MINIREVIEW Functional classification of scaffold proteins and related molecules La ´ szlo ´ Buday 1,2 and Pe ´ ter Tompa 1 1 Institute of Enzymology, Biological Research Center, Hungarian Academy of Sciences, Budapest, Hungary 2 Department of Medical Chemistry, Semmelweis University Medical School, Budapest, Hungary Introduction Cells use a multitude of signaling proteins to alter cel- lular behavior in response to changes in their external and internal environment. Because of the multiplicity and broad substrate specificity of signaling enzymes, it is of immense importance to understand how the cell achieves efficiency and accuracy in signaling. It is generally accepted that the primary mechanism is to promote the proximity of signaling enzymes by specific binding to a special class of regulatory proteins [1–4]. These proteins come under a variety of names, such as scaffold, adaptor, anchor and docking, but invariably function by ‘enforced proximity’, i.e. by binding at least two signaling enzymes together and directing, coordinating and regulating their action (Fig. 1). In Keywords adaptor; anchor; docking protein; kinase; multidomain protein; phosphatase; protein interaction; scaffold; signaling; structural disorder Correspondence L. Buday and P. Tompa, Institute of Enzymology, Hungarian Academy of Sciences, 1518 Budapest, P.O. Box 7, Hungary Fax: +361 466 5465 Tel: +361 279 3115; +361 279 3143 E-mail: buday@enzim.hu; tompa@enzim.hu (Received 14 May 2010, revised 3 August 2010, accepted 18 August 2010) doi:10.1111/j.1742-4658.2010.07864.x In this series of four minireviews the field of scaffold proteins and proteins of similar molecular ⁄ cellular functions is overviewed. By binding and bring- ing into proximity two or more signaling proteins, these proteins direct the flow of information in the cell by activating, coordinating and regulating signaling events in regulatory networks. Here we discuss the categories of scaffolds, anchors, docking proteins and adaptors in some detail, and using many examples we demonstrate that they cover a wide range of functional modes that appear to segregate into three practical categories, simple pro- teins binding two partners together (adaptors), larger multidomain proteins targeting and regulating more proteins in complex ways (scaffold ⁄ anchor- ing proteins) and proteins specialized to initiate signaling cascades by local- izing partners at the cell membrane (docking proteins). It will also be shown, however, that the categories partially overlap and often their names are used interchangeably in the literature. In addition, although not usually considered as scaffolds, several other proteins, such as regulatory proteins with catalytic activity, phosphatase targeting subunits, E3 ubiquitin ligases, ESCRT proteins in endosomal sorting and DNA damage sensors also func- tion by bona fide scaffolding mechanisms. Thus, the field is in a state of continuous advance and expansion, which demands that the classification scheme be regularly updated and, if needed, revised. Abbreviations AKAP, A-kinase anchoring protein; DAPP, dual-adapter for phosphotyrosine and 3-phosphoinositides; DD, death domain; FADD, Fas-associated protein with death domain; MAGUK, membrane-associated guanylate kinase; MYPT1, myosin phosphatase targeting subunit 1; Nck, non-catalytic region of tyrosine kinase adaptor protein; PDZ, post-synaptic density, disc large, zo-1 protein; PH, pleckstrin homology; PSD, post-synaptic density; RTK, receptor tyrosine kinase; SH2, Src homology 2; SH3, Src homology 3; SKAP, Src kinase-associated phosphoprotein. 4348 FEBS Journal 277 (2010) 4348–4355 ª 2010 The Authors Journal compilation ª 2010 FEBS the literature, the names are often used interchange- ably, which indicates that the functions of the proteins in the four categories overlap significantly and they are often difficult to distinguished. This article, and the following three in this minireview series provide a con- cise functional description of scaffold proteins and their kin, in general, and their distinct classes, in par- ticular, encompassing scaffolds [5], docking proteins [6], anchors [7] and adaptors (this article). We also point out overlaps and inconsistencies in terminology, to show that members of this entire functional class occupy positions along a rather contin- uous functional spectrum. We also describe targeting subunits of phosphatases, which have a function in close analogy with that of anchors. In addition, we also show that there are many other proteins that function using scaffolding mechanisms, and ought to be considered in extending the classification scheme of these regulatory proteins. It will be addressed