Structural basis for recruitment of CBP/p300 coactivators by STAT1 and STAT2 transactivation domains
2009; Springer Nature; Volume: 28; Issue: 7 Linguagem: Inglês
10.1038/emboj.2009.30
ISSN1460-2075
AutoresJonathan M. Wojciak, Maria A. Martinez‐Yamout, H. Jane Dyson, Peter E. Wright,
Tópico(s)Neuroblastoma Research and Treatments
ResumoArticle12 February 2009free access Structural basis for recruitment of CBP/p300 coactivators by STAT1 and STAT2 transactivation domains Jonathan M Wojciak Jonathan M Wojciak Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, CA, USA Search for more papers by this author Maria A Martinez-Yamout Maria A Martinez-Yamout Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, CA, USA Search for more papers by this author H Jane Dyson H Jane Dyson Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, CA, USA Search for more papers by this author Peter E Wright Corresponding Author Peter E Wright Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, CA, USA Search for more papers by this author Jonathan M Wojciak Jonathan M Wojciak Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, CA, USA Search for more papers by this author Maria A Martinez-Yamout Maria A Martinez-Yamout Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, CA, USA Search for more papers by this author H Jane Dyson H Jane Dyson Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, CA, USA Search for more papers by this author Peter E Wright Corresponding Author Peter E Wright Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, CA, USA Search for more papers by this author Author Information Jonathan M Wojciak1, Maria A Martinez-Yamout1, H Jane Dyson1 and Peter E Wright 1 1Department of Molecular Biology, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, CA, USA *Corresponding author. Department of Molecular Biology, MB2, The Skaggs Institute for Chemical Biology, The Scripps Research Institute, 10550 N Torrey Pines Road, La Jolla, CA 92037, USA. Tel.: +1 858 784 9721; Fax: +1 858 784 9822; E-mail: [email protected] The EMBO Journal (2009)28:948-958https://doi.org/10.1038/emboj.2009.30 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info CBP/p300 transcriptional coactivators mediate gene expression by integrating cellular signals through interactions with multiple transcription factors. To elucidate the molecular and structural basis for CBP-dependent gene expression, we determined structures of the CBP TAZ1 and TAZ2 domains in complex with the transactivation domains (TADs) of signal transducer and activator of transcription 2 (STAT2) and STAT1, respectively. Despite the topological similarity of the TAZ1 and TAZ2 domains, subtle differences in helix packing and surface grooves constitute major determinants of target selectivity. Our results suggest that TAZ1 preferentially binds long TADs capable of contacting multiple surface grooves simultaneously, whereas smaller TADs that are restricted to a single contiguous binding surface form complexes with TAZ2. Complex formation for both STAT TADs involves coupled folding and binding, driven by intermolecular hydrophobic and electrostatic interactions. Phosphorylation of S727, required for maximal transcriptional activity of STAT1, does not enhance binding to any of the CBP domains. Because the different STAT TADs recognize different regions of CBP/p300, there is a potential for multivalent binding by STAT heterodimers that could enhance the recruitment of the coactivators to promoters. Introduction Many signal-transduction pathways converge at transcriptional coactivator complexes, which function as nodes in complex protein–protein interaction networks that process cellular signals and regulate gene expression (Naar et al, 2001). One of the most extensively studied transcriptional coactivators, the CREB-binding protein (CBP) (Chrivia et al, 1993) and its paralogue p300 (Eckner et al, 1994), stimulate gene expression through histone acetyltransferase activity and integrate signals that promote cell growth, transformation, development and other essential cellular processes (Goodman and Smolik, 2000). Over 300 proteins have been classified as CBP/p300-binding partners (Kasper et al, 2006), and many of the interactions map to regions of CBP/p300 with remarkably high redundancy. These regions represent ‘gates’ where cellular signals converge and gene expression may be downregulated by competing nuclear factors or potentiated by cooperative recruitment of CBP/p300 and components of the transcriptional machinery. CBP and p300 are modular proteins that contain globular domains separated by long intrinsically disordered regions (Dyson and Wright, 2005). The TAZ1 and TAZ2 domains of CBP (Ponting et al, 1996) (sometimes named CH1 and CH3, respectively) mediate interactions with the transactivation domains (TADs) of numerous transcription factors. Sequence identity between CBP and p300 in these modules is high (>95%), suggesting conserved function, structure and binding specificity. Even though the TAZ1 and TAZ2 domains share the same fold, four α-helices stabilized by three zinc atoms (De Guzman et al, 2000, 2005), they are highly selective in their interactions and bind different subsets of transactivation motifs. To investigate the determinants of binding partner selectivity by the TAZ1 and TAZ2 domains of CBP, we characterized their interactions with selected TADs from the signal transducer and activator of transcription (STAT) family of proteins. The JAK–STAT pathway has a central function in the transmission of cytokine signals from the cell membrane to the nucleus and subsequent gene activation. Members of the STAT family participate in diverse biological processes and are constitutively activated in many human malignancies and diseases (O'Sullivan et al, 2007). STAT proteins are activated by tyrosine phosphorylation and homo- or heterodimerization through a conserved SH2 domain. The seven human STAT proteins contain a poorly conserved C-terminal TAD, which displays variable transactivational potency (Shen and Darnell, 2001) and directly interacts with CBP/p300 (Brierley and Fish, 2005). The TADs of STAT1 and STAT2 bind selectively to the TAZ2 and TAZ1 domains of CBP/p300, respectively (Bhattacharya et al, 1996; Zhang et al, 1996). The C-terminal TAD of the p65 subunit of the NF-κB complex also interacts with the TAZ1 domain (Gerritsen et al, 1997). As CBP/p300 appear to be expressed at limiting concentrations in most cells, these transcription factors may either compete for binding to a specific TAZ domain or cooperatively recruit CBP/p300 through simultaneous TAZ1 and TAZ2 binding. Both cases have been reported: the TADs of STAT2 and p65 compete for binding to the CH1 domain of p300 in vitro and in vivo (Hottiger et al, 1998), and the N- and C-terminal regions of CBP cooperate to mediate the STAT1/NF-κB synergistic expression of certain inflammatory genes (Hiroi and Ohmori, 2003). In addition to the NF-κB pathway, CBP/p300 represents a point of convergence between the JAK/STAT and other signalling pathways, including the Ras/AP-1 (Horvai et al, 1997), SMAD (Ghosh et al, 2001) and Src/Hif-1α pathways (Gray et al, 2005). To elucidate the structural basis for coordinated gene expression and CBP/p300 recruitment by STAT proteins, we have determined the structures of the TAZ1 domain in complex with the TAD of STAT2 and the TAZ2 domain in complex with the TAD of STAT1. Both TADs are intrinsically unstructured and undergo folding transitions upon binding to TAZ, leading to formation of an intermolecular hydrophobic core. Interestingly, although the binding mechanisms in the two complexes are similar and the TAZ1 and TAZ2 domains are topologically similar, the structures of the bound STAT1-TAD and STAT2-TAD are very different. The subtle differences in helix packing between TAZ1 and TAZ2 result in significantly different surface grooves for binding of targets, and appear to be the primary determinants of target selectivity. Results Binding affinity and selectivity Previous studies have mapped interactions between the C-terminal TADs of STAT1 (Zhang et al, 1996), STAT2 (Bhattacharya et al, 1996), STAT3 (Schuringa et al, 2001) and STAT4 (O'Sullivan et al, 2004) to regions of CBP encompassing the TAZ1 or TAZ2 domains (Figure 1A). To investigate the relative binding affinities of these TADs for the