A broad view of scaffolding suggests that scaffolding proteins can actively control regulation and signaling of multienzyme complexes through allostery

Biochimica et Biophysica Acta 1834 (2013) 820–829

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbapap

Review

A broad view of scaffolding suggests that scaffolding proteins can actively control regulation and signaling of multienzyme complexes through allostery☆ Ruth Nussinov a, b,⁎, Buyong Ma a, Chung-Jung Tsai a a Basic Research Program, SAIC-Frederick, Inc., Center for Cancer Research Nanobiology Program, Frederick National Laboratory for Cancer Research, National Cancer Institute, Frederick, MD 21702, USA b Sackler Inst. of Molecular Medicine, Department of Human Genetics and Molecular Medicine, Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel

a r t i c l e

i n f o

Article history: Received 16 October 2012 Received in revised form 19 December 2012 Accepted 21 December 2012 Available online 3 January 2013 Keywords: Multi-enzyme complex Allosteric Multiprotein Multi-protein Population shift Conformational selection

a b s t r a c t Enzymes often work sequentially in pathways; and consecutive reaction steps are typically carried out by molecules associated in the same multienzyme complex. Localization confines the enzymes; anchors them; increases the effective concentration of substrates and products; and shortens pathway timescales; however, it does not explain enzyme coordination or pathway branching. Here, we distinguish between metabolic and signaling multienzyme complexes. We argue for a central role of scaffolding proteins in regulating multienzyme complexes signaling and suggest that metabolic multienzyme complexes are less dependent on scaffolding because they undergo conformational control through direct subunit–subunit contacts. In particular, we propose that scaffolding proteins have an essential function in controlling branching in signaling pathways. This new broadened definition of scaffolding proteins goes beyond cases such as the classic yeast mitogen-activated protein kinase Ste5 and encompasses proteins such as E3 ligases which lack active sites and work via allostery. With this definition, we classify the mechanisms of multienzyme complexes based on whether the substrates are transferred through the involvement of scaffolding proteins, and outline the functional merits to metabolic or signaling pathways. Overall, while co-localization topography helps multistep pathways non-specifically, allosteric regulation requires precise multienzyme organization and interactions and works via population shift, either through direct enzyme subunit–subunit interactions or through active involvement of scaffolding proteins. This article is part of a Special Issue entitled: The emerging dynamic view of proteins: Protein plasticity in allostery, evolution and self-assembly. © 2013 Elsevier B.V. All rights reserved.

1. Introduction How are enzymatic actions in signaling pathways, and in metabolism, coordinated and controlled? Here, we focus on the important role of allostery in controlling multienzyme complexes and contrast allosteric mechanisms to other ways of regulation, such as subcellular co-compartmentalization topography. Our central thesis is that scaffolding proteins are central to the regulation of signaling multienzyme complexes. In contrast, metabolic multienzyme complexes are less dependent on scaffolding because they undergo more direct conformational control through subunit–subunit contacts. We assign scaffolding proteins a much more active, fine-tuning role than considered to date, one that actively involves conformational regulation

☆ This article is part of a Special Issue entitled: The emerging dynamic view of proteins: Protein plasticity in allostery, evolution and self-assembly. ⁎ Corresponding author at: Basic Research Program, SAIC-Frederick, Inc., Center for Cancer Research Nanobiology Program, Frederick National Laboratory for Cancer Research, National Cancer Institute, Frederick, MD 21702, USA. Tel.: +1 301 846 5579. E-mail address: [email protected] (R. Nussinov). 1570-9639/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bbapap.2012.12.014

through allostery. The yeast mitogen-activated protein kinase Ste5 can be viewed as a classic scaffold protein. Our view of scaffolding proteins is broader. We adopt a functional definition of scaffolding that also involves proteins such as E3 ligases, which have not been considered as scaffolding to date; but which work via allostery to facilitate the ubiquitin transfer reaction from the E2 enzyme to the substrate protein. Further, scaffolding proteins not only bias the conformational ensembles locally; via the cytoskeleton network, it can bias the ensembles across the cell. This new viewpoint is important because it designates scaffolding proteins as active key players in signaling pathways where multienzyme complexes invariably control pathway switching, often via post-translational modifications. 2. An overview: co-localization cannot provide effective answers to multienzyme regulation In large part, enzymes do not function autonomously [1,2]; instead they are integral elements of signaling or biochemical (metabolic) pathways, where a product of one reaction can serve as the substrate or

R. Nussinov et al. / Biochimica et Biophysica Acta 1834 (2013) 820–829

precursor for the next, with the end products achieved through multiple chemical steps (Fig. 1). Sequential enzymatic pathway steps require that enzymes be confined, caged in proximity to each other. A shared localization fastened at a specific cellular environment is of crucial importance: the enzymes are juxtapositioned proximal to the prior and following stages in the cellular network. Such an organization integrates the enzymes into the global cell responses to the environment. Co-localization also increases the effective local concentration of substrates/products and avoids the otherwise longrange diffusion–collision of the enzymes and their substrates. In signaling pathways, it makes possible an orchestrated translation of a signal into distinct events, particularly phosphorylation and ubiquitination. Related enzymatic steps are mainly carried out by multienzyme complexes which are anchored at a specific cellular sub-compartmentalization. A multienzyme complex can consist of multiple domains in a single polypeptide chain, distinct subunits, or both, that possess more than one catalytic site. The efficiency of the complex can be particularly high if the active sites of consecutive steps communicate directly, mediating the transfer of a molecule. Key questions include how are consecutive enzymatic steps accommodated? How can multienzyme pathways branch, and how is branching controlled? Why metabolic multienzyme complexes typically do not contain scaffolding proteins while signaling complexes do? On its own, co-localization is not able to provide effective answers. Here we propose and provide supporting data from the literature and our own work that allostery can play key roles in (i) controlling successive enzymatic steps in multienzyme complexes, and (ii) that it does so either via direct contact of sequestered, proximal enzymes, or via scaffolding proteins. We suggest (iii) that while in metabolic enzyme pathways sequestration and localization may be key elements this is not the case for multienzyme complexes in signaling pathways. In particular, we argue (iv) that scaffolding is much more prevalent than has been previously assumed; and enzyme domains, other enzymes, and other proteins to which the enzymes or the substrates attach, can also fulfill this role. Scaffolding facilitates effective control and switches which help in determining pathway branching; at the same time, it also allows combinatorial assembly of enzyme components, which expands their functional diversity. 3. Contrasting allosteric mechanisms to other ways of regulation of multienzyme complexes such as subcellular co-compartmentalization Allostery is a key regulator of protein activity; and thus of pathways and cell function [3–9]. Allostery is a cooperative event, linking perturbations at one (allosteric) site with their consequences at another (the active or binding site). Allosteric outcomes can be expressed by a larger

