in these cells contributing 24±5 pmol/min/mg to the total PDE activity.

0081
The PKA phosphorylation site conserved among long splice variants is indicated with red circles.
0019
Interaction of exogenous β1AR and PDE4D. (A, B) Co-IP of exogenous β1AR and Myc-tagged PDE4D splice variants expressed in HEK293 cells.

0411
(A, B) Co-IP of exogenous β1AR and Myc-tagged PDE4D splice variants expressed in HEK293 cells.

0019
(C) Shown is the co-IP of exogenous β1AR and PDE4D8-Myc from extracts of MEFs derived from mice deficient in β-arrestin 1 and 2 (βarr1/2KO) or from wild-type controls (WT-MEF).

0676
(D, E) PDE4D3, and Flag-tagged receptors, β1AR and β2AR, were affinity purified after baculovirus expression (see Supplementary Figure 1).
0411
HEK293 cells expressing exogenous β1AR and PDE4D8-Myc were treated with β-adrenergic agonists before cell lysis and IP of the β1AR. (A) Cells were treated for 10 min with 100 μM of the physiological β1AR agonist (−)-Norepinephrine or the stereoisomer (+)-Norepinephrine, which is not an efficient ligand for the β1AR.
0006
(A, B) Neonatal cardiac myocytes derived from mice deficient in β2AR were stimulated for 3 min with 100 nM ISO (A) and cells deficient in β1AR were treated for 3 min with 10 μM ISO (B). At the end of incubation, cells were lysed, PDE4D5, 8, and 9 were immunoprecipitated with variant-specific antibodies, and the PDE activity recovered in the IP pellet was measured.

0413
(A, B) Neonatal cardiac myocytes derived from mice deficient in β2AR were stimulated for 3 min with 100 nM ISO (A) and cells deficient in β1AR were treated for 3 min with 10 μM ISO (B). At the end of incubation, cells were lysed, PDE4D5, 8, and 9 were immunoprecipitated with variant-specific antibodies, and the PDE activity recovered in the IP pellet was measured.

0055
Shown is the average of five experiments; three experiments performed using myocytes deficient in β2AR and two experiments using cells deficient in β1AR.
0081
(E, F) Neonatal cardiac myocytes coexpressing a Flag-tagged β1AR and either GFP, a catalytically inactive PDE4D8 construct (PDE4D-DN; see also Supplementary Figure 5), or a catalytically inactive PDE3A1 (PDE3A1-DN) were subjected to α-Flag(M1)-IP, and the phosphorylation of the β1AR was subsequently detected in IB using a PKA-substrate-specific antibody.
0413
In the case of β1AR, a preformed complex with PDE4D8 that is likely responsible for controlling local cAMP concentration and PKA activity in the vicinity of the receptor under basal conditions is dissociated upon ligand binding.

0096
It remains to be determined to what extent PDE4D9, which is activated after both β1AR and β2AR stimulation (Figure 5) and which also showed interaction with β1AR in co-IPs of exogenous proteins (Figure 3A and B), can substitute for interaction with the βARs in vivo.
0411
AIRE is expressed in thymic medullary epithelial cells, where it promotes the expression of tissue-restricted antigens.
0411
AIRE promotes the thymic expression of many tissue-restricted antigens, enabling the negative selection of developing T cells and thus precluding self-reactivity (Anderson et al, 2002; Liston et al, 2003); however, the mechanisms are so far unknown.
0889
Our results, in agreement with recent studies of the DNMT3L and BHC80 PHD fingers (Lan et al, 2007; Ooi et al, 2007), show a new role for the PHD finger as an H3K4me0 reader.
0019
Western blot analysis of H3/AIRE–PHD1 complex formation by using antibodies for H3K4me1, H3K4me3 and H3K9me3 indicated that H3K4 trimethylation hinders interaction (Fig 1F), whereas H3K9 trimethylation does not.

0676
Western blot analysis of H3/AIRE–PHD1 complex formation by using antibodies for H3K4me1, H3K4me3 and H3K9me3 indicated that H3K4 trimethylation hinders interaction (Fig 1F), whereas H3K9 trimethylation does not.

0423
Binding experiments with amino-terminal histone H3 unmodified (H3K4me0) or modified (H3K4me1, H3K4me3 and H3K9me3) 20-mer peptides showed that these N-terminal residues of histone H3 are sufficient for binding to AIRE–PHD1 (Fig 1G).
0096
Notably, H3K4me3 did not show any significant interaction with AIRE–PHD1 in either binding assay.
0081
The mapping of the H3/AIRE interaction site uniquely to AIRE–PHD1 was further confirmed by NMR titrations of histone H3 peptides into AIRE–PHD2, which bound neither methylated nor H3K4me0 peptides (data not shown).
0114
We generated a model of AIRE–PHD1 complexed with the H3K4me0 peptide on the basis of the crystal structure of the BPTF–PHD finger bound to H3K4me3 and performed molecular dynamics calculations for 10 ns.

0104
Indeed, fluorescence spectroscopy and ITC assays showed that the alanine mutations R2A in the H3 peptide and D312A in AIRE–PHD1 markedly reduced the binding affinity (Table 1; Fig 4C) without affecting the protein fold (supplementary Fig S3 online).
0077
Remarkably, residues G305, G306 and G313 showed strong shifts when bound to H3K4me2 and disappeared completely from the NMR spectrum owing to line-shape broadening on binding to H3K4me3, indicating an involvement of this region in peptide binding.
0107
Although there are many similarities between these two structures and the AIRE–PHD1/H3K4me0 complex presented here, the AIRE–PHD1 finger differs in the additional recognition of the H3R2 side chain, which makes an important contribution to the high affinity of this interaction, as shown by our peptide mutagenesis experiments.

0676
Although there are many similarities between these two structures and the AIRE–PHD1/H3K4me0 complex presented here, the AIRE–PHD1 finger differs in the additional recognition of the H3R2 side chain, which makes an important contribution to the high affinity of this interaction, as shown by our peptide mutagenesis experiments.
0411
We have shown previously that transiently transfected AIRE enhances target gene expression in human embryonic kidney (HEK)293 cells (Pitkanen et al, 2005). So far, no cell line has been described with endogenous AIRE expression; therefore, we transfected HEK293 cells with an AIRE-encoding or control plasmid and generated stable cell lines called HEK-AIRE and HEK-control. We first tested HEK-AIRE compared with HEK-control cell lines for expression levels of tissue-restricted antigens that are downregulated in AIRE-deficient mouse thymic medullary epithelial cells (Derbinski et al, 2005). Indeed, the HEK-AIRE cell line showed enhanced expression of such antigens, including insulin, the principal autoantigen in type I diabetes (Babaya et al, 2005), involucrin and S100A8 (Fig 5A).

0413
So far, no cell line has been described with endogenous AIRE expression; therefore, we transfected HEK293 cells with an AIRE-encoding or control plasmid and generated stable cell lines called HEK-AIRE and HEK-control.

0019
Next, we studied in vivo histone binding by protein chromatin immunoprecipitation (ChIP) assays and observed that AIRE is found in complexes with a small fraction of histone H3 but not with H3K4me3.

0030
Although AIRE specificity towards chromatin might be influenced by other protein and DNA interactions (Ruan et al, 2007), the data presented here indicate that AIRE preferentially binds to and activates the promoters containing low levels of H3K4me3.

0411
On the basis of these results, we propose a speculative model for the regulation of tissue-restricted antigen expression in thymic epithelial cells (supplementary Fig S8 online). Normally, tissue-restricted antigens are silenced in immature thymic epithelial cells as they lack the active chromatin mark H3K4me3 on their promoters. During differentiation into mature thymic medullary epithelial cells, activation of AIRE expression (Kyewski & Klein, 2006) enables the read-out of non-methylated H3K4 as a signal to activate tissue-restricted antigen genes.
0065
NMR binding, fluorescence titration assays and isothermal titration calorimetry thermodynamic analysis.

0077
NMR binding, fluorescence titration assays and isothermal titration calorimetry thermodynamic analysis. Details on NMR titrations, fluorescence spectroscopy and thermodynamic measurements are described in the supplementary information online.
0096
(C) A similar experiment to (B) but using GST-PHD proteins, detected by western blot using anti-H3 (top) or Coomassie staining (bottom). (D) GST-PHD1 and GST-AIRE, but not GST alone, interact with purified recombinant histone H3, visualized by Coomassie staining. (E) GST-PHD1, but not GST alone, interacts with native mononucleosomes detected by western blot against anti-H3 (middle) and anti-H2B (bottom). Equal input of GST proteins is shown with Ponceau red staining (top). (F) Interaction between GST-AIRE fusion proteins and whole histones, detected by anti-H3K4me1, anti-H3K4me3 and anti-H3K9me3. (G) Interaction between GST-AIRE–PHD1 fusion proteins and amino-terminal histone H3 peptides (H3K4me0, H3K4me1, H3K4me3 and H3K9me3), all detected by anti-GST. AIRE, autoimmune regulator; GST, glutathione-S-transferase.

0107
(F) Interaction between GST-AIRE fusion proteins and whole histones, detected by anti-H3K4me1, anti-H3K4me3 and anti-H3K9me3. (G) Interaction between GST-AIRE–PHD1 fusion proteins and amino-terminal histone H3 peptides (H3K4me0, H3K4me1, H3K4me3 and H3K9me3), all detected by anti-GST.
0107
(B) Interaction between AIRE–PHD1 (PHD1), AIRE–PHD1-D297A (D297A) mutant proteins and H3K4me0 peptide detected by anti-GST.

0107
N, Kd, ΔH and ΔS represent measured stoichiometric ratio, dissociation binding constant, differential enthalpy and differential entropy, respectively.
0019
AIRE, autoimmune regulator; ChIP, chromatin immunoprecipitation; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HEK, human embryonic kidney; H3K4me3, histone H3 trimethylated at lysine 4.
0413
(B) DNA ChIP analysis with anti-AIRE and IgG was performed from the stably transfected cells, as indicated.

0019
AIRE, autoimmune regulator; ChIP, chromatin immunoprecipitation; HEK, human embryonic kidney; PHD, plant homeodomain.
0081
Instead, ATR-dependent phosphorylation of serine 92 of Mcm2 is required for the recruitment of Plx1 to chromatin and for the recovery of DNA replication under stress.
0081
This is achieved through regulation of Chk1 phosphorylation. Plx1-mediated downregulation of the checkpoint requires ATM/ATR-mediated phosphorylation of serine 92 of Mcm2.
0081
However, although sufficiently high to be detected, the levels of Chk1 phosphorylation were not sufficient to inhibit DNA replication (Figure 1F). Instead, in these conditions, Plx1 depletion abolished DNA replication (Figure 1F).
0889
Compared to chemiluminescence, the detection with infrared dye is linear over a very wide range and amounts from 1 pg up to 10 ng can be detected more reliably (Chen et al, 2005).
0413
To test whether the increased Plx1 binding to chromatin was required for promoting DNA replication in the presence of stalled replication forks, we measured DNA replication in Plx1-depleted extracts treated with low amounts of aphidicolin.

0081
These templates mimic incompletely replicated DNA and induce Chk1 phosphorylation (Figure 3B), as described previously (Yoo et al, 2004a). Strikingly, phosphorylation of Chk1 was dramatically enhanced in Plx1-depleted extracts supplemented with pApT (Figure 3B).

0081
Again, Plx1 depletion led to a significant increase in Chk1 phosphorylation (Figure 3C). Supplementation of control extracts with recombinant Plx1 strongly suppressed Chk1 phosphorylation (Figure 3D).
0096
In extracts treated with aphidicolin, the binding of the Mcm complex to chromatin was not affected by Plx1 depletion, as shown by normal levels of Mcm7 detected on chromatin (Figure 4A). We instead observed a strong inhibition of Cdc45 binding to chromatin in the absence of Plx1, which could be restored by supplementing extracts with recombinant Plx1 (Figure 4A). Plx1 depletion did not impair Cdc45 binding in untreated extracts (Supplementary Figure 7).

0413
Significantly, the reduction of Cdk2 activity was more pronounced in the absence of Plx1. Treatment with caffeine restored Cdk2 activity to control levels in normal and Plx1-depleted extracts (Figure 4B), suggesting that downregulation of Cdk2 was mediated by hyperactivation of the checkpoint induced by the absence of Plx1.
0018
One of the truncated proteins contained the protein interaction domain known as the Polo box, which has been shown to interact with phosphorylated proteins (Elia et al, 2003).

0030
One of the truncated proteins contained the protein interaction domain known as the Polo box, which has been shown to interact with phosphorylated proteins (Elia et al, 2003).

0676
Proteins bound to the Polo box domain were then eluted using a phosphorylated peptide with a strong affinity for the Polo box, and subsequently subjected to matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) to identify the sequence of the interacting proteins.

0018
Proteins bound to the Polo box domain were then eluted using a phosphorylated peptide with a strong affinity for the Polo box, and subsequently subjected to matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) to identify the sequence of the interacting proteins.

0081
Proteins bound to the Polo box domain were then eluted using a phosphorylated peptide with a strong affinity for the Polo box, and subsequently subjected to matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) to identify the sequence of the interacting proteins.

0413
The binding of the Polo box to endogenous Mcm proteins was consistent with the ability of full-length Plx1 to interact with chromatin and Mcm7.

0081
Plx1 was unable to bind to chromatin in extracts depleted of the Mcm complex (Figure 5F), suggesting that the Mcm complex is the major binding site for Plx1 on chromatin during S-phase.
0096
The fact that Plx1 binding to the Mcm complex is enhanced by checkpoint activation suggests that post-translational modifications of the Mcm proteins induced by the activation of the DNA damage response could increase Plx1 affinity for the Mcm complex.

0676
The fact that Plx1 binding to the Mcm complex is enhanced by checkpoint activation suggests that post-translational modifications of the Mcm proteins induced by the activation of the DNA damage response could increase Plx1 affinity for the Mcm complex.

0081
Mcm2 is strongly phosphorylated on serine 92 by ATM/ATR (Yoo et al, 2004b), and its phosphorylation can be induced by pA/pT, aphidicolin or etoposide (Supplementary Figure 9A). To test whether phosphorylation of serine 92 of Mcm2 has a role in Plx1 recruitment to chromatin and in Plx1-mediated suppression of the checkpoint, we produced recombinant wild-type Mcm2 (Mcm2-WT) and Mcm2 in which serine 92 had been substituted with alanine (Mcm2-S92A) rendering it resistant to ATR-mediated phosphorylation (Supplementary Figure 9B).

0096
This procedure leads to the formation of a significant amount of endogenous Mcm complexes containing the exogenous Mcm2-WT or Mcm2-S92A proteins, which are able to assemble onto chromatin and support DNA replication, as shown previously (Yoo et al, 2004b). We confirmed that equivalent amounts of recombinant Mcm2-WT and Mcm2-S92A proteins could be loaded onto chromatin (Figure 6A).

0081
This suggests that Plx1 binding to the Mcm complex loaded onto chromatin requires ATR-mediated phosphorylation of serine 92.
0413
In mammalian cells, Plk1 kinase activity increases in M-phase (Golsteyn et al, 1995).

0676
Here, we show that Plx1 is required for DNA replication in the presence of stalled replication forks induced by different treatments such as the addition of aphidicolin and etoposide or by reducing the level of Mcm complex loaded onto chromatin using low amounts of recombinant geminin.

0107
As the interaction between Plx1 and the Mcm complex is mediated by the Polo box, which is a phospho-binding domain (Elia et al, 2003), it is plausible that direct phosphorylation of Mcm7 and/or Mcm2 by ATM/ATR could favour Plx1 interaction with the Mcm complex on chromatin. Indeed, we found that the interaction of Plx1 with Mcm proteins is increased by the induction of the checkpoint.

0081
These observations reveal an unanticipated function for the ATM/ATR-dependent phosphorylation of serine 92 of Mcm2, which appears to promote DNA replication under stress.
0030
Claspin inactivation by protein degradation restrains Chk1 activity and promotes recovery from the DNA replication checkpoint in mammalian organisms (Mailand et al, 2006; Peschiaroli et al, 2006). We have recently shown that Tipin, a factor critical for replication fork stabilization (Chou and Elledge, 2006), is required for Claspin binding to chromatin and for efficient Chk1 activation (Errico et al, 2007).

