Molecular Determinants of the Interaction of Mad with the PAH2 Domain of mSin3*

The Sin3 co-repressor acts as a protein scaffold to recruit transcription factors via its four highly homolo-gous paired amphipathic helix (PAH) domains. PAH2 has been shown to interact strongly with the Sin3 interacting domain (SID) of the tumor suppressor Mad. This PAH2/Mad complex has been studied extensively by NMR, but the molecular determinants that dictate the specificity of interaction remain to be elucidated. To uncover residues that convey the specificity of interaction between PAH2 and Mad, PAH2 residues contacted by the Mad-SID were introduced into the PAH1 domain of mSin3b and tested for gain-of-interaction in vivo in a yeast two-hybrid setting and further confirmed in a cell-free system. This approach led to the identification of PAH2-Phe-7 as a critical residue. Stabilization of the interaction between PAH1-Phe-7 and the Mad-SID was achieved by introducing Val-14 and Gln-39 into PAH1. Substitution of PAH2 residues contacted by the Mad-SID with their respective residues in PAH1 corrobo-rated and extended the critical role of Phe-7 and the stabilizing role of Val-14 and Gln-39. We conclude that Phe-7 is the critical determinant and provides the molecular specificity for the association between Sin3 and Mad in regulating cell growth and differentiation. The ordered and specific/selective recruitment of multipro-tein complexes with intrinsic chromatin remodelling or histone modifying activities to cis-acting DNA EcoRI fragment yielding PBS-PAH1-PAH2. Site-directed mutagenesis was performed on PBS-PAH1-PAH2 by using the QuikChange site-directed mutagenesis kit (Stratagene) following the manufacturer’s pro- tocol. All mutations were confirmed by DNA sequencing. Wild type or mutated PAH1 fragments of msin3b containing amino acids 25–107 were generated by PCR with the primer pair PAH1F97/ PAH1R97m containing, respectively, SalI and NotI restriction site overhangs. PAH1F97 was composed of 5 (cid:1) -GTGAGGTCGACCGGAG-GGCACGAGAAGCTG-3 (cid:1) , and PAH1R97m was composed of 5 (cid:1) -GTC- AGTAGCGGCCGCCTTGGGGATATCTATACGGTATCCA-3 (cid:1) . Wild type or mutated PAH2 fragments of msin3b comprising amino acids were generated by PCR using the PAH2F97/PAH2R97 primer pair containing SalI and NotI restriction sites. PAH2F97 was composed of 5 (cid:1) -GTGAGGTCGACCATGTCCTACAAGGAGGACAGA-G-3 (cid:1) , and PAH2R97 was composed of 5 (cid:1) -GTCAGTAGCGGCCGCAG- ACACAGGGCGCAGGAGTGA-3 (cid:1) . In-frame cloning of the amplified fragments was performed by using SalI/NotI cleaved pCP97 bait ve- ctor containing the GAL4/DNA-binding domain (40). We generated PAH1 and PAH2 PCR fragments by using primer pair PGEXPAH1- F/PGEXPAH1R or PGEXPAH2F/PGEXPAH2R, containing BamH1 and EcoR1 restriction sites. Amplified fragments were fused to GST using BamH1/EcoR1 sites available in PGEX-2T follows: PGEXPAH1F,

The ordered and specific/selective recruitment of multiprotein complexes with intrinsic chromatin remodelling or histone modifying activities to cis-acting DNA elements has emerged over the last decade as one of the major mechanisms to regulate transcription and cell fate determination (1). A plethora of studies have collectively shown that acetylation of specific Nterminal lysine residues of core histone tails by histone acetyl transferases (HATs) 1 creates a chromatin state which results in transcriptional activation (2). Removal of the acetyl moiety by histone deacetylases (HDACs) restores the compact state of chromatin, resulting in transcriptional repression. One of the players in this mechanism is the evolutionary highly conserved and ubiquitously expressed co-repressor Sin3. Sin3 acts as a scaffold protein in many complexes and presumably contrib-utes to repression of transcription by interacting with human class 1 or class 2 histone deacetylases, HDAC1 or HDAC2, or the Saccharomyces cerevisiae homolog Rpd3 (3)(4)(5)(6)(7)(8)(9). Purification of Sin3-containing complexes has revealed a core complex containing HDAC1/2, the pocket binding protein RBP1, Sds3, the histone binding proteins RbAp46 and RbAp48, and accessory proteins such as SAP30 and SAP130 (9 -13).
