an overview of recent advances in structural bioinformatics of protein–protein interactions and a...

Q6

Q2

Q1

lable at ScienceDirect

Progress in Biophysics and Molecular Biology xxx (2014) 1e10

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354

55

JPBM929_proof ■ 30 July 2014 ■ 1/10

Contents lists avai

Progress in Biophysics and Molecular Biology

journal homepage: www.elsevier .com/locate/pbiomolbio

565758596061626364
Original research 6566676869707172737475
An overview of recent advances in structural bioinformatics ofproteineprotein interactions and a guide to their principles

Govindarajan Sudha a, Ruth Nussinov b, c, **, Narayanaswamy Srinivasan a, *

a Molecular Biophysics Unit, Indian Institute of Science, Bangalore 560012, Indiab Cancer and Inflammation Program, Frederick National Laboratory for Cancer Research, Leidos Biomedical Research, Inc., National Cancer Institute,Frederick, MD 21702, USAc Sackler Institute of Molecular Medicine, Department of Human Genetics and Molecular Medicine, Sackler School of Medicine, Tel Aviv University,Tel Aviv 69978, Israel

7677
78798081828384858687888990
a r t i c l e i n f o

Article history:Available online xxx

Keywords:Proteineprotein interactionsProteineprotein complexesHomomeric proteinsOligomeric proteinsOligomersStructureEvolutionInteractionFunctionConformation

* Corresponding author. Cancer and InflammationLaboratory for Cancer Research, Leidos Biomedical ReInstitute, Frederick, MD 21702, USA.** Corresponding author.

E-mail addresses: [email protected] (G. Su(R. Nussinov), [email protected], [email protected]

http://dx.doi.org/10.1016/j.pbiomolbio.2014.07.0040079-6107/© 2014 Elsevier Ltd. All rights reserved.

9192939495

Please cite this article in press as: Sudha, G.,and a guide to their principles, Progress in B

a b s t r a c t

Rich data bearing on the structural and evolutionary principles of proteineprotein interactions arepaving the way to a better understanding of the regulation of function in the cell. This is particularly thecase when these interactions are considered in the framework of key pathways. Knowledge of the in-teractions may provide insights into the mechanisms of crucial ‘driver’ mutations in oncogenesis. Theyalso provide the foundation toward the design of proteineprotein interfaces and inhibitors that canabrogate their formation or enhance them. The main features to learn from known 3-D structures ofproteineprotein complexes and the extensive literature which analyzes them computationally andexperimentally include the interaction details which permit undertaking structure-based drug discovery,the evolution of complexes and their interactions, the consequences of alterations such as post-translational modifications, ligand binding, disease causing mutations, host pathogen interactions,oligomerization, aggregation and the roles of disorder, dynamics, allostery and more to the protein andthe cell. This review highlights some of the recent advances in these areas, including design, inhibitionand prediction of proteineprotein complexes. The field is broad, and much work has been carried out inthese areas, making it challenging to cover it in its entirety. Much of this is due to the fast increase in thenumber of molecules whose structures have been determined experimentally and the vast increase incomputational power. Here we provide a concise overview.

© 2014 Elsevier Ltd. All rights reserved.

9697
98 99
100101102103104105106107108109

1. The classical view of proteineprotein interactions

Understanding biological systems requires detailed knowledgeof cellular events at the detailed molecular level. This level includesthe physical interactions between macromolecules such as DNA,RNA and proteins and between these and their environment,including lipids, ions and second messengers, such as cAMP. Herewe focus on proteineprotein interactions which are responsible forcarrying out diverse processes in living systems. Structural andmechanistic features of proteineprotein interactions may be bestunderstood using the three-dimensional structures of the proteins

Program, Frederick Nationalsearch, Inc., National Cancer

dha), [email protected] (N. Srinivasan).

110111112113114115116

et al., An overview of recentiophysics and Molecular Bio

and their complexes. The structural database provides rich databoth of static crystal structures and their ensembles in solutions byNMR. Protein ensembles can also be glimpsed from collections ofcrystal structures of the same protein, however in different boundand unbound states and crystal forms. Even though the crystalenvironment captures only the state favored under specific crys-tallization conditions, these static structures still provide crucialinformation on the nature of the proteineprotein interactions. Avast majority of heterocomplexes with known 3D structures areheterodimers (Fig. 1). Therefore, there is a need to study the 3Dstructures of higher order heteromers, which often form thefunctional multiprotein assemblies in the cell. Structural bioinfor-matics of proteineprotein interactions, which deals with theanalysis of known 3D structures, has provided detailed informationon the underlying principles of structure, function and dysfunction,and evolution of proteineprotein complexes.

Proteins that are stable only in a proteineprotein complex formand remain together throughout their functional life time are

117118119

advances in structural bioinformatics of proteineprotein interactionslogy (2014), http://dx.doi.org/10.1016/j.pbiomolbio.2014.07.004

mailto:[email protected]




www.sciencedirect.com/science/journal/00796107

http://www.elsevier.com/locate/pbiomolbio

http://dx.doi.org/10.1016/j.pbiomolbio.2014.07.004



Fig. 1. Distribution of number of chaints in the heterocomplexes of known 3-Dstructure: The histogram shows the distribution of number of chaints in the hetero-complexes of known structure available in the Protein Data Bank (PDB (Berman et al.,2000)). The data to generate this figure corresponds to the number of chains in thebiological units as presented in the PDB.

G. Sudha et al. / Progress in Biophysics and Molecular Biology xxx (2014) 1e102

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

66676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130

JPBM929_proof ■ 30 July 2014 ■ 2/10

termed as ‘permanent’ proteineprotein complexes. On their own,these proteins are typically disordered; that is, they exist in a rangeof conformational states, with none of these having a sufficientlystable conformation to be captured in crystalline form. Proteine-protein complexes that interact with their partner for a brief periodof time to carry out a specific function and are stable in their freeform are termed ‘transient’ (Nooren and Thornton, 2003a). Onaverage, there are differences in the structures and chemicalcharacteristics of interfaces between permanent and transientproteineprotein complexes (De et al., 2005).

The evolution of the interfaces was suggested to be slower forpermanent proteineprotein complexes than for transient com-plexes (Mintseris and Weng, 2005). Transient proteineprotein in-terfaces show higher residue conservation than rest of the tertiarystructural surface (Choi et al., 2009; Mintseris and Weng, 2005;Valdar and Thornton, 2001). Physicochemical and geometricalcharacterization of protein interfaces have been extensively studiedthat are different from the rest of the surface (Jones and Thornton,1996) (De et al., 2005; Jones et al., 2000; Lo Conte et al., 1999;Sonavane and Chakrabarti, 2008). Differences in interfacial fea-tures have also been observed between permanent and transientproteineprotein complexes. Interface size (small interfaces intransient proteineprotein complexes versus large interfaces inpermanent complexes), area, polarity (polar interfaces in transientproteineprotein complexes versus non-polar interfaces in perma-nent complexes), shape complementarity, conformational changesupon binding, residue interface propensities and residue contactshave served as distinguishing features to predict and classify per-manent and transient proteineprotein complexes (Ansari andHelms, 2005; Bahadur et al., 2003; Block et al., 2006; De et al.,2005; Jones and Thornton, 1996; Keskin et al., 2008; Levy andPereira-Leal, 2008; Mintseris and Weng, 2003; Nooren andThornton, 2003b; Zhu et al., 2006).

A proteineprotein interface can be divided into core and rimwhich are buried in the interface and remain accessible to thesolvent, respectively (Bahadur et al., 2003). Interestingly, the coreand the rim differ in their amino acid composition and conservation(Janin et al., 2008). Another important approach to interface

Please cite this article in press as: Sudha, G., et al., An overview of recentand a guide to their principles, Progress in Biophysics and Molecular Bio

residue classification is based on the contributions to interactionenergy. The subset of interface residues that serve as major con-tributors to binding energy in proteineprotein interfaces (>2 kcal/mol) have been termed hot-spot residues (Bogan and Thorn, 1998).

Analysis of a large number of 3-D structures of proteineproteincomplexes revealed that, in general homologous proteineproteincomplex structures are conserved (Aloy et al., 2003). However, in-terfaces of distantly-related homologous proteins are usually nottopologically equivalent (Rekha et al., 2005). Further detailedanalysis showed that spatial orientations of interacting proteinswith respect to each other in some of the homologous proteine-protein complexes differ (Kim et al., 2006). Studies also showedthat interactions between proteins could often be predicted suc-cessfully if the proteins have high sequence similarities with pro-teins, which are known to interact with each other (Levy andPereira-Leal, 2008; Mika and Rost, 2006). Studies also showedthat structurally similar interfaces can bind proteins with differentbinding site structures and different functions (Tsai et al., 1996).This is accommodated through conserved interactions at similarinterface locations, despite having different partners (Keskin andNussinov, 2007). Even if the overall structures of the interactingchains are different, interface similarity may exist (Keskin andNussinov, 2005). While proteineprotein interfaces are typicallyhighly specific, there appear to be proteins with ‘promiscuous’binding characteristics (Schreiber and Keating, 2011). One way toachieve specificity is by utilizing different hotspot residues in theprotein interfaces (Gretes et al., 2009). Alternatively, differentconformations in the ensemble may be selected (Ma et al., 1999;Tsai et al., 1999a, b). Clusters of interacting residues have beenobserved in proteineprotein interfaces and cooperative in-teractions between residues in a cluster generate binding affinityand specificity (Reichmann et al., 2005). Below, we briefly discussrecent and emerging views in structural bioinformatics of pro-teineprotein interactions.