0413
Plx1-mediated inactivation of Chk1 at the onset of S-phase might happen at the level of replication origins following Mcm binding and might be due to Plx1-mediated phosphorylation of factors required for efficient Chk1 phosphorylation, such as Tipin (Errico et al, 2007). This event precedes the suppression of Chk1 phosphorylation at the transition into mitosis, which requires Plx1 binding to Claspin and probably a more robust signalling from Plx1 whose activity increases at mitosis onset in Xenopus (Kelm et al, 2002).
0413
During mitosis, Plx1 promotes Cdk1 activation through direct phosphorylation and maintenance of Cdc25 phosphatase activity, which dephosphorylates the inhibitory tyrosine 15 of Cdk1 (Kumagai and Dunphy, 1996; Liu and Maller, 2005). The checkpoint-induced downregulation of Cdk2 activity in egg extracts is mediated by the inactivating phosphorylation on tyrosine 15 of Cdk2/Cyclin E complex, which is under the control of Cdc25 (Costanzo et al, 2000). In the light of these observations, we cannot exclude that suppression of Cdk2 activity is also due to the direct role of Plx1 in controlling the activity of one of the Cdc25 phosphatases present in egg extracts.
0411
Downregulation of Plk1 expression leads to cell death in transformed cells (Liu and Erikson, 2003a, 2003b), suggesting that Plk1 is required for the cell cycle of tumour cells. The genome of tumour cells is highly unstable and has many DNA lesions that would normally interfere with replication progression.
0004
The eggs were collected in 90 mM NaCl. To prepare interphase extracts, the eggs were dejelled in a buffer (20 mM Tris, pH 8.5, 110 mM NaCl, 5 mM DTT) for 5 min, washed with ¼ Marc's modified Ringer (MMR) (5 × MMR: 100 mM HEPES, pH 7.5, 2 M NaCl, 10 mM KCl, 5 mM MgSO4, 10 mM CaCl2, 0.5 mM EDTA) and activated with 1 μg/ml calcium ionophore A23187 in MMR for 5 min. The activated eggs were washed with ¼ MMR and then washed three times with ice-cold S-buffer (0.25 M sucrose, 50 mM KCl, 2.5 mM MgCl2, 2 mM β-mercaptoethanol, 15 μg/ml leupeptin and 50 mM HEPES–KOH at pH 7.5).
0411
6xHistidine-tagged geminin (a gift from M Micheal) was expressed in BL21(DES) cells from a pET28 plasmid and purified on a nickel-NTA column according to standard protocols.

0426
Baculoviruses encoding 6xHistidine-tagged Plx1 (a gift from J Gautier), Mcm2-WT and Mcm2-S92A (a gift from W Dunphy) were used to infect Sf9 cells, which were expanded in large flasks using complete GRACE medium according to standard protocols (Invitrogen).
0040
If not otherwise indicated the replication reactions were stopped at 120 min with Stop buffer (1% SDS, 80 mM Tris, pH 8, 8 mM EDTA). For pulse labelling, aliquots of the replication reaction were incubated for 30 min in the presence of [α-32P]dATP. The samples were digested with 1 mg/ml proteinase K for 45 min, extracted with phenol/chloroform/isoamylalcohol, ethanol-precipitated and separated on 0.8% agarose gel.
0663
The anti-mouse secondary antibody, detecting Mcm7, had an emission at 680 nm (displayed in red in Figure 2A) and the anti-rabbit secondary antibody detecting Plx1 had an emission at 800 nm (displayed in green in Figure 2A).
0004
Nuclei were lysed for 1 h at 4°C in a lysis solution at pH 9.5 (2.5 M NaCl, 100 mM EDTA, 10 mM Tris base, 1% Triton X-100 and 10% dimethyl sulphoxide). The slides were washed three times in the electrophoresis buffer (100 mM Tris, 300 mM sodium acetate, adjusted to pH 8.3) and then equilibrated for 60 min in the same buffer.

0426
Images of 100 randomly selected ethidium bromide-stained cells were analysed from each coded slide at × 400 magnification. To measure the extent of DNA damage, image analysis software (Comet IV, Perceptive Instruments) was used. DNA damage was assessed by monitoring comet head length analysing 100 cells for each slide.
0858
(A) Immunodepletion of Plx1.

0081
(E) Phosphorylation of Chk1 serine 345 (p-CHK1) in extracts supplemented with 7000 nuclei/μl.
0096
(A) Cdc45 and Mcm7 binding to chromatin isolated from extracts that were untreated (CTRL), Plx1 depleted (ΔPlx1) or Plx1 depleted and reconstituted with recombinant Plx1 (+Plx1).

0413
(B) Nuclear Cdk2 activity measured by monitoring 32P incorporation in histone H1 after immunoprecipitating nuclear Cdk2 with anti-Xenopus Cdk2 antibodies from extracts that were untreated (CTRL), mock depleted (Mock) or Plx1 depleted (ΔPlx1) and treated with and without 50 μM aphidicolin (APH) or 5 mM caffeine (Caff), as indicated.
0104
After pull-down, proteins were consecutively eluted using 3 mg/ml of phosphopeptide solution (first elution=E1, second elution=E2, see Supplementary data for the sequence of the peptide and the Polo box) and subjected to MALDI-TOF analysis.

0004
The GST fusion protein containing the Polo box from Plk1 is shown at the bottom.

0040
The extract was untreated (C), mock depleted (M) or Mcm depleted (ΔMCM) using anti-Mcm3 antibodies and incubated with sperm nuclei for 60 min in the presence of 2.5 μM APH. Cytosol and chromatin were then subjected to western blot with anti-Mcm7 and anti-Plx1 antibodies.
0081
Mcm2 serine 92 phosphorylation is required for Plx1 function.

0096
(A) Chromatin binding of Plx1 in the presence of Mcm2-WT and Mcm2-S92A proteins. Sperm nuclei were incubated for 60 min in extracts supplemented with 200 ng/μl recombinant histidine-tagged Mcm2-WT and Mcm2-S92A proteins in the presence of 5 μM etoposide.
0426
The percentage increase in the head length of the comets subtracted from the average nuclear diameter of undepleted samples with high Mcm was used as a parameter to measure DSBs.

0413
In the presence of stalled replication forks, the ATR/ATM-dependent checkpoint activation promotes Plx1 binding to the Mcm complex close to stalled forks and attenuates the inhibitory activity of the checkpoint towards unfired origins.
0114
Here, we demonstrate the specific binding of the C-terminal acidic domain (AC-D) of the human TFIIEα subunit to the pleckstrin homology domain (PH-D) of the human TFIIH p62 subunit and describe the solution structures of the free and PH-D-bound forms of AC-D. Although the flexible N-terminal acidic tail from AC-D wraps around PH-D, the core domain of AC-D also interacts with PH-D.
0114
Following extensive phosphorylation of the C-terminal domain (CTD) of the largest subunit Rpb1 of Pol II by TFIIH, activated Pol II releases all GTFs except for TFIIF and proceeds to transcription elongation (Lu et al, 1992).
0030
In relation to transcriptional machinery, so far the structures of TBP (TATA box-binding protein) subunit (Nikolov et al, 1992) from TFIID, TBP–DNA (Kim et al, 1993a, 1993b), TBP–DNA–TFIIB (Nikolov et al, 1995), Pol II (Cramer et al, 2000, 2001) and Pol II–TFIIB (Bushnell et al, 2004) have been determined.
0107
With regard to the interaction between hTFIIE and hTFIIH, it has been shown that the C-terminal acidic region of hTFIIEα is necessary for native hTFIIH binding (Ohkuma et al, 1995) and hTFIIEα strongly binds to the p62 subunit of hTFIIH (Yamamoto et al, 2001; Okuda et al, 2004).

0114
Insight into the mechanism is gained from the work described here, which demonstrates that the C-terminal acidic domain (AC-D) of hTFIIEα containing the acidic region specifically binds to the N-terminal pleckstrin homology domain (PH-D) of the p62 subunit of hTFIIH.

0030
hTFIIEα AC-D is found to share its binding surface on p62 PH-D with the acidic transactivation domains (TADs) of tumour supressor protein p53 (Di Lello et al, 2006) and herpes simplex virus protein VP16 (Di Lello et al, 2005).
0423
Previous studies showed that hTFIIEα specifically bound to the p62 subunits of hTFIIH (Yamamoto et al, 2001; Okuda et al, 2004).

0676
To confirm this and to identify the AC-D-binding region in p62, we performed a GST pull-down assay using hTFIIEα AC-D and GST-fused p62 deletion mutants (Figure 2A).

0004
After purification by glutathione-Sepharose column chromatography, all samples containing the C-terminal region, namely full-length GST–p621−548, GST–p62109−548, GST–p62238−548 and GST–p62333−548, were considerably degraded or incompletely translated (data not shown).

0096
Though such instability of the C-terminal half of p62 has previously been reported (Jawhari et al, 2004), we found that full-length GST–p621−548, GST–p621−108, GST–p621−238 and GST–p621−333 bound to hTFIIEα AC-D, whereas no binding was observed with GST–p62109−548, GST–p62238−548, GST–p62333−548 and GST alone (Figure 2B).
0077
To obtain the p62 PH-D-bound structure of hTFIIEα AC-D using NMR, we performed NMR titration experiments in buffers both with and without 100 mM NaCl for both domains (Supplementary Figures 1–3). Although in the 100 mM NaCl buffer the dissociation constant (Kd) between AC-D and p62 PH-D was estimated from the titration plots as 376±81 nM (Supplementary Figure 1C) or 237±82 nM (Supplementary Figure 2C), the NaCl-free buffer NMR titration experiment showed much stronger binding affinity between AC-D and p62 PH-D because of the slow exchange timescale with a Kd below about 150 nM (Supplementary Figure 3).

0107
Although in the 100 mM NaCl buffer the dissociation constant (Kd) between AC-D and p62 PH-D was estimated from the titration plots as 376±81 nM (Supplementary Figure 1C) or 237±82 nM (Supplementary Figure 2C), the NaCl-free buffer NMR titration experiment showed much stronger binding affinity between AC-D and p62 PH-D because of the slow exchange timescale with a Kd below about 150 nM (Supplementary Figure 3).

0104
Although in the 100 mM NaCl buffer the dissociation constant (Kd) between AC-D and p62 PH-D was estimated from the titration plots as 376±81 nM (Supplementary Figure 1C) or 237±82 nM (Supplementary Figure 2C), the NaCl-free buffer NMR titration experiment showed much stronger binding affinity between AC-D and p62 PH-D because of the slow exchange timescale with a Kd below about 150 nM (Supplementary Figure 3).
0096
It is remarkable that in addition to the interaction involving the N-terminal flexible tail of hTFIIEα AC-D, its core structure also participates in the binding to p62 PH-D (Figure 3F).
0096
The ability of hTFIIEα mutants to bind to GST-tagged p62 PH-D was examined by in vitro binding assay (Figure 4B).

0096
We have shown previously that the N terminus of hTFIIEα is essential for binding to hTFIIEβ (Ohkuma et al, 1995). Consistent with these observations is the fact that none of the hTFIIEα mutations affected the binding of hTFIIEα to hTFIIEβ (Figure 4B, third column IIEβ, lanes 3–13).
0081
The mixture was analysed by SDS–PAGE and phosphorylated Rpb1 from Pol II was detected by autoradiography (Figure 4D). All mutants failed to stimulate CTD phosphorylation properly compared with wild type (Figure 4D, lanes 3–13 versus lane 2). Phosphorylation profiles of the above-described mutants (S365A, S365E, V372A, V372D, D380A and E383A) phosphorylated CTD but most of Rpb1 was detected at the hypo-phosphorylated IIa position (Figure 4D, lanes 3–8). In contrast, the p62 PH-D-binding defective mutants phosphorylated CTD only weakly (Figure 4D, lanes 9–12).
0030
Furthermore, the herpes simplex virus protein 16 (VP16) TAD also interacts with virtually identical sites of Tfb1 and p62 PH-D (Di Lello et al, 2005).

0081
Interestingly, the binding sites of p62 for p53 TAD2 and VP16 TAD significantly overlap with a part of the binding site for hTFIIEα AC-D.

0081
The binding site of p53 TAD2 peptide is disordered in an unbound state, but it forms a nine-residue amphipathic α-helix upon binding to Tfb1 PH-D and p62 PH-D.
0096
To analyse this interaction biochemically, several point mutants of p62 PH-D were created, bacterially expressed and used in binding studies with hTFIIEα AC-D and p531−73 (Figure 5C).

0096
As shown, K54, which forms the shallow pocket for F387 of hTFIIEα AC-D with its side chain interacting electrostatically with E389 of hTFIIEα AC-D, was the residue for which mutation to alanine had the largest effect as it prevented binding of all three hTFIIEα proteins tested (Figure 5C, lane 5). In addition, Q66, which also forms the same shallow pocket for F387 of hTFIIEα AC-D and the side chain of Q66 makes contact with F387 through amino-aromatic interaction, was shown to be essential for binding to hTFIIEα AC-D (Figure 5C, lane 6). The adjacent residues, V68, T74 and N76 of p62 PH-D as well as the N-terminal basic residues, K18 and K19, also affected binding but to a lesser extent (Figure 5C, lanes 3, 4, 7–9).

0096
In this case, Q66 was also central to the interaction but the essential binding residues were more widespread (Figure 5C, the bottom column, lanes 3 and 6–9).
0107
In the present study, the specific interaction between hTFIIEα AC-D and hTFIIH p62 PH-D was explored, and structures of both the free and PH-D-bound forms of hTFIIEα AC-D were determined. This is the first report of the structural determination of the complex describing the interaction between TFIIE and TFIIH at the molecular level.

0104
The result was that the NMR signals of STDE showed no significant changes and the Kd of 400±43 nM was almost the same as that estimated using hTFIIEα AC-D.

0081
We also examined the binding ability of peptide possessing only an N-terminal tail (AC-D381−394) (Supplementary Figure 5).

0104
hTFIIEα AC-D381−394 bound to p62 PH-D with a Kd of 2123±192 nM, which is about ∼6- to 9-fold weaker than that of hTFIIEα AC-D. Although the residues of p62 PH-D whose signals changed significantly were mostly consistent with the binding of AC-D and AC-D381−394, the extents of signal changes in the C-terminal region, to which I395, R432 and M433 of AC-D bind, were reduced.

0114
These results clearly indicate that hTFIIEα AC-D, which contains both the core structure and the flexible tail, is necessary and sufficient for the specific binding. This is an entirely new binding mode compared with the canonical binding modes found in some transcriptional activators or repressors, in which an intrinsically disordered region (Dyson and Wright, 2002) of each activation or repression domain binds to a target protein with part of the flexible region forming an ordered structure upon binding to the target. The core structure of hTFIIEα AC-D is essential for its binding to p62 PH-D in addition to its flexible arm.

0030
This is an entirely new binding mode compared with the canonical binding modes found in some transcriptional activators or repressors, in which an intrinsically disordered region (Dyson and Wright, 2002) of each activation or repression domain binds to a target protein with part of the flexible region forming an ordered structure upon binding to the target.
0114
As seen in the complex structure, the highly conserved F387 and V390 residues in the acidic region of hTFIIEα AC-D, not in the STDE region (Figure 1A) make a large contribution to binding.

0676
As seen in the complex structure, the highly conserved F387 and V390 residues in the acidic region of hTFIIEα AC-D, not in the STDE region (Figure 1A) make a large contribution to binding.
0107
For yeast TFIIEα (Tfa1), functional significance of the C-terminal region and specific interaction between Tfa1 and Tfb1 has been reported (Bushnell et al, 1996; Kuldell and Buratowski, 1997). We aligned sequences of TFIIEα AC-D from other species to ascertain whether the interaction observed in human is evolutionally conserved (Figure 1A).

0114
Considering that the main binding site of hTFIIEα AC-D is located on the N-terminal tail outside the core structure, Tfa1 does not seem to have a similar core AC-D structure, but the interaction with PH-D is likely to be conserved.
0096
The reason why transcription did not correlate well with the binding and CTD phosphorylation defects might be because there are more than 25 proteins involved in transcription, whereas only limited factors were used for both binding and phosphorylation studies (hTFIIEα AC-D mutants and p62 PH-D were used for binding studies and hTFIIEα AC-D mutants, hTFIIH and Pol II were used for CTD phosphorylation).
0096
In many cases, acidic TADs are disordered in an unbound form, but form an amphipathic helix upon binding to target proteins, for example, p53 TAD2–RPA70 (replication protein A 70) (Bochkareva et al, 2005), p53 TAD1–MDM2 (ubiquitin ligase) (Kussie et al, 1996), VP16 TAD–hTAFII31 (human TBP-associated factor) (Uesugi et al, 1997) complexes as well as the recently determined Tfb1 PH-D–p53 TAD2 complex (Di Lello et al, 2006).

0018
In many cases, acidic TADs are disordered in an unbound form, but form an amphipathic helix upon binding to target proteins, for example, p53 TAD2–RPA70 (replication protein A 70) (Bochkareva et al, 2005), p53 TAD1–MDM2 (ubiquitin ligase) (Kussie et al, 1996), VP16 TAD–hTAFII31 (human TBP-associated factor) (Uesugi et al, 1997) complexes as well as the recently determined Tfb1 PH-D–p53 TAD2 complex (Di Lello et al, 2006).