A large number of transcription factors have been described to recruit Sin3 to target genes. These include: Pf1, TIEG2, Opi1, Mnt, Mad family members, and S. cerevisiae Ume6 (6, 14 -23). The interaction between Mad and Sin3 is of great interest as it targets deacetylation and repression of transcription to genes required for proliferation of cells. Mad belongs to the basic helix-loop-helix leucine zipper protein family and is part of the Myc/Mad/Max network. The tumor suppressor Mad and the proto-oncogene Myc independently heterodimerizes with Max to form complexes capable of targeting the hexanucleotide core sequences 5Ј-CACGTG-3Ј enhancer box present on specific promoters such as cyclin D2, htert, and Id2 (24 -29). The Myc/Max heterodimer can recruit the TRRAP/GCN5 complex containing HAT activity in proliferating cells, as opposed to the Mad/Max heterodimer, which recruits the Sin3 complex triggering deacetylation of histones in differentiation (30 -32).
Interactions between Sin3 and transcription factors are mediated by four PAH domains within the Sin3 protein. Opi1, pf1, N-CoR, and the SMRT co-repressor have been described to interact with PAH1 (33,34). PAH2 has been reported to interact with Mad family members, Mnt, Ume6, as well as with TIEG2 and Pf1. So far, PAH3 has been shown to interact with SAP30. Recently glycosylation was linked to transcriptional repression by showing interaction between PAH4 and O-Glc-NAc transferase (35). A minimal conserved binding motif consisting of the N terminus residue 8 -20 of Mad1 was established to interact strongly to PAH2 (36). The interaction between the PAH2 domain of Sin3 with the SID of Mad has been characterized in detail by NMR. PAH2 folds as a wedged four-helix bundle structure that is stabilized upon complex formation with the Mad-SID, which adopts an amphipathic ␣-helix (37,38). Although the four domains have a high degree of homology, they seem to interact with distinct subsets of factors. Structural prerequisites of PAH interactors for a stable interaction with a PAH domain seem to be ␣-helicity and amphipathicity, but the structural studies do not shed light on the role of individual residues in binding and specificity.
Here, to uncover residues that convey the specificity of interaction between PAH2 and Mad, PAH2 residues contacted by the Mad-SID were introduced into the PAH1 domain of mSin3b and tested for gain of interaction. We report that the insertion of PAH2-specific residue Phe-7 in PAH1 conveyed an interaction with the Mad-SID. Stabilization of Mad-SID interaction with PAH1-Phe-7 occurred upon extra insertion of Val-14 and Gln-39. Strikingly, in PAH2, Phe-7 was further identified as a critical residue. Overall, we established Phe-7 as the most important molecular determinant of the interaction between PAH2 and Mad.
Cloning and Mutagenesis-Fragment containing the PAH1-PAH2 domain of pvzmsin3b was cloned into PBS-SK (Stratagene) as a SacII/ FIG. 1. Mad selectivity at the PAH1 and the PAH2 domains in vitro and in vivo. A, multiple sequence alignments of the PAH1 and the PAH2 domains of sin3. Alignment was generated with ClustalW and manually improved. Helical regions of the PAH2 domains are indicated with cylinders at the bottom. Residues mutated in this study are boxed. Black dots indicate PAH2 residues involved in contacting the Mad-SID; information was retrieved from NMR structure of the PAH2 domain of msin3a and msin3b (Protein Data Bank accession numbers 1G1E and 1E91, respectively). B, in vitro pull-down assay suggests a preferential binding of Mad to PAH2. Top panel, GST-PAH1 and GST-PAH2 bind separately to Mad. Excess amounts of GST-PAH1 and GST-PAH2 were incubated with IgG-Sepharose in the presence or absence of ProtG-Mad-SID(5-24) and washed extensively. B, bound material was eluted, separated on SDS-PAGE, and Coomassie Blue-stained. Bottom panel, Mad preferentially binds to PAH2 upon reducing amounts of Mad. C, in vivo interaction in a yeast two-hybrid assay suggests an exclusive interaction of Mad with PAH2. G4DBD-PAH1 or PAH2 were transformed in yeast with a prey containing G4AD fused to the thioredoxin A protein (G4AD-TRX). Mad-SID (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24) was inserted in the constrained active-site loop of thioredoxin to generate G4AD-TRX-Mad. Quantitative ␤-galactosidase assays were performed using five fresh transformants in triplicate.