821

enzyme population bearing distinct changes in the active site shape or dynamics [10–13]. The changes at the active site specify ligand selection and catalytic action [9]. Allostery can control function not only by changing the local shape (or dynamics) of the active site; but also by biasing the sampling of the three-dimensional space by the domains bearing the active sites [7,14] which are responsible for sequential enzymatic steps. These can be in the same or in different molecules in the multienzyme complex. Alternatively, allostery can work via a conformational change of these domains (Fig. 2). The importance of allosteric modulation in multienzyme complexes can hardly be over-emphasized, when we consider transfer of products/ substrates along enzymatic pathways. The allosteric modulators can be the substrate or product; a cofactor, protein, or a second messenger; and the modulation can take place via non-covalent or covalent binding events, such as phosphorylation, ubiquitination and neddylation [15]. The outcome can be a rotation, bringing the reactive, active site-linked substrate closer to its target, followed by transfer of the substrate from the catalytic site to the target. Such action can be mediated by an enzyme which works by creating a favorable environment, as in the case of ubiquitin transfer from E2 to the target in the cullin RING ligases [16]. Alternatively, conformationally-biased fluctuation of a highly flexible linker can also result in similar outcome. For allostery to be at play, precise physical interactions between the enzymes, or the enzymes and the scaffolding protein, are a prerequisite. Mutations at the protein– protein interface or far away can impede allosteric control. Similarly, mutations affecting allosteric propagation routes can obstruct allosteric regulation. The interactions between the enzymes can be short-lived; however, they should be for a sufficiently long time for the allosteric signal to go through. Over the years, allostery has been described as the linkage between two sites in the structure, with the allosteric event far away defining substrate specificity via the active site conformation and dynamics, and binding affinity. Multienzyme complexes show that the conformation and local fluctuations at the active site may not be the sole factor; the allosterically-governed positioning of the site, or of the domain bearing the site, can also play a key role on a global scale. Allostery can work by cooperatively enhancing the tendency of the site to re-orient in a productive direction [7]. This can facilitate transfer of substrates along enzymatic pathways in multienzyme complexes; it also poses the challenge of discovery of allosteric drugs that modulate function by biasing domain rotations. Collectively, here we propose that co-localization (or subcompartmentalization) on its own is unlikely to coordinate and control multienzyme function. Allosteric action, either directly between enzymes, or via scaffolding proteins, are key factors in

GPCR Repression

PKA ER-phosphorylation

REA

NCoR

ERα

Growth Factors

MEK

ERα

ERα

RAS

ERE

ERK

Agonist ligands

TRAP CoCoA

SRCs

Activation SRA PRMT

Fig. 1. An outline of estrogen receptor (ER) signaling as an example of the multiple chemical steps in signaling pathways. Estrogen receptors (Erα and ERβ) can be selectively activated by extracellular signals, hormone and co-factor binding events [89], and phosphorylation of the ER monomer. Examples of the extracellular signals are: (i) binding of dopamine and cAMP to GPCR can activate PKA; (ii) growth factors (GFs) activate their receptors with subsequent activation of the RAS–RAF–ERK pathway. Cofactors [90] like the nuclear receptor corepressor (NCoR) and the repressor of the estrogen receptor activity (REA) lead to repression of ER response elements (ERE). Examples of direct activators are the thyroid hormone receptor (TRAP), steroid receptor activator (SRA), and steroid receptor co-activators (SRCs). Secondary co-activators (like CoCoA and PRMT) also bind ERS indirectly through association with SRCs.

822

R. Nussinov et al. / Biochimica et Biophysica Acta 1834 (2013) 820–829

BS1 conformer1

BS1 conformer2

BS1 conformer3

BS2

BS2

conformer1

ligand1

ligand2

Conformational Selection

ligand3

BS2

Reaction1

ligand1

conformer2 BS2

ligand3

Reaction2

Reaction3

conformer3

Fig. 2. Allostery can control function not only by changing the local shape (or dynamics) of the active site; but also by biasing the sampling of three-dimensional space by the domains bearing the active sites [7,14] which are responsible for sequential enzymatic steps. As shown here, allostery can also work via a conformational change of these domains.

functional multienzyme (and broadly, multimolecular) complexes. The transient topography of the enzymes, the inter-domain linkages through flexible linkers and their organization through scaffolding proteins or domains, allow the cooperative allosteric control. Preferred organizations in co-localized multi-enzyme complexes are likely to be a theme in evolution. Metabolic and signaling pathways can accomplish chemical feats: they are highly specific and efficiently adapt the cell to a changing environment. Here we argue that these feats are achieved not only by subcellular localization and sequestration; the dynamic, and precise spatial arrangement of proximal multienzyme complexes, allows a coordinated allosteric regulation. We further propose that the role played by sequestration, where the multienzyme complexes are confined in their chamber can be crucial in metabolic pathways; however, scaffolding is the dominant venue in multienzyme complexes that function in signaling. Scaffolding permits conformational switches, joining and branching of cellular pathways, and can be governed by post-translational modification events acting on the scaffolding unit(s). This functional distinction leads us to classify multienzyme complexes into two categories: (1) There is no mediation by scaffolding proteins; metabolic multienzyme complexes usually belong to this category; (2) Scaffolding proteins are key players, as in signaling enzyme complexes. However, as we discuss below, this is not always the case. 4. Examples of metabolic and signaling multienzyme complexes There are many examples of multienzyme complexes; in particular those related to metabolic pathways. Among the most studied is the pyruvate dehydrogenase complex in the mitochondrial compartment which contains enzymes of the Krebs cycle, and other pathways, such as the fatty acid oxidation and amino acid metabolism (e.g. [17]). This highly flexible complex performs a central step in energy production, catalyzing the reaction that links glycolysis with the tricarboxylic acid (TCA) cycle, and a combination of crystallography, NMR spectroscopy and electron microscopy of the pyruvate dehydrogenase complex has provided detailed information [2,16,18]. The reaction is performed in three separate steps by three separate enzymes, but all three enzymes are linked efficiently together into one large multienzyme complex. The complex contains many copies of each of the three enzymes, E1 (24 copies); E2 (24 copies) dihydrolipoyl transacetylase, and E3 (12 copies) dihydrolipoyl dehydrogenase, all arranged in a regular polyhedron with 20 identical equi-lateral triangular faces, 30 edges and 12 vertices. Another example is the cytochrome p450 enzyme and fatty acid synthase, a multienzyme that plays a key role in fatty acid synthesis.