2. Recent and emerging views on proteineproteininteractions

2.1. Proteineprotein complexes are multifaceted

A grasp of the structural and evolutionary principles of pro-teineprotein interactions is essential to understand the roles ofproteins in the cell. Degeneracy is observed not only at the level ofprotein folds but also at the level of proteineprotein interfacestructures. This is due to the structural constraints of packing ofsecondary structural elements at the interface and functionalconstraints (Gao and Skolnick, 2010). Using available 3-D structuresof proteineprotein complexes, interfaces have been clustered and itwas proposed that the repertoire of structures of interfaces islimited (Cukuroglu et al., 2014). However, surprisingly the conser-vation of interfaces in evolutionarily-related proteineproteincomplexes does not always take place (Zhang et al., 2010), whichsuggests that interfaces are tuned for specific interactions, whichthen lead to specific cellular pathways. Alternate binding modes inhomologous proteineprotein complexes have been observed, withthe interfaces not entirely topologically equivalent (Fig. 2). (Hampand Rost, 2012; Kundrotas and Vakser, 2013), and there are ex-amples of proteins which can bind to different proteins with non-equivalent locations (Martin, 2010). There seems to be evolu-tionary ‘plasticity’ in homologous proteineprotein interfaces whichare manifested as different types of interface contacts especiallythose involving polar residues (Andreani et al., 2012). ‘Plasticity’reflects the presence of proteins as conformational ensembles, withdifferent conformations being selected followed by minor inducedfit optimization (Csermely et al., 2010). At the same time, we also


Fig. 2. An example of homologous proteineprotein complexes with different binding modes: (a) Erythropoietin-receptor complex (PDB code: 1eer) and (b) Granulocyte colony-stimulating factor e G-CSF receptor complex (PDB code: 1cd9) are homologous pairs of cytokine-receptor complexes with different binding modes. The cytokines are coloredin yellow while the receptors colored in red. (c) Structural superposition of homologous cytokines are shown with the interface residues colored as yellow sticks. The topologicallyequivalent non interface residues of the other homologous cytokine are shown as orange sticks. The types of the interface residues are drastically different from the topologicallyequivalent non interface residues as shown in the figure. The figures are generated using PyMOL (DeLano, 2002).

G. Sudha et al. / Progress in Biophysics and Molecular Biology xxx (2014) 1e10 3

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

66676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130

JPBM929_proof ■ 30 July 2014 ■ 3/10

observe conserved interactions and interface residues in homolo-gous proteineprotein complexes, which is the generally expectedsituation. Firstly, interface residues can be mutated during evolu-tion without affecting their interaction if the interaction is throughmain-chain atoms or b-carbons (Talavera et al., 2012). Secondly, itwas suggested that a subset of interface residues termed rigidinterface residues” are evolutionarily well conserved and con-formationally invariant in associated and unassociated forms ofproteineprotein complexes. This could be due to the dual functionof forming intra-and inter-protein contacts, as well as contacts withthe solvent, which provides stabilization to each side of the inter-face even in their unbound forms (Swapna et al., 2012a).

Proteineprotein interactions evolve. They are highly specific,with functionally required binding affinity and specificity. Paralo-gous proteins that bind to different proteins do so by employingsubfamily-specific residues at the interface (Aiello and Caffrey,2012). Rewiring of interaction specificity has been achieved bymutating some of the interfacial residues by residues from aninterface of another protein (Podgornaia et al., 2013). Informationon co-evolution between interacting proteins has been used tocarry out mutations which helped in altering the specificity (Chenget al., 2014). Specificity for the order of assembly of heteromericmulti-protein complexes may be dictated by protein interface size(Marsh et al., 2013), which acts to stabilize the ‘right’ conformer inthe interactions. However, other factors besides size may also be atplay, including specific stabilizing interactions, achieving the sameoutcome.

Clusters of conserved exposed hydrophobic and charged resi-dues in the uncomplexed form can account for most of the pro-teineprotein binding energy/hotspot residues in the complexedform (Agrawal et al., 2014). Small molecule binding sites in pro-teineprotein complexes share the proteineprotein interface hot-spots. Therefore, binding of small molecules to these hotspots couldinhibit the proteineprotein interactions (London et al., 2013;Thangudu et al., 2012) and constitute drug target. Hot spots arepre-organized in the unbound protein form, presenting reducedmobility (Kozakov et al., 2011; Ozbek et al., 2013). Their confor-mation in the unbound form resembles the one that they assume inthe bound state. Additionally, allosteric druggable hotspot regionscan also be used to modulate proteineprotein interactions (Ma andNussinov, 2013).

2.2. Proteineprotein interfaces can be altered e PTMs, smallmolecule binding and mutations

Chemical alterations in proteineprotein interfaces can bebeneficial or detrimental to proteineprotein complex formation.


Such changes at proteineprotein interfaces can occur by means ofpost-translational modifications such as phosphorylation, acetyla-tion and ubiquitination to either mediate or abrogate proteine-protein interactions (Beltrao et al., 2012; Nishi et al., 2011a;Nussinov et al., 2012). Cross-talk between phosphorylation andacetylation can also co-occur in the same protein interface (Beltraoet al., 2013; van Noort et al., 2012). These modified residues couldserve as interaction hotspots (Nishi et al., 2011a). They can also beintrinsically disordered in the phosphorylated or unphosphory-lated state for functional regulation (Nishi et al., 2013a). Allostericmodifications away from the proteineprotein interface can block ortune functional sites via conformational changes (Nussinov et al.,2012). Even though they are generally found to be evolutionarilymore conserved than other interface residues, they need not bealways conserved for signaling to take place (Nishi et al., 2011a; Tanet al., 2010).

Small molecules can bind to pockets within or close to theproteineprotein interface. This may or may not be due to pro-teineprotein packing. Packing may not be perfect within or at theperiphery of the interface resulting in the formation of pockets forligand binding. Control of proteineprotein interactions, enhancedstability and regulation of function may be achieved through bio-logically relevant ligand binding pockets (Gao and Skolnick, 2012).Interface residues which interact with the complementary proteinas well as small molecules are generally more conserved than otherinterface residues and tend to locate at pockets, which is alsowhereinteraction hotspots often are (Davis and Sali, 2010; Thanguduet al., 2012; Walter et al., 2013). In the complex, these pockets areoften, though not always, filled by the complementary proteinchain (Li et al., 2004). Filled pockets result in high packing density,thus making a residue a hot spot. A hot spot is typically a highlypacked residue in the interface. Since pockets are often identified atthe proteineprotein interface in the unbound or complexed forms,druggable sites could be identified. Moreover, engaging multiplepockets can be a productive strategy in inhibiting proteineproteininteractions (Fuller et al., 2009).

Mutations in proteineprotein interfaces can be manifested asdisease phenotypes. The structural changes following a mutationcould lead to loss of electrostatic interaction, reduction of the hy-drophobic effect, formation of steric clashes, changes in confor-mation, dynamics, and destabilization or over stabilization etc(David et al., 2012; Nishi et al., 2013b; Stefl et al., 2013; Teng et al.,2009). Disease-causing mutations of highly conserved residuesmanifest these detrimental effects (Talavera et al., 2012; Teng et al.,2009). Differences in the mutational microenvironments betweencancerous and neutral mutations have also been observed (Enginet al., 2013; Espinosa et al., 2014; Nishi et al., 2013b; Yates and



1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

66676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130

JPBM929_proof ■ 30 July 2014 ■ 4/10

Sternberg, 2013). Disease mutations can also result in gain of in-teractions leading to changes in partners, aggregation, and changesat the post-translational modification sites causing an order-disorder transition. Proteins whose mutations can provoke cancermay interact with partners through distinct interfaces in multi-interface hubs (Kar et al., 2009). It was also suggested that muta-tions on different interfaces are more likely to cause different dis-orders than those on same interface (Wang et al., 2012).

Alterations in proteineprotein interactions are common ininfection due to the interactions between proteins of pathogenand the host. Large scale comparisons of viral protein structureshave shown the mimicry of host protein structures at variousstructural levels which (Franzosa and Xia, 2011; Garamszegi et al.,2013; Itzhaki, 2011; Segura-Cabrera et al., 2013) appears to serveas a general strategy for successful hostepathogen interaction.Firstly, domain - domain interactions used in the host are alsoobserved in viruses showing mimicry of domains for competitiveinteraction with the host proteins (Itzhaki, 2011). Secondly, hostprotein interfaces that are mimicked are generally interaction hubproteins which are transient in nature (Franzosa and Xia, 2011).Interestingly, Hepatitis C viral protein mimics the interface ofhuman protein kinase without any gross structural similarity be-tween the host and the viral protein (Sudha et al., 2012). Thirdly,viruses can mimic linear motifs involved in domain-motif in-teractions. This is feasible for viruses owing to their small genomesize and convergent evolution (Gadkari and Srinivasan, 2010;Garamszegi et al., 2013). For example, human kinase substraterecognition motif is mimicked by a HCV protein (Sudha et al.,2012). Another strategy by viral proteins is to selectively targetthe host hub proteins for disrupting several host processes. Sinceviral proteins evolve faster than host proteins, there may exist an“evolutionary arms race” at the structural level for survival(Franzosa and Xia, 2011).

2.3. Proteineprotein complexes can be ordered and disordered

Function often requires protein self-assembly to form homo-oligomers. Homo-oligomers usually involve permanent in-teractions; however there may also exist weak dimers, which are inequilibrium with monomers. Interfaces of weak dimers are looselypacked resulting in low stability (Dey et al., 2010). The prevalence ofsymmetric homomers is due to the favorable interaction energyleading to structural and functional advantage despite the entropiccost of the association from disordered interacting pairs of mono-mers (Andre et al., 2008). Asymmetric organization is rare yetsignificant in homomers (Swapna et al., 2012c). The evolutionaryroute of homomer formation can be identified from interface size(Levy et al., 2008). Insertions or deletions of residues at the inter-action interface can be responsible for enabling or disablinghomomer formation. These regions have low aggregation pro-pensity, contain high proportion of polar residues, and largeinterface area which stabilizes the homomer (Hashimoto andPanchenko, 2010). Short insertions can act as spacers to fill cav-ities present at the interface which helps in oligomerization (Nishiet al., 2011b). Similar to transient proteineprotein complexes, ho-mologous homomers also display conservation of binding modeand oligomeric state (Dayhoff et al., 2010). Nevertheless, closelyrelated homomers with different oligomeric states are alsoobserved. Inter-subunit geometry, interface or allosteric mutations,functional and stability constraints can all contribute to changes inthe oligomeric state (Perica et al., 2012). Homomeric proteins havealso been designed artificially to a desired oligomeric state whichhas been used as protein nanomaterial (King et al., 2012). The dy-namics of the quaternary structure can be modulated by ligands,which can be exploited in drug discovery (Jaffe, 2013).