0030
In many cases, acidic TADs are disordered in an unbound form, but form an amphipathic helix upon binding to target proteins, for example, p53 TAD2–RPA70 (replication protein A 70) (Bochkareva et al, 2005), p53 TAD1–MDM2 (ubiquitin ligase) (Kussie et al, 1996), VP16 TAD–hTAFII31 (human TBP-associated factor) (Uesugi et al, 1997) complexes as well as the recently determined Tfb1 PH-D–p53 TAD2 complex (Di Lello et al, 2006).

0104
The Kd values for the binding of p53 TAD2 to p62 and Tfb1 PH-Ds determined by isothermal titration calorimetry (ITC) are 3175±570 and 391±74 nM, respectively (Di Lello et al, 2006). In the binding of VP16 TAD to Tfb1 PH-D, the Kd value estimated by NMR titration experiment was ∼4000–7000 nM (Di Lello et al, 2005). Compared with these Kd values, the binding of hTFIIEα AC-D to p62 PH-D is rather strong.

0077
In the binding of VP16 TAD to Tfb1 PH-D, the Kd value estimated by NMR titration experiment was ∼4000–7000 nM (Di Lello et al, 2005).

0107
In the binding of VP16 TAD to Tfb1 PH-D, the Kd value estimated by NMR titration experiment was ∼4000–7000 nM (Di Lello et al, 2005). Compared with these Kd values, the binding of hTFIIEα AC-D to p62 PH-D is rather strong.

0889
p62 has been shown to interact with the TADs of not only VP16 and p53 (Xiao et al, 1994) but also E2F-1 (Pearson and Greenblatt, 1997), and the oestrogen receptor α (ERα) (Chen et al, 2000).

0402
It can be imagined that if TFIIH is recruited by an activator near the promoter through its TAD then the recruited TFIIH could be captured by TFIIE instead of the activator to form the PIC. TFIIE regulates the enzymatic activities of TFIIH, which are necessary for the next stage after the PIC formation, that is, promoter melting or promoter clearance.
0104
The Kd values of p53 TAD2 to p62 PH-D are reported for the unphosphorylated form as 3175±570 nM, for the S46-phosphorylated form as 518±92 nM, for the T55-phosphorylated form as 457±75 nM and for both S46- and T55-phosphorylated form as 97±33 nM. Very recent ITC studies demonstrated the Kd value of hTFIIEα336−439 to p62 PH-D to be 45±25 nM (Di Lello et al, 2008).

0889
Very recent ITC studies demonstrated the Kd value of hTFIIEα336−439 to p62 PH-D to be 45±25 nM (Di Lello et al, 2008).
0411
The NdeI–BamHI fragment of mutated hTFIIEα cDNA was subcloned into the pET11d vector (Novagen) making the N-terminal hexa histidine-tagged hTFIIEα (6H–hTFIIEα) expression plasmid. The NdeI–BamHI fragment of mutated hTFIIH p62 cDNA was subcloned into the pET vector making the N-terminal FLAG-tagged hTFIIH p62 PH-D (FLAG–p62 PH) expression plasmid.
0004
The eluate was then applied onto Q-Sepharose (GE Healthcare). After digestion with thrombin to remove the 6H tag, sample was again loaded onto the Ni-NTA agarose column. Fractions passing through the column were concentrated and applied onto Superdex30 (GE Healthcare).

0071
After digestion with thrombin to remove the 6H tag, sample was again loaded onto the Ni-NTA agarose column. Fractions passing through the column were concentrated and applied onto Superdex30 (GE Healthcare).
0096
Recombinant point mutant hTFIIEα proteins were expressed in E. coli Rosetta™(DE3)pLysS (Novagen), and recombinant hTFIIH p62 point mutants were expressed in BL21(DE3)pLysS by induction with isopropyl-β-D-thiogalactopyranoside (Studier et al, 1990). The purification method of these recombinant proteins was as described previously (Watanabe et al, 2003).
0096
GST fusion proteins were used for protein interaction assays (Okamoto et al, 1998).

0004
GST fusion proteins were used for protein interaction assays (Okamoto et al, 1998).

0018
GST fusion proteins were used for protein interaction assays (Okamoto et al, 1998).

0030
GST fusion proteins were used for protein interaction assays (Okamoto et al, 1998).

0007
The bound proteins were released by boiling in SDS–PAGE loading buffer, separated by SDS–PAGE and detected by western blotting with anti-hTFIIEα rabbit antiserum (1:3000 dilution), anti-FLAG M2 monoclonal antibody (Sigma) and anti-p53 (DO-1) (Santa Cruz) using the enhanced chemiluminescence detection system (GE Healthcare).

0676
The bound proteins were released by boiling in SDS–PAGE loading buffer, separated by SDS–PAGE and detected by western blotting with anti-hTFIIEα rabbit antiserum (1:3000 dilution), anti-FLAG M2 monoclonal antibody (Sigma) and anti-p53 (DO-1) (Santa Cruz) using the enhanced chemiluminescence detection system (GE Healthcare).
0077
Left, superposition of the backbone heavy atoms of the 20 lowest energy NMR structures.
0114
The pleckstrin homology domain (PH-D) of the p62 subunit of hTFIIH directly binds to hTFIIEα AC-D.
0077
(A) Superposition of the backbone heavy atoms of the 20 lowest energy NMR structures.
0019
Interacting residues are indicated with the side chains and interaction areas are roughly marked in yellow.

0007
The top panel shows 5 ng of FLAG-tagged p62 PH-D mutant proteins (p62 PH-D mutants).

0007
Bound mutants were subjected to SDS–PAGE and detected by western blotting with anti-FLAG antibody.

0007
Bound proteins were subjected to SDS–PAGE and detected by western blotting with anti-FLAG, anti-p53 and anti-hTFIIEα antisera.
0413
Vav3 activated ERα partially via PI3K-Akt signaling and stimulated growth of breast cancer cells. Vav3 also potentiated EGF activity for cell growth and ERα activation in breast cancer cells.
0096
Vav proteins contain multiple function motifs and are involved in various cellular signaling processes, including cytoskeleton organization, calcium influx, phagocytosis, and cell transformation [3]. Vav proteins share a common structure, including a N-terminal calponin homology (CH) domain involved in Ca+2 mobilization and transforming activity, an acidic domain (AD) containing three regulatory tyrosines, a Dbl homology (DH) domain with a conserved region that promotes the exchange of GDP for GTP on Rac/Rho GTPases, a pleckstrin homology (PH) domain binding to PIP3 that enables its movement to the inner face of the plasma membrane, two Src-homology 3 (SH3) domains interacting with proteins containing proline-rich sequences, and a Src-homology 2 (SH2) domain interacting with proteins containing phosphorylated tyrosines [4,5].

0114
Vav proteins share a common structure, including a N-terminal calponin homology (CH) domain involved in Ca+2 mobilization and transforming activity, an acidic domain (AD) containing three regulatory tyrosines, a Dbl homology (DH) domain with a conserved region that promotes the exchange of GDP for GTP on Rac/Rho GTPases, a pleckstrin homology (PH) domain binding to PIP3 that enables its movement to the inner face of the plasma membrane, two Src-homology 3 (SH3) domains interacting with proteins containing proline-rich sequences, and a Src-homology 2 (SH2) domain interacting with proteins containing phosphorylated tyrosines [4,5].

0419
Vav proteins share a common structure, including a N-terminal calponin homology (CH) domain involved in Ca+2 mobilization and transforming activity, an acidic domain (AD) containing three regulatory tyrosines, a Dbl homology (DH) domain with a conserved region that promotes the exchange of GDP for GTP on Rac/Rho GTPases, a pleckstrin homology (PH) domain binding to PIP3 that enables its movement to the inner face of the plasma membrane, two Src-homology 3 (SH3) domains interacting with proteins containing proline-rich sequences, and a Src-homology 2 (SH2) domain interacting with proteins containing phosphorylated tyrosines [4,5].

0413
Vav proteins share a common structure, including a N-terminal calponin homology (CH) domain involved in Ca+2 mobilization and transforming activity, an acidic domain (AD) containing three regulatory tyrosines, a Dbl homology (DH) domain with a conserved region that promotes the exchange of GDP for GTP on Rac/Rho GTPases, a pleckstrin homology (PH) domain binding to PIP3 that enables its movement to the inner face of the plasma membrane, two Src-homology 3 (SH3) domains interacting with proteins containing proline-rich sequences, and a Src-homology 2 (SH2) domain interacting with proteins containing phosphorylated tyrosines [4,5].

0419
Upon phosphorylation of the tyrosines in the AD domain, the folding is opened and the DH domain is exposed. Thus, Vav protein is activated and interacts with substrate proteins, and the PH domain is exposed for PIP3 binding [6].

0096
Thus, Vav protein is activated and interacts with substrate proteins, and the PH domain is exposed for PIP3 binding [6].

0004
Thus, Vav protein is activated and interacts with substrate proteins, and the PH domain is exposed for PIP3 binding [6].

0018
Thus, Vav protein is activated and interacts with substrate proteins, and the PH domain is exposed for PIP3 binding [6].
0413
Recently, we and others found that Vav3 oncogene is overexpressed in androgen-independent prostate cancer cells, enhances androgen receptor (AR) activity, and stimulates androgen-independent growth in prostate cancer cells [25,26]. We further showed that Vav3, as a signal transducer, upregulates AR activity partially via PI3K-Akt signaling [25].

0413
Vav3 also potentiates EGF activity for cell growth and AR activation in prostate cancer cells.
0413
Vav3 stimulates growth of breast cancer cells and activates ERα partially via PI3K-Akt signaling. Vav3 potentiates EGF activity for cell growth and ERα activation in breast cancer cells.
0004
Briefly, aliquots of samples with the same amount of protein, determined using the Bradford assay (BioRad, Hercules, CA), were mixed with loading buffer (final concentrations of 62.5 mM Tris-HCl, pH 6.8, 2.3% SDS, 100 mM dithiothreitol, and 0.005% bromophenol blue), boiled, fractionated in a SDS-PAGE, and transferred onto a 0.45-um nitrocellulose membrane (BioRad).
0004
GST-Vav3-DH+PH and control GST vectors were transformed into BL21 bacteria, respectively (Protein Express, Inc.

0004
The transformed bacteria were cultured in L-Broth with addition of 100 uM of IPTG to induce GST-fusion protein expression. Then, the bacteria were harvested and subjected to GST fusion protein purification by Sonication and using Glutathione Sepharose 4B (Amersham Bioscience).

0096
Then, the bacteria were harvested and subjected to GST fusion protein purification by Sonication and using Glutathione Sepharose 4B (Amersham Bioscience).

0676
Then, the bacteria were harvested and subjected to GST fusion protein purification by Sonication and using Glutathione Sepharose 4B (Amersham Bioscience).
0413
For pull down reaction, 5~10 ug of GST or GST-Vav3-DH+PH was incubated with 1 mg of cell extracts from MCF7 cells in a binding buffer [20 mM of Tris.CL, PH.

0004
For pull down reaction, 5~10 ug of GST or GST-Vav3-DH+PH was incubated with 1 mg of cell extracts from MCF7 cells in a binding buffer [20 mM of Tris.CL, PH. 7.9; 300 mM of KCL; 0.05% of NP-40; 0.2 mM of EDTA; 20% of Glycerol; 1 mM of Dithiothritol; 1 mM of phenylmethylsulfonyl fluoride (PMSF), 1× of protease inhibitor cocktail (Roche Diagnostics)] for overnight [28,29].
0416
In contrast, Vav3 staining, detected in both cytoplasm and nucleus of the epithelial cells but not in stroma of breast tissues, was found in 35 out of 43 tumor tissue sections (35/43, 81%, p < 0.0001).
0411
Vav3 is involved in growth of breast cancer cells. (A) Expression analysis of Vav3 and ERα in breast cancer MCF7 and T47D cells and nontumoral breast epithelial MCF-10A cells. The cell extracts were prepared from T47D, MCF7, and MCF-10A cells and subjected to western blot analysis for Vav3 and ERα.

0411
(B) MCF7 cells were transiently transfected with Vav3 expression vector or control empty vector and then cultured in stripped medium in the absence or presence of EGF for 5 days, followed by MTT assay.

0104
The data was presented as absorbance at OD 570 nM.

0411
(C) MCF7 cells were transiently transfected with Vav3 expression vector or empty pHEF vector for 3 days, followed by cell extracts preparation and western blot analysis for Vav3.

0104
The data was presented as absorbance at OD 570 nM.
0411
We then determined whether knockdown expression of Vav3 inhibits growth of these breast cancer cells using Vav3 siRNA that has been characterized previously [25].

0413
Lipofectamine 2000 used for transfection of siRNA has been shown to knock down more than 80% activity of the endogenous gene in a panel of cells tested (Invitrogen).

0426
Lipofectamine 2000 used for transfection of siRNA has been shown to knock down more than 80% activity of the endogenous gene in a panel of cells tested (Invitrogen).

0411
A growth stimulatory effect in response to estrogen stimulation was observed in T47D and MCF7 cells (Figure 2Eand 2F). We found that knockdown expression of Vav3 significantly inhibited both estrogen-dependent and -independent growth in these breast cancer cells (Figure 2Eand 2F). At 1.2 pM of siRNA, the growth of T47D and MCF7 cells was inhibited by 64% and 68% in the absence of estrogen and 46% and 77% in the presence of estrogen relative to their respective controls.
0114
Previous studies have revealed that Vav3*, a Vav3 mutant with N-terminal domain deletion including the AD domain containing three tyrosine residues, is a constitutive active form and has much stronger oncogenic effect relative to that by wild type Vav3 [30]. Structure analysis of Vav protein suggested that in the non-phosphorylation state, Vav protein is folded, which is achieved by binding of tyrosines in the AD domain to the DH domain and binding of the CH domain to the C1 region [6].

0018
Phosphorylation of the tyrosines in the AD domain results in unfolding of Vav protein and exposure of the DH domain for interacting with substrate proteins.
0411
Hela cells were transiently cotransfected with a luciferase reporter driven by the estrogen response element (ERE), ERα expression vector, and either Vav3 or Vav3* expression vector.

0413
Upregulation of the endogenous ERα activity by Vav3 relative to the empty vector pHEF was also confirmed in MCF7 cells by stimulation of luciferase reporter expression driven by ERE (Figure 3B) and natural promoter of ERα target gene pS2 (Figure 3C). Vav3 also significantly enhanced ERα activity stimulated by sub-physiological concentrations of estrogen (10-11 and 10-10 M) (Figure 3B).

0402
Upregulation of the endogenous ERα activity by Vav3 relative to the empty vector pHEF was also confirmed in MCF7 cells by stimulation of luciferase reporter expression driven by ERE (Figure 3B) and natural promoter of ERα target gene pS2 (Figure 3C).

0411
Upregulation of the endogenous ERα activity by Vav3 relative to the empty vector pHEF was also confirmed in MCF7 cells by stimulation of luciferase reporter expression driven by ERE (Figure 3B) and natural promoter of ERα target gene pS2 (Figure 3C).
0411
(A) Hela cells (105 cells/well in 12-well plate) were cotransfected with ERE-Luc (0.5 ug), expression vectors for Vav3, Vav3*, or empty vector pHEF (200 ng) and ERα (50 ng), respectively. Then, the cells were treated without or with E2 (10-9 M) and without or with Tamoxifen. (B and C) MCF7 cells (105 cells/well in 12-well plate) were cotransfected with ERE-Luc (0.5 ug) (B) or pS2-Luc (0.5 ug) (C), and expression vector (0.25 ug) for Vav3*, or empty vector pHEF, respectively. Then, the cells were treated with E2.
0081
We next examined the domains of Vav3 required for upregulation of ERα activity by analyzing both wild type Vav3 and Vav3 deletion mutants (Figure 4A). We found that Vav3* demonstrated a much stronger activity for ERα activation relative to Vav3 as compared with the empty vector pHEF, supporting the previous observation that activity of Vav3 is subjected to regulation by phosphorylation (Figure 4Band Figure 5D) [6]. In addition, deletion of the DH domain, but not the SH domain or the CH domain, abolished Vav3 function for ERα activation (Figure 4B).