EcoRI fragment yielding PBS-PAH1-PAH2. Site-directed mutagenesis was performed on PBS-PAH1-PAH2 by using the QuikChange sitedirected mutagenesis kit (Stratagene) following the manufacturer's protocol. All mutations were confirmed by DNA sequencing.
Mad-SID containing protein sequence V 5 RMNIQMLLEAAD-YLERRER 24 was generated by annealing phosphorylated oligonucleotides containing RsrII overhangs. Cloning in-frame with GAL4 activation domain and thioredoxin A was done by using RsrII restriction sites in the prey vector pADTRX (40). All constructs were confirmed by automatic DNA sequencing.
Yeast Two-hybrid Assay-LiAc transformations were performed on S. cerevisiae Y190(Clontech) according to the Clontech yeast protocols handbook. Quantitative ␤-galactosidase assay was performed using O-nitrophenyl ␤-D-galactopyranoside (ONPG) as a substrate. Fresh transformants were grown overnight in 5 ml of S.D. (leu-trp) selective culture medium. 1.5 ml of cells were concentrated 5ϫ in Z buffer (60 mM Na 2 HPO 4 , 60 mM NaH 2 PO 4 , 10 mM KCl, 1 mM MgSO 4 , pH 7.0). Cells were resuspended in 100 l of Z buffer and freeze-thawed. 0.7 ml of Z buffer with 50 mM ␤-mercaptoethanol and 160 l (4 mg/ml) of ONPG was added to initiate the reaction. Incubation at 30°C for 30 min to 2 h was performed. Reaction was stopped with 0.4 ml of 1 M Na 2 CO 3 . ␤-Galactosidase units were calculated according to the formula ␤-galactosidase units (mol ONPG/min Ϫ1 ) ϭ 1000 ϫ A 420 /(time in min ϫ 0.1 ϫ 5 ϫ A 600 ). Experiments were repeated at least three times in triplicate.

Mad Preferentially Binds PAH2 in Vitro and Exclusively in
Vivo-The Sin3a and Sin3b co-repressors both contain four PAH domains mediating protein-protein interactions. Despite significant homology, PAH1, PAH2, PAH3, and PAH4 domains seem to associate with distinct proteins. Clustal alignment revealed a significant sequence identity between PAH1 and PAH2 domains from yeast to human of about ϳ45% (Fig. 1A). Contact residue analyses from the NMR structure of the PAH2 domain of mSin3a or mSin3b with the Mad-SID revealed a number of PAH2-specific residues. Some contact residues are conserved between PAH1 and PAH2 whereas others are distinct and may provide specificity to the interaction. Interestingly, they are exclusively located in helices 1 and 2 (Fig. 1A, boxed).
To assess the ability of PAH1 and PAH2 to interact with the Mad-SID, the domains were fused to GST, purified, and incubated with immobilized ProtG-Mad bound to IgG-Sepharose beads. When ProtG-Mad was in excess, both GST-PAH2 and GST-PAH1 were able to bind the Mad-SID (Fig. 1B). Incubating equal amounts of GST-PAH1 and GST-PAH2 with decreasing amounts of ProtG-Mad, i.e. creating a competitive setting, indicated that Mad-SID binds preferentially to PAH2 in vitro. To ascertain this differential interaction in vivo, we performed yeast two-hybrid experiments. The PAH1 and PAH2 domains of msin3b were fused to the DNA-binding domain of Gal4 (G4DBD) and used as baits. As a prey, a constrained Mad-SID displayed in the active-site loop of E. coli thioredoxin (TRX) molecule was used, fused to the gal4 activation domain (G4AD-TRX-Mad). PAH2 expression in combination with Mad resulted in a robust ␤-galactosidase level, whereas PAH1 yielded background levels (Fig. 1C). This activation mediated by G4DBD-PAH2 was strictly dependent upon the presence of the Mad-SID, as TRX alone resulted in background levels. Thus, the Mad-SID preferentially interacts with PAH2 in vitro and more selectively so in vivo.