Apart from proteins being well ordered, proteins or part of aprotein can be intrinsically disordered. Disordered states have beenobserved to fulfill key functions in the cell. The prevalence of polarinteractions and complementary electrostatic potentials help toachieve high specificity. Their multiple states coupled with theseresidues also permits interactions with multiple specific partners(Wong et al., 2013). Sequence correlations between amino acids ofthe same type were suggested to be enhanced in structurallydisordered proteins (Fong et al., 2009). Interactions in disorderedregions are less conserved than the ordered regions possibly owingto their higher capacity to interact with multiple partners andrewire interactions (Mosca et al., 2012). Disordered regions aremore common in symmetric homodimers than heterodimerswhich typically present small interfaces. The symmetric arrange-ment in homodimers was suggested to help bring the disorderedregions close in space which permits access to the interactingpartner thereby maintaining the function (Fong et al., 2009). Allo-steric mechanisms are also manifested in IDPs (intrinsically disor-dered proteins) (Ferreon et al., 2013). The binding partners of IDPsmay have different folds or with the same fold characterized by lowsequence identity (Hsu et al., 2013). Alternatively, spliced proteinsalso tend to be intrinsically disordered where different disorderedsegments can mediate different interactions resulting in newfunctions especially those involved in signaling pathways (Buljanet al., 2013; Hsu et al., 2013). Tissue specific exons in humans aregenerally disordered with distinct interacting partners in differenttissues (Buljan et al., 2012). Proteins belonging to various functionalclasses such as proteineprotein binding, phosphosphorylation,acetylation, metal binding, substrate/ligand binding, polymeriza-tion and transcriptional activation can be intrinsically disorderedproteins. Functions in these proteins could arise from a transition ofdisordered to ordered state, or function arising from disorderedstate (Sickmeier et al., 2007; Hsu et al., 2013; Nishi et al., 2013a).Disease associated mis-sense mutations can also be located in IDPswhich are deleterious (Vacic and Iakoucheva, 2012).

Transition of ordered to unordered assembly in proteineproteincomplexes causes a transition of function to dysfunction. Unor-dered assembly or non-specific interaction of proteins can result inaggregation, which is typically detrimental to the cell. An inter-esting observation showed that the interface regions are moreprone to aggregation than the exposed surface region, which maybe expected due to the higher fraction of hydrophobic and aromaticresidues in interfaces. However, uncontrolled assembly can beprevented by disulphide bonds and salt bridges to form specificinteractions (Pechmann et al., 2009). Highly abundant proteinsgenerally have less sticky surfaces preventing formation of non-functional self-interactions and are poorly conserved (Levy et al.,2012). Aggregation prone proteins are subjected to differentialtranscription, translation and degradation control. The concentra-tions of these regions are always kept under critical level requiredfor aggregation (Gsponer and Babu, 2012). Short living proteinsshow a higher aggregation propensity which is often associatedwith deposition diseases (De Baets et al., 2011). Generally, aggre-gation prone regions are protected by burial in the hydrophobiccore. However, mutations and stress can make these regions sol-vent exposed and leads to the formation of b-structure agglomer-ates resulting in a disease phenotype. Enrichment of chargedresidues at the flanking region of aggregation prone segments canhelp against aggregate formation serving as gatekeepers. Thesefactors increase the intrinsic solubility of otherwise aggregatingsequences (Beerten et al., 2012a,b). Better protection of aggregationprone regions and aggregation gate keeping ability are observedmore often in thermophilic proteins than in mesophilic proteins. Inspite of the fact that proteins of the same family from thermophilicand mesophilic organisms are homologous, interestingly



1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

66676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130

JPBM929_proof ■ 30 July 2014 ■ 5/10

aggregation prone regions appear not to be conserved (Thangakaniet al., 2012). Aggregation prone regions have been analyzed inrandomly generated amino acid sequences, monomeric proteins,intrinsically disordered proteins and catalytic residues. The aggre-gation propensities of monomeric proteins are lower than randomsequences (Buck et al., 2013). The likelihood of aggregation isgenerally not even in artificial sequences (Angyan et al., 2012).Aggregation in antibodies has been prevented by engineeringmutations in aggregation prone motifs which has led to enhancedstability in the IgG antibody (Chennamsetty et al., 2009).

2.4. Proteineprotein complexes are dynamic

Proteineprotein complexes are dynamic in nature. It is inter-esting to note that the evolutionary changes in the proteinsequence and structures are related to protein flexibility (Marshand Teichmann, 2014). The mobility of amino acids is inverselyproportional to amino acid conservation (Liu and Bahar, 2012).Evolution of functions has been shown to be related to changes inprotein dynamics (Lai et al., 2012). Catalytic loop motions help insubstrate recognition and binding (Kurkcuoglu et al., 2012). Regionsin the enzymewhich showco-evolution as well as highmobility arepredisposed to be substrate recognition sites. Similarity in thebinding pocket dynamics within a protein family is also observed(Lai et al., 2012). Enzymes bound to inhibitors and in unbound formshow that the ligands select the conformer that best matches thestructural and dynamic properties among the conformers (Bakanand Bahar, 2009). The mobility of amino acids at the dimericinterface is generally lower than other amino acids at the tertiarystructural surface (Zen et al., 2010). Conformational loop dynamicshas been used to understand binding promiscuity and specificity inEph-ephrin systems (Nussinov and Ma, 2012). Large conforma-tional changes are often observed upon binding. Relative solventaccessible surface area has been used to predict the magnitude ofbinding-induced conformational changes from single chains orproteineprotein complexes. Large conformational changes may beespecially significant in sequences enriched with intrinsicallydisordered regions, although lack of stable structures in such casesprevents direct comparisons and definitive assessments. Oligomerstake advantage of the intrinsic dynamics in the individual subunits.At the same time, oligomerization stiffens the interfacial regions ofthe subunits and was suggested to provide new cooperative modes(Marcos et al., 2011). Interactions may also entail minor, or evensubtle conformational changes, as can be seen from comparisons ofcorresponding bound and unbound protein structures.

In allostery the ligand or protein effector binds to a protein andchanges the structure and/or dynamics at a distant site which isoften a functional site (Tsai et al., 2008). Even though a metabolitebinds the enzyme at a site distant from the catalytic site, its bindingis coupled to the active site. Numerous atoms are involved in theinteraction between the two sites and the conformational changesin the protein structure (Manley et al., 2013).When surface sites arelinked to active sites, they can be preferred locations for emergenceof allosteric control and serve as hotspots for interaction (Reynoldset al., 2011). Sometimes substrates are transferred to the enzymewith help from scaffolding proteins through allosteric regulation(Nussinov et al., 2013). Allosteric signals are transmitted throughmultiple pre-existing pathways. Perturbations at any protein sitewill shift the pre-existing ensemble of pathways (del Sol et al.,2009).

Proteineprotein complexation causes structural changes notonly at the proteineprotein interface but also in regions away fromthe interface eliciting allostery. Allostery can be induced by anyprotein effector, and plays key roles in signaling systems (Kar et al.,2010; Swapna et al., 2012b). Allosteric events take place via the


dynamic conformational ensembles enabling information transferin signaling proteineprotein complexes (Kar et al., 2010; Motlaghet al., 2014). Since both tertiary and quaternary scales of motionsact interdependently, a global communication network had beendeveloped which integrates both tertiary (residue level) and qua-ternary (subunit level) structural changes which are both requiredfor allosteric communication. This approach can be used to designallostery in non-allosteric proteins (Daily and Gray, 2009).

Allostery can be induced by modulating the amplitude of ther-mal fluctuations around a mean structure rather than conforma-tional changes in the structure (McLeish et al., 2013). The CRP/FNRfamily of transcription factors shows only low frequency dynamicswithout the change in structure. Residues involved in allostericcontrol are conserved. Changes in the low frequency dynamicscorrelate with the allosteric effects of ligand binding (Rodgers et al.,2013). All protein conformations pre-exist and the ligand choosesthe most favored conformation. Upon ligand binding, populationshift in the conformational ensemble is observed thereby redis-tributing the conformational states. Upon binding a selectedconformation, optimization of side chain and back-bone in-teractions proceeds following the induced fit mechanism (Boehret al., 2009). Allosteric transition has been studied by varying thesize and interactions in the allosteric sites to construct a series ofenergy landscapes corresponding to effector bound and unboundstructures. Ligand induced cooperatively can measure how a givensite responds to the effector binding. These models have been usedto reproduce allosteric motion (Weinkam et al., 2012). Allostericdrug discovery, where the drug binds away from the native bindingsite andmodulates the native interactions holds promise. However,a main challenge is to identify the allosteric hotspots (Ma andNussinov, 2013). Aberrant allosteric actions can also result indiseased conditions (Nussinov and Tsai, 2013).

