Interfacial water molecules in SH3 interactions: Getting the full picture on polyproline recognition by protein–protein interaction domains

FEBS Letters 586 (2012) 2619–2630

journal homepage: www.FEBSLetters.org

Review

Interfacial water molecules in SH3 interactions: Getting the full picture on polyproline recognition by protein–protein interaction domains Ana Zafra-Ruano, Irene Luque ⇑ Department of Physical Chemistry and Institute of Biotechnology, Faculty of Sciences, University of Granada, 18071 Granada, Spain

a r t i c l e

i n f o

Article history: Received 23 March 2012 Revised 27 April 2012 Accepted 30 April 2012 Available online 11 May 2012 Edited by Marius Sudol, Gianni Cesareni, Giulio Superti-Furga and Wilhelm Just Keywords: SH3 domain Proline-rich ligand recognition Water-mediated interaction Protein–protein interaction module

a b s t r a c t The recognition of proline-rich sequences by protein–protein interaction modules is essential for many cellular processes. Nonetheless, in spite of the wealth of structural and functional information collected over the last two decades, polyproline recognition is still not well understood. The patent inconsistency between the generally accepted description of SH3 interactions, based primarily on the stacking of hydrophobic surfaces, and their markedly exothermic character is a clear illustration of the higher complexity of these systems. Here we review the structural and thermodynamic evidence revealing the need for a revision of the current binding paradigm, incomplete and clearly insufficient for a full understanding of binding affinity and specificity, to include interfacial water molecules as universal and relevant elements in polyproline recognition. Ó 2012 Federation of European Biochemical Societies. Published by Elsevier B.V.

1. Introduction The establishment of transient protein–protein interactions is essential for intracellular signaling. A subset of these interactions are mediated by modular domains [1,2] that recognize short linear motifs in their targets [3,4]. Contrary to most protein–protein interactions, implicating large and featureless surfaces with discontinuous binding epitopes, protein interaction modules recognize well defined and continuous sequences that bind to concave sites in the domain with small interaction surfaces, similar in size to those of protein enzymes. These properties facilitate the identification of relevant physiological targets and the design of small size inhibitors [5–7]. An important class of protein-interaction domains is constituted by protein modules recognizing solvent exposed proline-rich motifs. Amongst all types of short linear motifs, proline-rich sequences are the most widely distributed in the different genomes, from prokaryotes to eukaryotes [8,9], due, to a great extent, to the specific properties conferred by the proline amino acid [10–16]. Correspondingly, polyproline-recognition domains (PRD), which seem to have coevolved with the sequences they recognize [8], are also the most abundant protein modules in many proteomes

Abbreviations: ITC, isothermal titration calorimetry; Kd, dissociation constant; SH3, Src homology 3 domain ⇑ Corresponding author. Fax: +34 958 272 879. E-mail address: [email protected] (I. Luque).

Open access under CC BY-NC-ND license.

[17,18]. Several families of PRD have been described to date, including SH3 (Src-homology 3) domains, WW domains (named after two highly conserved tryptophan residues in the domain), EVH1 (enabled vasodilator-stimulated-protein homology) domains, GYF domains (so called due to their characteristic Gly-Tyr-Phe triad) and UEV (ubiquitin E2 variant) domains, together with the protein profilin and the recently identified CAP (cytoskeleton-associated protein)-Gly domain. A comparative analysis of the structure, function and ligand recognition properties of the different PRD families can be found in several reviews [14,17,19–22]. The different PRD families vary in size (ranging between the 35 residues of the WW domains and the 150 residues of the UEV domains), present different folds and recognize specific proline-rich ‘‘core motifs’’. In spite of this, they all share a very similar mechanism for proline recognition based on the establishment of hydrophobic interactions between key proline residues in the ligand and highly conserved hydrophobic pockets in the domain, leading to a significant level of cross-reactivity within and between the different PRD families. Nonetheless, because of the central role that PRD interactions play in the regulation of many cellular processes, including cell growth, differentiation, apoptosis, splicing, endosomal trafficking or transcription [17,21–23], their interactions with their physiological targets must be specific and well regulated. This apparent paradox between specificity and promiscuity in polyproline-recognition has been the subject of intense debate [17–19,21,23–27], highlighting the need for a profound understanding of the molecular basis for binding affinity and specificity

0014-5793 Ó 2012 Federation of European Biochemical Societies. Published by Elsevier B.V. Open access under CC BY-NC-ND license. http://dx.doi.org/10.1016/j.febslet.2012.04.057

2620

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

in polyproline-recognition. Such detailed knowledge is fundamental for the development of efficient and specific inhibitors of PRD interactions as potential therapeutic agents and useful tools for the investigation of cell physiology [22,28]. 2. Function and structure of SH3 domains Among the different PRD, SH3 domains are, by far, the most abundant in vertebrates (over 300 SH3 domains have been identified in the human proteome [18,29]) and the most extensively studied [19,30]. SH3 domains are small (about 60 amino acids) non-catalytic protein modules, found in multi-domain signaling proteins of different functions and architectures that generally act as anchor sites for the recruitment of substrates or the formation of supra-molecular complexes. In some cases, these domains play also an essential role in the regulation of the enzymatic activity of the proteins that contain them through the establishment of intramolecular interactions with other elements in the molecule [31,32]. This is the case of the tyrosine kinases of the Src family, involved in carcinogenesis [33]. In addition to cancer, SH3 domains are associated to other pathologies such as leukemia, Alzheimer disease, osteoporosis, inflammatory, allergic and asthmatic processes as well as viral infections, including AIDS and hepatitis C [17,19,21,34,35]. Because of their biological relevance, SH3 domains have been widely studied, both from structural and functional perspectives. Today, more than 400 structures of SH3 domains are available including NMR and X-ray structures of free domains, complexes and full-length proteins. SH3 domains fold into a common fivestranded b-barrel structure composed of two orthogonal anti-parallel b-sheets (see Fig. 1). The five b-strands are connected by three loops: a longer one, the –RT loop, which connects strands b1 and b2 adopting an irregular anti-parallel structure and two shorter loops, the n-Src and distal loops, connecting strands b2–b3 and b3–b4 respectively. Additionally, SH3 domains also contain a short 310 helix segment placed between strands b4 and b5 [20,35]. Many efforts have also been devoted to characterizing the stability, conformational equilibrium and folding mechanism of these domains, which have been shown to fold according to a two-state model in the absence of intermediates, both in equilibrium and kinetic studies [36,37]. b-Hairpins in the domains play a very relevant role in the determination of the folding mechanism, with

Fig. 1. Structure of SH3 domains. Cartoon representation of the SH3 structure of the c-Yes SH3 domain (Pdb code: 2hda). The five b strands are shown in yellow, the 310 helix in red and the n-Src, RT and distal loops in green. Conserved aromatic residues shaping the hydrophobic pockets for proline recognition are shown as sticks.

the folding nucleus being constituted by the distal loop and 310 helix regions [38–41]. In spite of the macroscopic simplicity of the folding process, NMR H/D exchange experiments have made patent the statistical nature of the native state conformational ensemble [42,43], which translates into the existence of regions of high structural stability (the central portions of the beta strands, the short 310 helix and the conserved part of the –RT loop) and highly flexible regions, (n-Src loop and the tip of the –RT loop). According to this, the SH3 binding site has a dual character, containing regions of high and low structural stability, which facilitates the transmission of long-range cooperative effects throughout the structure [44–46], of obvious functional relevance for a protein domain implicated in signal transduction. Most SH3 domains characterized to date recognize proline-rich sequences containing a uPpuP motif (where u and p are frequently hydrophobic and proline residues, respectively) [10,20, 25,35]. SH3 ligands bind in a PPII conformation to a flat and hydrophobic surface in the domain consisting of three shallow pockets. Each of the two uP moieties in the core motif packs tightly into one hydrophobic pocket, very specific for the recognition of proline and N-substituted amino acids [47], and formed by highly conserved aromatic residues on the surface of the domain (the two tyrosine residues in the conserved ALYDY motif at the base of the –RT loop, the first tryptophan residue of the WW motif at the boundary between the b3 strand and the n-Src loop, the proline residue at strand b4 and the tyrosine residue of the SNY motif at the 310 helix region). Proline staking into these hydrophobic pockets provides the basis for the general recognition of proline-rich sequences. Nonetheless, taking into account the wealth of prolinerich motifs in the proteomes (it has been estimated that about 25% of human proteins contain proline-rich regions [21]), selectivity for proline does not assure specificity between SH3 domains or other PRD families. In general, binding specificity seems to be mostly associated to the establishment of additional interactions between residues flanking the core motif in the ligand and a third pocket delimited by variable regions of the RT and n-Src loops [17,20,48,49]. Sequence and conformational variability within these loops seem to have been exploited by SH3 domains to modulate binding affinity and specificity [25,50–52]. This mechanism of binding is shared by most PRD families that contain one or two proline-recognition pockets combined with additional regions conferring binding specificity. Whether this ‘‘intrinsic’’ specificity determined by the peptide and domain sequences is enough for the required in vivo selectivity is under discussion. An additional level of specificity seems to be conferred by contextual factors, such as the establishment of tertiary contacts with other elements in the full-length protein (see [21,23,27] for a review), multidomain binding [31], dimerization [53,54] or cellular co-localization of proteins [55,56]. The pseudo-symmetry of the PPII conformation allows SH3 ligands to bind in a forward (N-to-C terminal in class I ligands) or reverse (C-to-N terminal in class II ligands) orientation with respect to the binding site [57]. Even though peptide array screening experiments seem to indicate a general preference for class II ligands [50], the preferred orientation depends on the characteristics of the SH3 binding site, being frequently determined by (a) the presence of a basic amino acid, located two residues Nterminal from the uPpuP motif in class I ligands and two residues C-terminal from the motif in class II ligands, that interacts with an acidic residue at the RT loop [57,58] and (b) the conformation of a conserved Trp side chain at the b3 strand in the domain [59]. In fact, most SH3 ligands identified to date correspond to the consensus sequences fn-R 3p 2u 1P0x1f2P2-fc and fc-R5f4P3u2p1P0w 1-fn for class I and class II ligands [20,60], respectively. There are other domains, such as Abl–SH3, that interact with ligands containing the canonical uPpuP motif but lacking positive residues in their se-

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

quences, which can be considered a second specificity group. Also, most interestingly, several SH3 domains have been identified that recognize atypical sequences in their targets (see [21,23] for reviews), including discontinuous epitopes [61–63], slight variations of the canonical motifs (PxxxPR [64], RKxxYxxY, RxxPxxxP [65]), or even non-PPII peptides containing the RxxK motif [66,67]. These are clear examples of how sequence variability within the RT and n-Src loops, not only allows for a certain level of binding specificity within canonical ligands, but can also be exploited to evolve new binding specificities [17,21,23,26,51].