0004
The truncated Vav3 protein with deletion of both PH and SH domains failed to activate ERα.
0411
(B) Hela cells (105 cells/well in 12-well plate) were cotransfected with ERE-Luc (0.5 ug) and expression vectors (200 ng) for Vav3, Vav3*, Vav3*-ΔDH, Vav3*-ΔSH, or empty vector pHEF, as well as expression vector (50 ng) for ERα, respectively.
0411
(A) Hela cells were cotransfected with ERE-Luc (0.25 ug), expression vectors ERα (25 ng), Vav3* or empty vector pHEF (50 ng), p85+p110 or empty vector pCR3.1 (50 ng). (B) Hela cells were cotransfected with ERE-Luc reporter (0.25 ug/well in 12-well plate), expression vectors ERα (25 ng), dominant-negative Akt expression vector or empty vector pCR3.1 (0.1 ug), and Vav3* expression vector or empty vector pHEF (0.1 ug), respectively. (C) T47D cells were cotransfected with ERE-Luc (0.5 ug) and expression vector for Vav3 or empty vector pHEF (0.25 ug). Then, the cells were treated with Wortmannin (0.5 um). (D) T47D cells were cotransfected with pS2-Luc (0.5 ug) and expression vector for Vav3 or empty vector pHEF (0.25 ug). Then, the cells were treated with EGF (20 ng/ml).
0411
Hela cells were transiently cotransfected with ERE-Luc reporter, and expression vectors ERα, Vav3*, and p85 (the regulatory subunit of PI3K) and p110 (the catalytic subunit of PI3K).

0413
As shown in Figure 5A, cotransfection of Vav3* or PI3K relative to the empty vector pHEF or pCR3.1 stimulated ERα activity. ERα activity was further enhanced in cells cotransfected with both Vav3* and PI3K.
0004
A nuclear localization signal (NLS) in the PH domain was shown to be solely responsible for the nucleus localization of Vav1 protein, indicating a role of Vav family protein as a transcription coregulator.

0419
A nuclear localization signal (NLS) in the PH domain was shown to be solely responsible for the nucleus localization of Vav1 protein, indicating a role of Vav family protein as a transcription coregulator. We have demonstrated that the DH domain of Vav3 is essential for AR and ERα activation (Figure 4) [25].

0081
We performed sequence analysis of Vav proteins and found that the DH domain of Vav3 contains three consensus sequences of the LXXLL motifs or NR boxes, which have been well characterized and involved in interaction with nuclear receptors (Figure 6A).

0096
We performed sequence analysis of Vav proteins and found that the DH domain of Vav3 contains three consensus sequences of the LXXLL motifs or NR boxes, which have been well characterized and involved in interaction with nuclear receptors (Figure 6A).

0419
In addition, homologous analysis of Vav3 and Vav1 genes identified a conserved NLS in the PH domain of Vav3 (Figure 6B).
0419
(B) A consensus sequence of the nucleus localization signal (NLS) in Vav3 was localized in the PH domain.
0107
We performed GST pull down experiment to confirm the interaction between Vav3 and ERα.

0004
A GST fusion protein including the DH and PH domain of Vav3 (GST-Vav3-DH+PH) was generated (Figure 7A). Cell extract derived from MCF7 cells was incubated with immobilized GST-Vav3-DH+PH fusion protein or GST protein.

0081
A GST fusion protein including the DH and PH domain of Vav3 (GST-Vav3-DH+PH) was generated (Figure 7A).

0419
A GST fusion protein including the DH and PH domain of Vav3 (GST-Vav3-DH+PH) was generated (Figure 7A).

0004
We found that GST-Vav3-DH+PH fusion protein, but GST protein, interacted with ERα (Figure 7C).

0419
In summary, we found that Vav3 contains NLS in the PH domain and three LXXLL motifs in the DH domain. The deletion mutation and functional analysis by luciferase reporter assay showed that the DH domain of Vav3 is essential for enhancing ERα activity and is involved in complex with ERα.
0004
(A) GST-Vav3-DH+PH fusion protein. (B) GST-Vav3-DH+PH fusion protein (lane 3 and 4) and control GST protein (lane 1 and 2) were subjected to pull down reaction in the absence (lane 1 and 3) and presence (land 2 and 4) of cell extract derived from MCF7 cells.
0411
The knockdown expression of Vav3 compromised both estrogen-stimulated and -independent growth of breast cancer cells.
0419
Vav3 is activated upon ligand stimulation of EGF, insulin, Ros, and IGF receptors and physically associates with a variety of signaling molecules, including Rac1, Cdc42, PI3K, Grb2, and PLC-γ, leading to alteration in cell morphology and cell transformation [33].
0004
A nuclear localization signal (NLS) in the PH domain is solely responsible for nucleus localization of Vav1 protein, indicating a role of Vav family proteins as a transcription coregulator.

0018
A nuclear localization signal (NLS) in the PH domain is solely responsible for nucleus localization of Vav1 protein, indicating a role of Vav family proteins as a transcription coregulator.

0419
A nuclear localization signal (NLS) in the PH domain is solely responsible for nucleus localization of Vav1 protein, indicating a role of Vav family proteins as a transcription coregulator. We found that Vav3 contains the LXXLL motifs in the DH domain and NLS in the PH domain.
0419
Our findings support the notion that Vav3 overexpression may play a role in breast cancer, based on the following reasons: 1) Vav3 is overexpressed and correlated with poorly differentiated tumors in human breast cancer; 2) Vav3 contains the LXXLL motifs and complexes with ERα; 3) Vav3 enhances ERα activity partially via the PI3K-Akt pathway; 4) Vav3 is a protein with multiple domains and functions, including the SH2 domain interacting with receptor protein tyrosine kinase, the PH domain binding PIP3 involved in association with the cell membrane, and the DH domain involved in interaction with ERα; 5) Vav3 potentiates EGF for cell growth and ERα activation.

0413
Our findings support the notion that Vav3 overexpression may play a role in breast cancer, based on the following reasons: 1) Vav3 is overexpressed and correlated with poorly differentiated tumors in human breast cancer; 2) Vav3 contains the LXXLL motifs and complexes with ERα; 3) Vav3 enhances ERα activity partially via the PI3K-Akt pathway; 4) Vav3 is a protein with multiple domains and functions, including the SH2 domain interacting with receptor protein tyrosine kinase, the PH domain binding PIP3 involved in association with the cell membrane, and the DH domain involved in interaction with ERα; 5) Vav3 potentiates EGF for cell growth and ERα activation.

0004
Our findings support the notion that Vav3 overexpression may play a role in breast cancer, based on the following reasons: 1) Vav3 is overexpressed and correlated with poorly differentiated tumors in human breast cancer; 2) Vav3 contains the LXXLL motifs and complexes with ERα; 3) Vav3 enhances ERα activity partially via the PI3K-Akt pathway; 4) Vav3 is a protein with multiple domains and functions, including the SH2 domain interacting with receptor protein tyrosine kinase, the PH domain binding PIP3 involved in association with the cell membrane, and the DH domain involved in interaction with ERα; 5) Vav3 potentiates EGF for cell growth and ERα activation.

0096
Our findings support the notion that Vav3 overexpression may play a role in breast cancer, based on the following reasons: 1) Vav3 is overexpressed and correlated with poorly differentiated tumors in human breast cancer; 2) Vav3 contains the LXXLL motifs and complexes with ERα; 3) Vav3 enhances ERα activity partially via the PI3K-Akt pathway; 4) Vav3 is a protein with multiple domains and functions, including the SH2 domain interacting with receptor protein tyrosine kinase, the PH domain binding PIP3 involved in association with the cell membrane, and the DH domain involved in interaction with ERα; 5) Vav3 potentiates EGF for cell growth and ERα activation.

0114
Our findings support the notion that Vav3 overexpression may play a role in breast cancer, based on the following reasons: 1) Vav3 is overexpressed and correlated with poorly differentiated tumors in human breast cancer; 2) Vav3 contains the LXXLL motifs and complexes with ERα; 3) Vav3 enhances ERα activity partially via the PI3K-Akt pathway; 4) Vav3 is a protein with multiple domains and functions, including the SH2 domain interacting with receptor protein tyrosine kinase, the PH domain binding PIP3 involved in association with the cell membrane, and the DH domain involved in interaction with ERα; 5) Vav3 potentiates EGF for cell growth and ERα activation.
0096
KL generated GST fusion protein for pull down analysis.

0004
KL generated GST fusion protein for pull down analysis.
0054
Although Bcl-XL and Bax are structurally similar, activated Bax forms large oligomers that permeabilize the outer mitochondrial membrane, thereby committing cells to apoptosis, whereas Bcl-XL inhibits this process.

0426
Although Bcl-XL and Bax are structurally similar, activated Bax forms large oligomers that permeabilize the outer mitochondrial membrane, thereby committing cells to apoptosis, whereas Bcl-XL inhibits this process.

0096
It has been difficult to sort out which interaction is important in cells, as all three proteins are present simultaneously. We examined the mechanism of Bax activation by tBid and its inhibition by Bcl-XL using full-length recombinant proteins and measuring permeabilization of liposomes and mitochondria in vitro.
0054
During development and under stress, cells can become committed to die via programmed cell death (apoptosis).

0426
During development and under stress, cells can become committed to die via programmed cell death (apoptosis).

0096
It has been difficult to sort out which interaction is important in cells, where multiple members of all three protein families are present simultaneously. Here, we describe an in vitro system containing the three recombinant proteins and the use of mutagenesis to selectively remove individual interactions. We show that Bcl-XL inhibits Bax by competing with it for binding to membranes, tBid, and activated Bax.
0096
Bcl-XL and Bax are structurally similar members of the Bcl-2 family of cell-death-related proteins, and they compete for binding to membranes, as well as to Bcl-2 family member tBid and activated Bax.
0096
Another large subgroup of the Bcl-2 family (BH3-only proteins; e.g., tBid) initiates apoptosis through binding to Bax and/or Bcl-XL. Even though Bcl-XL and Bax are structurally similar, experiments with protein fragments and peptides or full-length protein in the absence of membranes has led to the elaboration of models in which the functional relevance of binding partners for Bcl-XL differ.

0096
However, there is also considerable overlap between these two competing models: both models recognize that Bcl-XL directly binds to and inhibits a proapoptotic Bcl-2 family protein that is directly involved in membrane permeabilization, while other proapoptotic Bcl-2 family proteins indirectly induce apoptosis by binding to Bcl-XL and preventing this function.

0018
However, there is also considerable overlap between these two competing models: both models recognize that Bcl-XL directly binds to and inhibits a proapoptotic Bcl-2 family protein that is directly involved in membrane permeabilization, while other proapoptotic Bcl-2 family proteins indirectly induce apoptosis by binding to Bcl-XL and preventing this function.

0426
However, there is also considerable overlap between these two competing models: both models recognize that Bcl-XL directly binds to and inhibits a proapoptotic Bcl-2 family protein that is directly involved in membrane permeabilization, while other proapoptotic Bcl-2 family proteins indirectly induce apoptosis by binding to Bcl-XL and preventing this function.

0096
On the basis of these data and evidence from the literature, we have proposed a model termed “embedded together”, in which interaction of these proteins with each other changes after binding to the membrane as this causes conformational changes that alter and/or expose new binding surfaces [10,11].
0096
To examine the molecular mechanism of membrane permeabilization for full-length Bcl-XL and Bax, we used recombinant proteins and measured permeabilization of both liposomes and mitochondrial outer membranes in vitro. Recombinant tBid was used as an activator protein.
0096
To eliminate complications from all other known and unknown proteins, metabolites, and post-translational modifications that may contribute additional levels of regulation, we used a cell-free system composed of highly purified full-length recombinant proteins, without N- or C-terminal tags, and as a source of membranes either liposomes with mitochondria-like composition [16] or subcellular fractions containing mitochondria. As a representative activator protein, we used the caspase-8 cleavage product of Bid (tBid) to activate Bax.

0018
This protein drives Bax/Bak-dependent permeabilization of mitochondria [17,18] and also binds directly to antiapoptotic proteins such as Bcl-XL and Bcl-2 [19].
0426
The large increase in fluorescence that accompanies the release of the fluorophore/quencher pair 8-aminonaphthalene 1,3,6-trisulfonic acid (ANTS)/p-xylene-bis-pyridinium (DPX) from liposomes was used to measure Bax-dependent membrane permeabilization and its inhibition by Bcl-XL [7].

0096
As seen previously, single addition of recombinant Bax (100 nM) or tBid (20 nM) to liposomes had little effect, but in combination the two proteins caused an increase in fluorescence due to membrane permeabilization (Figure 1A).

0663
As seen previously, single addition of recombinant Bax (100 nM) or tBid (20 nM) to liposomes had little effect, but in combination the two proteins caused an increase in fluorescence due to membrane permeabilization (Figure 1A).
0071
Membrane-bound proteins were separated from soluble proteins by Sepharose CL-2B gel filtration chromatography.

0019
Individual fractions were analyzed by immunoblotting (IB) using Bid, Bax, or Bcl-XL antibodies, as indicated.
0019
(C) Mitochondria from bak knockout mice were incubated with tBid (20 nM), Bax (100 nM), and Bcl-XL (100 nM), as indicated. Permeabilization was assayed by pelleting the mitochondria and analyzing both the pellet (P) and the supernatant (S) fractions by immunoblotting using an α-cytochrome c antibody.
0071
In incubations of 20 nM tBid and liposomes, tBid bound effectively to liposomes as assayed by gel filtration chromatography (Figure 1B).
0413
At all effective concentrations of Bcl-XL, membrane binding by Bax was efficiently inhibited (Figures 1B and 3), as has been observed previously in cells [22] and for mitochondria [18]. While Bax membrane binding was inhibited, Bcl-XL had no effect on tBid binding to membranes (Figure 1B).
0019
(B) Mitochondria isolated from bak knockout mouse livers (bak –/– MLM) were incubated with 200 nM Bax, 250 pM tBid or tBid-mt1, and increasing concentrations of Bcl-XL or Bcl-XL Y101K, as indicated. Permeabilization was assayed by pelleting the mitochondria and analyzing both the pellet (P) and the supernatant (S) fractions by immunoblotting using an α-cytochrome c antibody.
0096
To assess the binding of recombinant Bax to mitochondria, the organelles were incubated with or without tBid or tBid and Bcl-XL, pelleted by centrifugation, and analyzed by immunoblotting (Figure 1D).
0104
The correlation between tBid concentration (20 nM) and the IC50 of ∼25 nM Bcl-XL for inhibition of liposome permeabilization is consistent with other published models suggesting Bcl-XL sequestration of Bax/Bak activators (in this case tBid) as one mechanism by which Bcl-XL could inhibit membrane binding by Bax.
0104
Alone, Bcl-XL showed minimal liposome binding, consistent with cytoplasmic or loosely membrane-bound localizations reported for Bcl-XL in live cells [23,24] (Figure 1B). However, 20 nM tBid caused migration of ∼80 nM Bcl-XL to liposomes both when the two proteins were added together (Figure 1B, quantified in Figure 4B) and when tBid was bound to liposomes before Bcl-XL was added (unpublished data).

0413
Alone, Bcl-XL showed minimal liposome binding, consistent with cytoplasmic or loosely membrane-bound localizations reported for Bcl-XL in live cells [23,24] (Figure 1B).
0104
This interaction was not dependent on the detergent used to solubilize the liposomes (unpublished data), and at concentrations of 100 nM Bcl-XL and 20 nM tBid the interaction was not affected by the addition of Bax (Figure 2A, left panel, lanes 1 and 2).

0107
In the absence of membranes, interaction between Bcl-XL and tBid could not be detected by co-immunoprecipitation (unpublished data).
0006
Samples were immunoprecipitated (IP) in either 2% CHAPS or 0.2% NP-40, as indicated, using an antibody with the indicated specificity and immunoblotted (IB) for the indicated protein.

0007
Samples were immunoprecipitated (IP) in either 2% CHAPS or 0.2% NP-40, as indicated, using an antibody with the indicated specificity and immunoblotted (IB) for the indicated protein.
0019
Interaction of this peripheral membrane (indicated by the shadow) Bax with membrane-bound tBid causes a further conformational change such that Bax integrates in the membrane in an oligomerization competent form (step 2).

0096
Conversely, cytoplasmic Bax may interact with other activator proteins to integrate into membranes, or spontaneously active Bax molecules may integrate into the membrane without binding an activator protein.

0018
Conversely, cytoplasmic Bax may interact with other activator proteins to integrate into membranes, or spontaneously active Bax molecules may integrate into the membrane without binding an activator protein.
0107
When tested in the absence of tBid, a stable interaction between Bax and Bcl-XL was not detected (Figure 2A, left panel, lane 3), suggesting that Bcl-XL does not sequester Bax in solution.

0054
However as expected, co-immunoprecipitation of Bcl-XL and Bax was observed in control experiments where membranes were solubilized with the nonionic detergent NP-40, known to induce a conformational change in Bax required for heterodimerization with Bcl-XL that is also seen in cells when apoptosis is induced [20,26].