that scaffold proteins and their kin not only bring signaling proteins together, but also regulate the flow of signal- ing information using a variety of mechanisms, owing to which proteins in this broad functional class no longer emerge as passive connectors, but active regula- tory elements shaping the response put out by signal- ing networks. We suggest that this active role may also be linked with extended structural disorder of the proteins [8–10], which enables dynamic communication between their distinct binding and functional elements (Fig. 1 and Table 1). It should also be made clear that because of space limitations this minireview is not comprehensive, and the reader is directed to many excellent recent reviews [1–4,11], and the other minireviews in this series [5–7] for further details and insight into this field of continu- ous advance. Scaffold/anchoring proteins Scaffold and anchoring proteins are reviewed in two of the other minireviews in this series [5,7]. Here we pro- vide a short overview of their most notable examples, to demonstrate that the distinction between these two categories is historical, rather than structural or functional. Scaffold proteins may be considered as the founders of this functional class, epitomizing the very essence of the action of signaling regulatory proteins [5]. They are defined as proteins organizing signaling complexes by binding at least two signaling enzymes together and promoting their communication by proximity. Classi- cally, they have been regarded as passive platforms for the assembly of signalosomes, but more recently it has become clear that they play rather active regulatory roles (Fig. 1 and Table 1). They may operate by four basic mechanisms, such as enforced proximity, combi- natorial use of elements, dynamic regulation and con- formational fine-tuning [4,5]. As demonstrated by one of the best characterized scaffolds, Ste5, which regu- lates the extracellular signal-regulated kinase ⁄ mitogen- activated protein kinase pathway in yeast, their func- tion may be to assemble complexes, enforce restricted intracellular localization, provide allosteric feedback and feed-forward regulation, and offer protection against degradation [3]. They may enable context- dependent fine-tuning of pre-existing signaling A B Adaptor A Docking P P P P C A B C Scaffold/anchor B Fig. 1. Mechanisms of scaffold proteins and their kin. This scheme outlines the most important aspects of the function of modular regu- latory proteins in the four (three) closely related categories. In all cat- egories, the proteins regulate signaling pathways by binding several of the components and targeting ⁄ regulating their action in complex ways (regulatory interactions ⁄ modifications marked by arrows). Within this generalized scheme, there are three distinct types of behavior. (A) Adaptors are usually small, and have two binding regions to target the action of two bound enzymes. (B) Scaf- folds ⁄ anchors are large multidomain proteins with a lot of structural disorder, able to bind and regulate several members of a signaling pathway. (C) Docking proteins have a similar structural and functional outline, their distinguishing feature being their ability to localize at the membrane next to an activating receptor, to which they bind in a phosphorylation-dependent manner. In the entire class of proteins, a high level of structural disorder (43.3% on average; Table 1) enables key functional attributes, such as binding of several partners, to be combined. A, B and C marks general proteins. P, phosphate. L. Buday and P. Tompa Functional classification of scaffold proteins FEBS Journal 277 (2010) 4348–4355 ª 2010 The Authors Journal compilation ª 2010 FEBS 4349 pathways or create new pathways using novel combi- nations of signaling elements. Scaffolds are extremely heterogeneous in structure and function, which may significantly overlap with those of other classes. A few selected examples (Table 1; cf. [5]) show that they often have separated binding domains for protein–pro- tein interactions, and also have a high level of struc- tural disorder, which may provide both short binding regions and flexibility for dynamic regulatory rear- rangements. These principles are also apparent in scaffolds that operate in neuronal and immune cells, which are highly polarized and form complex structures termed synapses. Synapses contain a plethora of receptors, ion channels and signaling proteins for sensing and pro- cessing extracellular signals, organized and connected to the cytoskeleton to form large complexes, such as the post-synaptic density (PSD) complex. The best known of the scaffolds are the membrane-associated guanylate