A highly regulated multienzyme complex mediates the catabolism of cellular fat stores [19,20]. The recently detailed multienzyme architecture of the eukaryotic fatty acid synthase revealed flexibly tethered ACP (acyl carrier protein) and thioesterase domains, whose structures were unresolved, possibly due to connecting flexible linkers. Of interest, in animals, but not yeast, all catalytic domains are present on one polypeptide [21]. Recently the RNA degradosome was reviewed [22], illustrating a massive multienzyme assembly that functions in bacterial RNA turnover, and post-transcriptional control of gene expression and serves also as a machine for processing structured RNA precursors during their maturation. Recruitment of components and cellular compartmentalization of the degradosome is mediated by small recognition domains in a natively unstructured segment within a scaffolding core, suggesting dynamic conformation, with the S1 and 5′ sensor domains undergoing a large conformational transition upon recognition of the 5′-monophosphate and RNA binding. Additional recent examples include the replisome [23]; the tryptophan synthase [24], the sixenzyme aggregate purinosome, which carries out the synthesis of IMP, a precursor of A and G [25,26], and a broad range of assemblies which consist of multienzyme complexes and regulatory elements [27]. The ubiquitin proteasome complex and the MAP kinase which are detailed below, provide examples of signaling multienzyme complexes. 5. The functional merit of co-localization of multienzymes Enzymes and substrates need to be brought together in time and space and depending on their functions, be (partially) segregated. Cellular strategies aim to optimize responsiveness to external signals; ensure that the cell does not waste resources; and that it channels flux through appropriate routes. Cellular localization and scaffolding of multidomain or multipolypeptide enzyme complexes and spatial isolation of enzymatic pathways at specific sites can address these aims [28]. Cells are large (on average ~ 10 μm–100 μm). Protein partners do not diffuse over the large cellular space and randomly collide; scaffolding, localization and compartmentalization increase the effective local concentration of components that fulfill the same function in a shared pathway, protect enzymes or unstable intermediates from harmful cellular conditions or competing reactions, protect the rest of the cell from toxic intermediates, and fix, optimize and preserve the spatial positioning of the pre-organized molecules. The importance of spatial proximity can be seen from the enhanced purinosome formation in purine-depleted medium [29]. Cellular micro-organization scenarios are possible because of the large number of copies of each protein, which are distributed in the relevant modules in the cell. A given protein may populate many components, or few. Such a functional organization can efficiently be

R. Nussinov et al. / Biochimica et Biophysica Acta 1834 (2013) 820–829

turned on or off by some signaling cue. The interplay between metabolic and signaling pathways can also be seen in the purinosome dynamics which are controlled by human protein kinase hCK2 phosphorylation [29,30], and whose assembly is controlled by GPCR, illustrating that GPCR can control formation of cytosolic multienzyme complexes [25,29]. Below, we provide some examples of co-localization and scaffolding. The first concerns lipid rafts, which are regions of membranes with distinct structural composition that appear to act as platforms to co-localize proteins involved in intracellular signaling pathways (e.g. [19]). A second relates to Src-family tyrosine kinases (SFKs), which function in a variety of signal transduction events, and localize to the cytoplasmic face of the plasma membrane. SFKs are activated by stimuli such as growth factors and adhesion proteins, and are involved in signaling events at the plasma membrane, resulting in cell proliferation, differentiation, migration, and cell-shape changes. SFK trafficking is specified by the palmitoylation state which anchor it to the membrane in the vicinity of its targets. A dually palmitoylated SFK, is directly targeted to the plasma membrane [31]. Further, SFKs are inactivated by phosphorylation of their C-terminal tyrosine by Csk. While SFKs are membrane-associated, Csk is a cytoplasmic protein. Membrane adaptors such as Csk Binding Protein (Cbp) or Caveolin-1 (Cav-1) localize and mediate SFK inhibition [32,33]. Multiprotein functional units which consist of multienzyme complexes, regulatory and signaling molecules, and substrates (e.g. DNA; RNA see [27]) have been termed ‘factories’. A transcription factory is an active gene transcription unit that is clustered in a discrete site within the eukaryotic nucleus. There are ~ 10,000 factories in the nucleoplasm of a HeLa cell, which reflect ~ 8000 RNA polymerase II holoenzyme factories and ~ 2000 RNA polymerase III factories. Since the majority of the active transcription units are associated with only one polymerase, one estimate suggests that a polymerase II factory contains on average ~8 polymerase holoenzymes. In erythroid cells and many tissues, there are only 100–300 RNA Pol II factories per nucleus. It was estimated that erythroid cells express at least 4000 genes, thus many genes probably share the same factory. The nucleolus is often considered to be a large RNA polymerase I factory, although this structure is the site of activities other than just rRNA transcription [34,35]. A recent review compiled experimental evidence in support of the transcription factory model. The transcription factory model has implications for the regulation of transcription initiation and elongation, for the organization of genes in the genome, for the co-regulation of genes and for genome instability [36]. Moreover, transcription factories are nuclear sub-compartments that remain even in the absence of transcription [37], poised to prime transcription prior to activation [38]. Factories consisting of multienzyme complexes are likely to be a general principle in the cell. The cytoskeleton, in particular the microtubule network, is also a spatial regulator, in addition to membranous organelles. It provides structural support; withstands mechanical stress; drives cell motility and forms tracks for motor-mediated transport. Microtubule-dependent organization of non-membranous components directs cellular function [39–41]. The cytoskeleton can also provide the scaffold for organizing protein complexes which share a function. Examples of multienzyme complex types modulated by the microtubule network include the purinosomes above; and the recruitment of the Rho guanine nucleotide exchange factor GEF-H1 [42]. The mitotic spindle can impose localization and concentration effects on molecules with binding affinity for microtubules, turning the cytoplasm from a free and fluid space into a structured milieu, which can organize with different morphologies and densities, and provide a spatial docking platform for molecules [41]. Such structured cellular platforms can spatially tether the multienzyme (or multimolecular) complexes, preorganizing the proteins for cooperative regulation of signaling and of enzymes' actions. Localization and scaffolding can also apply to the actin-binding molecules and the actin network. Recently, integrin has also been proposed to act through allosteric

823

regulation [43,44]. Thus, as we argue below, all are dynamic, and all bias the conformational ensembles in proteins; their complexes; and along cellular pathways. 6. The old view of rigid scaffolding cannot account for the regulation of pathway branching The importance of the localization of the multienzyme complexes, and their spatial pre-organization, which may also be shaped by structural scaffolding is clear; nonetheless, on their own neither explains how the catalytic reactions are coordinated, and how signals are communicated to the enzymatic complex to activate and control a series of pathway steps. Regulation can be the result of many factors, including the concentration of the proteins and of the substrates; and post-translational modifications at the binding sites (‘orthosteric’ PTMs); however, allostery can be expected to be a major player [45]. This is evident (1) from the large conformational and dynamic changes taking place as a consequence of protein binding or allosteric PTM conjugation; (2) from the sensitivity to the lengths of loops away from the binding sites, and of linkers; and (3) from the consequences of mutations which are far away from the active sites and can affect not only the host protein, but others in the complex as well. Below, we briefly describe the allosteric effect, and highlight how it can control multienzyme complexes and supplement it by some examples. We further emphasize that scaffolding proteins — including the cytoskeleton network — are not passive; beyond the preorganization of the complexes, scaffolding proteins actively mediate allosteric signaling through the (structured) cytoplasm and the nuclear spindle. Scaffolding proteins explain the coordination among multienzyme complexes and multimolecular assemblies, and how signaling can travel efficiently across the cell to coordinate actions. Scaffolding proteins help by biasing the distribution of the conformational ensembles locally, and via the cytoskeleton network, across the cell. 7. An overview of allostery Allosterism was first defined as a cooperative effect in 1965 [46]. Based on their observations on the hemoglobin tetramer, Monod, Wyman and Changeux posited that the protein has two states: a low-affinity state T and a high-affinity state R, where the T state is thermodynamically favored. At low amounts of bound ligand, the protein prefers the low-affinity T state; however, as the amount of bound ligand increases, the protein switches to the high-affinity state. This work showed that the affinity of hemoglobin's four binding sites for oxygen molecules is increased above that of the unbound hemoglobin when the first oxygen molecule binds, clearly demonstrating positive inter-site cooperativity. Since then the fundamental mechanistic basis of allostery has been increasingly — though not entirely — understood, as well as its potential functional outcomes. However, major questions remain; among these are how the inter-site communication takes place; the scenarios and the mechanisms through which allostery works in the cell; how the collective effect of multiple simultaneous allosteric events shape the protein; and in particular, prediction of allosteric outcomes. Allosteric cooperativity reflects the physical fact that proteins (and other biomacromolecules) exist as conformational ensembles whose distributions relate to the protein sequence, size, conformational properties (also the outcome of the sequence) and the environment. The free energy landscape [47–52] provides a snapshot of the distribution at a given time point under a specific set of conditions. A change in conditions, such as noncovalent or covalent binding events; mutations; or environmental factors such as changes in pH or light can be allosteric effectors. One such example is the light-driven conformational change that alters the oligomeric state of LOV (light, oxygen, voltage) protein domain in signaling proteins