3. Complexity in polyproline recognition by SH3 domains During the last two decades a wealth of structural and functional information has been collected for the different PRD families, especially for SH3 and WW domains, according to their potential as targets for the development of novel therapeutic agents against a variety of diseases [22,68,69]. Many efforts have been directed towards the design and identification of high-affinity ligands for SH3 domains [35,70,71] using different strategies, including screening of libraries of synthetic compounds [28,72,73], phage display techniques [26,74], the use of protein scaffolds to stabilize the PPII conformation [75] and structure-based rational design strategies directed to optimize peptide sequences [76,77], introduce nonpeptidic elements in the peptide ligands [78–80] or design smallmolecule inhibitors [81,82]. In most cases the results have been modest, although the use of combinatorial libraries of peptoids, with synthetic substitutes of key proline residues [47,83,84] and the combinatorial modification of peptide scaffolds [72,78] have provided some high-affinity ligands, with dissociation constants in the nM range. Today, after more than two decades of study, and in spite of the wealth of structural and functional information available, the design of high affinity and specificity inhibitors or modulators of polyproline recognition by SH3 domains or other PRD families remains quite a challenging task. This is partly due, as mentioned before, to the delicate balance between promiscuity and specificity characteristic of these interactions and to the low binding affinity of most SH3 natural ligands, usually ranging between 1 and 200 lM. These difficulties have been exacerbated by the lack of a profound and detailed understanding of the forces driving polyproline recognition. Traditionally, most identification and optimization strategies have been directed towards obtaining improvements in the binding affinity of the ligands. This has, indeed, been the case in most of the above-mentioned works on SH3 domains. However, considering binding affinity, determined by the Gibbs energy of binding (DG), as the only discriminating parameter in screening procedures or rational design strategies constitutes a very limited perspective of the binding process since the Gibbs energy alone does not provide information about the nature and magnitude of the forces driving the association. This information is, nonetheless, encoded in the enthalpic and entropic contributions to the binding Gibbs energy (DG = DH TDS), which arise from different types of interactions and, thus, report on the relative weight of the different intermolecular forces in a particular binding event [85,86]. Considering that the same binding affinity can arise from many different combinations of enthalpic (DH) and entropic (DS) contributions, knowledge of the thermodynamic signature of an interaction (the proportion in which the binding enthalpy and entropy contribute to the Gibbs energy) provides a very valuable insight into the forces governing the binding process [87–89]. Moreover, it is now well established that incorporating thermodynamic considerations into screening and rational design procedures can contribute to select the best initial leads and devise the best optimization strategies for the development of the most appropri-

2621

ate ligands in terms of binding affinity, selectivity [90–96] or even resistance profiles [97,98]. In this sense, the case of SH3 domains has been paradigmatic of how thermodynamic information can contribute to attain a better understanding of the binding process. According to the high quality structural information available, the recognition or proline-rich sequences by SH3 domains has been described as a process based primarily on the stacking of hydrophobic surfaces in the protein and ligand, with very few polar interactions. In most complexes, these are limited to a couple of highly conserved hydrogen bonds established between the carbonyl oxygens of proline residues in the ligand and the side chains of two conserved Trp and Tyr residues located at the b3 strand and the 310 helix, respectively, and, in some cases, to the conserved salt bridge at the specificity pocket. According to this description, the recognition of proline-rich sequences by SH3 domains would be expected to present a thermodynamic signature dominated by the hydrophobic effect, with the main driving force being a favorable entropic contribution associated to the increment of the conformational degrees of freedom of the solvating water molecules that are released into the bulk solvent upon formation of the complex. Correspondingly, the binding enthalpy would be expected to be unfavorable (positive) or, in an optimal situation, slightly favorable [89,90]. Nonetheless, all thermodynamic studies of SH3 interactions collected during the last two decades revealed that the recognition of proline-rich ligands by these domains is invariably driven by markedly negative (favorable) enthalpic contributions opposed by unfavorable binding entropies [27,52,67,75,99–108]. This thermodynamic behavior cannot be rationalized exclusively in terms of direct interactions between hydrophobic surfaces and, thus, reveals an underlying complexity in the recognition of proline-rich ligands by SH3 domains, not evident from the interpretation of static three dimensional structures. Significant efforts have been devoted to deciphering this apparent discrepancy between structure and thermodynamics. These studies have revealed a highly complex scenario to which several factors of different nature contribute significantly. In this sense, several studies have shown that the redistribution of the conformational ensemble of both, peptide and protein, upon formation of the complex has a significant impact on the binding energetics. In this way, the calorimetric and structural-thermodynamic analysis of the binding of the Sem-5 SH3 domain to its Sos ligand revealed that the adoption of the PPII conformation by the ligand upon binding results in unfavorable entropic contributions associated to the reduction in the conformational degrees of freedom of the ligand and favorable enthalpic effects, estimated to be on the range of 10 kJ mol 1[99]. Also, the rigidification of the RT and n-Src loops upon ligand binding and the corresponding reorganization of the intra- and inter-molecular network of hydrogen bond interactions have been postulated to contribute in a similar manner [43,52,109]. Along these lines, the NMR characterization of ligand binding to the SH3 domain of the Src tyrosine kinase revealed that the intercalation of the uP moieties from the ligand in the hydrophobic pockets of the domains results in the disruption of the intramolecular hydrogen bond network. These alterations propagate throughout the domain resulting in changes in backbone dynamics that have been postulated to contribute significantly to the exothermic character of the interaction [102,110]. In addition to these conformational effects, the thermodynamic and structural analysis of several complexes implicating the SH3 domain from the Abl tyrosine kinase brought to light additional effects associated to the presence of interfacial water molecules occluded at the binding site that seem to play a very relevant role in proline-rich ligand recognition by SH3 domains, highlighting the need for a revision of the current binding paradigm for these systems.

2622

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

4. Interfacial water molecules in Abl–SH3 complexes The cellular form of the Abelson leukemia virus, c-Abl, is a tightly regulated tyrosine kinase implicated in the development of chronic myelogenous leukemia. C-Abl becomes constitutively activated upon mutation or deletion of the SH3 domains [111]. Moreover, it has been shown that displacement of the SH3 intramolecular interaction with the linker sequence connecting the SH2 and the kinase domains enhances the inhibitory effect of the anti-tumor drug Gleevec [112]. Consequently the Abl–SH3 domain has been targeted for the design of novel ligands of high affinity and specificity of potential therapeutic applications [76,77]. Starting from the 3BP1 natural ligand (RAPTMPPPLPP), characterized by a dissociation constant of 34 lM for Abl–SH3 [113], the group of Luis Serrano was able to derive new ligands with increased affinity and specificity for the Abl–SH3 domain by rational design strategies. By introducing new favorable interactions with the RT loop, specific to Abl–SH3, and increasing the tendency of the free peptide to adopt the PPII conformation they arrived to the p40 (APTYSPPPPP) and p41 (APSYSPPPPP) sequences characterized by 2 lM binding affinities and a high selectivity for Abl–SH3 against Fyn–SH3 [76]. The thermodynamic study of the binding energetics of the p41 ligand and a set of related peptides to the SH3 domain of Abl revealed these interactions as a particularly striking example of the inconsistency existing between the canonical description of SH3 binding and its thermodynamic signature. In fact, the interactions of p41 related peptides with Abl–SH3 are characterized by extremely negative binding enthalpies ( 92 kJ mol 1 for p41) [100], which are notably bigger than the values reported for other SH3 complexes, typically ranging between 20 and 50 kJ mol 1. This is particularly striking considering the fact that the p41 peptide lacks polar or ionizable groups, present in most SH3 ligands in these thermodynamic studies. The large and negative binding enthalpy is partially compensated by highly unfavorable entropic contributions, resulting in a binding affinity within the range reported for other SH3 interactions. This thermodynamic signature cannot be easily reconciled with the canonical binding mode, especially considering the highly apolar character of the Abl–SH3/p41 binding interface (2.8 non-polar/polar ratio for the accessible surface area). In order to understand the molecular basis of the increment in binding affinity in the p41 optimization process and validate the design rationale, a detailed thermodynamic study was carried out with a set of p41 related peptides differing in unique and specific positions in the PPII region [100]. Some of these mutations implicated solvent exposed residues in the complex [114] not involved in any interactions with the domain, such as Ser5 and Pro8 in the p41 sequence (see Fig. 2). According to the canonical binding mode, the replacement of aliphatic residues by proline at these positions should result in more favorable entropic contributions due to the restrictions imposed on the conformational ensemble of the free ligand, favoring the adoption of a PPII conformation in the free state. No significant enthalpic effects are to be expected. This is what has been found in similar situations for other SH3 complexes [13]. Surprisingly, the analysis of the calorimetric results revealed a very different scenario, unveiling remarkable and surprising effects associated to these substitutions. Indeed, as expected, the introduction of proline residues at these positions led to an increase in the binding affinity. Nonetheless, as illustrated in Fig. 2, this increment was not associated to more favorable entropic contributions, but to a considerable increment (up to 15 kJ mol 1) in the negative binding enthalpy, particularly striking for residues not directly implicated in any interactions within the complex. Moreover, these effects were also found to be context-dependent, revealing some level of cooperativity between these two solvent exposed positions.