Exchange of Contact Residues between PAH2 and PAH1-The observed selective interaction between Mad and PAH2 in vivo in a yeast two-hybrid setting suggests that unique residues within PAH2 play a selective role in the binding of the Mad-SID. We used molecular modeling of PAH1 with WHAT IF (41) and superposition of the PAH1 model with the NMR structure of the PAH2 domain of msin3b with the Mad-SID to pinpoint potential selective PAH2 residues (Fig. 2) that may affect binding of Mad-SID in vivo. This revealed Glu-6, Phe-7, Ile-11, Val-14, Asn-15, Ile-17, Arg-29, Leu-35, His-36, and Gln-39 as potential selective modifiers. Therefore, we performed a systematic mutational analysis, replacing the respective PAH1 residues with the potentially selective PAH2 residues. Each mutant was then fused to the G4DBD and tested in the yeast two-hybrid setting.
FIG. 2. PAH1 molecular model derived from homology modeling of the NMR structure of the PAH2 domain of msin3b with the Mad-SID(5-24) points to potential PAH2 selective residues. PAH1 residues depicted at the interface with the Mad-SID were mutated into the respective PAH2 residues described in the table (left panel). The PAH1 model was generated with WHAT IF, and the molecular picture was generated with YASARA (42) and POVRAY. PAH1 is displayed in blue and Mad-SID(5-24) is displayed in orange. Side-chains of PAH1 residues mutated in this study are shown and labeled.
Single amino acid replacement in PAH1 revealed only one mutation, M35L, that yielded elevated levels of ␤-galactosidase expression, i.e. 5% of the level obtained with PAH2 and Mad (Fig. 3A). To identify additional residues, a second cycle of single amino acid substitutions was performed this time in the context of M35L. Some combinations, such as K36H with M35L, yielded reduced levels, whereas others seemed neutral. A significant enhancement of the interaction was observed with the double mutant containing V7F (Fig. 3B). Unexpectedly, two-hybrid assays performed in the absence of Mad or in the absence of a prey showed that PAH1-M35L had gained intrinsic transcriptional activity. On top of this, intrinsic transcriptional activity displayed by PAH1-M35L, the double-mutant PAH1-V7F/M35L yielded a significantly higher ␤-galactosidase level that was dependent upon the presence of the Mad-SID. Apparently, the intrinsic activity of PAH1-M35L had elevated the overall activity into a measurable range, thereby facilitating detection of the V7F effect in the yeast two-hybrid assays. As observed in Fig. 3A, the interaction of PAH1-V7F with Mad-SID was not sufficient to boost ␤-galactosidase levels above background.
Given the "disturbing" intrinsic transcriptional activity of PAH1-M35L, we decided to re-screen for second-site mutations, this time in the context of V7F. Combining PAH1-V7F with L14V and, to a lesser extent, K39Q, significantly boosted ␤-galactosidase levels, whereas all other combinations remained at background levels (Fig. 3, C and E). Control experiments showed that the ␤-galactosidase level was strictly dependent upon the presence of Mad-SID. To test whether the affinity of the mutated PAH1 for the Mad-SID could be further improved, third-site mutations were introduced, this time in the context of PAH1-V7F/L14V (Fig. 3, D and E). As expected from the second screen, K39Q increased the ␤-galactosidase level (L11I and K36H showed modest effects). Interaction of PAH1-V7F/ L14V/K39Q with Mad yielded ␤-galactosidase activity up to ϳ30% of that obtained with PAH2 and Mad in this yeast two-hybrid setting. Remarkably, PAH1-L14V/K39Q yielded background levels of ␤-galactosidase, thus reinforcing the critical role of Phe-7 and suggesting that Val-14 and Gln-39 contribute to the strength of the interaction, but do not represent the strongest determinants.
Phe-7 Is Critical in PAH2/Mad Interaction-To corroborate and extend the importance of PAH2 contact residues 7, 14, and 39 in the interaction with Mad-SID, these residues in PAH2 were mutated into the amino acid present at the respective positions in PAH1 and screened for their ability to interact with the Mad-SID. Strikingly, PAH2-F7V completely abolished the interaction with Mad-SID, further substantiating its critical role (Fig. 4). Single mutations of V14L and Q39K displayed only a reduced but appreciable level of interaction between PAH2 and Mad (17.49 and 8.71%, respectively, of PAH2/Mad wild-type activity). Remarkably, the double-mutant PAH2-V14L/Q39K still supported a low level but substantial interaction. When tested in the context of F7V, the V14L or K39Q alone or combined remained at a background level. We conclude that Phe-7 plays a key role in establishing the specific interaction between PAH2 and Mad-SID.