0426
However as expected, co-immunoprecipitation of Bcl-XL and Bax was observed in control experiments where membranes were solubilized with the nonionic detergent NP-40, known to induce a conformational change in Bax required for heterodimerization with Bcl-XL that is also seen in cells when apoptosis is induced [20,26].
0107
When tested in the presence of tBid (20 nM) and membranes, we did not detect an interaction between Bcl-XL (100 nM) and Bax (100 nM) (Figure 2A, left panel, lane 2).
0104
Consistent with this previous report, in our assay system tBid-mt1efficiently recruited Bax to membranes where Bax was activated (unpublished data), but tBid-mt1 did not coprecipitate with Bcl-XL (Figure 2B).

0096
As a negative control, we used Bcl-XL with a deletion of the BH4 domain (amino acids 4–24) that has been reported to remove binding to both Bax and the BH3-only protein Bad [29]. In our assay, ΔBH4 Bcl-XL did not coprecipitate tBid and bound very inefficiently to activated Bax (Figure 2D).
0096
By assay of ANTS/DPX release from liposomes for different combinations of these proteins, the residual function of Bcl-XL can be measured in the absence of stable interaction with tBid, Bax, or both.

0413
This indicates that a significant amount of Bcl-XL activity does not require Bcl-XL–tBid binding, as has been suggested previously by some [30] but not other models [3,31,32].
0096
In these incubations, substitution of the wild-type proteins with tBid-mt1, Bcl-XL Y101K, or both showed similar effects to those seen in bak –/– MLM and liposomes (Figure 3C). Thus, in contrast to widely promulgated models for the antiapoptotic mechanism of Bcl-XL, our results with purified proteins in the presence of relevant membrane targets indicate that Bcl-XL binding to Bax and tBid are both functionally relevant, but neither is paramount.
0096
Alternatively or in addition, Bcl-XL might directly inhibit Bax binding to membranes. To determine the mechanism(s) involved, we measured Bax liposome binding by gel filtration chromatography for reactions containing the different mutant proteins (Figure 4A).
0107
The absence of a stable interaction between tBid-mt1 and Bcl-XL significantly decreased Bcl-XL-mediated inhibition of Bax binding to membranes at all tBid concentrations assayed.
0018
It has been suggested that the multi-BH-region proapoptotic proteins Bax and Bak autoactivate after tBid (or another BH3-only protein) initiates the process and that autoactivation is inhibited by Bcl-2 [21,35].
0413
Similar to results obtained by examining liposome permeabilization, there is a residual activity of the tBid-mt1/Bcl-XL Y101K combination that prevented Bax binding to membranes even in the absence of a stable interaction of Bcl-XL Y101K with either tBid or Bax. This activity also was observed for a ΔBH4 mutant of Bcl-XL (Figure 4A, lanes 13, 14).
0054
At the onset of apoptosis in cells, Bcl-XL binds to membranes [23].

0426
At the onset of apoptosis in cells, Bcl-XL binds to membranes [23].

0096
Because the membrane appears to be the active locus, we used the mutant versions of Bcl-XL and tBid to examine the importance of stable binding to tBid or Bax for Bcl-XL to bind tightly to membranes (Figure 4B). In a negative control experiment without other added proteins, less than 10 nM of the added Bcl-XL (100 nM) bound to membranes (Figure 4B, lane 1).
0413
The functional importance of membrane binding by Bcl-XL is supported further by our observations that when Bcl-XL inhibits membrane permeability in assays containing tBid and Bax almost all of the Bcl-XL is membrane-bound and is in stoichiometric excess over membrane-bound tBid and Bax (Figure 4A and 4B, black bars), even when using mutants that prevent heterodimerization with one or both of the binding partners. Consistent with a role for membrane binding, removing the C-terminal tail that mediates membrane binding impaired but did not abolish Bcl-XL function (Figure S4), similar to the results obtained in cells using the same mutant [37]. However, removal of the C-terminal tail in the context of the tBid-mt1 and Bcl-XL Y101K mutations completely abolishes the remaining activity of Bcl-XL.
0019
Unlike the conformational change that accompanies tBid-induced insertion of Bax into membranes, the liposome-induced conformational change also disappears if liposomes are solubilized in CHAPS prior to immunoprecipitation (Figure 5A, compare lanes 1 and 3).
0006
Conformation-altered Bax was immunoprecipitated using the 6A7 antibody with or without the addition of 2% CHAPS to solubilize the liposomes prior to immunoprecipitation and analyzed by immunoblotting using an α-Bax antibody. The asterisk denotes the light chain of the 6A7 antibody.

0007
Conformation-altered Bax was immunoprecipitated using the 6A7 antibody with or without the addition of 2% CHAPS to solubilize the liposomes prior to immunoprecipitation and analyzed by immunoblotting using an α-Bax antibody. The asterisk denotes the light chain of the 6A7 antibody.
0096
To determine whether the effect of membrane-bound Bcl-XL on this transient, liposome-induced conformational change in Bax could be regulated, BH3 peptides were used to induce selectively Bcl-XL binding to membranes (Text S1). On the basis of the effects previously published for mutations in BH3 domains of proapoptotic proteins and peptides, we selected two peptides that both caused membrane binding by Bcl-XL (Figure S5A) but differed in their functional effects (Figure S5B).

0081
To determine whether the effect of membrane-bound Bcl-XL on this transient, liposome-induced conformational change in Bax could be regulated, BH3 peptides were used to induce selectively Bcl-XL binding to membranes (Text S1). On the basis of the effects previously published for mutations in BH3 domains of proapoptotic proteins and peptides, we selected two peptides that both caused membrane binding by Bcl-XL (Figure S5A) but differed in their functional effects (Figure S5B). One peptide, designated m1Bid BH3, containing a single mutation for a conserved and critical leucine residue, did not interfere with the antiapoptotic function of Bcl-XL.
0081
In the absence of Bcl-XL, neither peptide induced Bax membrane binding or Bax-dependent membrane permeabilization (unpublished data). When the m1Bid BH3 peptide triggered Bcl-XL binding to membranes, the Bcl-XL still inhibited the liposome-induced Bax conformational change (Figure 5F, top panel, lanes 2 and 3). In contrast, when the Bak BH3 peptide was added, membrane-bound Bcl-XL did not prevent the liposome-induced Bax conformational change (Figure 5F, bottom panel, lanes 2 and 3).

0413
Taken together, these results indicate that this activity is a regulatable function of membrane-bound Bcl-XL and is not merely the result of changes in the biophysical properties of liposomes after Bcl-XL binding (Figure 6D).
0054
Pioneering experiments to identify relevant binding partners for Bcl-2 and Bcl-XL using immunoprecipitation in transfected cells suggested a lack of correlation between Bax binding and inhibition of apoptosis, as only certain Bcl-XL point mutants that could no longer bind to Bax lost function [40].

0055
Pioneering experiments to identify relevant binding partners for Bcl-2 and Bcl-XL using immunoprecipitation in transfected cells suggested a lack of correlation between Bax binding and inhibition of apoptosis, as only certain Bcl-XL point mutants that could no longer bind to Bax lost function [40].

0413
Pioneering experiments to identify relevant binding partners for Bcl-2 and Bcl-XL using immunoprecipitation in transfected cells suggested a lack of correlation between Bax binding and inhibition of apoptosis, as only certain Bcl-XL point mutants that could no longer bind to Bax lost function [40].

0426
Pioneering experiments to identify relevant binding partners for Bcl-2 and Bcl-XL using immunoprecipitation in transfected cells suggested a lack of correlation between Bax binding and inhibition of apoptosis, as only certain Bcl-XL point mutants that could no longer bind to Bax lost function [40].

0096
Further analysis of two of these Bcl-XL mutants that did not bind to Bax (the F131V/D133A mutant that remained functional and the inactive G138E/R139L/I140N mutant) suggested that prevention of apoptosis required binding to the BH3-only proteins tBid, Bim, and Bad, a function that was specifically lost in the latter mutant [19]. However, as these experiments were conducted on whole cells where Bax was also present, it is possible that the lack of function of this mutant is caused by the loss of binding to both BH3-only proteins (e.g., tBid) and Bax , a result entirely consistent with our observations.

0413
However, as these experiments were conducted on whole cells where Bax was also present, it is possible that the lack of function of this mutant is caused by the loss of binding to both BH3-only proteins (e.g., tBid) and Bax , a result entirely consistent with our observations.

0426
Conversely, early work with the M97A/D98A Bid mutant that does not bind to Bcl-XL [27,30] that we have used in our study (tBid mt1) indicated that Bcl-XL inhibited the apoptosis caused by this Bid mutant, in cells and in purified mitochondria, implying that the interaction with activated Bax rather than Bid was critical to the antiapoptotic function of Bcl-XL in this context.

0019
Conversely, early work with the M97A/D98A Bid mutant that does not bind to Bcl-XL [27,30] that we have used in our study (tBid mt1) indicated that Bcl-XL inhibited the apoptosis caused by this Bid mutant, in cells and in purified mitochondria, implying that the interaction with activated Bax rather than Bid was critical to the antiapoptotic function of Bcl-XL in this context.

0054
Conversely, early work with the M97A/D98A Bid mutant that does not bind to Bcl-XL [27,30] that we have used in our study (tBid mt1) indicated that Bcl-XL inhibited the apoptosis caused by this Bid mutant, in cells and in purified mitochondria, implying that the interaction with activated Bax rather than Bid was critical to the antiapoptotic function of Bcl-XL in this context.
0018
Migration of the cytoplasmic fraction to membranes with insertion occurs during apoptosis [23,43,44] for Bcl-XL as well as other antiapoptotic proteins such as Mcl-1 [44] and Bcl-w [25,45]. For Bcl-w, interaction with a tethered Bim BH3 peptide displaces the C-terminal insertion sequence from the BH3 binding pocket within the protein triggering insertion of the protein into the membrane [25,46]. Although a structure has not been reported for Bcl-XL containing the C-terminal insertion sequence, the structure of the truncated protein is sufficiently similar to that of Bcl-w [46] and proapoptotic Bax [36] to suggest displacement of the insertion sequence of Bcl-XL as a mechanism that drives Bcl-XL into the membrane.

0096
Migration of the cytoplasmic fraction to membranes with insertion occurs during apoptosis [23,43,44] for Bcl-XL as well as other antiapoptotic proteins such as Mcl-1 [44] and Bcl-w [25,45]. For Bcl-w, interaction with a tethered Bim BH3 peptide displaces the C-terminal insertion sequence from the BH3 binding pocket within the protein triggering insertion of the protein into the membrane [25,46].

0081
We therefore determined whether peptides from the BH3 regions of Bid, Bim, Bad, Bax, and Bak caused insertion of Bcl-XL into liposomes, assayed by flotation on a sucrose gradient (Figure S6). All five peptides caused Bcl-XL to insert into membranes, while a mutant Bid peptide that fails to bind Bcl-XL [2] did not.
0019
Previous reports using transfected cells have indicated that the removal of the C-terminal insertion sequence of Bcl-XL severely impairs membrane insertion but has varying effects on Bcl-XL function, from moderate [29] to severe [50] reduction in function.

0426
Consistent with this variability, we have shown previously [37] that the function of ΔTM Bcl-XL compared to that of wild type is heavily dependent on the stimulus used to initiate apoptosis (and therefore possibly the BH3 protein(s) involved).

0054
However, in certain circumstances in cells ΔTM Bcl-XL may retain at least part of its antiapoptotic function in the cytoplasm depending on the location of relevant binding partners at the onset of apoptosis.

0413
However, in certain circumstances in cells ΔTM Bcl-XL may retain at least part of its antiapoptotic function in the cytoplasm depending on the location of relevant binding partners at the onset of apoptosis.

0426
However, in certain circumstances in cells ΔTM Bcl-XL may retain at least part of its antiapoptotic function in the cytoplasm depending on the location of relevant binding partners at the onset of apoptosis.
0096
The liposomes that we used to model physiologic membranes have a high intrinsic curvature and lipid composition that facilitate membrane binding of the recombinant proteins and induction of the 6A7 conformational change in Bax.

0054
In cells, these feature are likely represented by complex and dynamic physiologic processes such as mitochondrial membrane fission and fusion shown to be important in apoptosis [52–54] and the interaction of mitochondria with other membrane systems [39,52,55]. The major observations reported here were confirmed for cytochrome c release from mitochondria (Figure 3B–D), suggesting that they will be relevant in live cells.

0426
In cells, these feature are likely represented by complex and dynamic physiologic processes such as mitochondrial membrane fission and fusion shown to be important in apoptosis [52–54] and the interaction of mitochondria with other membrane systems [39,52,55]. The major observations reported here were confirmed for cytochrome c release from mitochondria (Figure 3B–D), suggesting that they will be relevant in live cells. Moreover, the simplicity and power of our in vitro system using recombinant proteins and membranes has allowed us to identify and measure functionally important and potentially “druggable” interactions important for the regulation of apoptosis (Figure 6).

0096
Moreover, the simplicity and power of our in vitro system using recombinant proteins and membranes has allowed us to identify and measure functionally important and potentially “druggable” interactions important for the regulation of apoptosis (Figure 6).
0676
Recombinant full-length human Bcl-XL (or Bcl-XL Y101K) with no additional amino acids was expressed in Escherichia coli as a C-terminal intein/chitin-binding domain fusion and purified by affinity chromatography on a chitin column followed by further purification on a phenyl-Sepharose column, similar to a method described previously [7] but with a final dialysis step to remove detergents.
0018
Samples were diluted to 1 mg/ml total protein (Bradford assay), incubated with purified proteins, added in the order Bcl-XL, Bax, and tBid, for 1 h at 30 °C, then centrifuged at 13,000 g for 10 min.

0096
Samples were diluted to 1 mg/ml total protein (Bradford assay), incubated with purified proteins, added in the order Bcl-XL, Bax, and tBid, for 1 h at 30 °C, then centrifuged at 13,000 g for 10 min.
0096
All sample components (buffers, liposomes, etc.) were added prior to the addition of recombinant proteins, which were added in the order Bcl-XL, Bax, and tBid.

0096
Membrane binding was measured by comparing the intensities of membrane-bound proteins (fractions 3 and 4) with total proteins (fractions 3 and 4 plus fractions 8–11).

0096
To assess the membrane binding and membrane insertion of Bax and Bcl-XL into mitochondria, purified proteins, added in the order Bcl-XL, Bax, and tBid, were incubated for 1 h at 30 °C with mitochondria (5 mg/ml) from bak –/– mice.
0007
Immunoprecipitation of Bcl-XL was performed using the polyclonal Bcl-XL antibody in assay buffer containing either 2% CHAPS or 0.2% NP-40.

0007
Immunoprecipitation using the conformation-specific 6A7 Bax antibody was performed on whole membranes and washed three times with assay buffer containing 2% CHAPS.
0018
elegans proteins share a common motif with Notch ligands from other species in a sequence defined here as the Delta and OSM-11 (DOS) motif. osm-11 loss-of-function defects in vulval development are exacerbated by loss of other DOS-motif genes or by loss of the Notch ligand DSL-1, suggesting that DOS-motif and DSL proteins act together to activate Notch signaling in vivo.
0018
elegans anterior pharynx defective-1 (APX-1) and DSL-1 are DSL domain–containing soluble proteins that function redundantly with LAG-2 during vulval development [35]. Noncanonical ligands for vertebrate Notch receptors have been identified, including Delta/notch-like EGF repeat containing protein (DNER), F3/contactin, and MAGP proteins [36–40], but functional C.
0081
elegans proteins, OSM-7 (T05D4.4), ZK507.4, K10G6.2, and K02F3.7, contain a signal peptide for secretion and a potential cEGF-1 domain [53] that is part of a conserved motif described below (Figure 2A).
0081
The signal peptide is shaded black, and putative O-linked glycosylation sites are indicated by vertical lines.
0114
The DOS motif overlaps with the first two EGF repeats of canonical Notch ligands and may define a unique subset of EGF repeats.
0081
elegans DOS-motif proteins are likely secreted based on the presence of a predicted N-terminal signal peptide. However, OSM-11 and DOS-3 also have a consensus proprotein convertase protease cleavage site and a C-terminal transmembrane domain, suggesting that they may be translated as transmembrane preproproteins prior to proteolytic processing and release of a soluble DOS protein.

0004
However, OSM-11 and DOS-3 also have a consensus proprotein convertase protease cleavage site and a C-terminal transmembrane domain, suggesting that they may be translated as transmembrane preproproteins prior to proteolytic processing and release of a soluble DOS protein.
0030
In known Notch ligands from Drosophila and vertebrates, the DOS motif is always located immediately following the DSL domain and overlapping the first two EGF repeats.

0114
In known Notch ligands from Drosophila and vertebrates, the DOS motif is always located immediately following the DSL domain and overlapping the first two EGF repeats.