Indeed, the thermodynamic behavior observed for the Abl–SH3 complexes is a clear indication of the complexity of these interactions to which additional factors, other than the direct interaction between the ligand and protein surfaces, must be contributing. With respect to the conformational effects described above, binding studies of a p41-containing miniprotein designed to stabilize the PPII conformation indicate that, in this particular case, the conformational reorganization of the ligand does not seem to contribute significantly to magnitude of the different thermodynamic parameters [75]. Interestingly, as illustrated in Fig. 3, inspection of the crystal structure of the Abl–SH3/p41 complex [114] revealed the presence of a set of fully buried interfacial water molecules distributed in two distinct hydration regions defined by different water-coordinating amino acids in the SH3 domains: a) waters 1–3 in Fig. 3 that mediate the interactions between residues in the n-Src loop and the specificity region of the ligand and b) waters 4 and 5 in Fig. 3 that bridge the interactions between residues Asn114 and Ser113 in the 310 helix region of the domain and the PPII region of p41. These residues, which are not in direct contact with the ligand, define an extended interaction surface characterized by a higher polar character than the canonical binding site, so that the ratio of apolar to polar surface area drops from 2.8 to 1.7. Interfacial water molecules will serve as adapters that fill empty spaces and optimize van der Waals interactions, satisfy the hydrogen bonding potential of the ligand and the binding site and assist in the dissipation of charges. All these terms might be expected to contribute favorably to the binding enthalpy [115,116]. Conversely, fixing a water molecule at the binding interface entails a considerable entropic penalty that counterbalances these favorable enthalpic effects, so that the impact of buried water molecules on the Gibbs energy of binding is, often, no more than modest. These effects are in agreement with the observed thermodynamic behavior and suggest that interfacial water molecules might be playing a very relevant role in the interaction, contributing significantly to the extremely negative binding enthalpy of the Abl–SH3/p41 complex. This hypothesis was confirmed by a structural, thermodynamic and molecular dynamics study of a set of conservative mutants of the Abl SH3 domain designed to perturb the water-mediated hydrogen bond network with a minimum impact on the structural and conformational properties of the domain. This analysis revealed significant enthalpic effects associated to the mutation of residues at position 114, at the base of the 310 helix, that coordinate water molecules at hydration sites 4 and 5 without being implicated in any direct contacts with the peptide ligand. A very good correlation was found between these enthalpic effects and the changes induced in the dynamic properties of water molecules at site 5. From these results, it was possible to tentatively estimate the contribution to the binding enthalpy of water molecules at this hydration site to be around 20 kJ mol 1, confirming a very substantial contribution of interfacial water molecules to the strongly exothermic character of Abl–SH3 interactions [117] and, probably, to the unexpected cooperative effects associated to the mutation of solvent exposed residues in the ligand. In summary, these results demonstrate that the currently accepted binding paradigm is incomplete and does not provide an adequate description of the binding of Abl–SH3 to the family of p41 peptides, which interact via two different mechanisms: the stacking of proline side-chains with the hydrophobic grooves in the binding site and the establishment of a robust and plastic network of water-mediated hydrogen bonds implicating residues in the periphery of the canonical binding site. This double binding mechanism is in better agreement with the thermodynamic behavior of the system and has important implications for molecular modeling and ligand design.

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

2623

Fig. 2. Thermodynamic effects of mutations at solvent exposed positions in p41-related peptides. Thermodynamic cycle summarizing the effects on the thermodynamic parameters of binding to the Abl–SH3 domain associated to substitutions at positions 5 (blue) and 8 (red) in the ligands. The values associated with the vertical arrows correspond to the substitutions at position 5 and those associated with the horizontal arrows to the substitutions at position 8. The inset corresponds to the cartoon and surface representation of the Abl–SH3/p41 complex. P41 ligand is shown as sticks. Mutations sites 5 and 8 are highlighted by blue and red cycles respectively. (Reproduced with permission from [100]).

Fig. 3. Water-mediated interactions at the Abl–SH3/p41 binding interface. The structure of the Abl–SH3 domain is shown in a gray cartoon representation. Residues defining the canonical binding site for polyproline recognition are shown as gray sticks. The structure of the p41 peptide is shown as cyan sticks. Fully buried water molecules at the binding interface are shown as green spheres. Peripheral water-coordinating residues in the 310 and n-Src regions are shown as purple and dark pink sticks respectively. Water-mediated hydrogen bonds are depicted as green lines. (Reproduced with permission from [86].)

5. Revision of the general binding paradigm for SH3 domains Taking into account that all SH3 complexes studied to date share the same thermodynamic profile, the question arises about the relevance of the Abl–SH3 dual binding mechanism for other SH3 domains. In Abl–SH3, the presence of most water molecules at the binding site is contingent upon ligand binding, being absent in the structures of free domains, with the exception of water molecules at hydration site 4, at the base of the 310 helix, which is occupied in all Abl–SH3 structures independently of their ligation state. Also, the properties of these water molecules, fully buried and characterized by very high residence times and low B factors,

are not affected by the binding of the ligand. This indicates that water molecules at this site are structural waters that should be considered as an integral part of the Abl–SH3 domain. Interestingly, these water molecules are coordinated by the backbone atoms of highly stable and well conserved residues (see Fig. 3), suggesting that equivalent water molecules would probably be found in many other SH3 domains. Analysis of the SH3 structural database, including close to 100 high resolution crystal structures of SH3 domains from different proteins including free domains, inter- and intra-molecular complexes with natural ligands, rationally designed and phage-display peptides as well as full-length target proteins, reveals that this is indeed the case [118]. As expected from the sequence conservation of the 310 helix region, water molecules at positions equivalent to hydration site 4 in Abl–SH3 were found in 77% of all SH3 structures, including complexes and free domains, with properties (solvent accessibility, B factors, and coordination patterns) very similar, in all cases, to that described for Abl–SH3 (see Fig. 4A, B). From these observations, it is clear that the water molecule at the base of the 310 helix is an invariant element in most SH3 domains and seems to be an integral part of the domain structure, independently of its ligation state. In fact, in the analyzed database, the absence of this hydration site was generally associated to anomalous sequences and conformations of SH3 domains and ligands. Consequently, this structural water molecule should be included in binding site descriptions for docking or rational design. A second structural water molecule, located at the n-Src loop, was found in about 60% (59% in free domains and 55% in complexes) of all SH3 structures (see Fig. 4C, D). These waters molecules, which are also fully buried and tightly bound, are implicated in the establishment of an average of 4 hydrogen bonds with residues within the loop, stabilizing its conformation. In fact, the presence of these waters has been found to be highly dependent on the length of the loop, being observed in 85% of the structures of n-Src loops between 5 and 8 residues and absent in structures with shorter or longer loops [118]. Even though, in most cases, these water molecules are not directly implicated in ligand recognition, they may play an important indirect role by regulating the dynamic and conformational properties of the n-Src loop, of relevance for binding specificity [25,27,52]. Globally, at least one structural water molecule was found in 93% of all free SH3 structures studied. In addition to these

2624

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

Fig. 4. Structural water molecules in SH3 domain structures. Cartoon representation of the SH3 domain structures colored in a rainbow scheme. Side chains of residues constituting the 310 helix hydration site are represented as sticks. Water molecules are depicted as purple non-bonded spheres (A) 310 helix structural waters in free SH3 domains, (B) 310 helix structural waters in SH3 complexes, (C) n-Src loop structural waters in free SH3 domains and (D) n-Src loop structural waters in SH3 complexes. (Reproduced with permission from [118].)

structural waters, other interfacial water molecules were consistently found in most SH3 complex structures, originating highly variable hydration patterns. As illustrated in Fig. 5, interfacial water molecules were found in all analyzed complex structures distributed throughout the binding interface, mediating the interactions between the ligand and the SH3 domain. As described for

Abl–SH3, water-mediated interactions frequently implicate residues in the periphery of the canonical binding site, different from those defining the hydrophobic pockets for proline recognition. This observation underlines the importance of considering interfacial water molecules for a correct definition of the complete SH3 binding interface.

Fig. 5. Cartoon representation of the superposed structures of the SH3 complexes. SH3 domains are colored in a rainbow scheme (blue: N-terminus; red: C-terminus) while ligands are represented in light pink. Buried interfacial water molecules are shown as purple non-bonded spheres. (Reproduced with permission from [118].)