kinase (MAGUK) proteins, which have vari- ous numbers of post-synaptic density, disc large, zo-1 protein (PDZ) domains, an Src homology (SH)3 domain and a C-terminal guanylate kinase domain [1]. The best studied MAGUK in T cells is DLG1, whereas in neurons it is PSD95, with notable similari- ties in function. PSD95 is the most abundant scaffold in the PSD, it has three PDZ domains, and binds a variety of receptors ⁄ channels, cytoskeletal and Table 1. Scaffold proteins and their kin. Representatives of the four categories of scaffold proteins and their relatives are presented. The length of the human (unless indicated otherwise) isoform is given, along with its typical domains, binding partners and percent structural disorder, as predicted by the IUPred algorithm [42]. Protein–protein interaction domains and interaction partners have been taken from the literature [1–7,22] and the list is not intended to be exhaustive. AKAP, A-kinase anchoring protein; Caskin 1, CASK-interacting protein 1; DED, death effector domain; DLG(1), discs, large homolog (1); Dok1, docking protein 1; ERK, extracellular signal-regulated kinase; FADD, Fas-associated protein with death domain; FRS, FGF receptor substrate; Gab,Grb2-associated binder; Grb2, growth factor receptor-bound protein 2; IRS, insulin receptor substrate; KSR, kinase suppressor of Ras; MYPT1, myosin phosphatase targeting subunit 1; Nck1, non-cata- lytic region of tyrosine kinase adaptor protein 1; PDZ, post-synaptic density, disc large, zo-1 protein (domain); PH, pleckstrin homology domain; PI3-kinase, phosphoinositide-3-kinase; PKA, protein kinase A; PKC, protein kinase C; PP1, protein phosphatase 1; PSD, post-synaptic density; PTB, phosphotyrosine binding domain; RACK, receptor for activated C kinase; SARA, SMAD-anchor for receptor activation; SH2, Src homology 2 domain; SH3, Src homology 3 domain; Ste5, sterile 5; TGFb, transforming growth factor beta. Protein Length (number of amino acids) Domain(s) Protein partners Disorder (%) Scaffold ste5 (yeast) 917 PHD, PH Ste4, Ste7, Ste20, Ste11, FUS3 275 (30.0) KSR 921 Protein kinase, zinc finger, Serpin Raf, MEK1, ERK, 14-3-3, PP2A 321 (34.9) PSD95 724 PDZ, SH3, GK glutamate (NMDA) receptor, K+ channels, nNOS, NGL-2, SALM2, ADAM22, SYNGAP 95 (13.1) DLG1 904 PDZ, SH3, GK, L27 Ezrin, CASK, PTEN 267 (29.5) Shank1 2161 ANK repeats, PDZ, SAM, SH3 IRSp53, Sharpin, Abp1, Homer 1515 (70.1) Caskin1 1431 ANK repeats, SAM, SH3 CASK, Nck, Abi2 996 (69.6) SARA 1425 FYVE-type zinc finger SMAD2, SMAD3, TGF-beta receptor 441 (30.9) Anchoring AKAP150, rat 714 – PKA, PKC, SAP97, glutamate (AMPA) receptor 611 (85.6) mAKAP 2319 Spectrin repeats PKA, Epac1, Rac1, ERK5 1091 (47.0) RACK1 317 WD repeats PKC, Src, integrin, dynamin1 2 (0.0) MYPT1 1030 Ankirin repeats PP1CA, ROCK1, Ezrin, Merlin, Synaptophysin 722 (70.1) Docking IRS1 1242 PH, PTB PI-P 3 , Grb2, Nck, PI 3-kinase, SHP2 971 (78.2) FRS2 508 PTB FGFR, NGFR, Shc1, Grb2, PTPN11 372 (73.2) Gab1 694 PH, MBD Grb2, c-Met, SHP2, PI 3-kinase, PLCc 544 (78.4) Dok1 481 PH, PTB Abl, Bcr-Abl, p120Ras-GAP 207 (43.0) Adaptor Grb2 217 SH2, SH3 EGFR, LAT, c-Met, Cbl, Dynamin, N-WASP, SAM68, SLP-76, Sos, Synapsin, Vav 0 (0) Nck1 377 SH2, SH3 Abl, Cbl, Dynamin, FAK, NAK, NAP1, N-WASP, PAK, PRK2, Sos 48 (12.7) FADD 208 DD, DED Fas receptor, TNFR-1, procaspase-8, procaspase-10 28 (13.5) Functional classification of scaffold proteins L. Buday and P. Tompa 4350 FEBS Journal 277 (2010) 4348–4355 ª 2010 The Authors Journal compilation ª 2010 FEBS signaling proteins (Table 1). Because of its multiplicity of binding partners, PSD95 functions in receptor clustering, targeting receptor action on Ras and Rho signaling, and regulating receptor modification under- scoring dynamics of synaptic function. A similar logic applies to DLG1 (synapse-associated protein-97), which regulates immunological synapses by physically linking T-cell receptor signaling to cytoskeletal rear- rangements. Table 1 lists a few further scaffolds, all of which are modular with several protein–protein interaction domains and motifs embedded in and ⁄ or connected by long disordered regions, such as Shank1 and Caskin1 in neuronal PSD or SARA in transform- ing growth factor beta signaling. Scaffolds epitomize the very essence of the entire family of signaling proteins, first appreciated in tyro- sine kinase signaling where adaptors, scaffolds or enzymes are recruited to the autophosphorylated receptor tyrosine kinases (RTKs). Cytosolic kinases and phosphatases, however, also require strict control of their activity and subcellular localization, which is achieved through anchoring proteins, originally defined as proteins targeting the action of protein kinase A (A-kinase