2.5. Sculpting proteineprotein interfaces can be advantageous

Making alterations in proteineprotein interfaces or proteininterface design helps in the generation of proteineprotein com-plexes with interaction specificity and improved affinity (Chen andKeating, 2012). Proteins with desired functions might be designedby means of computational grafting of functional motifs onto aprotein scaffold. This involves transplantation of both backboneand side chain of linear functional motifs onto scaffold proteins(Azoitei et al., 2012). Thermally stable protein scaffolds that mimicviral epitope structure were able to induce potent neutralizingantibodies (Correia et al., 2014). Also, epitope transplanted intoscaffolds can provide good affinity for antibodies (Ofek et al., 2010).Interface design has also generated a pH dependent IgG bindingprotein. Hotspot interactions are used to design the IgG bindingproteins which are extremely stable and heat resistant and thus canbe used for IgG affinity purification and diagnostic purposes(Strauch et al., 2014). Remodeling of loops near active sites hasintroduced specific enzymeesubstrate interactions making theredesigned enzyme more active than the native protein (Murphyet al., 2009). An enzyme which is naturally a hexamer has beenconverted into a homodimer using directed evolution. Mutations inthe designed homodimer open up the active site for the new sub-strate. However, the thermal stability can be compromised (Yipet al., 2011). Designed zinc-mediated protein interface forms acleft which creates enzyme active sites capable of hydrolysis (Deret al., 2012). An enzyme inhibitor has been designed which bindsto the active site of the enzyme. Protein interface design can bebased on interaction hotspots, shape complementary, and residuetypes resulting in high affinity binding with the enzyme (Prockoet al., 2013). Another approach in protein interface design isbased on a-helix mediated proteineprotein interactions (Azzarito



1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

66676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130

JPBM929_proof ■ 30 July 2014 ■ 6/10

et al., 2013). Design of interaction specificity in proteineproteininteractions has been carried out by rationally rewiring the inter-action interface of Thermotoga maritima two-component system toEscherichia coli two-component system. This shows how individualmutations can contribute to the rewiring of interaction specificity(Podgornaia et al., 2013). Designing many interaction partnerssimultaneously for a protein can provide insights into multi-specificity showing that there are distinct positions at the inter-face for affinity and interaction specificity (Fromer and Shifman,2009). Also, a de novo design of protein interactions based oninteraction hotspots within core secondary structural elements andvariable loops has been made (Fleishman et al., 2011). Design ofancestral homodimer that existed millions of years ago has pro-vided insights into the evolution of protein structure and function.Structures of dimeric galectins from ancestral fish were comparedwith the currently existing galectins. Interestingly, the hydrogenbonding pattern at the dimeric interface, carbohydrate binding siteof the ancestor is different from that of currently existing proteins(Konno et al., 2011).

At the same time it behooves us to note that interface grafting ischallenging. Transplanting amotif from one protein to another doesnot guarantee that the motif will not undergo changes in its newenvironment or that the protein environment will change followingthe grafting. The conformational ensemble will shift. The questionis to what extent the shift will alter the prevailing conformation.Currently such predictions are challenging to make.

2.6. Advances in methodology

Some of the recent improvements in the methodology of pro-teineprotein complex inhibition, prediction of interacting proteins,prediction of interface residues, modeling of proteineproteincomplexes, and proteineprotein docking are discussed brieflybelow (Kastritis et al., 2014; Malhotra et al., 2014; Rodrigues andBonvin, 2014; Shoemaker et al., 2013). It is a general notion thatproteineprotein interfaces are “non-druggable” as they tend to beflat, lacking druggable pockets. However, recent studies havechallenged this view and have changed the perspective aboutproteineprotein modulation by small molecules (Wells andMcClendon, 2007). It has been shown that the contact surfacesare flexible due to amino acid side chain motion and small loopmovements. These features have been captured by molecular dy-namics, even in the absence of ligands, showing transient openingof binding pockets. There have been several examples showing thatoptimised small molecules can bind with an affinity comparable tothat of the native protein. One such example is the ligand SP4206which binds close to the receptor region of IL2 - IL2 Receptor alpha.This ligand is highly specific and a tight binder (Wells andMcClendon, 2007). Scoring functions to assess the druggability ofproteineprotein interfaces using atomic structures have beendeveloped (Basse et al., 2013). Data collection of small moleculesinhibiting proteineprotein complexes and their molecular proper-ties can be a good resource for drug design studies (Buck et al.,2013; Higueruelo et al., 2013, 2009).

The traditional view of structure-based drug design used onlysingle protein-ligand structures. However, recent view has utilizedinformation from ensemble of protein conformations using mo-lecular dynamic simulations which reflect the flexibility of proteinsduring ligand binding, as in the case of HIV protease. This area isimportant for successful drug design (Meagher and Carlson, 2004).Yet another interesting discovery is the use of stable and cellpermeable stapled alpha helical peptides, which have been shownto successfully inhibit P53-HDMX and NOTCH transactivationcomplex (Chang et al., 2013). Previous interface residue predictionmethods were dependent on sequence information and structural


features in order to differentiate between interface residues fromsurface residues (Zhou and Qin, 2007). Recently a co-variancemethod has been adopted to identify the correlation betweenamino acid positions in interacting proteins using sequence infor-mation (Weigt et al., 2009). Evolutionary information derived fromhomologous domains in proteins with diverse architectures hasbeen used to predict domainedomain interfacial residues(Bhaskara et al., 2013). Prediction of interface residues in low res-olution cryoEM assemblies of Dengue viral coat proteins and cla-thrin vesicular assembly has been carried out based only on Caatom positions (Gadkari and Srinivasan, 2010, 2012; Gadkari et al.,2009). Prediction of interacting proteins has been achieved in astudy where the three dimensional structural information wasenhanced by information on functional sites with the quality of theprediction comparable to that of high throughput experiments.Application of this approach resulted in high confidence in-teractions in yeast and human (Zhang et al., 2012). A dockingapproach has been used to distinguish binding partners, since theyhave achieved good performance making their predictions offavorable models highly probable compared to a pool of non-binders (Wass et al., 2011). Proteineprotein docking to identifynative structures is challenging. One way to address this is byranking the docked poses based on structural interface parameters(Malhotra et al., 2014), energy and surface accessibility (Feliu et al.,2011). Attempts to improve proteineprotein docking have beenmade. Modeling a proteineprotein structure based on informationfrom 3-D structures of similar complexes is more reliable than blindproteineprotein docking. Structural comparison of the target sur-face to template proteineprotein interfaces followed by applicationof docking energy function is based on similar principle except thatthis strategy requires only the interface region rather than theentire structure (Tuncbag et al., 2011). Threading is anotherapproach to model proteineprotein complexes based on templateidentification (Mukherjee and Zhang, 2011). Data driven approachto docking evolutionarily conserved residues has also been useful(de Vries et al., 2010). Binding-induced conformational changespresent a great challenge in modeling proteineprotein complexesby docking. A flexible multi-domain docking has been devisedwhere a flexible partner is treated as an assembly of domains thatare docked simultaneously. The elastic network model predicts theextent of conformational change (Karaca and Bonvin, 2011). Sur-prisingly, in spite of the limited number of currently availableproteineprotein complexes, a study showed that templates can befound for complexes representing almost all known proteinepro-tein interactions (Kundrotas et al., 2012). Computational modelingalong with experimental information such as chemical shiftperturbation data with residual dipolar couplings have been usedto drive proteineprotein docking (van Zundert and Bonvin, 2014).Recently, methods presented in the critical assessment of predictedinteractions (CAPRI), predicted the position of water molecules inthe proteineprotein interface, estimated the relative binding af-finity, effect of point mutations on the stability of designed andnative proteineprotein interactions (Lensink and Wodak, 2013).

3. Outlook

A wealth of data on various features of proteineprotein com-plexes has been assembled in this review. We believe that there arestill many outstanding questions which are yet to be unraveled inthis vast area.

High throughout proteineprotein interaction datasets aregrowing rapidly. However, their completeness and the occurrenceof false positives are still major concerns. An integrated approachexploiting structural and evolutionary insights of proteineproteininteractions may improve the confidence in modeled


Q3,4


1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

66676869707172737475767778798081828384858687888990919293949596979899

100101102103104

JPBM929_proof ■ 30 July 2014 ■ 7/10

proteineprotein interactions. Further, it is important to understandwhy certain proteins do not interact with each other in spite of theirbeing co-localized. There are still template-based proteineproteincomplexes modelling approaches based on the premise thatinterface residues can be directly extrapolated from the templateproteineprotein complex to modeled proteineprotein complex.This notion may not always hold due to the high evolutionaryplasticity observed in protein interfaces, arguing for further im-provements in such template-based modeling methods. Anotherimportant aspect begging improvement is incorporation of proteinflexibility. To date, despite many efforts over years, proteineproteindocking methods are still largely rigid with only limited extent offlexibility. Since molecular surfaces are flexible, in the absence ofadditional biochemical data this may hamper accurate and reliabledocking. This area is challenging largely due to the high computa-tional costs of sampling proteineprotein modeling poses on theatomic scale.

Most of the time, the proteineprotein complexes being studiedextensively constitute part of a large macromolecular assembly inthe living system. The Cryo-EM (Cryo-Electron microscopy) tech-nique has provided low resolution maps of large assemblies. Asshown in Fig. 3, EM maps are commonly available in the resolutionrange of only �8Å e 12Å. Therefore, fitting the atomic levelstructures into EM maps is an important step in order to under-stand the molecular details of interactions in huge molecular as-semblies. Further, knowledge of the structural and evolutionaryfeatures of proteineprotein complexes such as conservation ofinterface residues and dynamics can be applied to improve thefitting of atomic level structures into the cryoEM density maps.Also, fitting of atomic level structures in cryoEM maps should takeinto account the conformational changes that occur betweenuncomplexed state (available at atomic resolution) and the com-plexed state (available at low resolution) in the multi-protein as-sembly. Detailed understanding of structural and functionalconstraints of quaternary structures of proteins as reflected in theirevolution needs to be explored in greater detail. These could ulti-mately provide a holistic view of the functioning of the protein inthe cell. Since evolutionary and structural dynamics are related,their combination can improve the confidence of protein function

Fig. 3. Distribution of resolution of single particle electron microscopy (EM) maps: Thehistogram shows the distribution of resolution of the single electron microscopy (EM)maps available in the Electron microscopy data bank (EMDB (Lawson et al., 2011)).


prediction. Protein interaction networks from various organismsprovide information on hub proteins coupling these together tobetter understand hub proteins regulation. A key question is howdoes a hub protein, with a shared binding site ‘know’which partnerto bind at any given time (Tsai et al., 2009).