isolated TAZ1 or TAZ2 domains, we performed in vitro pull-down experiments. Amino-acid sequences from the human STAT1 (residues 710–750), STAT2 (residues 769–851), STAT3 (residues 719–764) and STAT4 (residues 702–748) proteins (Figure 1B and C) were fused to the C terminus of the B1 domain from protein G (GB1). The GB1–STAT fusions were bound to IgG resin, incubated with either the isolated TAZ1 domain (residues 340–439) or TAZ2 domain (1764–1855) of CBP. All of the GB1–STAT fusion constructs interacted to some extent with TAZ2, but binding to the TAZ1 domain was more selective (Figure 2A). The GB1–STAT1 and GB1–STAT2 constructs showed higher affinity for TAZ2 compared with the GB1–STAT3 and GB1–STAT4 TADs. The TAZ1 domain showed preferential binding to the GB1–STAT2 TAD. A shorter construct of STAT2 (residues 786–838, C793H/C809S) used for structural studies retained high-affinity binding to TAZ1 but not to TAZ2, suggesting that the deleted region contains a TAZ2-binding site or enhances nonspecific binding to TAZ2. Figure 1.Characterization of STAT-TAD and TAZ domains. (A) Domain structure of mouse CBP showing the positions of the TAZ1 and TAZ2 domains in the mouse sequence. (B) Alignment of the human STAT1, STAT3 and STAT4 TAD amino-acid sequences. Bold-faced type highlights amino acids that are similar between family members. Asterisks indicate residues in the STAT1-TAD that directly contact TAZ2 in the complex structure. The grey bar denotes amino acids that form the helix in the bound STAT1-TAD. The circled ‘P’ denotes the position of the phosphorylation site. A dash in the sequence indicates a gap with no corresponding residue. (C) Amino-acid sequences of the human STAT2 C-terminal TAD (ctTAD) sequences and the truncated construct with two point mutations (nmrTAD). Underlined residues indicate an imperfect repeat in the STAT2-ctTAD sequence. A dash in the sequence indicates a gap with no corresponding residue. A dot in the STAT2-nmrTAD sequence indicates a position where the amino acid is identical to STAT2-ctTAD. Asterisks indicate residues in the STAT2-nmrTAD that contact TAZ1 in the complex structure. Grey bars denote the amino acids that comprise the three helical structures in the bound STAT2-nmrTAD. Download figure Download PowerPoint Figure 2.Interactions of TAZ and STAT domains. (A) SDS–PAGE analysis of mouse TAZ1 and TAZ2 assayed for interactions with the immobilized human STAT sequences shown in Figure 1B and C. GB1 (lanes 3 and 10), GB1–STAT1-TAD (lanes 4 and 11), GB1–STAT2-ctTAD (lanes 5 and 12), GB1–STAT2-nmrTAD (lanes 6 and 13), GB1–STAT3-TAD (lanes 7 and 14) and GB1–STAT4-TAD (lanes 8 and 15) bound to IgG Sepharose were incubated with purified TAZ1 (lanes 3–8) or TAZ2 (lanes 10–15). The bands at ∼25 and ∼50 kDa (lanes 3–8 and 10–15) correspond to IgG light and heavy chains. Lanes 2 and 9 contain free TAZ1 and TAZ2, respectively. (B) Results of pull-down assays showing that phosphorylation of S727 does not enhance the affinity of STAT1 TAD for CBP. TADs of STAT1 and MLL (used as a control for KIX binding) were bound to IgG Sepharose through an N-terminal GB1 fusion tag and binding of CBP domains TAZ1, KIX, TAZ2 and NCBD to STAT1 TAD (710–750), phosphoS727 and S727A STAT1 TADs was measured by pull-down assays monitored by reversed phase HPLC. Download figure Download PowerPoint To quantify the high-affinity interactions observed in the pull-down assay, we determined dissociation constants (Kd) using isothermal titration calorimetry (ITC). The complex formed between GB1–STAT1(710–750) and TAZ2 has an apparent Kd of 52±31 nM, whereas that between GB1–STAT2(786–838, C793H/C809S) and TAZ1 has a Kd of 58±3 nM. Binding to the non-cognate TAZ domains is ∼100-fold weaker; STAT1 binding to TAZ1 has a Kd of 4±1.4 μM and STAT2 binding to TAZ2 has a Kd of 5 μM. Serine phosphorylation within the STAT1, STAT3 and STAT4 TADs (Figure 1B) increases transcriptional activity and CBP/p300 recruitment in vivo (Wen et al, 1995; Decker and Kovarik, 2000) To determine the effect of phosphorylation on CBP binding, we expressed GB1–STAT1 TAD phosphorylated on S727 and measured binding affinities for TAZ1 and TAZ2 by ITC. No change