anchoring proteins; AKAPs), on specific substrates [7]. They are usually long proteins with a lot of structural disorder (Fig. 1 and Table 1), with func- tional attributes difficult to clearly distinguish from that of scaffolds (described in detail in ref. [7]). In addition to AKAPs, protein kinase C also interacts with a family of anchor proteins called receptors for activated C kinase. Cytosolic Ser ⁄ Thr phosphatases are also directed into signaling networks by mechanisms similar to that of protein kinase A and protein kinase C. For exam- ple, functional protein phosphatase 1 consists of a cat- alytic subunit and a regulatory subunit. The regulatory subunits target the catalytic subunit to specific cellular compartments and modulate substrate specificity [12]. One of the best characterized regulatory subunits is the myosin phosphatase target subunit (MYPT1). A num- ber of regulatory proteins, as well as Ser ⁄ Thr kinases, can associate with MYPT1, suggesting that similarly to AKAPs and receptors for activated C kinase, the function of MYPT1 is much broader than simply tar- geting protein phosphatase 1 to myosin II [13]. Docking proteins As outlined in the accompanying minireview by Brum- mer et al. [6], docking proteins were originally defined as accessory proteins in RTK signaling with a mem- brane-targeting region, a protein–protein interaction site for receptor binding and an extended region with several Tyr residues for receptor-dependent phosphory- lation (Fig. 1 and Table 1). This outline applies to all four major families of classical docking protein, such as the Grb2-associated binder ⁄ daughter of sevenless (Gab ⁄ DOS), insulin receptor substrate, FGF receptor substrate and docking protein families [6]. Docking proteins are recruited to the site of RTK activation at the plasma membrane, they reinforce binding by virtue of a receptor-binding domain (often a phosphotyro- sine-binding domain) and undergo multiple Tyr-phos- phorylation (there are more than five Tyr phosphorylation sites in at least one of the family members). Phosphorylation of Tyr residues is rather specific to certain RTKs, but may also proceed by cytoplasmic tyrosine kinases. Tyr-phosphorylated docking proteins recruit SH2-domain-containing sig- naling components to initiate specific signal cascades, and they coordinate and regulate Tyr kinase signaling events, and also display dynamic regulatory phenom- ena described in more detail for scaffold proteins (for details see ref. [6]). There are also other, atypical docking proteins [6], which lack the lipid-binding domain but have N-termi- nal domains ⁄ regions that help them localize at the plasma membrane next to activating receptors, and also contain several Tyr residues that undergo phos- phorylation and recruitment of signaling proteins. Although the function of these proteins (linker for activated T cells, Crk-associated substrate, SLP65) is closely related to other docking proteins, they appear to be better classified as scaffolds. Adaptor proteins The term adaptor protein is generally used for low molecular mass molecules that serve to link two func- tional members of a catalytic pathway (Fig. 1 and Table 1). Adaptors either possess two domains involved in protein–protein interactions or use two regions com- posed of two to three domains. The first group of adap- tor proteins identified was the family of SH2 ⁄ SH3 domain-containing proteins, including growth factor receptor-bound protein 2, Crk, CrkL and non-catalytic region of tyrosine kinase adaptor protein 1 (Nck) [14]. Their SH2 domain binds specific phosphotyrosine resi- dues on activated receptors or their substrates, whereas their SH3 domains bind proline-rich motifs on down- stream target proteins. Interestingly, this family of adaptor proteins contains only one SH2 domain, whereas they usually possess two or more SH3 domains. In theory, more than one SH3 domains may allow the adaptor to recruit several ligands separately; however, it seems from earlier studies that cooperation L. Buday and P. Tompa Functional classification of scaffold proteins FEBS Journal 277 (2010) 4348–4355 ª 2010 The Authors Journal compilation ª 2010 FEBS 4351 exists between the SH3 domains for ligand binding. For example, Nck adaptor contains one SH2 domain and three SH3 domains. Although individual SH3 domains of Nck were reported to be able to bind partners, such as p21 protein (Cdc42 ⁄ Rac)-activated kinase or PRK2, experimental data clearly showed that Nck constructs containing all three SH3 domains bind the protein part- ners with much higher affinity than the single SH3 domains [15,16]. Therefore, it is highly likely that even those adaptors which contain