0081
The first two EGF repeats of most Notch ligands differ from the remaining EGF repeats [27] (this study, Figure 2C).
0081
These proteins have a signal peptide sequence, and the DOS motif is located in the first two EGF repeats (Figure 2B).

0030
C901 is a predicted Drosophila protein of unknown function containing a DSL domain and multiple EGF repeats [56]; it is unclear whether C901 is a transmembrane DSL domain protein.
0055
These VPC fate decisions were assessed in osm-11(lf) animals and control animals at specific larval stages using the previously described green fluorescent protein (GFP) reporter constructs egl-17p::gfp, lin-11p::gfp, and lip-1p::gfp [60].
0104
In 71% of osm-11(lf) L4 animals, egl-17p::gfp expression in P5.p and/or P7.p descendants was lost, consistent with loss of 2° cell fates (unpublished data; n = 63). The aberrant egl-17p::gfp expression observed in osm-11(lf) animals suggests that 1° and 2° cell fates are not correctly specified in a fraction of osm-11(lf) animals, consistent with decreased Notch signaling.
0055
A transcriptional GFP reporter (osm-11p::gfp) was generated using the same upstream sequences used for osm-11 cDNA rescue. In animals harboring this transgene, GFP expression was observed in numerous unidentified cells during embryonic development from the comma stage onward (unpublished data). GFP expression was observed in the VPCs during larval development, as well as various hypodermal cells during larval stages (Figure 4).

0411
In animals harboring this transgene, GFP expression was observed in numerous unidentified cells during embryonic development from the comma stage onward (unpublished data). GFP expression was observed in the VPCs during larval development, as well as various hypodermal cells during larval stages (Figure 4).

0055
OSM-11 immunoreactivity was also observed in the seam cells of L1 larvae and adult animals (Figure 4A and 4D). In adult animals, osm-11p::gfp was expressed only in hypodermal seam cells in adult animals; hypodermal seam cell expression in adult animals was also confirmed with staining with OSM-11 antisera (Figure 4D). The larval hypodermal expression pattern of osm-11p::gfp is reminiscent of the osm-7p::gfp expression pattern described previously, but osm-7p::gfp expression in seam cells was not reported [52].

0411
OSM-11 immunoreactivity was also observed in the seam cells of L1 larvae and adult animals (Figure 4A and 4D). In adult animals, osm-11p::gfp was expressed only in hypodermal seam cells in adult animals; hypodermal seam cell expression in adult animals was also confirmed with staining with OSM-11 antisera (Figure 4D). The larval hypodermal expression pattern of osm-11p::gfp is reminiscent of the osm-7p::gfp expression pattern described previously, but osm-7p::gfp expression in seam cells was not reported [52]. OSM-11 protein was also expressed in the developing uterus of L4 larvae (Figure 4B) and in the spermatheca (Figure 4D); the LIN-12 Notch receptor plays a developmental role in these tissues as well [64], but only OSM-11 expression in VPCs was characterized further.
0411
(D) OSM-11 expression in seam cells and spermatheca in adult animals. An osm-11p::gfp reporter gene containing unc-54 3′ UTR sequences is expressed in adult seam cells (left); α-OSM-11 antisera was used to confirm seam cell and spermatheca expression (right). No OSM-11 was detected in neurons of larvae or adult animals (unpublished data); embryonic expression was not characterized.

0055
An osm-11p::gfp reporter gene containing unc-54 3′ UTR sequences is expressed in adult seam cells (left); α-OSM-11 antisera was used to confirm seam cell and spermatheca expression (right).

0426
An osm-11p::gfp reporter gene containing unc-54 3′ UTR sequences is expressed in adult seam cells (left); α-OSM-11 antisera was used to confirm seam cell and spermatheca expression (right).
0411
OSM-11 expression in Pn.p cells is consistent with a role for OSM-11 in initial cell fate specification.
0018
Because the predicted peptide sequence of OSM-11 contains a signal peptide, we tested whether OSM-11 is a secreted protein. When an osm-11 cDNA was expressed in Drosophila S2 tissue culture cells, OSM-11 protein accumulates in the media and not in cells (Figure 5A), consistent with OSM-11 acting in vivo as a soluble protein in the extracellular milieu.

0030
Because the predicted peptide sequence of OSM-11 contains a signal peptide, we tested whether OSM-11 is a secreted protein. When an osm-11 cDNA was expressed in Drosophila S2 tissue culture cells, OSM-11 protein accumulates in the media and not in cells (Figure 5A), consistent with OSM-11 acting in vivo as a soluble protein in the extracellular milieu.
0411
(A) Western blot of conditioned media from Drosophila S2 cells containing an OSM-11 cDNA expression construct or empty vector.
0018
To determine whether DOS-motif proteins have overlapping functions, we tested whether mutants defective in more than one DOS-motif protein had stronger vulval defects.
0423
We found that OSM-11 also interacted with LIN-12 extracellular EGF repeats 1 through 6 (Figure 8); OSM-11 did not interact with DSL-1 or LAG-2 ligands. We also confirmed previous studies [50] in which murine DLK1 EGF repeats 1 and 2 containing the DOS motif interacted specifically with murine Notch1 EGF repeats 12 and 13 in the same two-hybrid assay format (unpublished data).
0030
Dionne, et al., unpublished data), we also found that OSM-11 activates both LIN-12 and germline proliferation defective-1 (GLP-1) in the adult nervous system to regulate behavior.

0889
Dionne, et al., unpublished data), we also found that OSM-11 activates both LIN-12 and germline proliferation defective-1 (GLP-1) in the adult nervous system to regulate behavior.
0114
elegans has two Notch receptors (lin-12 and glp-1), ten DSL domain proteins that lack DOS motifs [35] and five DOS-motif proteins without DSL domains (this study).
0411
OSM-11 mRNA localization by in situ hybridization is consistent with expression in VPCs and hypoderm in young larvae and in seam cells in adult animals (see NEXTDB, http://nematode.lab.nig.ac.jp/db2/ShowCloneInfo.php?clone=59g10; Y.
0018
The osm-11 cDNA clone was obtained by PCR from the Vidal laboratory ORFeome cDNA library [94] and agrees exactly with the predicted sequence in WormBase and at NCBI. osm-11 cDNA constructs used herein for rescue contained the unc-54 3′ UTR, whereas genomic rescue clones contained the osm-11 3′ UTR.

0018
Proteolysis of membrane-bound DLK1 yields the soluble protein originally known as fetal antigen 1 (FA1). A murine DLK1 cDNA fragment that encodes the DLK1 FA1 protein isoform was used in C.
0081
briggsae homologs of OSM-11 and used to generate the final DOS-motif consensus (Figure 2).
0018
The only exceptions to this were the sequences used for MmJagged1 and HsJagged1; in these proteins, a gap between EGF repeats 1 and 2 contained cysteine and tryptophan residues that followed the spacing of the DOS motif consensus sequence but were clearly not part of EGF repeat 2. Amino acid sequence from the gap instead of from EGF repeat 2 was used for these two proteins. Two outgroups were used in the alignment: EGF repeats 1 and 2 from CeAPX-1 and CeLAG-2, which lack the SELCT motif and have been previously shown to be phylogenetically distinct from EGF repeats 1 and 2 of other DSL ligands [27]; and EGF repeats 3 and 4 (EGF3–4) of selected DOS motif–containing proteins (using the same N- and C-terminal boundaries as above), as examples of canonical EGF repeats.
0018
The adaptor protein paxillin contains five conserved leucine-rich (LD) motifs that interact with a variety of focal adhesion proteins, such as α-parvin.
0030
Cell adhesion and migration are coordinated by dynamic membrane-associated protein assemblies called focal adhesions (FAs). With over 150 components, FAs play a crucial role in transmitting signals bidirectionally across the cell membrane and provide a physical link between integrin receptors and the actin cytoskeleton (Zaidel-Bar et al., 2007).

0889
With over 150 components, FAs play a crucial role in transmitting signals bidirectionally across the cell membrane and provide a physical link between integrin receptors and the actin cytoskeleton (Zaidel-Bar et al., 2007).

0030
LD-binding proteins include the kinases FAK (Turner et al., 1999), ILK (Nikolopoulos and Turner, 2001), and PAK3 (Hashimoto et al., 2001), the Arf-GAP PKL (Turner et al., 1999), the antiapoptotic protein Bcl-2 (Sheibani et al., 2008), the papillomavirus oncoprotein E6 (Tong et al., 1997), and the cytoskeletal proteins vinculin (Turner et al., 1999) and α-parvin (Nikolopoulos and Turner, 2000). While several of these proteins, such as FAK, PKL, and vinculin, bind LD motifs through parallel α-helical bundles (Gao et al., 2004; Hoellerer et al., 2003; Liu et al., 2002; Scheswohl et al., 2008; Schmalzigaug et al., 2007; Zhang et al., 2008), others employ topologically distinct modules, such as the kinase domain of ILK (Tu et al., 2001), the BH4 domain of Bcl-2 (Sorenson, 2004), and the calponin homology (CH) domain of α-parvin (Nikolopoulos and Turner, 2000).

0018
LD-binding proteins include the kinases FAK (Turner et al., 1999), ILK (Nikolopoulos and Turner, 2001), and PAK3 (Hashimoto et al., 2001), the Arf-GAP PKL (Turner et al., 1999), the antiapoptotic protein Bcl-2 (Sheibani et al., 2008), the papillomavirus oncoprotein E6 (Tong et al., 1997), and the cytoskeletal proteins vinculin (Turner et al., 1999) and α-parvin (Nikolopoulos and Turner, 2000).

0889
LD-binding proteins include the kinases FAK (Turner et al., 1999), ILK (Nikolopoulos and Turner, 2001), and PAK3 (Hashimoto et al., 2001), the Arf-GAP PKL (Turner et al., 1999), the antiapoptotic protein Bcl-2 (Sheibani et al., 2008), the papillomavirus oncoprotein E6 (Tong et al., 1997), and the cytoskeletal proteins vinculin (Turner et al., 1999) and α-parvin (Nikolopoulos and Turner, 2000). While several of these proteins, such as FAK, PKL, and vinculin, bind LD motifs through parallel α-helical bundles (Gao et al., 2004; Hoellerer et al., 2003; Liu et al., 2002; Scheswohl et al., 2008; Schmalzigaug et al., 2007; Zhang et al., 2008), others employ topologically distinct modules, such as the kinase domain of ILK (Tu et al., 2001), the BH4 domain of Bcl-2 (Sorenson, 2004), and the calponin homology (CH) domain of α-parvin (Nikolopoulos and Turner, 2000).
0030
α-parvin (Olski et al., 2001), also known as actopaxin (Nikolopoulos and Turner, 2000) or CH-ILKBP (Tu et al., 2001), is part of the ILK signaling complex (Tu et al., 2001), plays an essential role in adhesion-dependent PKB/Act activation (Fukuda et al., 2003) and in the regulation of actin organization (LaLonde et al., 2005) and Rac activation (LaLonde et al., 2005; Zhang et al., 2004). It belongs to the highly conserved parvin family that shares a common architecture composed of a variable N-terminal region followed by two CH domains (Olski et al., 2001).

0889
α-parvin (Olski et al., 2001), also known as actopaxin (Nikolopoulos and Turner, 2000) or CH-ILKBP (Tu et al., 2001), is part of the ILK signaling complex (Tu et al., 2001), plays an essential role in adhesion-dependent PKB/Act activation (Fukuda et al., 2003) and in the regulation of actin organization (LaLonde et al., 2005) and Rac activation (LaLonde et al., 2005; Zhang et al., 2004). It belongs to the highly conserved parvin family that shares a common architecture composed of a variable N-terminal region followed by two CH domains (Olski et al., 2001).

0030
However, the primary sequences of both CH domains of α-parvin are highly diverged from the typical type-1 and type-2 CH domains found in ABDs; they have therefore been classified as type-4 and type-5 CH domains (Gimona et al., 2002). An ability to recognize paxillin LD motifs has only been reported for the type-5 CH domains (Nikolopoulos and Turner, 2000; Yoshimi et al., 2006).

0889
However, the primary sequences of both CH domains of α-parvin are highly diverged from the typical type-1 and type-2 CH domains found in ABDs; they have therefore been classified as type-4 and type-5 CH domains (Gimona et al., 2002). An ability to recognize paxillin LD motifs has only been reported for the type-5 CH domains (Nikolopoulos and Turner, 2000; Yoshimi et al., 2006).
0030
While this manuscript was in preparation, an independent study reported the NMR structure of the C-terminal CH domain of α-parvin in complex with a 10-residue peptide derived from the paxillin LD1 motif (Wang et al., 2008).

0081
While this manuscript was in preparation, an independent study reported the NMR structure of the C-terminal CH domain of α-parvin in complex with a 10-residue peptide derived from the paxillin LD1 motif (Wang et al., 2008).

0114
While this manuscript was in preparation, an independent study reported the NMR structure of the C-terminal CH domain of α-parvin in complex with a 10-residue peptide derived from the paxillin LD1 motif (Wang et al., 2008).

0889
While this manuscript was in preparation, an independent study reported the NMR structure of the C-terminal CH domain of α-parvin in complex with a 10-residue peptide derived from the paxillin LD1 motif (Wang et al., 2008).
0081
Three cocrystal structures of α-parvin-CHC with 20-residue peptides representing LD1, LD2, and LD4 allow us to characterize the interaction at atomic resolution and to highlight binding-induced conformational changes in α-parvin-CHC. This analysis together with NMR studies of a spin-labeled LD1 peptide supports the surprising finding that LD motifs can associate with a single binding site bidirectionally.
0114
We initially attempted to solve the crystal structure of full-length human α-parvin, but no crystals could be obtained.

0030
This fragment includes residues 242–372 and thereby spans the entire C-terminal CH domain as defined by SMART (Schultz et al., 1998) and a portion of the inter-CH domain linker (Figure 1A).

0889
This fragment includes residues 242–372 and thereby spans the entire C-terminal CH domain as defined by SMART (Schultz et al., 1998) and a portion of the inter-CH domain linker (Figure 1A).

0114
The crystal structure of α-parvin-CHC at 1.05 Å resolution was determined by molecular replacement using an ensemble of homologous type-1 CH domain structures as search model.

0114
The refined structural model includes α-parvin residues 246–372 (molecule A) or 247–372 (molecule B) and represents the first high-resolution crystal structure of a type-5 CH domain (Figure 1B).

0030
In spite of very low levels of sequence conservation (≤26% identity) compared to canonical type-1 CH domains (see Figure S1 available online), α-parvin-CHC exhibits a typical CH domain core composed of four main α-helices (αA, αC, αE, and αG; nomenclature from Djinovic Carugo et al., 1997), which are connected by loops and shorter helical elements (αD and αF).

0081
In spite of very low levels of sequence conservation (≤26% identity) compared to canonical type-1 CH domains (see Figure S1 available online), α-parvin-CHC exhibits a typical CH domain core composed of four main α-helices (αA, αC, αE, and αG; nomenclature from Djinovic Carugo et al., 1997), which are connected by loops and shorter helical elements (αD and αF).

0889
In spite of very low levels of sequence conservation (≤26% identity) compared to canonical type-1 CH domains (see Figure S1 available online), α-parvin-CHC exhibits a typical CH domain core composed of four main α-helices (αA, αC, αE, and αG; nomenclature from Djinovic Carugo et al., 1997), which are connected by loops and shorter helical elements (αD and αF).

0114
Secondary structure matching (Krissinel and Henrick, 2004) of α-parvin-CHC with its closest homolog, the type-1 CH domain of α-actinin 3 (PDB: 1WKU), yields an RMSD of 1.19 Å in 103 equivalent Cα-positions.

0114
Another feature peculiar to this type-5 CH domain is the loop between helices C and E, which contains a 3 (or 4)-residue insertion (relative to type-1 CH domains) that is conserved throughout the parvin family (Figures 1C and S1). However, conformational differences in this region may not be significant, since the C/E-loop structure varies between α-parvin-CHC molecules both in the same and different crystal forms and is involved in crystal packing (data not shown).
0096
On the basis of primary sequence comparison with other LD-binding proteins, such as FAK and vinculin, and mutation studies, the LD-binding site (or “paxillin binding subdomain” [PBS]) of α-parvin was mapped to residues 274–291 (Nikolopoulos and Turner, 2000) (i.e., the A/C-loop of α-parvin-CHC) (Figure 1B).

0081
On the basis of primary sequence comparison with other LD-binding proteins, such as FAK and vinculin, and mutation studies, the LD-binding site (or “paxillin binding subdomain” [PBS]) of α-parvin was mapped to residues 274–291 (Nikolopoulos and Turner, 2000) (i.e., the A/C-loop of α-parvin-CHC) (Figure 1B).