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

A third conserved hydration site, equivalent to site 5 in Abl– SH3, was occupied in approximately one third of the complex structures. This water molecule, which in the Abl–SH3/p41 complex was found to interact mostly with the ligand through interactions established with the carbonyl oxygens of P6 and p7 in the peptide, is implicated in an average of 3 hydrogen bonds with atoms in the protein and ligand and with the structural water molecules at the base of the 310 helix. In this case, a higher variability exists in the location and interaction pattern of this water molecule that seems to be dependent mostly on the length of the side chains at position 114 in the domain and the nature of the central amino acids in the uPpuP core motif of the ligand. These factors determine whether the hydrogen bonding potential of carboxyl atoms in the peptide is satisfied by the establishment of direct hydrogen bonds with the side chains at position 114 in the domain or through water-mediated interactions [118]. These patterns provide a rationalization for the puzzling thermodynamic results associated to mutations at solvent-exposed positions in the p41 ligand [100]. In addition to the conserved hydration sites at the polyprolinerecognition region, interfacial water molecules were very frequently observed mediating the interactions between the RT and n-Src loops and the specificity region of the ligand, giving rise to highly variable hydration patterns, very dependent on the sequence of ligands and loops. In general, waters at the RT loop region were frequently found mediating the interactions between charged residues in the peptide and the protein. This explains why no water molecules in the vicinity of the RT loop were observed in the Abl–SH3 complexes. In spite of the high variability in this region, no significant differences in hydration patterns were found associated to ligand orientation. Conserved water molecules in the polyproline recognition region were equally found for class I and class II ligands and a similar variability in water configurations was observed for both classes at the specificity region. In summary, the analysis of the SH3 structural database clearly shows that the presence of interfacial water molecules is a universal factor in SH3 interactions and, consequently, the dual binding mechanism described for Abl–SH3, far from being an exception, is a general feature of these domains. This new description of SH3 binding is in better agreement with the thermodynamic properties of these systems. In this context, it becomes apparent that ignoring the role of interfacial water molecules in polyproline recognition by SH3 domains would preclude a full and deep understanding of the molecular basis of binding affinity and specificity. Consequently, the current description of SH3 complexes needs to be replaced by a new binding paradigm that incorporates interfacial waters as relevant elements in the mechanism of polyproline recognition by these domains.

6. Implications for target identification and rational design This dual binding paradigm opens new perspectives for rational design and target identification for SH3 domains, since, as it has been demonstrated in numerous studies, the efficiency of docking [119–126] or rational design [127–130] procedures can be significantly enhanced if detailed information about of the role and weight of each water molecule at the binding interface is incorporated into binding site descriptions (see [130,131] for detailed reviews). In this context, important questions arise that would need to be answered in order to take full advantage of the second level of interactions provided by the water-mediated hydrogen bonds in SH3 domains. Is the impact of all interfacial water molecules on the thermodynamic signature of the binding process equally relevant for all SH3 complexes? What is the actual weight of these interfacial waters on binding affinity? Do they have a real

2625

impact on SH3 binding specificity? Could new rational design strategies be devised to improve binding affinity or specificity by engineering new interactions with interfacial waters of by their displacement from the binding interface? Should the hydration profile of a particular domain be considered for docking and ligand identification protocols? To answer these questions a profound understanding of the properties and role of each specific water molecule would be required. With respect to the binding energetics, it is well established that the presence of interfacial water molecules will alter the thermodynamic signature of the binding process, resulting in entropic penalties arising from the reduction in the configurational degrees of freedom of the occluded water molecules, enthalpic benefits associated to the establishment of water-mediated hydrogen bonds and more negative binding heat capacities [132–136]. Nonetheless, to date, estimating the actual contribution of a particular water molecule to the different thermodynamic parameters remains a considerable challenge. In general, the contribution of each water molecule would be expected to vary according to its degree of immobilization and the number and quality of its interactions. In this sense, theoretical studies applying the theory of inhomogenous fluids predict the entropic cost for water immobilization to range between 0 and 41 kJ mol 1[137,138]. Also, according to these studies, the presence of interfacial water molecules always results in an enthalpic benefit that, in some cases is modest ( 8 kJ mol 1 for a water molecule in the cyclophilin-cyclosporin A complex [139]), while in others can be very relevant ( 63 kJ mol 1 for a conserved interfacial water in HIV-1 protease complexes [140]). These theoretical predictions are in good agreement with experimental observations that show a wide spectrum of entropic and enthalpic effects depending on the specific characteristics of the system. Experimentally, it is observed that the displacement of water molecules from the binding interface, either by modification of the ligand or mutation of the protein, is associated to variable entropic benefits combined with enthalpic penalties that range between 9.5 kJ mol 1 [141] and 30 kJ mol 1 [142]. It is interesting to point out that the 20 kJ mol 1 estimated to be contributed by the waters at site 5 to the binding enthalpy of p41/Abl– SH3 [117] falls well within this range. In spite of the considerable magnitude of the entropic and enthalpic effects, the net impact of interfacial water molecules on the binding affinity is not as significant and can be, also, very variable. In this sense, studies carried out with different systems (i.e. HIV-1 protease [143], scytalone dehydratase [144], OppA peptide binding protein [145], FK506 binding protein [146], ConA [147], major urinary protein MUP-1 [142]) reveal that water displacement is sometimes energetically favorable, resulting in improved binding affinities, and sometimes unfavorable. Moreover, in some cases, strong enthalpy/entropy compensation dampens all effects, resulting in very modest or even negligible changes on the binding affinity. This variability indicates that, in order to efficiently incorporate water-mediated interactions rational design, a good understanding of the properties of the different interfacial waters and the quality of the interactions they mediate is required. This would allow us to adequately select which water molecules should be targeted for displacement and which could be considered as good anchor sites for the engineering of new, energetically favorable, interactions. Within the last years, several methods have been developed to easily classify interfacial waters, such as Consolv [148], Waterscore [149], HINT/RANK [150,151] or GRID [152], which are generally based on geometrical studies of molecular surfaces, number of hydrogen bonds, crystallographic B factors and proximity to polar residues. These type of analysis indicate that the contribution to the binding Gibbs energy of each water molecule would strongly depend on its environment, so that tightly bound water molecules

2626

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

are typically found in polar cavities, fully buried and establishing at least three hydrogen bonds with the protein and the ligand [153]. Nonetheless, the predictive capacity of these methodologies is still quite limited. Alternative, more exhaustive and time-consuming approaches, including free energy calculations such as the double decoupling method [154,155], or molecular dynamics simulations can provide a more reliable description of the role of interfacial waters in a particular system. In this sense, the detailed structural, thermodynamic and molecular dynamic analysis carried out with Abl–SH3 complexes provided a very detailed description of the different hydration sites, showing that not all sites at the SH3 interface are equivalent. This is further sustained by the analysis of the SH3 structural database that confirms the existence of three types of water molecules in these systems with distinct properties (accessibility, B factors and number of interactions): (a) tightly bound structural water molecules, conserved in most SH3 structures and present in free domains as integral part of their structure. These water molecules do not interact significantly with the ligand; (b) conserved water molecules in the polyproline region mediating the interactions with the ligand and characterized by high residence times, such as waters at site 5 in Abl–SH3, which were shown to interact mostly with the peptide ligands [117] and (c) waters at the specificity region, more mobile and exposed. Which of these water molecules should be targeted for rational design? In general, two different scenarios can be considered to include interfacial waters in rational design (a) modification of the ligand to optimize its interactions with interfacial water molecules, aimed at increasing binding affinity by improving the binding enthalpy or (b) introduction of new moieties in the ligand that displace binding site waters present in the free protein or in related complexes, taking advantage of the entropic benefit of releasing a trapped water molecule. In any case, in order to overcome enthalpy/entropy compensation effects and obtain a net improvement on the binding affinity, it is important to select carefully the targeted waters. In the first case, a significant improvement on binding affinity would only be achieved if the newly engineered interactions can overcome any entropic penalty associated to the additional restrictions on the configurational degrees of freedom imposed on the water molecule as well as in the ligand and the protein. These enthalpy/entropy compensation effects can be minimized if the new interactions are directed against highly structured and stable elements in the binding site [88]. In this sense, the structural water molecules at the base of the 310 helix, tightly coordinated by backbone atoms of the most structurally stable residues in the domain [42,45] constitute a good anchor site against which to design new interactions, i.e. hydrogen bonds, with reduced associated enthalpy–entropy compensation effects, and, thus, with good chances of leading to significant improvements in binding affinity. On the contrary, in the second scenario, it is important to select which water molecules can be easily displaced from the binding interface. In this sense, water molecules at the polyproline regions (waters at sites equivalent to position 5 in Abl–SH3) seem to be interacting predominantly with the ligand with higher conformational flexibility. In this sense, the introduction of new moieties in the ligand to displace these waters from the binding site and establish interactions with the structural water at the 310 helix could be energetically favorable and lead to improvements on binding affinity. In any case, the new moieties in the ligand should be carefully design to establish highly optimized hydrogen bonds in order to compensate for the loss of water-mediated interactions. In addition to modulate binding affinity, water molecules in biomolecular interfaces can also play a very relevant role in binding specificity. Because of their small size and their ability to establish multiple hydrogen bonds as donors or acceptors, interfacial waters are very versatile elements in protein recognition. In some systems they contribute to narrow binding specificity, as happens

with the recognition of phosphotyrosine peptides by SH2 domains, where water-mediated hydrogen bonds are essential to maintain the specificity for a glutamic side chain two positions C-terminal from the phosphotyrosine residue in the ligands [156]. Nonetheless, in other systems, interfacial waters confer great adaptability to molecular surfaces allowing a considerable degree of promiscuity in the interactions. The OppA peptide binding protein is a paradigmatic example of this situation. This protein can recognize oligopeptidic chains between 2 and 5 amino acids independently of their sequence by combining two levels of interactions: direct, conserved, interactions between the protein and the backbone of the peptide ligand and the insertion of the peptide side chains into hydrophobic pockets that modify their water content depending on the specific sequence of the ligand. In this way, water molecules modulate protein binding specificity by occupying in each case only a few positions from a set of well-defined, highly conserved and favorable hydration sites at the surface of the protein [135,157]. This situation is very similar to that observed for SH3 domains. The high variability of hydration patterns found in the specificity region of SH3 complexes appears to indicate that interfacial water molecules might be acting as adaptors to facilitate the recognition of diverse sequences, contributing to the versatility and promiscuity of these domains. Consequently, interfacial waters seem to be also of great relevance to fully understand binding specificity and optimize target identification in SH3 domains.