two or three tandem SH3 domains actually link only two members of a catalytic pathway. SH2 ⁄ SH3 domain-containing adaptors link down- stream target molecules to the membrane-bound recep- tor. In some cases, the adaptor may recruit binding partners directly to the plasma membrane through its lipid-binding domain. In hematopoietic cells, pleckstrin homology domain (PH)-containing adaptor molecules provide important links between phosphoinositide- 3 kinase and lymphocyte function. For example, recruitment of Src kinase-associated phosphoprotein (SKAP) and B-lymphocyte adapter molecule of 32 kDa (Bam32) ⁄ dual-adapter for phosphotyrosine and 3-phos- phoinositides (DAPP) to the plasma membrane of acti- vated lymphocytes is driven by lipid products generated through the action of phosphoinositide-3 kinase [17]. Whereas SKAP contains a PH domain on the N-termi- nus and an SH3 domain on the C-terminus, Bam32 ⁄ DAPP1 possesses an SH2 domain and a PH domain. Therefore, although both adaptors could bind phospho- inositide in the plasma membrane, SKAP recruits pro- line-rich target molecules, whereas Bam32 ⁄ DAPP1 may associate with phosphotyrosine-containing proteins. Possessing a PH domain for membrane association, this subfamily of adaptors, including SKAP, Bam32 and SH2B proteins, performs a bona fide adaptor function, they nevertheless link downstream targets directly to the plasma membrane DAPP [17]. The majority of adaptor molecules are implicated in RTK signaling, however, other cell-surface receptors also utilize adaptors possessing specific modular domains. Death receptors, members of the tumor necrosis factor receptor superfamily, possess a cytoplas- mic death domain (DD). They transmit signals through apical protein complexes, which are nucleated by the death domain adaptors Fas-associated protein with death domain (FADD) and tumor necrosis factor receptor type 1-associated death domain protein (TRADD). FADD is a protein containing two structur- ally similar protein motifs, the N-terminal death effec- tor domain and the C-terminal DD. The primary role of FADD in death receptor signaling is to recruit initia- tor procaspase 8 and procaspase 10 to death receptors [18]. The principle by which the SH2 ⁄ SH3 domain-con- taining adaptors and FADD function is very similar, and the modular domains of growth factor receptor- bound protein 2 and FADD are commutable. This was demonstrated by creating an artificial signaling pathway in which the SH2 domain of growth factor receptor- bound protein 2 was fused to the death effector domain of FADD. The chimeric adaptor protein could reroute RTK signals to induce procaspase activation and cell death [19]. Classification of scaffolds and its limitations The functions of the above four closely related catego- ries of signaling regulatory proteins show significant similarities and overlap (Fig. 1 and Table 1), but appear to segregate into three broad functional catego- ries: simple proteins binding two partners together (adaptors), larger multidomain proteins targeting and regulating more proteins in complex ways (scaf- folds ⁄ anchoring proteins) and proteins specialized to initiate signaling cascades by localizing partners at the cell membrane (docking proteins). This classification has both advantages and limita- tions. Its major advantage is simplification, because it enables one to think in terms of a few concepts instead of a practically unmanageable number of individual observations. This very fact, however, also makes it inherently limited, because of the necessary neglect of many details. If such exceptions to the ‘rule’ of the system become too numerous, the system has to be improved ⁄ refined to comply with progress in the field. Here we list a few such notable exceptions that may eventually demand the extension of the classification scheme. A diagnostic mark of the limitations inherent in the scheme is that very often the names are used inter- changeably, without much attempt to clarify where the given protein belongs to. For example, anchoring pro- teins are often also termed scaffolds [4,7], the term scaf- folds and adaptors are sometimes used interchangeably [1,4], the distinction between docking and adaptor is somewhat arbitrary [6], and docking proteins may also be termed adaptors and scaffolds [2]. There are many individual proteins called by different names in differ- ent articles, such as MyD88 scaffold and adaptor [3], linker for activated T cells (LAT) scaffold [3] and dock- ing protein [6], GAB docking protein [6] and scaffold [4], mAKAP anchor protein [7] and scaffold [4], MAG- UK proteins, which ‘anchor’ receptors at the synapse, scaffold [1] and adaptor [20], FGF receptor