Similar approaches can be employed to understand if annotatedsteps in the metabolic pathway are sequential or simultaneous,although since these often involve enzymatic reactions, differentconsiderations may apply.

Even though several principles guide the design of inhibitors forproteineprotein complexes, a combined approach which considersthe structural dynamics of proteineprotein interface, interactionhotspots, ‘rigid’ interface residues, post-translational modificationsites and allosteric sites could eventually be adopted. This is how-ever challenging because of their variability, temporal occurrencesand combinations which can be expected to lead to differentinterface conformations. Notwithstanding, these may enhance orattenuate the affinity and specificity of the designed molecule.

Ultimately, understanding biological processes requires themolecular details of proteineprotein interactions. It is of para-mount importance due to the diverse applications in translationalresearch in the area of drug discovery, protein interface design, andmost of all the fundamental understanding which bears on all ofthese.

105106107

Acknowledgments

We thank lab members for discussions and suggestions. We alsothank Prof. Tom Blundell and Dr. Harry Jubb for their criticalcomments and suggestions. G.S is supported by a fellowship fromthe Department of Biotechnology, India. This research is supportedby the Department of Biotechnology, Government of India. Thisproject has been funded inwhole or in part with Federal funds fromthe Frederick National Laboratory for Cancer Research, NationalInstitutes of Health, under contract HHSN261200800001E. Thisresearch was supported [in part] by the Intramural Research Pro-gram of NIH, Frederick National Lab, Center for Cancer Research.The content of this publication does not necessarily reflect theviews or policies of the Department of Health and Human Services,nor does mention of trade names, commercial products or orga-nizations imply endorsement by the US Government.

108109110111112113114115116117118119120121122123124125126127128129130

References

Agrawal, N.J., Helk, B., Trout, B.L., 2014. A computational tool to predict the evolu-tionarily conserved protein-protein interaction hot-spot residues from thestructure of the unbound protein. FEBS Lett. 588, 326e333.

Aiello, D., Caffrey, D.R., 2012. Evolution of specific protein-protein interaction sitesfollowing gene duplication. J. Mol. Biol. 423, 257e272.

Aloy, P., Ceulemans, H., Stark, A., Russell, R.B., 2003. The relationship betweensequence and interaction divergence in proteins. J. Mol. Biol. 332, 989e998.

Andre, I., Strauss, C.E., Kaplan, D.B., Bradley, P., Baker, D., 2008. Emergence ofsymmetry in homooligomeric biological assemblies. Proc. Natl. Acad. Sci. U. S. A.105, 16148e16152.

Andreani, J., Faure, G., Guerois, R., 2012. Versatility and invariance in the evolutionof homologous heteromeric interfaces. PLoS Comput. Biol. 8, e1002677.

Angyan, A.F., Perczel, A., Gaspari, Z., 2012. Estimating intrinsic structural prefer-ences of de novo emerging random-sequence proteins: is aggregation the mainbottleneck? FEBS Lett. 586, 2468e2472.

Ansari, S., Helms, V., 2005. Statistical analysis of predominantly transient protein-protein interfaces. Proteins 61, 344e355.

Azoitei, M.L., Ban, Y.E., Julien, J.P., Bryson, S., Schroeter, A., Kalyuzhniy, O., Porter, J.R.,Adachi, Y., Baker, D., Pai, E.F., Schief, W.R., 2012. Computational design of high-affinity epitope scaffolds by backbone grafting of a linear epitope. J. Mol. Biol.415, 175e192.

Azzarito, V., Long, K., Murphy, N.S., Wilson, A.J., 2013. Inhibition of alpha-helix-mediated protein-protein interactions using designed molecules. Nat. Chem.5, 161e173.

Bahadur, R.P., Chakrabarti, P., Rodier, F., Janin, J., 2003. Dissecting subunit interfacesin homodimeric proteins. Proteins 53, 708e719.


Q5


1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

66676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130

JPBM929_proof ■ 30 July 2014 ■ 8/10

Bakan, A., Bahar, I., 2009. The intrinsic dynamics of enzymes plays a dominant rolein determining the structural changes induced upon inhibitor binding. Proc.Natl. Acad. Sci. U. S. A. 106, 14349e14354.

Basse, M.J., Betzi, S., Bourgeas, R., Bouzidi, S., Chetrit, B., Hamon, V., Morelli, X.,Roche, P., 2013. 2P2Idb: a structural database dedicated to orthosteric modu-lation of protein-protein interactions. Nucleic Acids Res. 41, D824eD827.

Beerten, J., Jonckheere, W., Rudyak, S., Xu, J., Wilkinson, H., De Smet, F.,Schymkowitz, J., Rousseau, F., 2012a. Aggregation gatekeepers modulate proteinhomeostasis of aggregating sequences and affect bacterial fitness. Protein Eng.Des. Sel. 25, 357e366.

Beerten, J., Schymkowitz, J., Rousseau, F., 2012b. Aggregation prone regions andgatekeeping residues in protein sequences. Curr. Top. Med. Chem. 12,2470e2478.

Beltrao, P., Albanese, V., Kenner, L.R., Swaney, D.L., Burlingame, A., Villen, J.,Lim, W.A., Fraser, J.S., Frydman, J., Krogan, N.J., 2012. Systematic functionalprioritization of protein posttranslational modifications. Cell 150, 413e425.

Beltrao, P., Bork, P., Krogan, N.J., van Noort, V., 2013. Evolution and functional cross-talk of protein post-translational modifications. Mol. Syst. Biol. 9, 714.

Berman, H.M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T.N., Weissig, H.,Shindyalov, I.N., Bourne, P.E., 2000. The protein data bank. Nucleic Acids Res. 28,235e242.

Bhaskara, R.M., Padhi, A., Srinivasan, N., 2013. Accurate prediction of interfacialresidues in two-domain proteins using evolutionary information: implicationsfor three-dimensional modeling. Proteins.

Block, P., Paern, J., Hullermeier, E., Sanschagrin, P., Sotriffer, C.A., Klebe, G., 2006.Physicochemical descriptors to discriminate protein-protein interactions inpermanent and transient complexes selected by means of machine learningalgorithms. Proteins 65, 607e622.

Boehr, D.D., Nussinov, R., Wright, P.E., 2009. The role of dynamic conformationalensembles in biomolecular recognition. Nat. Chem. Biol. 5, 789e796.

Bogan, A.A., Thorn, K.S., 1998. Anatomy of hot spots in protein interfaces. J. Mol. Biol.280, 1e9.

Buck, P.M., Kumar, S., Singh, S.K., 2013. On the role of aggregation prone regions inprotein evolution, stability, and enzymatic catalysis: insights from diverse an-alyses. PLoS Comput. Biol. 9, e1003291.

Buljan, M., Chalancon, G., Dunker, A.K., Bateman, A., Balaji, S., Fuxreiter, M.,Babu, M.M., 2013. Alternative splicing of intrinsically disordered regions andrewiring of protein interactions. Curr. Opin. Struct. Biol. 23, 443e450.

Buljan, M., Chalancon, G., Eustermann, S., Wagner, G.P., Fuxreiter, M., Bateman, A.,Babu, M.M., 2012. Tissue-specific splicing of disordered segments that embedbinding motifs rewires protein interaction networks. Mol. Cell. 46, 871e883.

Chang, Y.S., Graves, B., Guerlavais, V., Tovar, C., Packman, K., To, K.H., Olson, K.A.,Kesavan, K., Gangurde, P., Mukherjee, A., Baker, T., Darlak, K., Elkin, C.,Filipovic, Z., Qureshi, F.Z., Cai, H., Berry, P., Feyfant, E., Shi, X.E., Horstick, J.,Annis, D.A., Manning, A.M., Fotouhi, N., Nash, H., Vassilev, L.T., Sawyer, T.K.,2013. Stapled alpha-helical peptide drug development: a potent dual inhibitorof MDM2 and MDMX for p53-dependent cancer therapy. Proc. Natl. Acad. Sci. U.S. A. 110, E3445eE3454.

Chen, T.S., Keating, A.E., 2012. Designing specific protein-protein interactions usingcomputation, experimental library screening, or integrated methods. ProteinSci. 21, 949e963.

Cheng, R.R., Morcos, F., Levine, H., Onuchic, J.N., 2014. Toward rationally redesigningbacterial two-component signaling systems using coevolutionary information.Proc. Natl. Acad. Sci. U. S. A. 111, E563eE571.

Chennamsetty, N., Helk, B., Voynov, V., Kayser, V., Trout, B.L., 2009. Aggregation-prone motifs in human immunoglobulin G. J. Mol. Biol. 391, 404e413.

Choi, Y.S., Yang, J.S., Choi, Y., Ryu, S.H., Kim, S., 2009. Evolutionary conservation inmultiple faces of protein interaction. Proteins 77, 14e25.

Correia, B.E., Bates, J.T., Loomis, R.J., Baneyx, G., Carrico, C., Jardine, J.G., Rupert, P.,Correnti, C., Kalyuzhniy, O., Vittal, V., Connell, M.J., Stevens, E., Schroeter, A.,Chen, M., Macpherson, S., Serra, A.M., Adachi, Y., Holmes, M.A., Li, Y., Klevit, R.E.,Graham, B.S., Wyatt, R.T., Baker, D., Strong, R.K., Crowe Jr., J.E., Johnson, P.R.,Schief, W.R., 2014. Proof of principle for epitope-focused vaccine design. Nature507, 201e206.

Csermely, P., Palotai, R., Nussinov, R., 2010. Induced fit, conformational selection andindependent dynamic segments: an extended view of binding events. TrendsBiochem. Sci. 35, 539e546.

Cukuroglu, E., Gursoy, A., Nussinov, R., Keskin, O., 2014. Non-redundant unique inter-face structures as templates formodeling protein interactions. PLoSOne 9, e86738.