7. Polyproline recognition by other domain families Even though the structural information available is much more limited for the other families of polyproline recognition domains, water mediated interactions were also systematically found in all high resolution crystal structures available for these systems (13 structures of WW domains from different specificity classes, 6 structures of UEV domains and 10 structures of EVH-1 domains, including free domains and complexes) [118]. As illustrated in Fig. 6 for WW and UEV domains, the situation is very similar to that described for SH3 domains, with some hydration sites corresponding to structural water molecules present and conserved in free domains and complexes and others being only occupied upon binding of the ligand. In all cases, as in SH3 domains, these watermediated interactions result in extended binding interfaces with higher polar character. This is in agreement with the fact that, to our knowledge, all thermodynamic studies reported for polyproline recognition by PRD reveal a common thermodynamic signature, dominated by favorable enthalpic interactions. It is interesting to point out that this exothermic character is inherent of proline-rich motifs, as illustrated by the fact that the binding of short peptides corresponding to the PPPY core motif for type 1 WW domains is characterized by a binding enthalpy similar to that found for longer peptides containing charged or polar residues in their sequences [158]. In any case, to this moment, the structural information for these systems is insufficient to derive any specific patterns. Obviously, the actual relevance of this mechanism for all PRD families will need to be confirmed as the structural and thermodynamic databases increase. In conclusion, the current evidence indicate that the dual binding mechanism, combining hydrophobic interactions with the establishment of strong networks of water-mediated hydrogen bonds, is probably general, not only to SH3 domains, but also to most families of polyproline recognition modules. This seems to indicate that the tendency of proline-rich sequences to be highly hydrated, illustrated by the solid and extensive hydration structure of the collage triple helices [159,160], has been exploited by nature to favor the adaptability and plasticity of the different families of protein modules for the recognition of proline-rich targets. In this

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

2627

Fig. 6. Water-mediated interaction patterns in WW and UEV domains. WW and UEV domains are shown in white, ligands are depicted as marine blue sticks and relevant protein residues are represented as cyan sticks. Interfacial waters are shown as non-bonded spheres (waters in complex structures are colored in blue tones, while waters in free structures are colored in red, orange, yellow and purple). Hydrogen bonds are depicted as discontinuous lines following the same color scheme. (a) Class IV Pin1–WW domain, free (1PIN, 2Q5A, 2F21, 2ITK, 1ZCN) and in complex with a phosphoserine peptide. (b) Class I dystrophin–WW domain free (1EG3) and in complex with a bdystroglican peptide (1EG4). (c) Class II FE65–WW domain free (2IDH) and in complex with proline-rich peptide from Mena (2OEI). (d) Tsg101–UEV domain free (3OBS, 2FOR) and in complex with proline-rich viral late domain sequences (3OBQ, 3OBU, 3OBX). Hydrogen bond patterns common to most structures (including the free domains) are shown as discontinuous marine blue lines while interactions specific to particular structures are shown in their respective colors. (Reproduced with permission from [118].)

context, understanding the role of interfacial water molecules may provide important clues for deciphering the specificity versus promiscuity paradox in polyproline recognition. Acknowledgements This work was supported by Grants BIO2009-13261-CO2 from the Spanish Ministry of Science and Innovation, FEDER Funds and Grant CVI-5915 from the Andalusian Government. Ana Zafra-Ruano is supported by a FPI predoctoral research contract from the Spanish Ministry of Science and Technology. References [1] Pawson, T., Raina, M. and Nash, P. (2002) Interaction domains: from simple binding events to complex cellular behavior. FEBS Lett. 513, 2–10. [2] Pawson, T. and Scott, J.D. (1997) Signaling through scaffold, anchoring, and adaptor proteins. Science 278, 2075–2080. [3] Weatheritt, R., Luck, K., Petsalaki, E., Davey, N. and Gibson, T. (2012) The identification of short linear motif-mediated interfaces within the human interactome. Bioinformatics 28, 976–982. [4] Gould, C.M. et al. (2010) ELM: the status of the 2010 eukaryotic linear motif resource. Nucleic Acids Res. 38, D167–D180. [5] Xu, Z., Hou, T., Li, N., Xu, Y. and Wang, W. (2012) Proteome-wide detection of Abl1 SH3-binding peptides by integrating computational prediction and peptide microarray. Mol Cell Proteomics 11, O111 010389.

[6] Obenauer, J.C. and Yaffe, M.B. (2004) Computational prediction of protein– protein interactions. Methods Mol. Biol. 261, 445–468. [7] Lam, H.Y., Kim, P.M., Mok, J., Tonikian, R., Sidhu, S.S., Turk, B.E., Snyder, M. and Gerstein, M.B. (2010) MOTIPS: automated motif analysis for predicting targets of modular protein domains. BMC Bioinform. 11, 243. [8] Rubin, G.M. et al. (2000) Comparative genomics of the eukaryotes. Science 287, 2204–2215. [9] Chandra, B.R., Gowthaman, R., Akhouri, R.R., Gupta, D. and Sharma, A. (2004) Distribution of proline-rich (PxxP) motifs in distinct proteomes: functional and therapeutic implications for malaria and tuberculosis. Protein Eng. Des. Sel. 17, 175–182. [10] Zarrinpar, A. and Lim, W.A. (2000) Converging on proline: the mechanism of WW domain peptide recognition. Nat. Struct. Biol. 7, 611–613. [11] Creamer, T.P. (1998) Left-handed polyproline II helix formation is (very) locally driven. Proteins 33, 218–226. [12] Creamer, T.P. (2000) Side-chain conformational entropy in protein unfolded states. Proteins 40, 443–450. [13] Ferreon, J.C. and Hilser, V.J. (2003) The effect of the polyproline II (PPII) conformation on the denatured state entropy. Protein Sci. 12, 447–457. [14] Kay, B.K., Williamson, M.P. and Sudol, M. (2000) The importance of being proline: the interaction of proline-rich motifs in signaling proteins with their cognate domains. FASEB J. 14, 231–241. [15] Stapley, B.J. and Creamer, T.P. (1999) A survey of left-handed polyproline II helices. Protein Sci. 8, 587–595. [16] Vijayakumar, M., Qian, H. and Zhou, H.X. (1999) Hydrogen bonds between short polar side chains and peptide backbone: prevalence in proteins and effects on helix-forming propensities. Proteins 34, 497–507. [17] Zarrinpar, A., Bhattacharyya, R.P. and Lim, W.A. (2003) The structure and function of proline recognition domains. Sci. STKE 2003, RE8. [18] Castagnoli, L. et al. (2004) Selectivity and promiscuity in the interaction network mediated by protein recognition modules. FEBS Lett. 567, 74–79.