824

R. Nussinov et al. / Biochimica et Biophysica Acta 1834 (2013) 820–829

associated with circadian rhythms through formation of a cysteinyl– flavin adduct. The adduct generates a new hydrogen bond network that releases the N-terminus from the protein core and restructures an acceptor pocket for binding of the terminus on the opposite subunit of the dimer, explaining how members of a large family of photoreceptors convert light responses to alterations in protein– protein interactions [53]. Allosteric effectors change the distributions of the conformational states in the ensemble: states may be destabilized or stabilized as a consequence of the allosteric perturbation. A population shift requires climbing over barriers which separate the states. A population shift will take place either when the first state gets destabilized by the perturbation event or a next state gets stabilized. In our example above, light-driven adduct formation stabilizes a state where a hydrogen bond is formed, which then leads to the subsequent conformational changes. Allosterism can elicit minor or major conformational change; it can also be expressed in changes in the dynamics. The time frame of the shift depends on the barrier height. The higher the barrier the slower the conformational change. Large conformational changes, such as those related to domain reorientations, typically require crossing higher barriers. While allosteric phenomena reflect a fundamental physical law, evolution has exploited them for function in multiple ways. Below, we will show how evolution tapped allostery to accomplish functional modulation of multienzyme complexes via different routes and different effectors. The allosteric mechanisms taking place in multienzyme complexes depend on their function. In general, in metabolic pathways, the product of one catalytic step serves as a substrate for the next. Under such circumstances, the enzymes are generally in (transient) contact. The catalytic reaction in the first enzyme perturbs its structure; and this perturbation propagates, via a population shift through the enzyme–enzyme interface, which may act to prime the successive enzyme in the pathway, coordinating it with the preceding step. The substrate may be passed directly between the active sites, as in the case of the pyruvate dehydrogenase complex; or via a ‘release–enter’, ‘opening– closing’, enzymatic steps. Such coordination requires that the enzymes be accurately juxtaposed with respect to each other. A mutation in the interface may adversely affect the interaction, and thus the propagation of the allosteric signals. On the other hand, in signaling pathways, typically the allosteric wave proceeds via scaffolding proteins. Scaffolding proteins can recruit the enzymes, or act as a conformational ON/OFF switch. A switch mechanism can activate/deactivate an enzyme, or serve as a branching point in the selection of a successive enzyme recruitment and activation. The signal may initiate upstream; or by binding or post-translational modification events of the scaffolding proteins. Scaffolding proteins do not simply serve to tether enzymes together; their dynamic nature argues that they are the ones that regulate and control their actions. Below, we provide examples illustrating a few mechanisms through which they may work. 8. Allosteric mechanisms in multienzyme complexes To classify the mechanisms of allosteric regulation in multienzyme complexes, we distinguish between two types of cases. In Type I, the activated substrate is transferred directly from one active site to another; in Type II, the scaffolding action mediates catalysis; however, catalysis will proceed also in its absence, albeit at (much) lower efficiency. This action can also be fulfilled by an enzyme which does not directly participate in the catalytic reaction but its involvement facilitates it. Type II works via a conformational biasing mechanism; that is, a population shift. In each type the enzymes can either be on distinct chains; or be domains on the same chain. The scaffolding unit can similarly also be provided by a domain linked to the two enzyme domains, or by a distinct chain. Distinct chains for enzymes and scaffolding proteins allow combinatorial assembly. This complexity is advantageous since the same molecules can act on more diverse substrates, in different tissues

and under different conditions. Distinct chains can be expected to present broad control mechanisms, and specificity switches; both are likely to rely on allosterism. Allosteric regulation can take place through a trigger by a preceding catalytic reaction in one enzyme, or through noncovalent or covalent attachment of an allosteric effector to the scaffolding protein, which could be post-translational modification [45]. Below, we provide examples for these mechanistic types, based on enzyme complexes active in ubiquitination and phosphorylation through the MAP kinase signaling pathway. These examples relate to enzyme complexes which act in major pathways and can have multiple substrates. As such, they also allow addressing the question of selectivity; where data are available, we highlight how specificity can be achieved: either through different modules of the scaffolding proteins, or a switch that controls their action. Additional key examples relate to β-arrestin scaffolds which mediate signaling. Overall, scaffolding emerges as playing premier role in allosteric regulation of enzyme complexes in signaling. While sequestration helps to segregate the enzymes, and co-localization assembles them in proximity to each other, the scaffolds link them and conformationally control the reactions. For those cases where there is no apparent scaffold, allostery can either take place directly, or not be involved in a significant way. Future studies may unravel additional proteins playing a scaffolding role by connecting these complexes to the network, which would add a layer of allosteric regulation and provide checkpoints to decide the enzymes' actions. 9. Examples of allosteric mechanisms in multienzyme complexes Below, we provide examples illustrating how allostery can regulate multienzyme complexes. 9.1. Multiprotein and multidomain enzyme complexes in the ubiquitin– proteasome system (UPS) The UPS offers mechanistic prototypes of allosteric regulation of multienzyme complexes [54–64]. The UPS regulates protein degradation, signaling, cell cycle control and development. The ubiquitination of a target protein via the UPS is highly regulated and involves several steps. The ubiquitin is first activated by ubiquitin-activating enzyme E1; it is then transferred to the ubiquitin-conjugating enzyme E2; and to the E3 ubiquitin ligase where it is tagged onto the substrate (Fig. 3). The poly-ubiquitin labeled substrate dissociates, and is degraded by the proteasome. E3 ligases consist of Homologues of E6AP Carboxy Terminus (HECT) E3s and Really Interesting New Gene (RING)/U-box E3s. The U-box domain constitutes a relatively small family of E3s and is similar to the structure of the RING domain with the exception that it lacks the conserved histidine and cysteine residues. HECT E3s transfer ubiquitin via an E3-ubiquitin thioester intermediate, while RING/U-box E3s do not form such intermediates. HECT is a domain of ~350 amino acids. HECT E3s have a bilobal structure, in which the N-terminal lobe contains the E2 binding site. The ubiquitin is transferred from the E2 to a conserved active site Cys residue in the C-terminal lobe of HECT E3 to form the thioester bond, with subsequent transfer from E3 to the substrate protein. A different scenario is presented by RING E3s. These systems are divided into RING E3s which have RING-finger E2-binding domain and substrate-binding domain on the same polypeptide; and multi-protein Cullin-RING Ligases (CRLs), which contain four protein chains: RING-Box proteins (RBX), whose RING domain binds E2; cullin, a scaffold protein; an adaptor protein, which connects the substrate binding protein to the cullin scaffold; and the substrate binding proteins. In the RING E3s, the thioesterified ubiquitin is transferred directly from the E2 to a lysine side-chain on the substrate with no direct E3 involvement via a thioester bond to ubiquitin. Instead, E3 allosterically controls the ubiquitin transfer reaction. We classify the E2-HECT type E3s as Type I; and the multiprotein CRLs as Type II.