in affinity was observed relative to the non-phosphorylated TAD (Kd=87±20 nM and 3.7±0.6 μM for the TAZ2 and TAZ1 domains, respectively). These results were confirmed by pull-down assays that showed that phosphorylation does not change the affinity for the KIX domain, the binding site of the phosphorylated CREB TAD (Parker et al, 1996) and impairs binding to the NCBD (Figure 2B). Mapping STAT TADs for structure determination The minimal high-affinity binding site for TAZ1 on the STAT2 TAD was identified by screening the interactions of a set of truncated TAD constructs with 15N-labelled TAZ1 using 2D 1H–15N HSQC spectra. The region from G748 to the C terminus (F851) was mapped and a construct spanning residues V786-S838 found to be optimal for NMR analysis. Two mutations (C793H and C809S) were made to enhance sample stability; these cysteines are poorly conserved among mammalian STAT2 proteins and the mutations result in a sequence identical to that from pig. As the full-length TAD of STAT1 is relatively short, we generated only two constructs, one starting at residue S710 and extending to residue S745 and the other from S710 to the C terminus, V750. Both peptides bind with similar affinity to TAZ2, suggesting that the C-terminal five amino acids are not essential for complex formation. Nevertheless, the longer construct (S710-V750) was used for structure determination. The isolated STAT1-TAD and STAT2-TAD are intrinsically unstructured; the amide proton resonances occupy narrow chemical shift ranges in 1H–15N HSQC spectra (Figure 3). Upon addition of the cognate TAZ domain, the amide proton resonances disperse and exhibit chemical shifts indicative of a stable, folded structure. Both complexes are in slow exchange on the NMR timescale. The amide proton resonances of residues 720–750 in the STAT1-TAD and residues L791-M816 of the STAT2 TAD show the largest deviations from random coil chemical shifts. The C terminus of STAT2-TAD (P817-S838) undergoes only small chemical shift perturbations and 15N heteronuclear NOEs show that it remains flexible in the complex. Figure 3.NMR spectra of free and bound STAT-TAD. (A) 1H–15N HSQC spectrum of STAT1-TAD free (black) and bound to TAZ2 (red). (B) 1H–15N HSQC spectrum of STAT2-TAD free (black) and bound to TAZ1 (red). Selected cross-peaks of the bound protein are labelled. Download figure Download PowerPoint Structure determination The structures of the STAT1-TAD:TAZ2 and STAT2-TAD:TAZ1 complexes were determined using heteronuclear multidimensional NMR. To minimize ambiguity due to spectral overlap, several samples were prepared in which one component was uniformly 15N- or 15N/13C-labelled and the other component was at natural isotopic abundance. This labelling scheme enabled chemical shift assignments for >90% of the 1H, 13C and 15N resonances for both components of each complex. Experimentally derived distance and dihedral angle restraints were used to calculate and refine an ensemble of 20 structures for the STAT2-TAD:TAZ1 (Figure 4A) and STAT1-TAD:TAZ2 (Figure 5A) complexes. Each TAZ domain contains four α-helices, designated H1–H4, and three zinc atoms and constitutes a pre-formed scaffold for ligand binding (De Guzman et al, 2005). The structures in complex with the STAT2-TAD and STAT1-TAD can be superimposed on those of the respective free proteins with r.m.s. deviations of 1.8 Å for TAZ1 and 1.5 Å for TAZ2 for backbone heavy atoms (N, Cα, C′). Figure 4.NMR solution structure of the TAZ1:STAT2-TAD. (A) Ensemble of the 20 lowest energy structures superimposed on the backbone atoms of TAZ1:STAT2-TAD (TAZ1, residues Pro347-Ala372 and Cys384-Asn434; STAT2-TAD, residues Asp790-Met816). The backbones of TAZ1 and STAT2-TAD are shown in blue and green, respectively. The boundaries of the well-structured region of STAT2-TAD, and the N and C termini of TAZ1 are labelled. The backbone atoms of the flexible C-terminal region of the STAT2-TAD (residues Pro817-Ser835) are shown in light green, and the flexible loop of TAZ1 (residues 373–383) is shown in light blue. (B) Ribbon diagram showing the backbone of the lowest energy structure with zinc atoms and secondary structure elements of TAZ1 