Structural Determinants of the Mad-SID: Constraints versus Helical Propensity-In our assays, the Mad-SID was conformationally constrained by insertion into the surface loop of the TRX that is between two cysteines. In its native context, Mad-SID is located at the N terminus of Mad and can freely adopt any structure. Therefore, we evaluated the interaction properties of the PAH domains with the Mad-SID when displayed in a more flexible environment. To achieve this, a stop codon was inserted downstream of the Mad sequence before the second cysteine, thus creating a C terminus. In this C-terminal, "flexible" context, Mad-SID strongly interacted with PAH2 but not detectably with PAH1, and hence the specificity of interaction was not altered in this context (Fig. 5A). The single-mutants PAH1-L14V and PAH1-K39Q remained at background levels (Fig. 5B), whereas a modest but reproducible interaction was obtained with PAH1-V7F. This modest gain of interaction is probably due to unleashing free helical properties. Similarly, the double-mutants PAH1-V7F/L14V and V7F/K39Q displayed a 3-fold higher activity, as compared with the constrained context. The PAH1-L14V/K39Q, lacking the critical V7F, remained at background level. The flexible triple mutant displayed roughly equal ␤-galactosidase activity (to ϳ30%) as compared with the constrained version. These results show that free helicity plays an important role in establishing an interaction with Mad, but without a phenylalanine at position 7, unleashing of free helical properties is insufficient to support an interaction.
The interaction of the flexible Mad-SID with single PAH2 mutants V14L and K39Q or their combination yielded a 5-fold higher level of ␤-galactosidase, as compared with the constrained setting (Fig. 5C). Mutation of F7V again abolished the interaction, as observed in constrained conditions. Noticeably, the gain of free movement by Mad-SID tolerated and partially compensated for the loss of interaction because of V14L and Q39K, but not when the mutation F7V was present.
To extend and validate in a cell-free system the observed role of PAH residues 7, 14, and 39 in the interaction with Mad, we used the IgG pull-down assay. PAH1 and PAH2 mutants were fused to GST, purified, and incubated with fixed amounts of immobilized ProtG-Mad bound to IgG-Sepharose beads (Fig.  5D). In the context of PAH1, Phe-7 partially facilitated binding to Mad-SID, which was further enhanced with the inclusion of Val-14 and/or Gln-39. Val-14 in the context of PAH1 displayed some binding to the Mad-SID, whereas Gln-39 was unable to support an interaction. Overall, the in vitro data show that Phe-7 is the residue that contributes most for binding, followed by Val-14 and Gln-39. In the reverse situation, mutation in PAH2 of Phe-7 abolished the interaction in vitro, as observed in the yeast two-hybrid setting. Mutating Gln-39 in PAH2 slightly affected the interaction, whereas no effect could be observed when Val-14 was mutated. Combining Phe-7 and Gln-39 mutations yielded background levels as observed in vivo. Mutations combining Phe-7 and Val-14 in PAH2 resulted in a substantial interaction, whereas no effect could be observed from combining Val-14 and Gln-39. We conclude that the phenylalanine at position 7 is the specific, critical molecular determinant of complex formation between PAH2 and Mad-SID in vivo in a yeast two-hybrid assay, as well as in vitro in a direct interaction assay. DISCUSSION In this study, we employed NMR structural information and homology modeling to direct a mutagenesis screen and identify amino acids conveying specificity of PAH2/Mad-SID interaction. Introduction of PAH2 contact residues Phe-7, Val-14, and Gln-39 at their corresponding position in PAH1 enabled a specific interaction with the Mad-SID in vivo and in vitro. Phe-7 at the PAH1 domain restores on its own an interaction with the Mad-SID in a free helical environment, whereas Val-14 and Gln-39 stabilize this interaction. Furthermore, we showed that mutation of Val-14 and Gln-39 in PAH2 were not critical for the interaction with Mad-SID, as opposed to the change of Phe-7 into a Val, which abolished the interaction. Overall, these results demonstrate that, to a large extent, Phe-7 conveys the specificity of the interaction between PAH2 and Mad-SID in vivo in a yeast two-hybrid setting and in vitro in a pull-down assay.