R. Nussinov et al. / Biochimica et Biophysica Acta 1834 (2013) 820–829

825

Ub

box Skp1 or EC Cullin

Cullin

A

B

E2 RBX

Skp1 or EC

RBX

S 50-60Å Ub E2

Ub E2 RBX

box

S

Ub Ub S

box

U 60-90 Å

b

Skp1 or EC Cullin

C

D Ub S

RBX

β-TrCP1

Ub E2

Skp1

Cul1/Cul5/Cul4A E Ub Ub E2

SOCS2

RBX

S

Elongin B/C

Cul1/Cul5/Cul4A F Ub

DDB1

Ub E2

RBX

SV5V

S

Cul1/Cul5/Cul4A

Fig. 3. The strong correlations in the motions between the linker and the substrate binding domain in the bound state, suggest allosteric effects with the linker further rotating the substrate binding domain following substrate binding for optimal ubiquitin transfer position. (A) Prior to binding to other E3 modules, the linker is flexible. (B) In the favored E3-bound conformation the substrate binding domain is rotated on the linker to the optimal position to facilitate ubiquitin transfer. (C) The linker rotation facilitates the poly-ubiquitin-labeled substrate dissociation from the E3 ligase. (D–F) illustrate that different cullins (Cul1, Cul5, Cul4A) mediate ubiquitination of different substrates via different substrate binding proteins, with each cullin leading to different preferred distance ranges in the molecular dynamics simulations. (G) This schematic diagram provides an illustration of the sampling of the space by both flexible arms, shown by the different positions of the box domain and the ubiquitin and the E2 and the cullin (left handside), the outcome of the flexibilities of the proteins, particularly the hinges of the linkers, and the preferred orientation which is the outcome of the allosteric effect; i.e. the population shift. Adapted from [20], with permission.

826

R. Nussinov et al. / Biochimica et Biophysica Acta 1834 (2013) 820–829

G

S

box

RBX

Adaptor

RBX

Ub E2

box

Adaptor

S

Ub E2

S

U b

Ub

S

Cullin

Cullin Fig. 3 (continued).

Crystal structures of the components of both the multiprotein (Type II) and the multidomain (Type II) RING complexes illustrate that ubiquitin transfer faces a problem: the distance between the E2-bound ubiquitin and the substrate to which it should be transferred is large. This raises the question of how the catalytic reaction between two active sites which are far away bridges the distance for the ubiquitin transfer to take place [14,20,65,66]. A model [67] of the multichain SCF Skp2 (Skp–Cullin–F box protein, where the F-box protein is Skp2)–Rbx–E2 complex, which the crystallographers built by superimposing the Cul1–Rbx1–Skp1–F box on the Skp1–Skp2 complex [68] and docking the UbcH7 E2 onto the Rbx1 RING domain [69], indicated that the distance between the ubiquitin E2 active site cysteine and the tip of the Skp2 substrate binding protein is 50 Å [67,70] (Fig. 3). Including the bound peptide derived from the p27 substrate complex increased the distance between the E2 active site and the substrate binding site to 59 Å [70]. An additional model built by crystallographers [71] for the substrate binding protein β-Trcp1, led to a similar 59 Å measurement [70]. The CRL is a machine, which can be viewed as containing two arms: one consists of the adaptor protein, the substrate binding protein, and the bound peptide derived from the substrate; and the second the RING box protein, and E2-ubiquitin. The two arms are connected by the scaffolding protein cullin (Fig. 3). Explicit solvent molecular dynamic simulations with the nine available substrate binding protein complexes in the PDB, the Rbx, and the three available cullin structures, Cul1, Cul4A, and Cul5 [14,20,65,66] indicated that all linkers in all simulated proteins in the E3 complex are flexible; and that their flexibility is allosterically regulated by the formation of the complex, that is, the binding events of the substrate binding protein to the substrate and to its adaptor; and the binding of the RBx to cullin. The two arms sampled space broadly, resulting in distances as short as 9 Å between the ubiquitin E2 active site cysteine and the substrate; with the largest linkers' rotations mostly in the same direction. These rotations shrink the distance, to accurately position the substrate for ubiquitination (Fig. 3A,B). The cullins' flexibility was observed to allow increasing the distance to as much as 100 Å, which leaves sufficient space for the polyubiquitinated substrate. The human genome encodes six canonical human cullin proteins (Cul1, Cul2, Cul3, Cul4A, Cul4B, and Cul5) and three atypical cullin proteins. Among the six, the sequence homology is high, and the structures are expected to be conserved. The cullins include two domains: the N-terminal (NTD) and C-terminal (CTD) domains. The NTD consists of three repeats. The CTD is globular with a cullin homology domain. Via repeat 1, the NTD recognizes different adaptor proteins, whereas the CTD binds to Rbx proteins. Of particular interest, the degrees of flexibilities are distinct among the different cullins (Fig. 3D–F), and at least for Cul1 can be changed by deletion of the long loop (which is absent in Cul4A) in the N-terminal domain, suggesting that the loop may have an allosteric functional role. CRLs are modified by allosteric NEDD8 and Cand1 binding. The CTD can

form a covalent bond with NEDD8 ubiquitin-like protein (neddylation). Neddylation increases the flexibility of the Rbx protein and the cullin CTD and, in this way, confers specificity. Cand1 blocks the cullinbinding site to NEDD8, thus inhibiting neddylation. Furthermore, mutation of lysine, the cullin residue to which NEDD8 covalently attaches, to arginine dramatically reduces CRL conformational changes, suggesting that the acceptor lysine allosterically regulates CRLs. Phosphorylation-dependent activation of c-Cbl, a single chain multidomain RING E3 ubiquitin ligase, attenuates receptor tyrosine kinase (RTK) downstream signal transduction. Cbls have a highly conserved N-terminal SH2-containing tyrosine kinase-binding domain (TKBD), a linker helix region (LHR), a RING domain, and a variable proline-rich region. c-Cbl (and b-Cbl) also contains a highly variable C-terminal extension which has a role in dimerization, and in the binding of ubiquitinated proteins and substrate recruitment following phosphorylation. TKBD binds to phosphotyrosine motifs in RTKs or tyrosine kinases such as Zap-70 kinase; the proline-rich region recruits proteins containing an SH3 domain. The LHR and RING domain are involved in recruiting E2s and mediating target ubiquitination. Four available crystal structures provide an illustration of the combination of two allosteric actions. The binding of substrate peptides (e.g. ZAP-70) to the TKBD; the binding of E2; and Tyr371 phosphorylation, collectively lead to c-Cbl activation [72]. In the absence of E2 and the TKBD-substrate, the c-Cbl

3FZE

4F2H

Ser770

Co-activator loop The ‘minimal scaffold’ domain of Ste5 Fig. 4. Allosteric regulation in the ‘minimal scaffold’ domain of Ste5. A Ser770Asn mutation disrupts the autoinhibition of the Fus co-activator function of Ste5 [81,83]. Two Ste5 von Willebrand type A domain structures are superimposed to show conformational dynamics (red ribbon, PDB code: 3FZE; yellow ribbon, PDB code: 4F2H).