labelled. The ribbon is traced from residues Thr344-Ser436 of TAZ1 (blue) and Glu788-Met816 of the STAT2-TAD (green), with the C-terminal tail shown in light green. (C–E) Close-up views showing TAZ1 (yellow surface) and STAT2-TAD (white backbone), with the side chains of residues that form intermolecular contacts. The colour scheme for the side chains is as follows: acidic residues (Asp and Glu), red; basic residues (Arg, Lys and His), blue; neutral polar residues (Ser, Thr and Gln) and backbone carbonyl atoms, cyan; hydrophobic residues (Ala, Leu, Ile, Phe, Tyr, Val, Met and Pro), green. Download figure Download PowerPoint Figure 5.NMR solution structure of the TAZ2:STAT1-TAD. (A) Ensemble of the 20 lowest energy structures superimposed on the backbone atoms of TAZ2:STAT1-TAD (TAZ2, residues Gln1766-Arg1852; STAT1-TAD, residues Asp721-Asn748). The backbones of TAZ2 and STAT1-TAD are shown in purple and yellow, respectively. The N and C termini of TAZ2 and STAT1-TAD are labelled. (B) Ribbon diagram showing the backbone of the lowest energy structure with zinc atoms and secondary structure elements of TAZ2 labelled. (C–E) Close-up views showing TAZ2 (yellow surface) and STAT1-TAD (white backbone), with the side chains of residues that form intermolecular contacts. The colour scheme for the side chains is the same as for Figure 4. The position of the side chain of S727, the phosphorylation site characterized in Figure 1, is shown in magenta. Download figure Download PowerPoint Structure of the bound STAT2-TAD In the STAT2-TAD:TAZ1 complex, the STAT2-TAD wraps entirely around helix H1 of TAZ1, and contacts amino acids along a spiral groove formed by the packing of helices H1/H3, H1/H2 and H1/H4 (Figure 4B). The structure of STAT2-TAD is predominantly extended, with a short alpha helix (H793-H796) and another 310 helix (P802-F806). The N-terminal region of the STAT2-TAD recognizes the surface of TAZ1 formed by helix H1 and the Zn3-binding site (Figure 4C), where the side chains of E788, D790 and D794 form intermolecular electrostatic interactions with the side chains of R350, R423 and H417. The hydrophobic groove containing the side chains of I353, I414 and V428 in TAZ1 presents a favourable surface for interactions with L791, L795 and L798 of the STAT2-TAD. The walls of this groove are formed by H417 and P427, which contact P792 and P789. The side chain of M803 also contributes to this intermolecular hydrophobic core. Residues P802-F806 form a short helical structure, with the aromatic ring of F806 stacked between two imidazole rings (H407 and H383) and in contact with the methyl groups of L359 and M387 (Figure 4D). The side chain of R807 forms a salt bridge with the carboxylate of E348, whereas the backbone is positioned by a hydrogen bond between the carbonyl oxygen and the Nε of Q355. Residues N808 and S809 are solvent exposed and form a turn that positions residues V810-M816 for interactions with a groove formed by helix H1, the Zn1-binding site and the N terminus of helix H4 in TAZ1 (Figure 4E). The side chains from residues V358, L359, L361, L381, W418, P431, V432 and A435 form a large hydrophobic surface that accommodates the side chains of V810, I812, I815 and M816 of the STAT2 TAD. The C terminus (P817-S838) of the bound STAT2 is more flexible, yet residues in this region contact the TAZ1 domain. The side chains of V829 and Y833 pack against an exposed hydrophobic patch formed by the side chains of L391, M394 and I415, and K419 forms an electrostatic interaction with D830 in the majority of structures in the ensemble. These interactions are not essential for TAZ1 binding but apparently contribute to the overall stability of the complex. Structure of the bound STAT1-TAD The structure of STAT1-TAD bound to TAZ2 differs dramatically from that of STAT2-TAD bound to TAZ1. Unlike the extended surface over which the STAT2-TAD binds TAZ1, the STAT1-TAD is localized to one face of TAZ2 and adopts a single well-defined helix (residues P728-V738) that is longer than any of the STAT2 helices. This helix binds in a large hydrophobic surface created by the packing of helices H1, H2 and H3 of TAZ2 (Figure 5B). Residues D721-S740 