2628

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

[19] Mayer, B.J. (2001) SH3 domains: complexity in moderation. J. Cell Sci. 114, 1253–1263. [20] Ball, L.J., Kuhne, R., Schneider-Mergener, J. and Oschkinat, H. (2005) Recognition of proline-rich motifs by protein–protein-interaction domains. Angew. Chem., Int. Ed. Engl. 44, 2852–2869. [21] Li, S.S. (2005) Specificity and versatility of SH3 and other proline-recognition domains: structural basis and implications for cellular signal transduction. Biochem. J. 390, 641–653. [22] Freund, C., Schmalz, H.G., Sticht, J. and Kuhne, R. (2008) Proline-rich sequence recognition domains (PRD): ligands, function and inhibition. Handb. Exp. Pharmacol., 407–429. [23] Kaneko, T., Li, L. and Li, S.S. (2008) The SH3 domain–a family of versatile peptide- and protein-recognition module. Front. Biosci. 13, 4938–4952. [24] Ladbury, J.E. and Arold, S. (2000) Searching for specificity in SH domains. Chem. Biol. 7, R3–R8. [25] Kaneko, T., Sidhu, S.S. and Li, S.S. (2011) Evolving specificity from variability for protein interaction domains. Trends Biochem. Sci. 36, 183–190. [26] Panni, S., Dente, L. and Cesareni, G. (2002) In vitro evolution of recognition specificity mediated by SH3 domains reveals target recognition rules. J. Biol. Chem. 277, 21666–21674. [27] Ladbury, J.E. and Arold, S.T. (2011) Energetics of Src homology domain interactions in receptor tyrosine kinase-mediated signaling. Methods Enzymol. 488, 147–183. [28] Oneyama, C., Nakano, H. and Sharma, S.V. (2002) UCS15A, a novel small molecule, SH3 domain-mediated protein–protein interaction blocking drug. Oncogene 21, 2037–2050. [29] Karkkainen, S. et al. (2006) Identification of preferred protein interactions by phage-display of the human Src homology-3 proteome. EMBO Rep. 7, 186– 191. [30] Macias, M.J., Wiesner, S. and Sudol, M. (2002) WW and SH3 domains, two different scaffolds to recognize proline-rich ligands. FEBS Lett. 513, 30–37. [31] Arold, S.T., Ulmer, T.S., Mulhern, T.D., Werner, J.M., Ladbury, J.E., Campbell, I.D. and Noble, M.E. (2001) The role of the Src homology 3-Src homology 2 interface in the regulation of Src kinases. J. Biol. Chem. 276, 17199–17205. [32] Barila, D. and Superti-Furga, G. (1998) An intramolecular SH3-domain interaction regulates c-Abl activity. Nat. Genet. 18, 280–282. [33] Miyazaki, K., Senga, T., Matsuda, S., Tanaka, M., Machida, K., Takenouchi, Y., Nimura, Y. and Hamaguchi, M. (1999) Critical amino acid substitutions in the Src SH3 domain that convert c-Src to be oncogenic. Biochem. Biophys. Res. Commun. 263, 759–764. [34] Chin, J. et al. (2005) Fyn kinase induces synaptic and cognitive impairments in a transgenic mouse model of Alzheimer’s disease. J. Neurosci. 25, 9694– 9703. [35] Dalgarno, D.C., Botfield, M.C. and Rickles, R.J. (1997) SH3 domains and drug design: ligands, structure, and biological function. Biopolymers 43, 383–400. [36] Viguera, A.R., Martinez, J.C., Filimonov, V.V., Mateo, P.L. and Serrano, L. (1994) Thermodynamic and kinetic analysis of the SH3 domain of spectrin shows a two-state folding transition. Biochemistry 33, 2142–2150. [37] Cobos, E.S., Filimonov, V.V., Vega, M.C., Mateo, P.L., Serrano, L. and Martinez, J.C. (2003) A thermodynamic and kinetic analysis of the folding pathway of an SH3 domain entropically stabilised by a redesigned hydrophobic core. J. Mol. Biol. 328, 221–233. [38] Martinez, J.C. and Serrano, L. (1999) The folding transition state between SH3 domains is conformationally restricted and evolutionarily conserved. Nat. Struct. Biol. 6, 1010–1016. [39] Martinez, J.C., Pisabarro, M.T. and Serrano, L. (1998) Obligatory steps in protein folding and the conformational diversity of the transition state. Nat. Struct. Biol. 5, 721–729. [40] Riddle, D.S., Grantcharova, V.P., Santiago, J.V., Alm, E., Ruczinski, I. and Baker, D. (1999) Experiment and theory highlight role of native state topology in SH3 folding. Nat. Struct. Biol. 6, 1016–1024. [41] Viguera, A.R., Jimenez, M.A., Rico, M. and Serrano, L. (1996) Conformational analysis of peptides corresponding to beta-hairpins and a beta-sheet that represent the entire sequence of the alpha-spectrin SH3 domain. J. Mol. Biol. 255, 507–521. [42] Sadqi, M., Casares, S., Abril, M.A., Lopez-Mayorga, O., Conejero-Lara, F. and Freire, E. (1999) The native state conformational ensemble of the SH3 domain from alpha-spectrin. Biochemistry 38, 8899–8906. [43] Ferreon, J.C., Volk, D.E., Luxon, B.A., Gorenstein, D.G. and Hilser, V.J. (2003) Solution structure, dynamics, and thermodynamics of the native state ensemble of the Sem-5 C-terminal SH3 domain. Biochemistry 42, 5582–5591. [44] Luque, I. and Freire, E. (2000) Structural stability of binding sites: consequences for binding affinity and allosteric effects. Proteins Suppl. 4, 63–71. [45] Casares, S., Lopez-Mayorga, O., Vega, M.C., Camara-Artigas, A. and ConejeroLara, F. (2007) Cooperative propagation of local stability changes from lowstability and high-stability regions in a SH3 domain. Proteins 67, 531–547. [46] Manson, A., Whitten, S.T., Ferreon, J.C., Fox, R.O. and Hilser, V.J. (2009) Characterizing the role of ensemble modulation in mutation-induced changes in binding affinity. J. Am. Chem. Soc. 131, 6785–6793. [47] Nguyen, J.T., Turck, C.W., Cohen, F.E., Zuckermann, R.N. and Lim, W.A. (1998) Exploiting the basis of proline recognition by SH3 and WW domains: design of N-substituted inhibitors. Science 282, 2088–2092. [48] Feng, S., Kasahara, C., Rickles, R.J. and Schreiber, S.L. (1995) Specific interactions outside the proline-rich core of two classes of Src homology 3 ligands. Proc. Natl. Acad. Sci. U S A 92, 12408–12415.

[49] Mayer, B.J. and Saksela, K. (2005) SH3 Domains in: Modular Protein Domains (Cesareni, G., Sudol, M. and Pawson, T., Eds.), pp. 37–58, Wiley-VCH. [50] Wu, C. et al. (2007) Systematic identification of SH3 domain-mediated human protein–protein interactions by peptide array target screening. Proteomics 7, 1775–1785. [51] Hiipakka, M. and Saksela, K. (2007) Versatile retargeting of SH3 domain binding by modification of non-conserved loop residues. FEBS Lett. 581, 1735–1741. [52] Arold, S., O’Brien, R., Franken, P., Strub, M.P., Hoh, F., Dumas, C. and Ladbury, J.E. (1998) RT loop flexibility enhances the specificity of Src family SH3 domains for HIV-1 Nef. Biochemistry 37, 14683–14691. [53] Kishan, K.V., Scita, G., Wong, W.T., Di Fiore, P.P. and Newcomer, M.E. (1997) The SH3 domain of Eps8 exists as a novel intertwined dimer. Nat. Struct. Biol. 4, 739–743. [54] Mongiovi, A.M., Romano, P.R., Panni, S., Mendoza, M., Wong, W.T., Musacchio, A., Cesareni, G. and Di Fiore, P.P. (1999) A novel peptide–SH3 interaction. EMBO J. 18, 5300–5309. [55] Sancier, F. et al. (2011) Specific oncogenic activity of the Src-family tyrosine kinase c-Yes in colon carcinoma cells. PLoS ONE 6, e17237. [56] Summy, J.M., Sudol, M., Eck, M.J., Monteiro, A.N., Gatesman, A. and Flynn, D.C. (2003) Specificity in signaling by c-Yes. Front. Biosci. 8, s185–s205. [57] Feng, S., Chen, J.K., Yu, H., Simon, J.A. and Schreiber, S.L. (1994) Two binding orientations for peptides to the Src SH3 domain: development of a general model for SH3–ligand interactions. Science 266, 1241–1247. [58] Lim, W.A., Richards, F.M. and Fox, R.O. (1994) Structural determinants of peptide-binding orientation and of sequence specificity in SH3 domains. Nature 372, 375–379. [59] Fernandez-Ballester, G., Blanes-Mira, C. and Serrano, L. (2004) The tryptophan switch: changing ligand-binding specificity from type I to type II in SH3 domains. J. Mol. Biol. 335, 619–629. [60] Aasland, R. et al. (2002) Normalization of nomenclature for peptide motifs as ligands of modular protein domains. FEBS Lett. 513, 141–144. [61] Ghose, R., Shekhtman, A., Goger, M.J., Ji, H. and Cowburn, D. (2001) A novel, specific interaction involving the Csk SH3 domain and its natural ligand. Nat. Struct. Biol. 8, 998–1004. [62] Kami, K., Takeya, R., Sumimoto, H. and Kohda, D. (2002) Diverse recognition of non-PxxP peptide ligands by the SH3 domains from p67(phox), Grb2 and Pex13p. EMBO J. 21, 4268–4276. [63] Lewitzky, M., Harkiolaki, M., Domart, M.C., Jones, E.Y. and Feller, S.M. (2004) Mona/Gads SH3C binding to hematopoietic progenitor kinase 1 (HPK1) combines an atypical SH3 binding motif, R/KXXK, with a classical PXXP motif embedded in a polyproline type II (PPII) helix. J. Biol. Chem. 279, 28724– 28732. [64] Moncalian, G., Cardenes, N., Deribe, Y.L., Spinola-Amilibia, M., Dikic, I. and Bravo, J. (2006) Atypical polyproline recognition by the CMS N-terminal Src homology 3 domain. J. Biol. Chem. 281, 38845–38853. [65] Tian, L., Chen, L., McClafferty, H., Sailer, C.A., Ruth, P., Knaus, H.G. and Shipston, M.J. (2006) A noncanonical SH3 domain binding motif links BK channels to the actin cytoskeleton via the SH3 adapter cortactin. FASEB J. 20, 2588–2590. [66] Liu, Q., Berry, D., Nash, P., Pawson, T., McGlade, C.J. and Li, S.S. (2003) Structural basis for specific binding of the Gads SH3 domain to an RxxK motif-containing SLP-76 peptide: a novel mode of peptide recognition. Mol. Cell. 11, 471–481. [67] Seet, B.T., Berry, D.M., Maltzman, J.S., Shabason, J., Raina, M., Koretzky, G.A., McGlade, C.J. and Pawson, T. (2007) Efficient T-cell receptor signaling requires a high-affinity interaction between the Gads C-SH3 domain and the SLP-76 RxxK motif. EMBO J. 26, 678–689. [68] Vidal, M., Gigoux, V. and Garbay, C. (2001) SH2 and SH3 domains as targets for anti-proliferative agents. Crit. Rev. Oncol. Hematol. 40, 175–186. [69] Smithgall, T.E. (1995) SH2 and SH3 domains: potential targets for anti-cancer drug design. J. Pharmacol. Toxicol. Methods 34, 125–132. [70] Garbay, C., Liu, W.Q., Vidal, M. and Roques, B.P. (2000) Inhibitors of Ras signal transduction as antitumor agents. Biochem. Pharmacol. 60, 1165–1169. [71] Lulf, S., Horenkamp, F.A., Breuer, S. and Geyer, M. (2011) Nef surfaces: where to interfere with function. Curr. HIV Res. 9, 543–551. [72] Li, H. and Lawrence, D.S. (2005) Acquisition of Fyn-selective SH3 domain ligands via a combinatorial library strategy. Chem. Biol. 12, 905–912. [73] Oneyama, C., Agatsuma, T., Kanda, Y., Nakano, H., Sharma, S.V., Nakano, S., Narazaki, F. and Tatsuta, K. (2003) Synthetic inhibitors of proline-rich ligandmediated protein–protein interaction: potent analogs of UCS15A. Chem. Biol. 10, 443–451. [74] Tong, A.H. et al. (2002) A combined experimental and computational strategy to define protein interaction networks for peptide recognition modules. Science 295, 321–324. [75] Cobos, E.S., Pisabarro, M.T., Vega, M.C., Lacroix, E., Serrano, L., Ruiz-Sanz, J. and Martinez, J.C. (2004) A miniprotein scaffold used to assemble the polyproline II binding epitope recognized by SH3 domains. J. Mol. Biol. 342, 355–365. [76] Pisabarro, M.T. and Serrano, L. (1996) Rational design of specific high-affinity peptide ligands for the Abl-SH3 domain. Biochemistry 35, 10634–10640. [77] Pisabarro, M.T., Ortiz, A.R., Viguera, A.R., Gago, F. and Serrano, L. (1994) Molecular modeling of the interaction of polyproline-based peptides with the Abl-SH3 domain: rational modification of the interaction. Protein Eng. 7, 1455–1462. [78] Lawrence, D.S. (2005) Signaling protein inhibitors via the combinatorial modification of peptide scaffolds. Biochim. Biophys. Acta 1754, 50–57.