R. Nussinov et al. / Biochimica et Biophysica Acta 1834 (2013) 820–829

prefers a closed conformation, in which the E2-binding surface of the RING subdomain interacts with the TKBD. Thus, the allosteric effect elicited by the binding of the ZAP-70 peptide leads to a conformational change which partially opens the RING subdomain, and predisposes it for E2 binding. The E2 binding induces the RING subdomain to adopt an open conformation. Tyr371 in LHR mediates its binding to the TKBD, keeping the c-Cbl in an inactive state; phosphorylation of Tyr371 activates c-Cbl by releasing it from TKBD. This results in a large conformational change, where the LHR domain undergoes ~180° rotation that brings the RING subdomain and E2 closer to the substrate, which shortens the distance between the catalytic cysteine of E2 to the ZAP-70 peptide from ~67 Å to ~28 Å. This facilitates the catalytic reaction of the ubiquitin transfer. Thus, the HECT example above relates to Type I mechanism, with the activated substrate being transferred from E2 to E3, with no scaffolding unit and no apparent allosteric mediation taking place. The multi-protein Cullin-RING Ligases (CRLs) provide an example of Type II where the substrate was passed from the E2 enzyme directly to the substrate, helped by the E3 allosteric involvement, which also creates a favorable environment for catalysis; the single chain multidomain c-Cbl also belongs to Type II. In CRLs, in the absence of an E3, the reactivity of E2-Ub conjugates toward the substrate protein is significantly reduced; a similar observation can be expected in the absence of the scaffolding domains in c-Cbl. Thus, these E3 ligases not only bind E2-Ub conjugate and substrate, but also enhance the direct transfer of Ub to a target lysine [73–75]. In our multiprotein Type II example, the allosteric trigger is provided by noncovalent binding of the E2 to the E3 scaffolding module of the Rbx-cullin and by the covalent post-translation modification of the NEDD8 attachment. Specificity is achieved through the variability of the adapter proteins which are bound to the different cullins, whose selectivity appears to at least partially relate to their loops and linkers which determine the cullin flexibility. In the single polypeptide Type II example, the scaffolding unit is a domain, and the trigger is the Tyr371 phosphorylation of this LHR domain, which enhances the catalytic activity. Sumoylation, the covalent attachment of SUMO (small ubiquitin-like modifier) to proteins, can change the proteins' intracellular localization, interaction patterns and other post-translational events. In sumoylation, E2 ligase can function without E3 enzymes, even though with lower reaction efficiency. Available crystal structures [76,77] coupled with molecular dynamic simulations [78,79] help in understanding the two mechanisms which were proposed for sumoylation. In both the first step involves E2 conjugation to SUMO; however, whereas in one mechanism E2-SUMO forms a complex with the target and E3, followed by SUMO transfer to the target, in the second, E2-SUMO binds to the target protein and SUMO transfer takes place without E3. Allostery is important in both mechanisms. In the first, E3 ligase triggers E2 target recognition and catalysis by shifting the equilibrium and enhancing populations: E2 bound to SUMO appears to be already pre-organized for the transfer of SUMO to a target protein and E3 binding further stabilizes the conformations, shifting the ensemble and thus increasing the efficiency of the sumoylation; in the second, basal mechanism, the pre-existence of conformational states can explain the experimental observations that sumoylation can occur without E3. Thus, in the presence of E3, it acts as an allosteric effector of E2, and in its absence, SUMO acts as the allosteric effector. E3 allosterically pre-organizes the structural environment for catalysis by restricting the conformational space. Sumoylation provides a clear example of Type II. In both mechanisms proposed for sumoylation the first step involves E2 conjugation to SUMO; however, the subsequent sequence of events differs: in the first, E2-SUMO forms a complex with the target protein and E3, followed by SUMO transfer to the target. In the second, E2-SUMO binds to the target and SUMO transfer takes place without E3. Here, E3 can be viewed as the scaffolding protein. The dynamics of E2-SUMO–Target in the absence and presence of E3 suggested that two different allosteric sites regulate the ligase activity: in the presence of E3, the E2's loop 2; in the absence