of the STAT1-TAD form a wedge, with its vertex at P725 that penetrates into a deep hydrophobic cavity formed by the intersection of helices H1, H2 and H3 and packs against the side chains of M1799, V1802, L1826 and the aliphatic portion of Q1822 (Figure 5C). The side chain of Q1822 forms a buried hydrogen bond with the backbone carbonyl group of M726 and the methionine side chain projects deeply into the TAZ2 domain to contact the methyl groups of A1825 and L1826. The N-terminal region of the STAT1-TAD adopts an extended backbone conformation that is stabilized by hydrophobic contacts from L724 to P1818 and V1819 and an intermolecular salt bridge between the side chains of D721 and R1811. Within the amphipathic α-helix between residues P728-V738, the side chains of F731, V734, I737 and V738 form a hydrophobic face that contacts the side chains of I1773, A1825, L1826 and Y1829 located in helices H1 and H3 (Figure 5D). The opposite face of the helix is negatively charged and consists of E729, E730, D732 and E733, which form electrostatic interactions with the guanidinium groups of R1769, R1770 and R1775 from helix H1 of TAZ2. The helix terminates at G739, which forms a turn and positions the C-terminal portion of the STAT1-TAD to rest in the cleft between the Zn2-binding site, helix H3 and helix H4 (Figure 5E). A hydrophobic patch formed by the side chains of T1813, L1824, I1847, L1851 and F1843 and the aliphatic portions of K1821 and K1850 are contacted by the side chains of V741, F743, M746 and M747. Residues at the N terminus of STAT1 (S710-D721) are disordered and do not contact TAZ2 in the structures. Both TADs bury considerable surface area upon complex formation, ∼2300 Å2 for TAZ1:STAT2-TAD and ∼1600 Å2 for TAZ2:STAT1-TAD. These extensive interfaces encompass numerous highly specific intermolecular interactions that constitute the molecular basis for high-affinity binding and selective recruitment of the CBP TAZ2 and TAZ1 domains, respectively, by the STAT1 and STAT2 TADs. Complementary electrostatics mediate recognition of acidic TADs The surfaces of both TAZ domains are highly electropositive (Figure 6), favouring binding of acidic TADs. The electrostatic interactions between the TAZ2 domain and the STAT1-TAD are localized to one face of the TAZ2 scaffold (Figure 6A) and arise exclusively from acidic residues in STAT1-TAD contacting basic residues in TAZ2. The opposite surface of TAZ2 is composed of mostly uncharged amino acids and appears less competent for binding negatively charged TADs. By contrast, the surface of the TAZ1 domain has a more even distribution of positive charges (Figure 6B). The STAT2-TAD wraps around the TAZ1 domain, forming multiple electrostatic interactions with surface positive charges. In addition, the positively charged side chain of R807 in STAT2-TAD forms an electrostatic interaction with the negatively charged side chain of E348 of TAZ1. The preferential binding of acidic activation domains by TAZ domains and the evidence for charge complementarity of the bound STAT1-TAD and STAT2-TAD suggest that electrostatic interactions have a key function in determining the specificity of TAD recognition, and the intermolecular hydrophobic core that forms upon TAD folding provides stability to the complex. Figure 6.Electrostatic potentials on the surfaces of the TAZ domains. Positively charged surfaces are shown in blue, and negatively charged in red. (A) TAZ2 in complex with the STAT1-TAD (yellow backbone). (B) TAZ1 in complex with STAT2-TAD (green backbone). The left and right images represent a 180° rotation around the vertical axis in the plane of the page. (C) Superposition of the structure of STAT2-TAD (green) on the TAZ2:STAT1-TAD complex (transparent surface plus yellow backbone). (D) Superposition of the structure of STAT1-TAD (yellow) on the TAZ1:STAT2-TAD complex (transparent surface plus green backbone). Download figure Download PowerPoint Structural basis for TAD recognition Aside from specific intermolecular side chain interactions, which predicate the precision of the sequence specificity, the structural differences between TAZ1 and TAZ2 provide a further basis for binding partner