Daily, M.D., Gray, J.J., 2009. Allosteric communication occurs via networks of tertiaryand quaternary motions in proteins. PLoS Comput. Biol. 5, e1000293.

David, A., Razali, R., Wass, M.N., Sternberg, M.J., 2012. Protein-protein interactionsites are hot spots for disease-associated nonsynonymous SNPs. Hum. Mutat.33, 359e363.

Davis, F.P., Sali, A., 2010. The overlap of small molecule and protein binding siteswithin families of protein structures. PLoS Comput. Biol. 6, e1000668.

Dayhoff, J.E., Shoemaker, B.A., Bryant, S.H., Panchenko, A.R., 2010. Evolution ofprotein binding modes in homooligomers. J. Mol. Biol. 395, 860e870.

De Baets, G., Reumers, J., Delgado Blanco, J., Dopazo, J., Schymkowitz, J., Rousseau, F.,2011. An evolutionary trade-off between protein turnover rate and proteinaggregation favors a higher aggregation propensity in fast degrading proteins.PLoS Comput. Biol. 7, e1002090.

De, S., Krishnadev, O., Srinivasan, N., Rekha, N., 2005. Interaction preferences acrossprotein-protein interfaces of obligatory and non-obligatory components aredifferent. BMC Struct. Biol. 5, 15.


de Vries, S.J., van Dijk, M., Bonvin, A.M., 2010. The HADDOCK web server for data-driven biomolecular docking. Nat. Protoc. 5, 883e897.

del Sol, A., Tsai, C.J., Ma, B., Nussinov, R., 2009. The origin of allosteric functionalmodulation: multiple pre-existing pathways. Structure 17, 1042e1050.

DeLano, W.L., 2002. The pyMOL Molecular Graphics System. DeLano scientific, PaloAlto, CA.

Der, B.S., Edwards, D.R., Kuhlman, B., 2012. Catalysis by a de novo zinc-mediatedprotein interface: implications for natural enzyme evolution and rationalenzyme engineering. Biochemistry 51, 3933e3940.

Dey, S., Pal, A., Chakrabarti, P., Janin, J., 2010. The subunit interfaces of weaklyassociated homodimeric proteins. J. Mol. Biol. 398, 146e160.

Engin, H.B., Guney, E., Keskin, O., Oliva, B., Gursoy, A., 2013. Integrating structure toprotein-protein interaction networks that drive metastasis to brain and lung inbreast cancer. PLoS One 8, e81035.

Espinosa, O., Mitsopoulos, K., Hakas, J., Pearl, F., Zvelebil, M., 2014. Deriving a mu-tation index of carcinogenicity using protein structure and protein interfaces.PLoS One 9, e84598.

Feliu, E., Aloy, P., Oliva, B., 2011. On the analysis of protein-protein interactions viaknowledge-based potentials for the prediction of protein-protein docking.Protein Sci. 20, 529e541.

Ferreon, A.C., Ferreon, J.C., Wright, P.E., Deniz, A.A., 2013. Modulation of allostery byprotein intrinsic disorder. Nature 498, 390e394.

Fleishman, S.J., Corn, J.E., Strauch, E.M., Whitehead, T.A., Karanicolas, J., Baker, D.,2011. Hotspot-centric de novo design of protein binders. J. Mol. Biol. 413,1047e1062.

Fong, J.H., Shoemaker, B.A., Garbuzynskiy, S.O., Lobanov, M.Y., Galzitskaya, O.V.,Panchenko, A.R., 2009. Intrinsic disorder in protein interactions: insights from acomprehensive structural analysis. PLoS Comput. Biol. 5, e1000316.

Franzosa, E.A., Xia, Y., 2011. Structural principles within the human-virus protein-protein interaction network. Proc. Natl. Acad. Sci. U. S. A. 108, 10538e10543.

Fromer, M., Shifman, J.M., 2009. Tradeoff between stability and multispecificity inthe design of promiscuous proteins. PLoS Comput Biol. 5, e1000627.

Fuller, J.C., Burgoyne, N.J., Jackson, R.M., 2009. Predicting druggable binding sites atthe protein-protein interface. Drug. Discov. Today 14, 155e161.

Gadkari, R.A., Srinivasan, N., 2010. Prediction of protein-protein interactions indengue virus coat proteins guided by low resolution cryoEM structures. BMCStruct. Biol. 10, 17.

Gadkari, R.A., Srinivasan, N., 2012. Protein-protein interactions in clathrin vesicularassembly: radial distribution of evolutionary constraints in interfaces. PLoS One7, e31445.

Gadkari, R.A., Varughese, D., Srinivasan, N., 2009. Recognition of interaction inter-face residues in low-resolution structures of protein assemblies solely from thepositions of C(alpha) atoms. PLoS One 4, e4476.

Gao, M., Skolnick, J., 2010. Structural space of protein-protein interfaces is degen-erate, close to complete, and highly connected. Proc. Natl. Acad. Sci. U. S. A. 107,22517e22522.

Gao, M., Skolnick, J., 2012. The distribution of ligand-binding pockets aroundprotein-protein interfaces suggests a general mechanism for pocket formation.Proc. Natl. Acad. Sci. U. S. A. 109, 3784e3789.

Garamszegi, S., Franzosa, E.A., Xia, Y., 2013. Signatures of pleiotropy, economy andconvergent evolution in a domain-resolved map of human-virus protein-pro-tein interaction networks. PLoS Pathog. 9, e1003778.

Gretes, M., Lim, D.C., de Castro, L., Jensen, S.E., Kang, S.G., Lee, K.J., Strynadka, N.C.,2009. Insights into positive and negative requirements for protein-protein in-teractions by crystallographic analysis of the beta-lactamase inhibitory proteinsBLIP, BLIP-I, and BLP. J. Mol. Biol. 389, 289e305.

Gsponer, J., Babu, M.M., 2012. Cellular strategies for regulating functional andnonfunctional protein aggregation. Cell. Rep. 2, 1425e1437.

Hamp, T., Rost, B., 2012. Alternative protein-protein interfaces are frequent excep-tions. PLoS Comput. Biol. 8, e1002623.

Hashimoto, K., Panchenko, A.R., 2010. Mechanisms of protein oligomerization, thecritical role of insertions and deletions in maintaining different oligomericstates. Proc. Natl. Acad. Sci. U. S. A. 107, 20352e20357.

Higueruelo, A.P., Jubb, H., Blundell, T.L., 2013. TIMBAL V2: Update of a DatabaseHolding Small Molecules Modulating Protein-protein Interactions. Database(Oxford) 2013, bat039.

Higueruelo, A.P., Schreyer, A., Bickerton, G.R., Pitt, W.R., Groom, C.R.,Blundell, T.L., 2009. Atomic interactions and profile of small molecules dis-rupting protein-protein interfaces: the TIMBAL database. Chem. Biol. Drug.Des. 74, 457e467.

Hsu, W.L., Oldfield, C.J., Xue, B., Meng, J., Huang, F., Romero, P., Uversky, V.N.,Dunker, A.K., 2013. Exploring the binding diversity of intrinsically disorderedproteins involved in one-to-many binding. Protein Sci. 22, 258e273.

Itzhaki, Z., 2011. Domain-domain interactions underlying herpesvirus-human pro-tein-protein interaction networks. PLoS One 6, e21724.

Jaffe, E.K., 2013. Impact of quaternary structure dynamics on allosteric drug dis-covery. Curr. Top. Med. Chem. 13, 55e63.

Janin, J., Bahadur, R.P., Chakrabarti, P., 2008. Protein-protein interaction and qua-ternary structure. Q. Rev. Biophys. 41, 133e180.

Jones, S., Marin, A., Thornton, J.M., 2000. Protein domain interfaces: characteriza-tion and comparison with oligomeric protein interfaces. Protein Eng. 13, 77e82.

Jones, S., Thornton, J.M., 1996. Principles of protein-protein interactions. Proc. Natl.Acad. Sci. U. S. A. 93, 13e20.

Kar, G., Gursoy, A., Keskin, O., 2009. Human cancer protein-protein interactionnetwork: a structural perspective. PLoS Comput. Biol. 5, e1000601.



1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

66676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130

JPBM929_proof ■ 30 July 2014 ■ 9/10

Kar, G., Keskin, O., Gursoy, A., Nussinov, R., 2010. Allostery and population shift indrug discovery. Curr. Opin. Pharmacol. 10, 715e722.

Karaca, E., Bonvin, A.M., 2011. A multidomain flexible docking approach to deal withlarge conformational changes in the modeling of biomolecular complexes.Structure 19, 555e565.

Kastritis, P.L., Rodrigues, J.P., Bonvin, A.M., 2014. HADDOCK2P2I: a biophysicalmodel for predicting the binding affinity of protein-protein interaction in-hibitors. J. Chem. Inf. Model 54, 826e836.

Keskin, O., Gursoy, A., Ma, B., Nussinov, R., 2008. Principles of protein-protein in-teractions: what are the preferred ways for proteins to interact? Chem. Rev. 108,1225e1244.

Keskin, O., Nussinov, R., 2005. Favorable scaffolds: proteins with different sequence,structure and function may associate in similar ways. Protein Eng. Des. Sel. 18,11e24.

Keskin, O., Nussinov, R., 2007. Similar binding sites and different partners: impli-cations to shared proteins in cellular pathways. Structure 15, 341e354.

Kim, W.K., Henschel, A., Winter, C., Schroeder, M., 2006. The many faces of protein-protein interactions: a compendium of interface geometry. PLoS Comput. Biol.2, e124.

King, N.P., Sheffler, W., Sawaya, M.R., Vollmar, B.S., Sumida, J.P., Andre, I., Gonen, T.,Yeates, T.O., Baker, D., 2012. Computational design of self-assembling proteinnanomaterials with atomic level accuracy. Science 336, 1171e1174.

Konno, A., Kitagawa, A., Watanabe, M., Ogawa, T., Shirai, T., 2011. Tracing pro-tein evolution through ancestral structures of fish galectin. Structure 19,711e721.