827

of E3, the Leu65-Arg70 region of SUMO [78,79]. Thus, the catalytic reaction of transferring SUMO to the target can still be maintained, albeit at lower, basal efficiency. 9.2. The MAP kinase multienzyme complex Pheromone from the opposite mating type (α-factor) in yeast binds to a G-protein-coupled receptor. The activated G-protein stimulates a MAPK module comprised of the p21 (Cdc42)-associated kinase-Ste20, MAPKKK-Ste11, MAPKK-Ste7, and two MAPKs, Fus3 and Kss1 [80]. Kss1 and Fus3 are very similar to each other. In the budding yeast, the MAPKKK-Ste11 undergoes autophosphorylation. The phosphorylated MAPKKK activates the MAPKK-Ste7, which in turn activates the mating-specific MAPK-Fus3. Activation of Fus3 leads to mating response. Activation of Kss1, which takes place under starvation, triggers filamentous growth. The yeast MAPK cascade is organized via scaffolding protein Ste5, which has separate binding sites for MAPKKK-Ste11, MAPKK-Ste7, and MAPK-Fus3. Kss1 does not bind directly to Ste5; but is in this complex through binding to Ste7. Beyond linking the three kinases, scaffolding protein Ste5 acts as an allosteric activator of the MAPK Fus3, rendering it competent to be a kinase substrate for signal transmission [81,82] (Fig. 4). Ste5 not only brings the protein kinases of the mitogenactivated protein kinase (MAPK) pathway into close proximity; it channels and transmits kinase communication, illustrating the active role played by scaffolding proteins via propagation of allosteric perturbation [7]. The Ste5ms domain changes the activity of MAPKK Ste7 toward MAPK-Fus3 and MAPK-Kss1, boosting the Ste7 activity for Fus3 ~ 5000 fold; but not toward MAPK-Kss1 [80]. The Ste5ms domain has two interfaces: one for binding Ste7; the other specific for Fus3. Ste7 binds strongly to both Ste5 domain and Fus3, tethering two proteins that normally interact only very weakly. However, even though Fus3 and Ste7 bind tightly, the phosphorylation site on Fus3 is inaccessible to Ste7; binding of Ste5ms leads to a conformational change, exposing the Fus3 phosphorylation site. Ste5m coactivator loop promotes Fus3 phosphorylation by Ste7 [81,82]. Most recently [82,83], structural and functional studies of Ste5 also showed that in addition, it allosterically prevents misactivation: under basal conditions, an intramolecular interaction of its PH domain and a von Willebrand type A domain hinders the coactivation of MAPK Fus3; binding of Ste5 triggers a conformational change which releases this autoinhibition. When the Ste5 binding site is removed from Ste11, MAPK-Kss1 is preferentially phosphorylated, indicating that Ste5 has two distinct functions: to activate MAPKKK-Ste11 and to release the Ste5 autoinhibition to allow Ste7 to activate Fus3 [82,83]. This observation provides compelling illustration that a scaffolding protein is flexible and contains a broad conformational ensemble. Not only does it undergo a population shift which propagates the signal from MAPKK-Ste7 to MAPK-Fus3, it can act as a checkpoint in deciding if and when a signal can be transmitted, in this case to mating output. However, it is still unclear why during starvation, Kss1 activation by Ste7 does not require Ste5 activation via alpha-factor binding to GPCR, as in mating pathway. The MAP kinase falls into Type II. However, it also further clearly illustrates the switch mechanism of the scaffolding protein Ste5, and its conformational control of the activation of Fus3 versus the MAPK-Kss1. Both noncovalent binding and phosphorylation of Ste7 act as allosteric effectors. Additional Type II examples relate to the β-arrestin scaffold. β-Arrestins are multifunctional adaptor proteins, which mediate GPCR desensitization (they sterically hinder the interaction of G-protein to agonist-activated seven-transmembrane receptors), endocytosis, and alternate signaling pathways of seven membrane-spanning receptors (7MSRs). The β-arrestin scaffold mediates signaling cascades involving multienzyme complexes. The β-arrestins not only bring elements of specific signaling pathways into close proximity; they actively channel

828

R. Nussinov et al. / Biochimica et Biophysica Acta 1834 (2013) 820–829

and regulate the signals, and undergo phosphorylation and ubiquitination. Among these, it mediates the extracellular signal-related kinase (ERK) activation. ERK1/2 and Raf-1 bind β-arrestin2 directly, and indirectly the mitogen-activated protein kinase kinase MEK-1; it acts as a scaffold for c-Jun N-terminal kinase (JNK) 3 activation, straddling via direct interaction JNK3 and apoptosis signaling kinase 1 (ASK1), and indirectly MAP kinase kinase 4 (MKK4). The β-arrestin scaffold also functions in Akt regulation, where β-arrestin2 associates with Akt and PP2A. β-Arrestins can function to activate signaling cascades independently of G protein activation, by serving as multiprotein scaffolds [84,85]. 10. Conclusions Multienzyme complexes are common in the cell. They fulfill a broad range of functions. It is well established that enzymes catalyzing successive reactions in pathways are often co-localized, and frequently are also connected by scaffolding proteins. Here we propose that allostery plays a key role in the presence and in the absence of scaffolding proteins. In the absence of scaffolding proteins, the precise, often short-lived, topographical organization of the enzymes in the complexes allows allosteric propagation to transmit through the enzyme–enzyme interface and in this way to coordinate the enzymatic reactions, priming the successive enzyme through the conformational perturbation created by the preceding reaction. On its own, mere co-localization is non-specific, and cannot achieve such coordination. However, the presence of the scaffolding protein allows more complex pathway control: scaffolding proteins can decide pathway unification and branching. This fundamental functional difference between the functions in the presence (or absence) of scaffolding proteins in the multienzyme complexes leads us to propose a classification of multienzyme complexes based on the scaffolding proteins (or domains). For the most part, metabolic multienzyme complexes do not contain scaffolding proteins; however, signaling multienzyme complexes usually do. We further propose that even metabolic multienzymes may be linked to the cellular network and to signaling proteins via scaffolding proteins, to transmit the cue and turn them ON/OFF. Scaffolding proteins are essential in signal transfer, and in actively manipulating the signal. These actions can be elicited by binding events to the scaffolding proteins; prior enzymatic reactions; and combinatorial post-translational modifications, such as phosphorylation, ubiquitination, and acetylation, many of which are allosteric. Above all, our central premise here is that scaffolding proteins are active components of multienzyme complexes; much more so than considered by the classic view. We present a new view which ascribes these proteins a cardinal role in the control of signaling pathways. This, scaffolding protein-centric view posits that via (orthosteric, and allosteric) post-translational modifications and cofactor binding, these proteins can regulate the cellular network. Allostery plays a key role in dynamically and efficiently controlling signals across the cell [9,86]. It does so via population shift [9,87]. Allostery is selective: the perturbations elicited by specific effectors lead to increased (or, decreased) populations of distinct conformational states whose active sites favor (disfavor) certain, specific partners [88]. Here, we highlight the role of allostery in multienzyme complexes. Allostery is a fundamental functional phenomenon in the cell whose roles, consequences and actions are challenging to predict; nonetheless, it is there. Acknowledgements This project has been funded in whole or in part with Federal funds from the National Cancer Institute, National Institutes of Health, under contract number HHSN261200800001E. The content of this publication does not necessarily reflect the views or policies of the

Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. This research was supported (in part) by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research.