0030
However, we previously demonstrated that LD-binding to the FAT domain of FAK does not reside in a local peptide sequence such as the PBS (Hoellerer et al., 2003) and thus investigated the interaction of α-parvin-CHC with LD motifs using solution NMR.

0019
However, we previously demonstrated that LD-binding to the FAT domain of FAK does not reside in a local peptide sequence such as the PBS (Hoellerer et al., 2003) and thus investigated the interaction of α-parvin-CHC with LD motifs using solution NMR.

0889
However, we previously demonstrated that LD-binding to the FAT domain of FAK does not reside in a local peptide sequence such as the PBS (Hoellerer et al., 2003) and thus investigated the interaction of α-parvin-CHC with LD motifs using solution NMR.

0081
1H-15N HSQC monitored titrations of 15N-enriched α-parvin-CHC were performed with peptides representing all five paxillin LD motifs. Each peptide was found to induce resonance-specific chemical shift perturbations (Figure 2 and data not shown), indicating an interaction with α-parvin-CHC. Global fitting of the resulting binding curves shows that the affinities of individual LD motifs for α−parvin-CHC differ substantially (Table 2; Figure S2) with LD1 being the highest affinity ligand followed by LD4 and LD2.

0081
The overall pattern of chemical shift perturbations induced by saturating concentrations of different LD peptides is very similar (Figure 2), suggesting that all five LD motifs interact with α-parvin-CHC in a similar fashion.
0081
In all three cases, continuous positive difference density was identified into which the peptide ligands could be placed unambiguously (Figure S3).

0081
The remaining C-terminal residues of the LD peptides appear disordered, presumably because they do not form contacts with α-parvin-CHC (Figure 3B). All three LD motifs bind to the same binding site on α-parvin-CHC formed by the N-linker helix, the N-terminal part of helix A and the C-terminal part of helix G, which is consistent with our results from chemical shift mapping (Figure 2). We thus conclude that the PBS region previously identified on α-parvin is not directly involved in the interaction with LD peptides.
0081
Despite their antiparallel orientations, however, the binding modes of different LD peptides are very similar (Figure 3A, Figure S4). As the result of a slight rotation around the helical axis of the bound peptides, the character and position of side chains facing α-parvin-CHC is largely preserved, the same hydrophobic pockets are occupied, and a similar amount of surface area (∼500 Å2) becomes buried.

0081
The side chains of their conserved leucine residues in position 0, +3, +4, and +7 (Figure 3B) interact with a hydrophobic patch on the surface of the CHC domain formed by residues from the N-linker helix (A249, F250, L253, A257), helix A (V263, V264, and L268) and helix G (Y362 and F365) (Figure 3A). In addition, two positively charged residues (K260 and R369) of α-parvin-CHC are in close proximity to negatively charged residues of the bound LD peptides.
0081
Consistent with their similar binding modes (Figure S4), all three LD peptides, irrespective of directionality, impose similar conformational changes in the N-terminal region of α-parvin-CHC, whereas the protein core remains virtually unperturbed (Figure 4).

0096
Consistent with their similar binding modes (Figure S4), all three LD peptides, irrespective of directionality, impose similar conformational changes in the N-terminal region of α-parvin-CHC, whereas the protein core remains virtually unperturbed (Figure 4).
0004
As measured by NMR, none of these substitutions substantially altered the binding affinity for α-parvin-CHC, indicating that the predicted electrostatic contacts are energetically neutral under the assay conditions (50 mM sodium phosphate and 100 mM NaCl [pH 6.9]) used.

0096
As measured by NMR, none of these substitutions substantially altered the binding affinity for α-parvin-CHC, indicating that the predicted electrostatic contacts are energetically neutral under the assay conditions (50 mM sodium phosphate and 100 mM NaCl [pH 6.9]) used. Therefore, analysis of LD1 peptide variants did not define the binding mode of LD1 in solution.

0081
Likewise, comparison of the chemical shift perturbations imposed by oppositely aligned peptides on backbone amide resonances of α-parvin-CHC (Figure 2) is inconclusive. We note that the peptide side-chains contacting the LD-binding site of α-parvin-CHC are similar in either orientation (Figure S4), as is the conformational rearrangement induced by peptide binding, so that the perturbation patterns of either unidirectional binding mode (or a bidirectional mixture) may be indistinguishable.
0077
This phenomenon is also illustrated in Figure 5, using two diagnostic NMR signals as examples: the resonance originating from residue 257 is predicted to experience strong PRE in the forward mode (calculated distance from spin-label 10.7 Å) but should only be weakly affected in the backward mode (calculated distance 20.2 Å).

0104
However, our experimental data show that both resonances undergo significant broadening, suggesting that a simple unidirectional model may be insufficient to describe LD1 binding in solution. To quantify this observation, we calculated the linear correlation coefficients, R, between the experimental and the simulated PRE data for various ratios of forward-to-backward binding in a bidirectional mixture (Figure S5B). The resulting curve is bell shaped, indicating that the experimental data are more consistent with a bidirectional than a unidirectional binding model.
0081
Both binding curves could be fitted by a single site model, suggesting that α-parvin contains only one LD-binding site located within the CHC fragment (Figure S6).
0889
This binding site is consistent with the one described in a recent solution NMR structure of α-parvin-CHC in complex with an LD1 peptide (Wang et al., 2008).

0030
This binding site is consistent with the one described in a recent solution NMR structure of α-parvin-CHC in complex with an LD1 peptide (Wang et al., 2008).

0081
This binding site is consistent with the one described in a recent solution NMR structure of α-parvin-CHC in complex with an LD1 peptide (Wang et al., 2008).

0104
This binding site is consistent with the one described in a recent solution NMR structure of α-parvin-CHC in complex with an LD1 peptide (Wang et al., 2008).

0030
The significantly lower value of ∼1.2 mM recently measured (Wang et al., 2008) might be due to the restricted length of the peptide (10 versus 20 residues) used in their study.

0889
The significantly lower value of ∼1.2 mM recently measured (Wang et al., 2008) might be due to the restricted length of the peptide (10 versus 20 residues) used in their study.
0030
The FAT domain of FAK contains two independent LD binding sites, which bind equally well to LD2 and LD4 in vitro (Hoellerer et al., 2003; Gao et al., 2004), but it also interacts to some degree with the other LD motifs (M.K.H., unpublished results). Recent NMR studies (Zhang et al., 2008) have also demonstrated that the single LD binding site of PKL/GIT1 binds to both LD4 (Turner et al., 1999) and LD2. Taken together, these observations suggest that paxillin LD motifs are promiscuous protein interaction modules.

0114
The FAT domain of FAK contains two independent LD binding sites, which bind equally well to LD2 and LD4 in vitro (Hoellerer et al., 2003; Gao et al., 2004), but it also interacts to some degree with the other LD motifs (M.K.H., unpublished results).

0889
The FAT domain of FAK contains two independent LD binding sites, which bind equally well to LD2 and LD4 in vitro (Hoellerer et al., 2003; Gao et al., 2004), but it also interacts to some degree with the other LD motifs (M.K.H., unpublished results). Recent NMR studies (Zhang et al., 2008) have also demonstrated that the single LD binding site of PKL/GIT1 binds to both LD4 (Turner et al., 1999) and LD2.
0030
On the basis of our results for α-parvin, we can make two general predictions with respect to LD recognition by CH domains: First, as noted previously (Wang et al., 2008), we propose that the presence of an N-linker helix is a prerequisite for the interaction with LD motifs.

0889
On the basis of our results for α-parvin, we can make two general predictions with respect to LD recognition by CH domains: First, as noted previously (Wang et al., 2008), we propose that the presence of an N-linker helix is a prerequisite for the interaction with LD motifs.

0030
Interestingly, γ-parvin was shown to associate with paxillin in vivo (Yoshimi et al., 2006). Available experiments with β-parvin, however, have proved negative to date (Yamaji et al., 2004), although this protein is more similar in sequence to α-parvin than is γ-parvin (Figure 1C).

0889
Interestingly, γ-parvin was shown to associate with paxillin in vivo (Yoshimi et al., 2006). Available experiments with β-parvin, however, have proved negative to date (Yamaji et al., 2004), although this protein is more similar in sequence to α-parvin than is γ-parvin (Figure 1C).
0030
The structure of α-parvin-CHC is topologically distinct from other LD-binding domains, such as the FAT domain of FAK; it associates with a single LD motif across three oblique helices, whereas FAT accommodates two LD motifs in parallel fashion on opposite sites of its 4-helical bundle (Hoellerer et al., 2003).

0114
The structure of α-parvin-CHC is topologically distinct from other LD-binding domains, such as the FAT domain of FAK; it associates with a single LD motif across three oblique helices, whereas FAT accommodates two LD motifs in parallel fashion on opposite sites of its 4-helical bundle (Hoellerer et al., 2003). This suggests that recognition of paxillin LD motifs does not depend on a conserved fold.

0081
The structure of α-parvin-CHC is topologically distinct from other LD-binding domains, such as the FAT domain of FAK; it associates with a single LD motif across three oblique helices, whereas FAT accommodates two LD motifs in parallel fashion on opposite sites of its 4-helical bundle (Hoellerer et al., 2003).

0889
The structure of α-parvin-CHC is topologically distinct from other LD-binding domains, such as the FAT domain of FAK; it associates with a single LD motif across three oblique helices, whereas FAT accommodates two LD motifs in parallel fashion on opposite sites of its 4-helical bundle (Hoellerer et al., 2003).

0096
In both cases, LD motifs adopt an α-helical conformation when bound, which allows the conserved leucine residues on one face of the helix to interact with the hydrophobic binding site. Notably, the role of electrostatic contacts in LD recognition might differ between the two proteins: while substitutions of the conserved aspartates (D+1) of LD2 and LD4 to alanine were found to abolish the interaction of N-terminal paxillin fragments with FAT (Brown et al., 1996; Scheswohl et al., 2008; Thomas et al., 1999), our study and others (Wang et al., 2008) show that the equivalent substitutions in the isolated LD1 peptides do not significantly perturb α-parvin-CHC binding in vitro.

0889
Notably, the role of electrostatic contacts in LD recognition might differ between the two proteins: while substitutions of the conserved aspartates (D+1) of LD2 and LD4 to alanine were found to abolish the interaction of N-terminal paxillin fragments with FAT (Brown et al., 1996; Scheswohl et al., 2008; Thomas et al., 1999), our study and others (Wang et al., 2008) show that the equivalent substitutions in the isolated LD1 peptides do not significantly perturb α-parvin-CHC binding in vitro.

0019
Notably, the role of electrostatic contacts in LD recognition might differ between the two proteins: while substitutions of the conserved aspartates (D+1) of LD2 and LD4 to alanine were found to abolish the interaction of N-terminal paxillin fragments with FAT (Brown et al., 1996; Scheswohl et al., 2008; Thomas et al., 1999), our study and others (Wang et al., 2008) show that the equivalent substitutions in the isolated LD1 peptides do not significantly perturb α-parvin-CHC binding in vitro.

0030
Notably, the role of electrostatic contacts in LD recognition might differ between the two proteins: while substitutions of the conserved aspartates (D+1) of LD2 and LD4 to alanine were found to abolish the interaction of N-terminal paxillin fragments with FAT (Brown et al., 1996; Scheswohl et al., 2008; Thomas et al., 1999), our study and others (Wang et al., 2008) show that the equivalent substitutions in the isolated LD1 peptides do not significantly perturb α-parvin-CHC binding in vitro.

0107
Notably, the role of electrostatic contacts in LD recognition might differ between the two proteins: while substitutions of the conserved aspartates (D+1) of LD2 and LD4 to alanine were found to abolish the interaction of N-terminal paxillin fragments with FAT (Brown et al., 1996; Scheswohl et al., 2008; Thomas et al., 1999), our study and others (Wang et al., 2008) show that the equivalent substitutions in the isolated LD1 peptides do not significantly perturb α-parvin-CHC binding in vitro.
0030
Bidirectional protein-ligand interactions are widespread among PPII (poly-proline type-2) helical ligands, as this conformation has a pseudo C2-rotational symmetry perpendicular to its long axis (Ball et al., 2005).

0889
Bidirectional protein-ligand interactions are widespread among PPII (poly-proline type-2) helical ligands, as this conformation has a pseudo C2-rotational symmetry perpendicular to its long axis (Ball et al., 2005).

0030
To our knowledge, this phenomenon has only been reported for the following: the recruitment of the histone deactelylase-associated Sin3 corepressor by the HBP1 and Mad1 repressors, respectively (Brubaker et al., 2000; Spronk et al., 2000; Swanson et al., 2004), the association of FAT with paxillin LD motifs and the endocytosis motif of CD4 (Garron et al., 2008), and the interaction between AKAP-IS and the RIIa subunit of PKA (Gold et al., 2006).

0107
To our knowledge, this phenomenon has only been reported for the following: the recruitment of the histone deactelylase-associated Sin3 corepressor by the HBP1 and Mad1 repressors, respectively (Brubaker et al., 2000; Spronk et al., 2000; Swanson et al., 2004), the association of FAT with paxillin LD motifs and the endocytosis motif of CD4 (Garron et al., 2008), and the interaction between AKAP-IS and the RIIa subunit of PKA (Gold et al., 2006).

0889
To our knowledge, this phenomenon has only been reported for the following: the recruitment of the histone deactelylase-associated Sin3 corepressor by the HBP1 and Mad1 repressors, respectively (Brubaker et al., 2000; Spronk et al., 2000; Swanson et al., 2004), the association of FAT with paxillin LD motifs and the endocytosis motif of CD4 (Garron et al., 2008), and the interaction between AKAP-IS and the RIIa subunit of PKA (Gold et al., 2006).
0077
Small changes in a variety of factors, such as experimental conditions or the length of the paxillin fragment studied, might influence the observed binding orientation in NMR, where a mean is detected.

0030
This could explain a similar PRE experiment performed with a shorter LD1 variant (residues 3–13), which was interpreted as unidirectional forward-type binding and used to restrain the solution structure of the LD1/α-parvin-CHC complex (Wang et al., 2008).

0889
This could explain a similar PRE experiment performed with a shorter LD1 variant (residues 3–13), which was interpreted as unidirectional forward-type binding and used to restrain the solution structure of the LD1/α-parvin-CHC complex (Wang et al., 2008).
0104
Although our PRE experiments suggest bidirectional binding in solution, the refinement of different models against our X-ray data strongly indicate a predominance of the forward binding mode in the α-parvin-CHC/LD1 crystal. At the same time, the TLS-corrected individual isotropic temperature factors of the LD-peptides are not atypical of those expected for weakly bound ligands (data not shown): deviation from such behavior would be expected if the crystal contained significant static disorder in peptide binding.

0104
The capacity of such relatively weak crystal-packing forces to bring about homogenization of orientation is consistent with our finding that the two different orientations are approximately isoenergetic in solution.
0030
Although α−parvin has been shown to bind to F-actin with an affinity comparable to other CH domain-containing actin-binding proteins (Nikolopoulos and Turner, 2000; Olski et al., 2001), the canonical model of actin recognition may not apply to this protein. Usually CH domains bind F-actin through a tandem array of an N-terminal type-1 CH domain containing two actin-binding sites (ABS1 and ABS2) and a C-terminal type-2 CH domain contributing a third actin-binding site (ABS3) (Gimona et al., 2002).

0018
Although α−parvin has been shown to bind to F-actin with an affinity comparable to other CH domain-containing actin-binding proteins (Nikolopoulos and Turner, 2000; Olski et al., 2001), the canonical model of actin recognition may not apply to this protein.

0889
Although α−parvin has been shown to bind to F-actin with an affinity comparable to other CH domain-containing actin-binding proteins (Nikolopoulos and Turner, 2000; Olski et al., 2001), the canonical model of actin recognition may not apply to this protein. Usually CH domains bind F-actin through a tandem array of an N-terminal type-1 CH domain containing two actin-binding sites (ABS1 and ABS2) and a C-terminal type-2 CH domain contributing a third actin-binding site (ABS3) (Gimona et al., 2002).

0081
Usually CH domains bind F-actin through a tandem array of an N-terminal type-1 CH domain containing two actin-binding sites (ABS1 and ABS2) and a C-terminal type-2 CH domain contributing a third actin-binding site (ABS3) (Gimona et al., 2002).

0114
Usually CH domains bind F-actin through a tandem array of an N-terminal type-1 CH domain containing two actin-binding sites (ABS1 and ABS2) and a C-terminal type-2 CH domain contributing a third actin-binding site (ABS3) (Gimona et al., 2002).