Kozakov, D., Hall, D.R., Chuang, G.Y., Cencic, R., Brenke, R., Grove, L.E., Beglov, D.,Pelletier, J., Whitty, A., Vajda, S., 2011. Structural conservation of druggable hotspots in protein-protein interfaces. Proc. Natl. Acad. Sci. U. S. A. 108,13528e13533.

Kundrotas, P.J., Vakser, I.A., 2013. Protein-protein alternative binding modes do notoverlap. Protein Sci. 22, 1141e1145.

Kundrotas, P.J., Zhu, Z., Janin, J., Vakser, I.A., 2012. Templates are available to modelnearly all complexes of structurally characterized proteins. Proc. Natl. Acad. Sci.U. S. A. 109, 9438e9441.

Kurkcuoglu, Z., Bakan, A., Kocaman, D., Bahar, I., Doruker, P., 2012. Coupling be-tween catalytic loop motions and enzyme global dynamics. PLoS Comput. Biol.8, e1002705.

Lai, J., Jin, J., Kubelka, J., Liberles, D.A., 2012. A phylogenetic analysis of normalmodes evolution in enzymes and its relationship to enzyme function. J. Mol.Biol. 422, 442e459.

Lawson, C.L., Baker, M.L., Best, C., Bi, C., Dougherty, M., Feng, P., van Ginkel, G.,Devkota, B., Lagerstedt, I., Ludtke, S.J., Newman, R.H., Oldfield, T.J., Rees, I.,Sahni, G., Sala, R., Velankar, S., Warren, J., Westbrook, J.D., Henrick, K.,Kleywegt, G.J., Berman, H.M., Chiu, W., 2011. EMDataBank.org: unified dataresource for CryoEM. Nucleic Acids Res. 39, D456eD464.

Lensink, M.F., Wodak, S.J., 2013. Docking, scoring, and affinity prediction in CAPRI.Proteins 81, 2082e2095.

Levy, E.D., Boeri Erba, E., Robinson, C.V., Teichmann, S.A., 2008. Assembly reflectsevolution of protein complexes. Nature 453, 1262e1265.

Levy, E.D., De, S., Teichmann, S.A., 2012. Cellular crowding imposes global con-straints on the chemistry and evolution of proteomes. Proc. Natl. Acad. Sci. U. S.A. 109, 20461e20466.

Levy, E.D., Pereira-Leal, J.B., 2008. Evolution and dynamics of protein interactionsand networks. Curr. Opin. Struct. Biol. 18, 349e357.

Li, X., Keskin, O., Ma, B., Nussinov, R., Liang, J., 2004. Protein-protein interactions:hot spots and structurally conserved residues often locate in complementedpockets that pre-organized in the unbound states: implications for docking.J. Mol. Biol. 344, 781e795.

Liu, Y., Bahar, I., 2012. Sequence evolution correlates with structural dynamics. Mol.Biol. Evol. 29, 2253e2263.

Lo Conte, L., Chothia, C., Janin, J., 1999. The atomic structure of protein-proteinrecognition sites. J. Mol. Biol. 285, 2177e2198.

London, N., Raveh, B., Schueler-Furman, O., 2013. Druggable protein-protein inter-actionsefrom hot spots to hot segments. Curr. Opin. Chem. Biol. 17, 952e959.

Ma, B., Kumar, S., Tsai, C.J., Nussinov, R., 1999. Folding funnels and binding mech-anisms. Protein Eng. 12, 713e720.

Ma, B., Nussinov, R., 2013. Druggable orthosteric and allosteric hot spots to targetprotein-protein interactions. Curr. Pharm. Des..

Malhotra, S., Sankar, K., Sowdhamini, R., 2014. Structural interface parameters arediscriminatory in recognising near-native poses of protein-protein interactions.PLoS One 9, e80255.

Manley, G., Rivalta, I., Loria, J.P., 2013. Solution NMR and computational methods forunderstanding protein allostery. J. Phys. Chem. B 117, 3063e3073.

Marcos, E., Crehuet, R., Bahar, I., 2011. Changes in dynamics upon oligomerizationregulate substrate binding and allostery in amino acid kinase family members.PLoS Comput. Biol. 7, e1002201.

Marsh, J.A., Hernandez, H., Hall, Z., Ahnert, S.E., Perica, T., Robinson, C.V.,Teichmann, S.A., 2013. Protein complexes are under evolutionary selection toassemble via ordered pathways. Cell 153, 461e470.

Marsh, J.A., Teichmann, S.A., 2014. Parallel dynamics and evolution: proteinconformational fluctuations and assembly reflect evolutionary changes insequence and structure. Bioessays 36, 209e218.

Martin, J., 2010. Beauty is in the eye of the beholder: proteins can recognize bindingsites of homologous proteins in more than one way. PLoS Comput. Biol. 6,e1000821.


McLeish, T.C., Rodgers, T.L., Wilson, M.R., 2013. Allostery without conformationchange: modelling protein dynamics at multiple scales. Phys. Biol. 10,056004.

Meagher, K.L., Carlson, H.A., 2004. Incorporating protein flexibility in structure-based drug discovery: using HIV-1 protease as a test case. J. Am. Chem. Soc.126, 13276e13281.

Mika, S., Rost, B., 2006. Protein-protein interactions more conserved within speciesthan across species. PLoS Comput. Biol. 2, e79.

Mintseris, J., Weng, Z., 2003. Atomic contact vectors in protein-protein recognition.Proteins 53, 629e639.

Mintseris, J., Weng, Z., 2005. Structure, function, and evolution of transient andobligate protein-protein interactions. Proc. Natl. Acad. Sci. U. S. A. 102,10930e10935.

Mosca, R., Pache, R.A., Aloy, P., 2012. The role of structural disorder in the rewiringof protein interactions through evolution. Mol. Cell. Proteomics 11. M111014969.

Motlagh, H.N., Wrabl, J.O., Li, J., Hilser, V.J., 2014. The ensemble nature of allostery.Nature 508, 331e339.

Mukherjee, S., Zhang, Y., 2011. Protein-protein complex structure predictions bymultimeric threading and template recombination. Structure 19, 955e966.

Murphy, P.M., Bolduc, J.M., Gallaher, J.L., Stoddard, B.L., Baker, D., 2009. Alteration ofenzyme specificity by computational loop remodeling and design. Proc. Natl.Acad. Sci. U. S. A. 106, 9215e9220.

Nishi, H., Fong, J.H., Chang, C., Teichmann, S.A., Panchenko, A.R., 2013a. Regulation ofprotein-protein binding by coupling between phosphorylation and intrinsicdisorder: analysis of human protein complexes. Mol. Biosyst. 9, 1620e1626.

Nishi, H., Hashimoto, K., Panchenko, A.R., 2011a. Phosphorylation in protein-proteinbinding: effect on stability and function. Structure 19, 1807e1815.

Nishi, H., Koike, R., Ota, M., 2011b. Cover and spacer insertions: small non-hydrophobic accessories that assist protein oligomerization. Proteins 79,2372e2379.

Nishi, H., Tyagi, M., Teng, S., Shoemaker, B.A., Hashimoto, K., Alexov, E., Wuchty, S.,Panchenko, A.R., 2013b. Cancer missense mutations alter binding properties ofproteins and their interaction networks. PLoS One 8, e66273.

Nooren, I.M., Thornton, J.M., 2003a. Diversity of protein-protein interactions. EMBOJ. 22, 3486e3492.

Nooren, I.M., Thornton, J.M., 2003b. Structural characterisation and functional sig-nificance of transient protein-protein interactions. J. Mol. Biol. 325, 991e1018.

Nussinov, R., Ma, B., 2012. Protein dynamics and conformational selection in bidi-rectional signal transduction. BMC Biol. 10, 2.

Nussinov, R., Ma, B., Tsai, C.J., 2013. A broad view of scaffolding suggests thatscaffolding proteins can actively control regulation and signaling of multien-zyme complexes through allostery. Biochim. Biophys. Acta 1834, 820e829.

Nussinov, R., Tsai, C.J., 2013. Allostery in disease and in drug discovery. Cell 153,293e305.

Nussinov, R., Tsai, C.J., Xin, F., Radivojac, P., 2012. Allosteric post-translationalmodification codes. Trends Biochem Sci. 37, 447e455.

Ofek, G., Guenaga, F.J., Schief, W.R., Skinner, J., Baker, D., Wyatt, R., Kwong, P.D., 2010.Elicitation of structure-specific antibodies by epitope scaffolds. Proc. Natl. Acad.Sci. U. S. A. 107, 17880e17887.

Ozbek, P., Soner, S., Haliloglu, T., 2013. Hot spots in a network of functional sites.PLoS One 8, e74320.

Pechmann, S., Levy, E.D., Tartaglia, G.G., Vendruscolo, M., 2009. Physicochemicalprinciples that regulate the competition between functional and dysfunctionalassociation of proteins. Proc. Natl. Acad. Sci. U. S. A. 106, 10159e10164.

Perica, T., Chothia, C., Teichmann, S.A., 2012. Evolution of oligomeric state throughgeometric coupling of protein interfaces. Proc. Natl. Acad. Sci. U. S. A. 109,8127e8132.

Podgornaia, A.I., Casino, P., Marina, A., Laub, M.T., 2013. Structural basis of a ratio-nally rewired protein-protein interface critical to bacterial signaling. Structure21, 1636e1647.

Procko, E., Hedman, R., Hamilton, K., Seetharaman, J., Fleishman, S.J., Su, M.,Aramini, J., Kornhaber, G., Hunt, J.F., Tong, L., Montelione, G.T., Baker, D., 2013.Computational design of a protein-based enzyme inhibitor. J. Mol. Biol. 425,3563e3575.

Reichmann, D., Rahat, O., Albeck, S., Meged, R., Dym, O., Schreiber, G., 2005. Themodular architecture of protein-protein binding interfaces. Proc. Natl. Acad. Sci.U. S. A. 102, 57e62.