substrate docking protein [6] and ‘scaffold adaptor’ [21]. These Functional classification of scaffold proteins L. Buday and P. Tompa 4352 FEBS Journal 277 (2010) 4348–4355 ª 2010 The Authors Journal compilation ª 2010 FEBS and many other examples in the literature demonstrate that the terms cannot be unequivocally separated. A different type of limitation is the exclusion of pro- teins with enzymatic activity. Although it serves the simple purpose of separating two basic functions in signal transduction – enzymes that act and proteins that orchestrate their action [22] – it is clear that it cannot be done for many relevant proteins that have both enzymatic domains and regulatory protein–pro- tein interaction domains. In effect, there are several proteins considered as scaffolds that have enzymatic activity, such as MAGUK proteins already discussed among scaffolds [3], RNAse E [23], MEKK1 of the JNK signaling pathway [24], RTKs themselves [3] or integrin-linked kinase [25]. Actually, there is little func- tional distinction between a scaffold protein binding a given enzyme partner and a scaffold that has an enzy- matic domain of the same type. An additional complicating factor is that there are many other signaling pathways beyond kinase cascades, which are rather neglected in this context. For example, transfer of the small protein ubiquitin proceeds via a cascade, from E1 ubiquitin-activating enzyme to E2 ubiquitin-conjugating enzyme to a specific substrate with the intervention of a targeting-type of interaction mediated by E3 ubiquitin ligases. Because ubiquitinated target proteins may be targeted for degradation or sig- naling activation [26], the operation of E3 proteins, such as MDM2 [27] and Siah-2 [28], can be best ratio- nalized by the scaffold concept. The assembly of Esc- herichia coli RNA degradosome by RNAse E [23] is also of very similar molecular logic. It has been sug- gested that there are also adaptors that are important in pathways activated by internal signals, such as DNA damage, e.g. MDC1 in DNA doule-strand breaks [20] and BRCA1 in many cellular pathways including DNA repair [29]. In a similar vein, trafficking of signaling proteins is itself dependent on adaptors associated with protein sorting in endosomes, such as ESCRT [30]. In addition, both proteins and also metabolites can be scaffolded, as demonstrated by Pmel17, a physiological amyloid that ‘scaffolds’ and ‘sequesters’ toxic interme- diates during the biosynthesis of the pigment melanin in melanocytes in the skin [31]. Of further note, many fully or largely disordered pro- teins noted for their assembly function (see below) are not usually considered to be scaffold proteins, although they do bind and orchestrate the action of several sig- naling partners, such as caldesmon (Ca 2+ ⁄ calmodulin, F-actin, myosin, tropomyosin), SIBLING proteins (integrin, complement factor H, CD44, fibronectin), or RNAPII CTD (capping, splicing and polyadenylation factors). The molecular mechanism and function of these proteins also comply with the scaffolding princi- ples outlined in this article. Scaffold proteins and structural disorder The decision on what is included among scaffolds can also be approached from a structural point of view, because in all categories – with the possible exception of adaptors – the proteins have a very high level of func- tion-related structural disorder (Table 1). It has recently been recognized that a significant proportion of eukary- otic genomes encodes for proteins (IDPs) or regions of proteins (IDRs) that lack a well-defined 3D structure under native, functional conditions [10,32–34]. Struc- tural disorder abounds in proteins of regulatory and sig- naling function, and it is also closely correlated with disease, such as cancer and neurodegenerative disorders. The molecular function of IDPs ⁄ IDRs may stem either from recognition of partner molecules via short motifs [35,36] or disordered domains [37], from regulatory post-translational modification and also from providing ‘entropic-chain’ functions, such as linkers and segments contributing entropic exclusion and force generation. As a result, the molecular function of scaffolds corre- sponds to the ‘assembler’ function of IDPs, i.e. they have been described to assemble complexes [9,34,37]. The role of structural disorder in scaffold-type func- tions is also apparent in hub proteins, which have been found to have a large number of binding partners in high-throughput studies of protein–protein interactions in the interactome [38]. In particular, ‘party’ hubs have been defined as those being able to bind their partners simultaneously, which is the very essence of the action of scaffolds. Hub