An drophilic-hydrophilic contacts with other PAH2 residues, such as His-36, Thr-37, Tyr-38, and Gln-42. Therefore, Gln-39 could have an important role in charge maintenance. A lysine muta-tion would lead to a unilateral positive charge and, as a consequence, these hydrophilic-hydrophilic interactions could be affected and disturb the folding of PAH2.
Careful inspection of the NMR structure with respect to Val-14 residue as well as CSU analysis show intermolecular contacts with conserved residues Leu-12, Ala-15, Ala-16, and Leu-19 of the Mad-SID. It has already been remarked that the short side-chain of Ala-15 and Ala-16 of the Mad-SID might play a crucial role in allowing the maintenance of the PAH2/ Mad-SID interface. Val-14 fits in a hole formed by Ala-15, Ala-16, and Leu-19 of the Mad-SID (38). Replacement of Val-14 with a leucine would increase the side-chain length. This increase could generate a distance as close as 2.4 Å between Leu-14 in PAH2 and Ala-15 in the Mad-SID. It suggests that Leu-14 bumps into the Mad-SID, and thus explains our results with the PAH2-V14L mutation. The PAH1 molecular model depicts a distance as close as 2.2 Å between Leu-14 and Ala-15 and 2.6 Å between Leu-14 and Ala-15. Proper contact distances can be restored when introducing the PAH1-L14V mutation. Distance can be measured as close as 3.4 Å between Val-14 and Ala-15 and 4.4 Å between Val-14 and Ala-15 (Fig. 6), overall strengthening the stabilizing role of Val-14.
Strikingly CSU analysis of PAH2 residue Phe-7 depicts a contact with Mad-SID residue Tyr-18. Careful inspection of the surrounding of PAH1-Phe-7 in our molecular model suggests contact with the Mad-SID aromatic residue Tyr-18. A distance as close as 3.7 Å can be measured between those residues (Fig.  6). Based upon the results from our experiments, structural FIG. 5. Free Helicity of Mad is a molecular determinant for residues 7, 14, and 39 of the PAH1 and PAH2 domains of mSin3b. The indicated PAH mutants fused to G4DBD were transformed in yeast with G4AD-TRX-Mad or a C-terminal-truncated variant G4AD-TRX-Madstop and subjected to quantitative ␤-galactosidase assay. ␤-Galactosidase units activity and standard error are indicated. Results are the average of at least five assays. A, free helicity stabilizes interaction of PAH2 with Mad. B, free helicity provides a 3-fold increase in the levels of PAH1-V7F/L14V and PAH1-V7F/K39Q and detection of PAH1-V7F. C, free helicity compensates for V14L and Q39K effects at the PAH2 domain of mSin3b. D, IgG pull-down confirms critical role of Phe-7 in vitro. Equal amounts of GST-PAH1-or GST-PAH2-mutated proteins were incubated with fixed amounts of ProtG-Mad-SID(5-24) bound to IgG-Sepharose and washed extensively. Bound material was eluted, separated on SDS-PAGE, and Coomassie Blue-stained.
FIG. 6. Model of interaction between PAH1-V7F/L14V and Mad SID. PAH1 molecular model was generated with WHAT IF and further mutated toward PAH2 specificity to generate PAH1-V7F/L14V. Molecular graphics were generated using YASARA and POVRAY. PAH1-V7F/L14V is displayed in blue, and Mad-SID(5-24) is displayed in orange. Distances between presumed contact residues are indicated in Å. information, and modeling, a stacking interaction between the aromatic rings of Phe-7 and Tyr-18 is most likely the key to the complex formation between PAH2 and Mad.
In conclusion, interaction between the Sin3 co-repressor, the tumor suppressor Mad, and other PAH interactors seems to be characterized by conserved properties such as amphipathicity and ␣-helicity and reach their specificity at a molecular level through residues such as Phe-7 that is present only at the PAH2 domain of Sin3. Open questions remain on understanding the specificity of PAH1, PAH3, and PAH4 and their interacting partners. Consensus sequences for PAH-interacting protein are still unsolved and subject to many questions. Ultimately, a definition of the specificity of every PAH domain can provide a new way to identify the residues required and needed in amphipathic ␣-helix motifs of Sin3-interacting members, enlarging our molecular view of the role of the Sin3 co-repressor complex in transcriptional repression and cell growth.