Rekha, N., Machado, S.M., Narayanan, C., Krupa, A., Srinivasan, N., 2005. Interactioninterfaces of protein domains are not topologically equivalent across familieswithin superfamilies: implications for metabolic and signaling pathways. Pro-teins 58, 339e353.

Reynolds, K.A., McLaughlin, R.N., Ranganathan, R., 2011. Hot spots for allostericregulation on protein surfaces. Cell 147, 1564e1575.

Rodgers, T.L., Townsend, P.D., Burnell, D., Jones, M.L., Richards, S.A., McLeish, T.C.,Pohl, E., Wilson, M.R., Cann, M.J., 2013. Modulation of global low-frequencymotions underlies allosteric regulation: demonstration in CRP/FNR familytranscription factors. PLoS Biol. 11, e1001651.

Rodrigues, J.P., Bonvin, A.M., 2014. Integrative computational modeling of proteininteractions. FEBS J..

Schreiber, G., Keating, A.E., 2011. Protein binding specificity versus promiscuity.Curr. Opin. Struct. Biol. 21, 50e61.

Segura-Cabrera, A., Garcia-Perez, C.A., Guo, X., Rodriguez-Perez, M.A., 2013. A viral-human interactome based on structural motif-domain interactions captures thehuman infectome. PLoS One 8, e71526.



12345678910111213141516171819202122232425262728293031323334353637383940

414243444546474849505152535455565758596061626364656667686970717273747576777879

JPBM929_proof ■ 30 July 2014 ■ 10/10

Shoemaker, B., Wuchty, S., Panchenko, A.R., 2013. Computational large-scale map-ping of protein-protein interactions using structural complexes. Curr. Protoc.Protein Sci. 73. Unit 3 9.

Sickmeier, M., Hamilton, J.A., LeGall, T., Vacic, V., Cortese, M.S., Tantos, A., Szabo, B.,Tompa, P., Chen, J., Uversky, V.N., Obradovic, Z., Dunker, A.K., 2007. DisProt: thedatabase of disordered proteins. Nucleic Acids Res. 35, D786eD793.

Sonavane, S., Chakrabarti, P., 2008. Cavities and atomic packing in protein structuresand interfaces. PLoS Comput Biol. 4, e1000188.

Stefl, S., Nishi, H., Petukh, M., Panchenko, A.R., Alexov, E., 2013. Molecular mecha-nisms of disease-causing missense mutations. J. Mol. Biol. 425, 3919e3936.

Strauch, E.M., Fleishman, S.J., Baker, D., 2014. Computational design of a pH-sensitive IgG binding protein. Proc. Natl. Acad. Sci. U. S. A. 111, 675e680.

Sudha, G., Yamunadevi, S., Tyagi, N., Das, S., Srinivasan, N., 2012. Structural andmolecular basis of interaction of HCV non-structural protein 5A with humancasein kinase 1alpha and PKR. BMC Struct. Biol. 12, 28.

Swapna, L.S., Bhaskara, R.M., Sharma, J., Srinivasan, N., 2012a. Roles of residues inthe interface of transient protein-protein complexes before complexation. Sci.Rep. 2, 334.

Swapna, L.S., Mahajan, S., de Brevern, A.G., Srinivasan, N., 2012b. Comparison oftertiary structures of proteins in protein-protein complexes with unboundforms suggests prevalence of allostery in signalling proteins. BMC Struct. Biol.12, 6.

Swapna, L.S., Srikeerthana, K., Srinivasan, N., 2012c. Extent of structural asymmetryin homodimeric proteins: prevalence and relevance. PLoS One 7, e36688.

Talavera, D., Williams, S.G., Norris, M.G., Robertson, D.L., Lovell, S.C., 2012. Evolv-ability of yeast protein-protein interaction interfaces. J. Mol. Biol. 419, 387e396.

Tan, C.S., Jorgensen, C., Linding, R., 2010. Roles of “junk phosphorylation” inmodulating biomolecular association of phosphorylated proteins? Cell Cycle 9,1276e1280.

Teng, S., Madej, T., Panchenko, A., Alexov, E., 2009. Modeling effects of human singlenucleotide polymorphisms on protein-protein interactions. Biophys. J. 96,2178e2188.

Thangakani, A.M., Kumar, S., Velmurugan, D., Gromiha, M.S., 2012. How do ther-mophilic proteins resist aggregation? Proteins 80, 1003e1015.

Thangudu, R.R., Bryant, S.H., Panchenko, A.R., Madej, T., 2012. Modulating protein-protein interactions with small molecules: the importance of binding hotspots.J. Mol. Biol. 415, 443e453.

Tsai, C.J., del Sol, A., Nussinov, R., 2008. Allostery: absence of a change in shape doesnot imply that allostery is not at play. J. Mol. Biol. 378, 1e11.

Tsai, C.J., Kumar, S., Ma, B., Nussinov, R., 1999a. Folding funnels, binding funnels, andprotein function. Protein Sci. 8, 1181e1190.

Tsai, C.J., Lin, S.L., Wolfson, H.J., Nussinov, R., 1996. A dataset of protein-proteininterfaces generated with a sequence-order-independent comparison tech-nique. J. Mol. Biol. 260, 604e620.

Tsai, C.J., Ma, B., Nussinov, R., 1999b. Folding and binding cascades: shifts in energylandscapes. Proc. Natl. Acad. Sci. U. S. A. 96, 9970e9972.

Tsai, C.J., Ma, B., Nussinov, R., 2009. Protein-protein interaction networks: how can ahub protein bind so many different partners? Trends Biochem Sci. 34, 594e600.

Tuncbag, N., Gursoy, A., Nussinov, R., Keskin, O., 2011. Predicting protein-proteininteractions on a proteome scale by matching evolutionary and structuralsimilarities at interfaces using PRISM. Nat. Protoc. 6, 1341e1354.


Vacic, V., Iakoucheva, L.M., 2012. Disease mutations in disordered region-seexception to the rule? Mol. Biosyst. 8, 27e32.

Valdar, W.S., Thornton, J.M., 2001. Protein-protein interfaces: analysis of amino acidconservation in homodimers. Proteins 42, 108e124.

van Noort, V., Seebacher, J., Bader, S., Mohammed, S., Vonkova, I., Betts, M.J.,Kuhner, S., Kumar, R., Maier, T., O'Flaherty, M., Rybin, V., Schmeisky, A., Yus, E.,Stulke, J., Serrano, L., Russell, R.B., Heck, A.J., Bork, P., Gavin, A.C., 2012. Cross-talk between phosphorylation and lysine acetylation in a genome-reducedbacterium. Mol. Syst. Biol. 8, 571.

van Zundert, G.C., Bonvin, A.M., 2014. Modeling protein-protein complexes usingthe HADDOCK Webserver “Modeling protein complexes with HADDOCK”.Methods Mol. Biol. 1137, 163e179.

Walter, P., Metzger, J., Thiel, C., Helms, V., 2013. Predicting where small moleculesbind at protein-protein interfaces. PLoS One 8, e58583.

Wang, X., Wei, X., Thijssen, B., Das, J., Lipkin, S.M., Yu, H., 2012. Three-dimensionalreconstruction of protein networks provides insight into human genetic dis-ease. Nat. Biotechnol. 30, 159e164.

Wass, M.N., Fuentes, G., Pons, C., Pazos, F., Valencia, A., 2011. Towards the predictionof protein interaction partners using physical docking. Mol. Syst. Biol. 7, 469.

Weigt, M., White, R.A., Szurmant, H., Hoch, J.A., Hwa, T., 2009. Identification ofdirect residue contacts in protein-protein interaction by message passing. Proc.Natl. Acad. Sci. U. S. A. 106, 67e72.

Weinkam, P., Pons, J., Sali, A., 2012. Structure-based model of allostery predictscoupling between distant sites. Proc. Natl. Acad. Sci. U. S. A. 109, 4875e4880.

Wells, J.A., McClendon, C.L., 2007. Reaching for high-hanging fruit in drug discoveryat protein-protein interfaces. Nature 450, 1001e1009.

Wong, E.T., Na, D., Gsponer, J., 2013. On the importance of polar interactions forcomplexes containing intrinsically disordered proteins. PLoS Comput. Biol. 9,e1003192.

Yates, C.M., Sternberg, M.J., 2013. The effects of non-synonymous single nucleotidepolymorphisms (nsSNPs) on protein-protein interactions. J. Mol. Biol. 425,3949e3963.

Yip, S.H., Foo, J.L., Schenk, G., Gahan, L.R., Carr, P.D., Ollis, D.L., 2011. Directed evo-lution combined with rational design increases activity of GpdQ toward a non-physiological substrate and alters the oligomeric structure of the enzyme.Protein Eng. Des. Sel. 24, 861e872.

Zen, A., Micheletti, C., Keskin, O., Nussinov, R., 2010. Comparing interfacial dynamicsin protein-protein complexes: an elastic network approach. BMC Struct. Biol. 10,26.

Zhang, Q.C., Petrey, D., Deng, L., Qiang, L., Shi, Y., Thu, C.A., Bisikirska, B., Lefebvre, C.,Accili, D., Hunter, T., Maniatis, T., Califano, A., Honig, B., 2012. Structure-basedprediction of protein-protein interactions on a genome-wide scale. Nature 490,556e560.

Zhang, Q.C., Petrey, D., Norel, R., Honig, B.H., 2010. Protein interface conservationacross structure space. Proc. Natl. Acad. Sci. U. S. A. 107, 10896e10901.

Zhou, H.X., Qin, S., 2007. Interaction-site prediction for protein complexes: a criticalassessment. Bioinformatics 23, 2203e2209.

Zhu, H., Domingues, F.S., Sommer, I., Lengauer, T., 2006. NOXclass: prediction ofprotein-protein interaction types. BMC Bioinform. 7, 27.

80


an overview of recent advances in structural bioinformatics of protein–protein interactions and a...

Documents