proteins have an elevated level of dis- order [39], which is also the case with the examples cited here (average disorder, 43.3%; Table 1), and also previ- ous findings that scaffold proteins constitute one of the most disordered functional categories [8,40] and the average disorder correlates with the number of subunits of multiprotein complexes [41]. In all, structural disor- der seems to be closely associated with several attributes of scaffold function, such as the ability of binding multi- ple partners, mediating their complex and transient interactions and themselves undergoing a complex array of regulatory post-translational modifications (Fig. 1). Conclusion: where is the field of scaffolds headed? From all the considerations described, it seems that a useful practical functional classification of scaffold proteins and their kin can be given. At first sight, all L. Buday and P. Tompa Functional classification of scaffold proteins FEBS Journal 277 (2010) 4348–4355 ª 2010 The Authors Journal compilation ª 2010 FEBS 4353 the proteins enlisted represent variations on a common theme, binding signaling proteins together to direct and control the flow of information in the cell. This basic theme segregates in a rather consistent manner into three coherent categories. Scaffold ⁄ anchor pro- teins usually bind more than two signaling components together and regulate their activity in complex and dynamic ways, involving activation and repression of activity. Adaptor proteins are usually smaller and their function is simpler, connecting two partners together. Docking proteins distinguish themselves by assembling signaling complexes at the plasma membrane in a Tyr- phosphorylation-dependent way. There are many pro- teins excluded from this scheme, although they do act via very closely related mechanisms. Their inclusion following careful judgment of their functional modes, possibly leading to an extension of the classification scheme, should be considered. Acknowledgements This work is supported by grants OTKA K60694 and NK71582 from the Hungarian Scientific Research Fund and ETT 245 ⁄ 2006 from the Hungarian Ministry of Health (for PT), the Miha ´ ly Pola ´ nyi Program (Agency for Research Fund Management and Research Exploitation, KPI) and a ‘Lendu ¨ let’ grant from the Hungarian Academy of Sciences (for LB). References 1 Feng W & Zhang M (2009) Organiztaion and dynamics of PDZ-domain-related supramodules in the postsynap- tic density. Nat Rev Neurosci 10, 87–99. 2 Pawson T & Scott JD (1997) Signaling through scaffold, anchoring, and adaptor proteins. Science 278, 2075–2080. 3 Shaw AS & Filbert EL (2009) Scaffold proteins and immune-cell signalling. Nat Rev Immunol 9, 47–56. 4 Zeke A, Lukacs M, Lim WA & Remenyi A (2009) Scaffolds: interaction platforms for cellular signalling circuits. Trends Cell Biol 19, 364–374. 5 Alexa A, Varga J & Reme ´ nyi A (2010) Scaffolds are ‘active’ regulators of signaling modules. FEBS J 277, 4376–4382. 6 Brummer T, Schmitz-Peiffer C & Daly RJ (2010) Dock- ing proteins. FEBS J 277, 4356–4369. 7 Logue JS & Scott JD (2010) Organizing signal transduc- tion through A-kinase anchoring proteins (AKAPs). FEBS J 277, 4370–4375. 8 Balazs A, Csizmok V, Buday L, Rakacs M, Kiss R, Bokor M, Udupa R, Tompa K & Tompa P (2009) High levels of structural disorder in scaffold proteins as exemplified by a novel neuronal protein, CASK-interac- tive protein 1. FEBS J 276, 3744–3756. 9 Tompa P (2005) The interplay between structure and function in intrinsically unstructured proteins. FEBS Lett 579, 3346–3354. 10 Tompa P (2009) Structure and Function of Intrinsically Disordered Proteins . CRC Press, Boca Raton, FL. 11 Ramirez F & Albrecht M (2009) Finding scaffold proteins in interactomes. Trends Cell Biol 20, 2–4. 12 Shi Y (2009) Assembly and structure of protein phosphatase 2A. Sci China C Life Sci 52, 135–146. 13 Matsumura F & Hartshorne DJ (2008) Myosin phosphatase target subunit: many roles in cell function. Biochem Biophys Res Commun 369, 149–156. 14 Buday L (1999) Membrane-targeting of signalling molecules by SH2 ⁄ SH3 domain-containing adaptor proteins. Biochim Biophys Acta 1422, 187–204. 15 Buday L, Wunderlich L & Tamas P (2002) The Nck family of adapter proteins: regulators of actin cytoskeleton. Cell Signal 14 , 723–731. 16 Wunderlich L, Goher A, Farago A, Downward J & Buday L (1999) Requirement of multiple SH3 domains of Nck for ligand binding. Cell Signal 11, 253–262. 17 Zhang TT, Li H, Cheung SM, Costantini JL, Hou S, Al-Alwan M & Marshall AJ (2009) Phosphoinosi- tide 3-kinase-regulated adapters in lymphocyte activation. Immunol Rev 232, 255–272. 