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630 [79] Mayer, J.P. and Dimarchi, R.D. (2005) Drugging the undruggable. Chem. Biol. 12, 860–861. [80] Vidal, M., Liu, W.Q., Lenoir, C., Salzmann, J., Gresh, N. and Garbay, C. (2004) Design of peptoid analogue dimers and measure of their affinity for Grb2 SH3 domains. Biochemistry 43, 7336–7344. [81] Inglis, S., Jones, R., Fritz, D., Stojkoski, C., Booker, G. and Pyke, S. (2005) Synthesis of 5-, 6- and 7-substituted-2-aminoquinolines as SH3 domain ligands. Org. Biomol. Chem. 3, 2543–2557. [82] Inglis, S.R., Stojkoski, C., Branson, K.M., Cawthray, J.F., Fritz, D., Wiadrowski, E., Pyke, S.M. and Booker, G.W. (2004) Identification and specificity studies of small-molecule ligands for SH3 protein domains. J. Med. Chem. 47, 5405–5417. [83] Nguyen, J.T., Porter, M., Amoui, M., Miller, W.T., Zuckermann, R.N. and Lim, W.A. (2000) Improving SH3 domain ligand selectivity using a non-natural scaffold. Chem. Biol. 7, 463–473. [84] Aghazadeh, B. and Rosen, M.K. (1999) Ligand recognition by SH3 and WW domains: the role of N-alkylation in PPII helices. Chem. Biol. 6, R241–R246. [85] Freire, E. (2009) A thermodynamic approach to the affinity optimization of drug candidates. Chem Biol Drug Des 74, 468–472. [86] Luque, J. (2011) Biophysics of Protein–Protein Interactions in: Protein Surface Recognition. Approaches for Drug Discovery (Giralt, E., Peczuh, M.W. and Salvatella, X., Eds.), pp. 23–45, John Wiley & Sons. [87] Freire, E. (2008) Do enthalpy and entropy distinguish first in class from best in class? Drug Discov. Today 13, 869–874. [88] Lafont, V., Armstrong, A.A., Ohtaka, H., Kiso, Y., Mario Amzel, L. and Freire, E. (2007) Compensating enthalpic and entropic changes hinder binding affinity optimization. Chem. Biol. Drug Des. 69, 413–422. [89] Velazquez Campoy, A. and Freire, E. (2005) ITC in the post-genomic era...?Priceless. Biophys. Chem. 115, 115–124. [90] Velazquez-Campoy, A., Luque, I. and Freire, E. (2001) The application of thermodynamic methods in drug design. Thermochim. Acta 380, 217–227. [91] Schon, A., Lam, S.Y. and Freire, E. (2011) Thermodynamics-based drug design: strategies for inhibiting protein–protein interactions. Future Med. Chem. 3, 1129–1137. [92] Schon, A., Madani, N., Smith, A.B., Lalonde, J.M. and Freire, E. (2011) Some binding-related drug properties are dependent on thermodynamic signature. Chem. Biol. Drug Des. 77, 161–165. [93] Ladbury, J.E., Klebe, G. and Freire, E. (2010) Adding calorimetric data to decision making in lead discovery: a hot tip. Nat. Rev. Drug Discov. 9, 23–27. [94] Kawasaki, Y. and Freire, E. (2011) Finding a better path to drug selectivity. Drug Discov. Today 16, 985–990. [95] Chaires, J.B. (2008) Calorimetry and thermodynamics in drug design. Annu. Rev. Biophys. 37, 135–151. [96] Holdgate, G.A. and Ward, W.H. (2005) Measurements of binding thermodynamics in drug discovery. Drug Discov. Today 10, 1543–1550. [97] Ohtaka, H., Velazquez-Campoy, A., Xie, D. and Freire, E. (2002) Overcoming drug resistance in HIV-1 chemotherapy: the binding thermodynamics of Amprenavir and TMC-126 to wild-type and drug-resistant mutants of the HIV-1 protease. Protein Sci. 11, 1908–1916. [98] Freire, E. (2002) Designing drugs against heterogeneous targets. Nat. Biotechnol. 20, 15–16. [99] Ferreon, J.C. and Hilser, V.J. (2004) Thermodynamics of binding to SH3 domains: the energetic impact of polyproline II (PII) helix formation. Biochemistry 43, 7787–7797. [100] Palencia, A., Cobos, E.S., Mateo, P.L., Martinez, J.C. and Luque, I. (2004) Thermodynamic dissection of the binding energetics of proline-rich peptides to the Abl-SH3 domain: implications for rational ligand design. J. Mol. Biol. 336, 527–537. [101] Renzoni, D.A. et al. (1996) Structural and thermodynamic characterization of the interaction of the SH3 domain from Fyn with the proline-rich binding site on the p85 subunit of PI3-Kinase. Biochemistry 35, 15646–15653. [102] Wang, C., Pawley, N.H. and Nicholson, L.K. (2001) The role of backbone motions in ligand binding to c-Src SH3 domain. J. Mol. Biol. 313, 873–887. [103] Wittekind, M. et al. (1994) Orientation of peptide fragments from Sos proteins boud to the N-terminal SH3 domain of Grb2 determined by NMR spectroscopy. Biochemistry 33, 13531–13539. [104] Rubini, C. et al. (2010) Recognition of lysine-rich peptide ligands by murine cortactin SH3 domain: CD, ITC, and NMR studies. Biopolymers 94, 298–306. [105] Arold, S.T. and Baur, A.S. (2001) Dynamic Nef and Nef dynamics: how structure could explain the complex activities of this small HIV protein. Trends Biochem. Sci. 26, 356–363. [106] Chan, B. et al. (2003) SAP couples Fyn to SLAM immune receptors. Nat. Cell. Biol. 5, 155–160. [107] Aitio, O. et al. (2008) Structural basis of PxxDY motif recognition in SH3 binding. J. Mol. Biol. 382, 167–178. [108] McDonald, C.B., Seldeen, K.L., Deegan, B.J. and Farooq, A. (2009) SH3 domains of Grb2 adaptor bind to PXpsiPXR motifs within the Sos1 nucleotide exchange factor in a discriminate manner. Biochemistry 48, 4074–4085. [109] Ferreon, J.C. and Hilser, V.J. (2003) Ligand-induced changes in dynamics in the RT loop of the C-terminal SH3 domain of Sem-5 indicate cooperative conformational coupling. Protein Sci. 12, 982–996. [110] Cordier, F., Wang, C., Grzesiek, S. and Nicholson, L.K. (2000) Ligand-induced strain in hydrogen bonds of the c-Src SH3 domain detected by NMR. J. Mol. Biol. 304, 497–505. [111] Brasher, B.B., Roumiantsev, S. and Van Etten, R.A. (2001) Mutational analysis of the regulatory function of the c-Abl Src homology 3 domain. Oncogene 20, 7744–7752.