0114
Interestingly, the crystal structure of α−parvin-CHC presented here shows that part of the putative ABS2 region is obstructed by the N-linker helix.
0030
CH domains in intact ABDs often interact across a sizeable interface, which can modulate their ability to interact with F-actin (Gimona et al., 2002). It remains to be established whether such intramolecular interactions occur to regulate the association of α-parvin with F-actin and other binding partners, such as paxillin, TESK1 (LaLonde et al., 2005), and ILK (Tu et al., 2001).

0889
CH domains in intact ABDs often interact across a sizeable interface, which can modulate their ability to interact with F-actin (Gimona et al., 2002). It remains to be established whether such intramolecular interactions occur to regulate the association of α-parvin with F-actin and other binding partners, such as paxillin, TESK1 (LaLonde et al., 2005), and ILK (Tu et al., 2001).
0030
Several ABD containing proteins are capable of forming dimers, thereby cross-linking actin filaments (Gimona et al., 2002). On the basis of gel filtration experiments with the N-terminal CH domain of α-parvin, Wang et al.

0889
Several ABD containing proteins are capable of forming dimers, thereby cross-linking actin filaments (Gimona et al., 2002). On the basis of gel filtration experiments with the N-terminal CH domain of α-parvin, Wang et al.

0071
On the basis of gel filtration experiments with the N-terminal CH domain of α-parvin, Wang et al.
0107
Taken together, we have presented a comprehensive structural characterization of the interaction between paxillin LD motifs and α-parvin, which has revealed a surprising degree of promiscuity, both in terms of LD motif selectivity and binding directionality.

0030
Such features may allow the assembly of various constitutionally and conformationally distinct protein complexes with specific signaling properties.
0004
α-parvin-CHC was purified as follows: cleared cell lysate (in 75 mM Tris [pH 8.0], 200 mM NaCl, 5 mM β-mercaptoethanol, 0.4% Triton X-100, 2 mM EDTA, 5 mM benzamidine, and protease inhibitor cocktail [Roche]) was applied to glutathione sepharose 4B (GE Healthcare) in binding buffer (20 mM Tris [pH 8.0], 150 mM NaCl, and 2 mM DTT) washed with 20 mM Tris (pH 8.0), 1 M NaCl, 2 mM DTT, 2 mM EDTA, and 5 mM benzamidine, and was eluted with 50 mM glutathione in binding buffer (pH 8.0). After cleavage with recombinant human rhinovirus 3C-protease, the sample was subjected to size exclusion chromatography (Superdex 75, GE Healthcare) in 25 mM Tris (pH 8.0), 150 mM NaCl, 2 mM DTT, and 2 mM EDTA.

0071
α-parvin-CHC was purified as follows: cleared cell lysate (in 75 mM Tris [pH 8.0], 200 mM NaCl, 5 mM β-mercaptoethanol, 0.4% Triton X-100, 2 mM EDTA, 5 mM benzamidine, and protease inhibitor cocktail [Roche]) was applied to glutathione sepharose 4B (GE Healthcare) in binding buffer (20 mM Tris [pH 8.0], 150 mM NaCl, and 2 mM DTT) washed with 20 mM Tris (pH 8.0), 1 M NaCl, 2 mM DTT, 2 mM EDTA, and 5 mM benzamidine, and was eluted with 50 mM glutathione in binding buffer (pH 8.0). After cleavage with recombinant human rhinovirus 3C-protease, the sample was subjected to size exclusion chromatography (Superdex 75, GE Healthcare) in 25 mM Tris (pH 8.0), 150 mM NaCl, 2 mM DTT, and 2 mM EDTA.
0004
To purify full-length α-parvin, cleared cell lysate in 200 mM potassium phosphate (pH 8.0), 10 mM NaCl, 5 mM β-mercaptoethanol, 0.4% Triton X-100, 5 mM benzamidine, 2 mM EDTA, and protease inhibitor cocktail (Roche) was applied to glutathione sepharose 4B (GE Healthcare), washed with 200 mM potassium phosphate, 10 mM NaCl, and 4 mM DTT, and α−parvin was released by cleavage with recombinant human rhinovirus 3C-protease at 4°C over night. After elution with 200 mM potassium phosphate, 10 mM NaCl, 4 mM DTT, and 2.5% glycerol, the protein was dialyzed into 25 mM potassium phosphate (pH 8.0), 1.5 mM NaCl, 2 mM DTT, and 2.5% (v/v) glycerol for subsequent anion exchange chromatography (MonoQ, GE Healthcare) and gradient elution with 250 mM potassium phosphate (pH 8.0), 15 mM NaCl, 250 mM KCl, 2 mM DTT, and 2.5% glycerol.

0071
To purify full-length α-parvin, cleared cell lysate in 200 mM potassium phosphate (pH 8.0), 10 mM NaCl, 5 mM β-mercaptoethanol, 0.4% Triton X-100, 5 mM benzamidine, 2 mM EDTA, and protease inhibitor cocktail (Roche) was applied to glutathione sepharose 4B (GE Healthcare), washed with 200 mM potassium phosphate, 10 mM NaCl, and 4 mM DTT, and α−parvin was released by cleavage with recombinant human rhinovirus 3C-protease at 4°C over night.
0053
For fluorescence anisotropy studies, LD1 with 5-carboxyfluorescein (5-FAM) attached to the ɛ-amino group of the C-terminal lysine was used.
0030
Manual model building was performed with COOT (Emsley and Cowtan, 2004), and refinement was performed with REFMAC5 (Murshudov et al., 1997); occupancies of alternate conformers were refined with PHENIX (Afonine et al., 2007).

0889
Manual model building was performed with COOT (Emsley and Cowtan, 2004), and refinement was performed with REFMAC5 (Murshudov et al., 1997); occupancies of alternate conformers were refined with PHENIX (Afonine et al., 2007).
0004
For PRE measurements, 1H-15N HSQC experiments were recorded on 230 μM 15N-enriched α-parvin-CHC in the presence of 250 μM PROXYL-labeled LD1 peptide in 50 mM sodium phosphate and 100 mM NaCl (pH 6.9) in the absence and presence of 5 mM ascorbate at 25°C.

0030
For the simulation of PRE effects in both unidirectional binding modes, the crystal structures of α-parvin-CHC bound to LD1 and LD4 were protonated with the program PDB 2PQR (Dolinsky et al., 2004).

0889
For the simulation of PRE effects in both unidirectional binding modes, the crystal structures of α-parvin-CHC bound to LD1 and LD4 were protonated with the program PDB 2PQR (Dolinsky et al., 2004).

0030
The resulting values were used as approximation of the distances, r, of the unpaired electron of the PROXYL moiety to the backbone amide protons and to derive residue-specific PRE-values, I/I0, according to (Jain et al., 2001; Johnson et al., 1999) (for details, see Figure S5).

0889
The resulting values were used as approximation of the distances, r, of the unpaired electron of the PROXYL moiety to the backbone amide protons and to derive residue-specific PRE-values, I/I0, according to (Jain et al., 2001; Johnson et al., 1999) (for details, see Figure S5).
0114
The binding orientation of LD1 seen in the crystal structure is denoted “forward.”.
0114
The yeast Ctk1 kinase associates with elongation complexes and phosphorylates serine 2 in the YSPTSPS repeats of the Rpb1 C-terminal domain, a modification that couples transcription to mRNA 3′-end processing.
0889
Analyses with purified mammalian factors suggested that only TFIID and TFIIA remain at the promoter (Reinberg et al, 1987). A more recent in vitro study using yeast extracts found that TFIID, TFIIA, TFIIH, TFIIE, and Mediator remain behind at the promoter in a ‘scaffold complex' primed for rapid reinitiation (Yudkovsky et al, 2000; Hahn, 2004).

0676
A more recent in vitro study using yeast extracts found that TFIID, TFIIA, TFIIH, TFIIE, and Mediator remain behind at the promoter in a ‘scaffold complex' primed for rapid reinitiation (Yudkovsky et al, 2000; Hahn, 2004). TFIIB and TFIIF dissociate from the promoter-bound complex, whereas RNApII moves into the elongation phase.

0889
The signals for these transitions remain unclear, although recent experiments suggest that both Kin28 (TFIIH) and Srb10 (Mediator) kinases are important for scaffold formation (Yudkovsky et al, 2000).
0019
Chromatin immunoprecipitation (ChIP) experiments show that Ctk1 associates with RNApII throughout elongation (Kim et al, 2004a).

0114
Similar to Cdk9, Ctk1 couples transcription with polyadenylation by phosphorylating the C-terminal domain (CTD) of Rpb1, the largest subunit of RNApII (Ahn et al, 2004).
0676
Consistent with this, when ChIP was carried out using strains deleted for the other subunits of the Ctk1 complex, including Ctk2 (cyclin subunit) or Ctk3 (accessory factor), TBP occupancy was again increased in the coding region of PMA1 (Supplementary Figure S1).

0081
In accordance with this report, Sua7 cross-linked with region 7, downstream of the PMA1 poly(A) sites in a WT strain (Figure 1C).
0081
To test whether the extended pattern of basal factor cross-linking was completely coincident with RNApII, ChIP experiments were carried out at the termination site for PMA1 (Figure 3).
0081
The upstream activating sequence (UAS; Figure 1A) of the three genes used in our studies (PMA1, ADH1, and PYK1) contain a binding site for repressor-activator protein 1 (Rap 1), a transactivator for many promoters in exponentially growing cells (Rao et al, 1993; Lieb et al, 2001). The UAS from all three genes show strong Rap1 cross-linking, and this signal remains appropriately localized in ctk1Δ cells (Figure 3C).

0096
The upstream activating sequence (UAS; Figure 1A) of the three genes used in our studies (PMA1, ADH1, and PYK1) contain a binding site for repressor-activator protein 1 (Rap 1), a transactivator for many promoters in exponentially growing cells (Rao et al, 1993; Lieb et al, 2001).

0030
The upstream activating sequence (UAS; Figure 1A) of the three genes used in our studies (PMA1, ADH1, and PYK1) contain a binding site for repressor-activator protein 1 (Rap 1), a transactivator for many promoters in exponentially growing cells (Rao et al, 1993; Lieb et al, 2001).

0413
The upstream activating sequence (UAS; Figure 1A) of the three genes used in our studies (PMA1, ADH1, and PYK1) contain a binding site for repressor-activator protein 1 (Rap 1), a transactivator for many promoters in exponentially growing cells (Rao et al, 1993; Lieb et al, 2001). The UAS from all three genes show strong Rap1 cross-linking, and this signal remains appropriately localized in ctk1Δ cells (Figure 3C).

0426
The upstream activating sequence (UAS; Figure 1A) of the three genes used in our studies (PMA1, ADH1, and PYK1) contain a binding site for repressor-activator protein 1 (Rap 1), a transactivator for many promoters in exponentially growing cells (Rao et al, 1993; Lieb et al, 2001). The UAS from all three genes show strong Rap1 cross-linking, and this signal remains appropriately localized in ctk1Δ cells (Figure 3C).
0081
As serine 2 phosphorylation is known to be required for cotranscriptional 3′-end processing (Ahn et al, 2004), we also tested for the polyadenylation factor Rna14 as a positive control (Figure 4B).

0081
At the nonpermissive temperature (37°C), where the only source of Rpb1 is the thermostable S2A mutant, there was no indication of basal factor spreading (Figure 4D). Taken together, these results argue that CTD serine 2 phosphorylation does not regulate the dissociation of basal transcription factors from elongating RNApII.
0081
Two mutant alleles were employed: ctk1-D324N is a catalytic site mutant that shows no kinase activity in vitro and confers a slow-growth, cold-sensitive phenotype in vivo, whereas ctk1-T338A is a weakened kinase mutated in the T-loop threonine that accepts an activating phosphorylation (Ostapenko and Solomon, 2003). Transformation of empty vector into a ctk1Δ strain produced no change in TBP cross-linking, which was still seen in the PMA1 coding region (Figure 4E).
0081
Although serine 2 phosphorylation of the Rpb1 CTD mediates the coupling of transcription with polyadenylation at 3′-ends (Ahn et al, 2004; Ni et al, 2004), our data suggest that Ctk1 also affects the transition from transcription initiation to elongation at 5′-ends of genes by promoting dissociation of basal factors from polymerase.
0411
Cells deficient in Ctk1 were deleted at the genomic locus but covered by pCTK1, a low-copy URA3 plasmid containing WT Ctk1. Before our analyses, cells were grown overnight on YPD, streaked onto 5-FOA to select for those who lost the covering plasmid (confirmed by slow growth) and immediately used for subsequent experiments.

0426
Before our analyses, cells were grown overnight on YPD, streaked onto 5-FOA to select for those who lost the covering plasmid (confirmed by slow growth) and immediately used for subsequent experiments.
0019
Chromatin immunoprecipitation was performed as described previously (Ahn et al, 2004).

0676
For precipitation of TAP-tagged proteins, 10 μl of rabbit IgG agarose (Sigma) was incubated with chromatin solution overnight at 4°C.

0019
For Rpb3 immunoprecipitation, antibody (1Y26, NeoClone) was preincubated with protein G-sepharose (Amersham).

0007
For Rpb3 immunoprecipitation, antibody (1Y26, NeoClone) was preincubated with protein G-sepharose (Amersham).

0019
The signal for each specific gene primer in the immunoprecipitation was then divided by this ratio to convert the signal to normalized units. This value was divided by the immunoprecipitation signal of the non-transcribed control product to determine the fold enrichment of the ChIP over background signals.
0004
Dynabeads M-280 Streptavidin (Dynal) were concentrated with a magnetic particle concentrator (MPC) (Dynal) and washed twice in Buffer T (10 mM Tris (pH 7.5), 1 mM EDTA, 1 M NaCl).

0071
Dynabeads M-280 Streptavidin (Dynal) were concentrated with a magnetic particle concentrator (MPC) (Dynal) and washed twice in Buffer T (10 mM Tris (pH 7.5), 1 mM EDTA, 1 M NaCl).

0411
Dynabeads were then incubated with 8.6 ng biotinylated template per μg of bead in Buffer T for 30 min at room temperature with constant agitation.

0040
Immobilized templates were washed once in Buffer T and blocked in blocking buffer (1 ml/mg beads) for 15 min at room temperature. Blocking buffer consists of transcription buffer (10 mM HEPES (pH 7.6), 100 mM potassium glutamate, 10 mM magnesium acetate, 5 mM EGTA, 3.5% glycerol) containing 60 mg/ml casein, 5 mg/ml polyvinylpyrrolidone, and 2.5 mM DTT.
0004
The supernatant was then removed and incubated with protein G-sepharose (Amersham) overnight at 4°C with anti-Rpb3 antibody in buffer E (20 mM HEPES (pH 8), 350 mM NaCl, 10% glycerol, 0.1% Tween 20).

0071
The supernatant was then removed and incubated with protein G-sepharose (Amersham) overnight at 4°C with anti-Rpb3 antibody in buffer E (20 mM HEPES (pH 8), 350 mM NaCl, 10% glycerol, 0.1% Tween 20).
0019
The abnormal cross-linking of basal transcription factors in ctk1Δ cells is confirmed by independent immunoprecipitation of TAP-tagged TFIIF subunits.

0019
IgG agarose was used for immunoprecipitation of TAP-tagged proteins.
0019
(C) Occupancies of Rap1, Rpb1, and TBP at the indicated regions of genes were determined using the indicated polyclonal antibodies in both WT (YSB726) and ctk1Δ (YSB854) backgrounds.
0081
PCR products from (A), (B), (D), and (E) are shown as in Figure 1B. (B) Serine 2 phosphorylation of the Rpb1 CTD does not regulate the 5′ transitions.

0019
Protein A- or protein G-sepharose was used for TBP or Rpb3 immunoprecipitation and rabbit IgG agarose was used for RNA14–TAP pull-down.

0096
Protein A- or protein G-sepharose was used for TBP or Rpb3 immunoprecipitation and rabbit IgG agarose was used for RNA14–TAP pull-down.

0676
Protein A- or protein G-sepharose was used for TBP or Rpb3 immunoprecipitation and rabbit IgG agarose was used for RNA14–TAP pull-down.
0412
(D) Efficiency of scaffold formation was measured by quantitation of band intensities from lane 3 or 7 in WT and ctk1Δ cells (ImageJ v1.32).

0006
The cell extracts from either WT (YSB726) or ctk1Δ (YSB854) were incubated with the HIS4-immobilized templates and Gal4-VP16 for 40 min. To measure the association between RNApII and the basal transcription factors after termination, the supernatant was removed after incubation with NTPs for 2 min and immunoprecipitated with anti-Rpb3 antibody.

0007
(F) Coimmunoprecipitated proteins with Rpb3 from WT and ctk1Δ cells were determined by western blot using antibodies against non-scaffold (Tfg2 or Sua7) and scaffold components (TBP, Tfb1, Kin28, or Tfa2).