18 Wilson TR, Redmond KM, McLaughlin KM, Crawford N, Gately K, O’Byrne K, Le-Clorrenec C, Holohan C, Fennell DA, Johnston PG et al. (2009) Procaspase 8 overexpression in non-small-cell lung cancer promotes apoptosis induced by FLIP silencing. Cell Death Differ 16 , 1352–1361. 19 Howard PL, Chia MC, Del Rizzo S, Liu FF & Pawson T (2003) Redirecting tyrosine kinase signaling to an apoptotic caspase pathway through chimeric adaptor proteins. Proc Natl Acad Sci USA 100, 11267–11272. 20 Pawson T (2007) Dynamic control of signaling by modular adaptor proteins. Curr Opin Cell Biol 19, 112–116. 21 Gotoh N (2008) Regulation of growth factor signaling by FRS2 family docking ⁄ scaffold adaptor proteins. Cancer Sci 99, 1319–1325. 22 Bhattacharyya RP, Remenyi A, Yeh BJ & Lim WA (2006) Domains, motifs, and scaffolds: the role of modular interactions in the evolution and wiring of cell signaling circuits. Annu Rev Biochem 75, 655–680. 23 Worrall JA, Gorna M, Crump NT, Phillips LG, Tuck AC, Price AJ, Bavro VN & Luisi BF (2008) Reconstitu- tion and analysis of the multienzyme Escherichia coli RNA degradosome. J Mol Biol 382, 870–883. 24 Su YC, Han J, Xu S, Cobb M & Skolnik EY (1997) NIK is a new Ste20-related kinase that binds NCK and Functional classification of scaffold proteins L. Buday and P. Tompa 4354 FEBS Journal 277 (2010) 4348–4355 ª 2010 The Authors Journal compilation ª 2010 FEBS MEKK1 and activates the SAPK ⁄ JNK cascade via a conserved regulatory domain. EMBO J 16 , 1279–1290. 25 Legate KR, Montanez E, Kudlacek O & Fassler R (2006) ILK, PINCH and parvin: the tIPP of integrin signalling. Nat Rev Mol Cell Biol 7, 20–31. 26 Conaway RC, Brower CS & Conaway JW (2002) Emerging roles of ubiquitin in transcription regulation. Science 296, 1254–1258. 27 Iwakuma T & Lozano G (2003) MDM2, an introduc- tion. Mol Cancer Res 1, 993–1000. 28 Habelhah H, Frew IJ, Laine A, Janes PW, Relaix F, Sassoon D, Bowtell DD & Ronai Z (2002) Stress- induced decrease in TRAF2 stability is mediated by Siah2. EMBO J 21, 5756–5765. 29 Mark WY, Liao JC, Lu Y, Ayed A, Laister R, Szymczyna B, Chakrabartty A & Arrowsmith CH (2005) Characterization of segments from the central region of BRCA1: an intrinsically disordered scaffold for multiple protein-protein and protein-DNA interactions? J Mol Biol 345, 275–287. 30 Wollert T & Hurley JH (2010) Molecular mechanism of multivesicular body biogenesis by ESCRT complexes. Nature 464, 864–869. 31 Fowler DM, Koulov AV, Balch WE & Kelly JW (2007) Functional amyloid – from bacteria to humans. Trends Biochem Sci 32, 217–224. 32 Dunker AK, Brown CJ, Lawson JD, Iakoucheva LM & Obradovic Z (2002) Intrinsic disorder and protein function. Biochemistry 41, 6573–6582. 33 Dyson HJ & Wright PE (2005) Intrinsically unstruc- tured proteins and their functions. Nat Rev Mol Cell Biol 6, 197–208. 34 Tompa P (2002) Intrinsically unstructured proteins. Trends Biochem Sci 27, 527–533. 35 Fuxreiter M, Simon I, Friedrich P & Tompa P (2004) Preformed structural elements feature in partner recog- nition by intrinsically unstructured proteins. J Mol Biol 338, 1015–1026. 36 Fuxreiter M, Tompa P & Simon I (2007) Local structural disorder imparts plasticity on linear motifs. Bioinformatics 23, 950–956. 37 Tompa P, Fuxreiter M, Oldfield CJ, Simon I, Dunker AK & Uversky VN (2009) Close encounters of the third kind: disordered domains and the interactions of proteins. Bioessays 31, 328–335. 38 Han JD, Bertin N, Hao T, Goldberg DS, Berriz GF, Zhang LV, Dupuy D, Walhout AJ, Cusick ME, Roth FP et al. (2004) Evidence for dynamically organized modularity in the yeast protein-protein interaction network. Nature 430, 88–93. 39 Dosztanyi Z, Chen J, Dunker AK, Simon I & Tompa P (2006) Disorder and sequence repeats in hub proteins and their implications for network evolution. J Prote- ome Res 5, 2985–2995. 40 Cortese MS, Uversky VN & Keith Dunker A (2008) Intrinsic disorder in scaffold proteins: getting more from less. Prog Biophys Mol Biol 98, 85–106. 41 Hegyi H, Schad E & Tompa P (2007) Structural disorder promotes assembly of protein complexes. BMC Struct Biol 7, 65. 42 Dosztanyi Z, Csizmok V, Tompa P & Simon I (2005) IUPred: web server for the prediction of intrinsically unstructured regions of proteins based on estimated energy content. Bioinformatics 21, 3433–3434. L. Buday and P. Tompa Functional classification of scaffold proteins FEBS Journal 277 (2010) 4348–4355 ª 2010 The Authors Journal compilation ª 2010 FEBS 4355 . MINIREVIEW Functional classification of scaffold proteins and related molecules La ´ szlo ´ Buday 1,2 and Pe ´ ter Tompa 1 1 Institute of Enzymology,. this series of four minireviews the field of scaffold proteins and proteins of similar molecular ⁄ cellular functions is overviewed. By binding and bring- ing