2629

[112] Azam, M., Latek, R.R. and Daley, G.Q. (2003) Mechanisms of autoinhibition and STI-571/imatinib resistance revealed by mutagenesis of BCR-ABL. Cell 112, 831–843. [113] Viguera, A.R., Arrondo, J.L., Musacchio, A., Saraste, M. and Serrano, L. (1994) Characterization of the interaction of natural proline-rich peptides with five different SH3 domains. Biochemistry 33, 10925–10933. [114] Pisabarro, M.T., Serrano, L. and Wilmanns, M. (1998) Crystal structure of the abl-SH3 domain complexed with a designed high-affinity peptide ligand: implications for SH3-ligand interactions. J. Mol. Biol. 281, 513–521. [115] Luque, I. and Freire, E. (2002) Structural parameterization of the binding enthalpy of small ligands. Proteins 49, 181–190. [116] Velazquez-Campoy, A., Luque, I., Todd, M.J., Milutinovich, M., Kiso, Y. and Freire, E. (2000) Thermodynamic dissection of the binding energetics of KNI272, a potent HIV-1 protease inhibitor. Protein Sci. 9, 1801–1809. [117] Palencia, A., Camara-Artigas, A., Pisabarro, M.T., Martinez, J.C. and Luque, I. (2010) Role of interfacial water molecules in proline-rich ligand recognition by the Src homology 3 domain of Abl. J. Biol. Chem. 285, 2823–2833. [118] Martin-Garcia, J.M., Ruiz-Sanz, J. and Luque, I. (2012) Interfacial water molecules in SH3 interactions: a revised paradigm for polyproline recognition. Biochem. J. 442, 443–451. [119] Minke, W.E., Diller, D.J., Hol, W.G. and Verlinde, C.L. (1999) The role of waters in docking strategies with incremental flexibility for carbohydrate derivatives: heat-labile enterotoxin, a multivalent test case. J. Med. Chem. 42, 1778–1788. [120] Rarey, M., Kramer, B. and Lengauer, T. (1999) Docking of hydrophobic ligands with interaction-based matching algorithms. Bioinformatics 15, 243–250. [121] Lie, M.A., Thomsen, R., Pedersen, C.N., Schiott, B. and Christensen, M.H. (2011) Molecular docking with ligand attached water molecules. J. Chem. Inf. Model. 51, 909–917. [122] Thilagavathi, R. and Mancera, R.L. (2010) Ligand–protein cross-docking with water molecules. J. Chem. Inf. Model. 50, 415–421. [123] Roberts, B.C. and Mancera, R.L. (2008) Ligand-protein docking with water molecules. J. Chem. Inf. Model. 48, 397–408. [124] van Dijk, A.D. and Bonvin, A.M. (2006) Solvated docking: introducing water into the modelling of biomolecular complexes. Bioinformatics 22, 2340– 2347. [125] Verdonk, M.L., Chessari, G., Cole, J.C., Hartshorn, M.J., Murray, C.W., Nissink, J.W., Taylor, R.D. and Taylor, R. (2005) Modeling water molecules in protein– ligand docking using GOLD. J. Med. Chem. 48, 6504–6515. [126] Schnecke, V. and Kuhn, L.A. (2000) Virtual Screeing with solvation and ligand-induced complementarity. Perspect. Drug Discov. 20, 171–190. [127] Mancera, R.L. (2002) De novo ligand design with explicit water molecules: an application to bacterial neuraminidase. J. Comput. Aided Mol. Des. 16, 479– 499. [128] Garcia-Sosa, A.T. and Mancera, R.L. (2006) The effect of a tightly bound water molecule on scaffold diversity in the computer-aided de novo ligand design of CDK2 inhibitors. J. Mol. Model. 12, 422–431. [129] Garcia-Sosa, A.T., Firth-Clark, S. and Mancera, R.L. (2005) Including tightlybound water molecules in de novo drug design. Exemplification through the in silico generation of poly(ADP-ribose)polymerase ligands. J. Chem. Inf. Model. 45, 624–633. [130] Mancera, R.L. (2007) Molecular modeling of hydration in drug design. Curr. Opin. Drug Discov. Devel. 10, 275–280. [131] de Beer, S.B., Vermeulen, N.P. and Oostenbrink, C. (2010) The role of water molecules in computational drug design. Curr. Top Med. Chem. 10, 55–66. [132] Dunitz, J.D. (1995) Win some, lose some: enthalpy–entropy compensation in weak intermolecular interactions. Chem. Biol. 2, 709–712. [133] Dunitz, J.D. (1994) The entropic cost of bound water in crystals and biomolecules. Science 264, 670. [134] Cooper, A. (2005) Heat capacity effects in protein folding and ligand binding: a re-evaluation of the role of water in biomolecular thermodynamics. Biophys. Chem. 115, 89–97. [135] Tame, J.R., Sleigh, S.H., Wilkinson, A.J. and Ladbury, J.E. (1996) The role of water in sequence-independent ligand binding by an oligopeptide transporter protein. Nat. Struct. Biol. 3, 998–1001. [136] Ladbury, J.E. (1996) Just add water! The effect of water on the specificity of protein–ligand binding sites and its potential application to drug design. Chem. Biol. 3, 973–980. [137] Lazaridis, T. (1998) Inhomogeneous fluid approach to solvation thermodynamics. 1 Theory. J. Phys. Chem. B 102, 3531–3541. [138] Lazaridis, T. (1998) Inhomogeneous fluid approach to solvation thermodynamics. 2. Applications to simple fluids. J. Phys. Chem. B 102, 3542–3550. [139] Li, Z. and Lazaridis, T. (2006) Thermodynamics of buried water clusters at a protein–ligand binding interface. J. Phys. Chem. B 110, 1464–1475. [140] Li, Z. and Lazaridis, T. (2003) Thermodynamic contributions of the ordered water molecule in HIV-1 protease. J. Am. Chem. Soc. 125, 6636–6637. [141] Li, Z. and Lazaridis, T. (2005) The effect of water displacement on binding thermodynamics: concanavalin A. J. Phys. Chem. B 109, 662–670. [142] Sharrow, S.D., Edmonds, K.A., Goodman, M.A., Novotny, M.V. and Stone, M.J. (2005) Thermodynamic consequences of disrupting a water-mediated hydrogen bond network in a protein:pheromone complex. Protein Sci. 14, 249–256. [143] Lam, P.Y. et al. (1994) Rational design of potent, bioavailable, nonpeptide cyclic ureas as HIV protease inhibitors. Science 263, 380–384.

2630

A. Zafra-Ruano, I. Luque / FEBS Letters 586 (2012) 2619–2630

[144] Chen, J.M., Xu, S.L., Wawrzak, Z., Basarab, G.S. and Jordan, D.B. (1998) Structure-based design of potent inhibitors of scytalone dehydratase: displacement of a water molecule from the active site. Biochemistry 37, 17735–17744. [145] Sleigh, S.H., Seavers, P.R., Wilkinson, A.J., Ladbury, J.E. and Tame, J.R. (1999) Crystallographic and calorimetric analysis of peptide binding to OppA protein. J. Mol. Biol. 291, 393–415. [146] Connelly, P.R. et al. (1994) Enthalpy of hydrogen bond formation in a protein–ligand binding reaction. Proc. Natl. Acad. Sci. U S A 91, 1964–1968. [147] Clarke, C., Woods, R.J., Gluska, J., Cooper, A., Nutley, M.A. and Boons, G.J. (2001) Involvement of water in carbohydrate–protein binding. J. Am. Chem. Soc. 123, 12238–12247. [148] Raymer, M.L., Sanschagrin, P.C., Punch, W.F., Venkataraman, S., Goodman, E.D. and Kuhn, L.A. (1997) Predicting conserved water-mediated and polar ligand interactions in proteins using a K-nearest-neighbors genetic algorithm. J. Mol. Biol. 265, 445–464. [149] Garcia-Sosa, A.T., Mancera, R.L. and Dean, P.M. (2003) WaterScore: a novel method for distinguishing between bound and displaceable water molecules in the crystal structure of the binding site of protein–ligand complexes. J. Mol. Model. 9, 172–182. [150] Kellogg, G.E., Semus, S.F. and Abraham, D.J. (1991) HINT: a new method of empirical hydrophobic field calculation for CoMFA. J. Comput. Aided Mol. Des. 5, 545–552. [151] Chen, D.L. and Kellogg, G.E. (2005) A computational tool to optimize ligand selectivity between two similar biomacromolecular targets. J. Comput. Aided Mol. Des. 19, 69–82.

[152] Goodford, P.J. (1985) A computational procedure for determining energetically favorable binding sites on biologically important macromolecules. J. Med. Chem. 28, 849–857. [153] Lu, Y., Wang, R., Yang, C.Y. and Wang, S. (2007) Analysis of ligand-bound water molecules in high-resolution crystal structures of protein–ligand complexes. J. Chem. Inf. Model. 47, 668–675. [154] Gilson, M.K., Given, J.A., Bush, B.L. and McCammon, J.A. (1997) The statisticalthermodynamic basis for computation of binding affinities: a critical review. Biophys. J. 72, 1047–1069. [155] Hamelberg, D. and McCammon, J.A. (2004) Standard free energy of releasing a localized water molecule from the binding pockets of proteins: doubledecoupling method. J. Am. Chem. Soc. 126, 7683–7689. [156] Henriques, D.A. and Ladbury, J.E. (2001) Inhibitors to the Src SH2 domain: a lesson in structure–thermodynamic correlation in drug design. Arch. Biochem. Biophys. 390, 158–168. [157] Tame, J.R., Murshudov, G.N., Dodson, E.J., Neil, T.K., Dodson, G.G., Higgins, C.F. and Wilkinson, A.J. (1994) The structural basis of sequence-independent peptide binding by OppA protein. Science 264, 1578–1581. [158] Iglesias-Bexiga, M., Palencia, A., Blanco, F.J., Macias, M.J., Cobos, E.S. and Luque, I. (2012). Plasticity in viral PPxY Late domain recognition by the third WW domain of human Nedd4. In preparation. [159] Bella, J., Brodsky, B. and Berman, H.M. (1995) Hydration structure of a collagen peptide. Structure 3, 893–906. [160] Melacini, G., Bonvin, A.M., Goodman, M., Boelens, R. and Kaptein, R. (2000) Hydration dynamics of the collagen triple helix by NMR. J. Mol. Biol. 300, 1041–1049.