recognition and discrimination. Comparison of the free TAZ1 and TAZ2 structures reveals that, even though the overall topologies are very similar, there are significant differences in the orientation of helix H4 (De Guzman et al, 2000, 2005). Helix H4 of TAZ1 forms part of the binding surface for the C-terminal activation domains of HIF-1α and CITED2 (De Guzman et al, 2005) and appears also to be a major determinant for recognition of the STAT2-TAD. The C terminus of helix H4 packs against helix H1 in TAZ1, forming a hydrophobic pocket that binds residues I812-M816 of the STAT2-TAD. Both surfaces are dramatically different when helix H4 is positioned as in the TAZ2 structure (Figure 6C and D). In addition to differences in the orientation of helix H4, differences in the packing of helix H3 also appear to have a function in binding partner discrimination by TAZ domains. The helix H2–H3 packing angle differs by ∼35° in the structures of TAZ1 and TAZ2 (De Guzman et al, 2000, 2005). The position of H3 in TAZ2 creates an exposed hydrophobic pocket contacted by the LPMSP motif in STAT1-TAD (Figure 6C). In TAZ1, the change in the packing angle of helices H3 and H1 occludes this binding pocket (Figure 6D). This is consistent with the binding assays, which show only very weak interaction between TAZ1 and the STAT1, STAT3 and STAT4 TADs containing the LPMSP motif (Figure 2A). A model of the STAT2-TAD superimposed onto the structure of TAZ2 (Figure 6C) predicts a more favourable interaction devoid of such steric clashes, and this too is consistent with the ability of GB1–STAT2 to pull-down TAZ2 in vitro. The helix packing in TAZ2 creates a contiguous U-shaped interface for protein binding. The centre of this interface contains a hydrophobic groove that contacts the helix (residues P728-V738) in the bound STAT1-TAD structure (Figure 6A). The groove is formed by the side chains of R1775 and K1832, which are ∼20 Å apart in the structure, and the side chains of R1769 and K1821, which are ∼18 Å apart. The structure and charge distribution of this groove present a favourable surface to bind a negatively charged, amphipathic helix. The groove is wide, suggesting it could accommodate a helix with bulky side chains (such as F743 in STAT1-TAD) or slightly different orientations without causing steric clashes. Adjacent to the helix-binding groove is a deep and narrow hydrophobic cleft that contacts the LPMSP motif in the STAT1-TAD:TAZ2 complex and favours binding of extended structure rather than helix. The combination of these features presents a contiguous surface that binds and stabilizes the compact structure of the bound STAT1-TAD. The helices in TAZ1 pack to form a spiral groove in the surface that supports coupled folding and binding of the STAT2-TAD. Although the dimensions of this groove vary, it is in general narrower than the helix-binding groove and shallower than the LPMSP-binding cleft of TAZ2. Unlike the relatively flat protein-binding surface on TAZ2, the groove in TAZ1 wraps around the structure presenting an extensive concave surface for protein binding. This feature induces the STAT2-TAD to adopt an extended structure, with three short helical turns that contact isolated regions along the spiral groove. Discussion The structures of the STAT1 and STAT2 complexes reported here reveal different modes of binding of transcriptional activation domains by the TAZ1 and TAZ2 domains of CBP/p300. The distinct topologies of the TAZ1 and TAZ2 domains appear to form the basis for binding partner selectivity (De Guzman et al, 2005). The surface of the TAZ1 domain contains interlinked hydrophobic grooves that favour interactions with long regions of intrinsically disordered TADs that can form one or more local helical segments upon binding and also interact with the narrower grooves of TAZ1 in relatively extended conformations. Many of these grooves are absent from the surface of the TAZ2 domain. The relative affinities with which an intrinsically disordered TAD binds to TAZ1 or TAZ2 depend upon its ability to fold and form the appropriate stabilizing interactions, as discussed in detail below. TAZ recognition by other binding par
Referência(s)