References [1] R. Caspi, T. Altman, K. Dreher, C.A. Fulcher, P. Subhraveti, I.M. Keseler, A. Kothari, M. Krummenacker, M. Latendresse, L.A. Mueller, Q. Ong, S. Paley, A. Pujar, A.G. Shearer, M. Travers, D. Weerasinghe, P. Zhang, P.D. Karp, The MetaCyc database of metabolic pathways and enzymes and the BioCyc collection of pathway/genome databases, Nucleic Acids Res. 40 (2012) D742–D753. [2] P. Romero, J. Wagg, M.L. Green, D. Kaiser, M. Krummenacker, P.D. Karp, Computational prediction of human metabolic pathways from the complete human genome, Genome Biol. 6 (2005) R2. [3] N.M. Goodey, S.J. Benkovic, Allosteric regulation and catalysis emerge via a common route, Nat. Chem. Biol. 4 (2008) 474–482. [4] Q. Cui, M. Karplus, Allostery and cooperativity revisited, Protein Sci. 17 (2008) 1295–1307. [5] D.M. Leitner, Energy flow in proteins, Annu. Rev. Phys. Chem. 59 (2008) 233–259. [6] M. Fuxreiter, I. Simon, S. Bondos, Dynamic protein–DNA recognition: beyond what can be seen, Trends Biochem. Sci. 36 (2011) 415–423. [7] B. Ma, C.J. Tsai, T. Haliloglu, R. Nussinov, Dynamic allostery: linkers are not merely flexible, Structure 19 (2011) 907–917. [8] C.J. Tsai, R. Nussinov, Gene-specific transcription activation via long-range allosteric shape-shifting, Biochem. J. 439 (2011) 15–25. [9] B. Ma, R. Nussinov, Amplification of signaling via cellular allosteric relay and protein disorder, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 6887–6888. [10] K. Gunasekaran, B. Ma, R. Nussinov, Is allostery an intrinsic property of all dynamic proteins? Proteins 57 (2004) 433–443. [11] C.J. Tsai, A. del Sol, R. Nussinov, Allostery: absence of a change in shape does not imply that allostery is not at play, J. Mol. Biol. 378 (2008) 1–11. [12] S.R. Tzeng, C.G. Kalodimos, Protein dynamics and allostery: an NMR view, Curr. Opin. Struct. Biol. 21 (2011) 62–67. [13] A. del Sol, C.J. Tsai, B. Ma, R. Nussinov, The origin of allosteric functional modulation: multiple pre-existing pathways, Structure 17 (2009) 1042–1050. [14] J. Liu, R. Nussinov, Molecular dynamics reveal the essential role of linker motions in the function of cullin-RING E3 ligases, J. Mol. Biol. 396 (2010) 1508–1523. [15] C.J. Tsai, A. Del Sol, R. Nussinov, Protein allostery, signal transmission and dynamics: a classification scheme of allosteric mechanisms, Mol. Biosyst. 5 (2009) 207–216. [16] J.N. Pruneda, P.J. Littlefield, S.E. Soss, K.A. Nordquist, W.J. Chazin, P.S. Brzovic, R.E. Klevit, Structure of an E3:E2 Ub complex reveals an allosteric mechanism shared among RING/U-box ligases, Mol. Cell 47 (2012) 933–942. [17] S.J. Mihalik, S.F. Michaliszyn, J. de las Heras, F. Bacha, S. Lee, D.H. Chace, V.R. DeJesus, J. Vockley, S.A. Arslanian, Metabolomic profiling of fatty acid and amino acid metabolism in youth with obesity and type 2 diabetes: evidence for enhanced mitochondrial oxidation, Diabetes Care 35 (2012) 605–611. [18] T. Izard, A. Aevarsson, M.D. Allen, A.H. Westphal, R.N. Perham, A. de Kok, W.G. Hol, Principles of quasi-equivalence and Euclidean geometry govern the assembly of cubic and dodecahedral cores of pyruvate dehydrogenase complexes, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 1240–1245. [19] J.C. Holthuis, C. Ungermann, Cellular microcompartments constitute general sub-organellar functional units in cells, Biol. Chem. (in press). [20] J. Liu, R. Nussinov, Flexible cullins in cullin-RING E3 ligases allosterically regulate ubiquitination, J. Biol. Chem. 286 (2011) 40934–40942. [21] M. Leibundgut, T. Maier, S. Jenni, N. Ban, The multienzyme architecture of eukaryotic fatty acid synthases, Curr. Opin. Struct. Biol. 18 (2008) 714–725. [22] M.W. Gorna, A.J. Carpousis, B.F. Luisi, From conformational chaos to robust regulation: the structure and function of the multi-enzyme RNA degradosome, Q. Rev. Biophys. 45 (2012) 105–145. [23] A.M. van Oijen, J.J. Loparo, Single-molecule studies of the replisome, Annu. Rev. Biophys. 39 (2010) 429–448. [24] M.F. Dunn, D. Niks, H. Ngo, T.R. Barends, I. Schlichting, Tryptophan synthase: the workings of a channeling nanomachine, Trends Biochem. Sci. 33 (2008) 254–264. [25] F. Verrier, S. An, A.M. Ferrie, H. Sun, M. Kyoung, H. Deng, Y. Fang, S.J. Benkovic, GPCRs regulate the assembly of a multienzyme complex for purine biosynthesis, Nat. Chem. Biol. 7 (2011) 909–915. [26] Y. Deng, J. Gam, J.B. French, H. Zhao, S. An, S.J. Benkovic, Mapping protein–protein proximity in the purinosome, J. Biol. Chem. 19 (2012) 38442–38448. [27] K. Van Bortle, V.G. Corces, Nuclear organization and genome function, Annu. Rev. Cell Dev. Biol. 28 (2012) 163–187. [28] C.M. Agapakis, P.M. Boyle, P.A. Silver, Natural strategies for the spatial optimization of metabolism in synthetic biology, Nat. Chem. Biol. 8 (2012) 527–535. [29] K. Mohr, E. Kostenis, Signaling pathways: purinosomes under GPCR control, Nat. Chem. Biol. 7 (2011) 860–861. [30] S. An, M. Kyoung, J.J. Allen, K.M. Shokat, S.J. Benkovic, Dynamic regulation of a metabolic multi-enzyme complex by protein kinase CK2, J. Biol. Chem. 285 (2010) 11093–11099. [31] I. Sato, Y. Obata, K. Kasahara, Y. Nakayama, Y. Fukumoto, T. Yamasaki, K.K. Yokoyama, T. Saito, N. Yamaguchi, Differential trafficking of Src, Lyn, Yes and Fyn is specified by the state of palmitoylation in the SH4 domain, J. Cell Sci. 122 (2009) 965–975.

R. Nussinov et al. / Biochimica et Biophysica Acta 1834 (2013) 820–829 [32] A.T. Place, Z. Chen, F.R. Bakhshi, G. Liu, J.P. O'Bryan, R.D. Minshall, Cooperative role of caveolin-1 and C-terminal Src kinase binding protein in C-terminal Src kinasemediated negative regulation of c-Src, Mol. Pharmacol. 80 (2011) 665–672. [33] T. Kanou, C. Oneyama, K. Kawahara, A. Okimura, M. Ohta, N. Ikeda, Y. Shintani, M. Okumura, M. Okada, The transmembrane adaptor Cbp/PAG1 controls the malignant potential of human non-small cell lung cancers that have c-src upregulation, Mol. Cancer Res. 9 (2011) 103–114. [34] C.S. Osborne, L. Chakalova, K.E. Brown, D. Carter, A. Horton, E. Debrand, B. Goyenechea, J.A. Mitchell, S. Lopes, W. Reik, P. Fraser, Active genes dynamically colocalize to shared sites of ongoing transcription, Nat. Genet. 36 (2004) 1065–1071. [35] L.S. Shopland, C.V. Johnson, M. Byron, J. McNeil, J.B. Lawrence, Clustering of multiple specific genes and gene-rich R-bands around SC-35 domains: evidence for local euchromatic neighborhoods, J. Cell Biol. 162 (2003) 981–990. [36] H. Sutherland, W.A. Bickmore, Transcription factories: gene expression in unions? Nat. Rev. Genet. 10 (2009) 457–466. [37] J.A. Mitchell, P. Fraser, Transcription factories are nuclear subcompartments that remain in the absence of transcription, Genes Dev. 22 (2008) 20–25. [38] C. Ferrai, S.Q. Xie, P. Luraghi, D. Munari, F. Ramirez, M.R. Branco, A. Pombo, M.P. Crippa, Poised transcription factories prime silent uPA gene prior to activation, PLoS Biol. 8 (2010) e1000270. [39] Y. Zheng, A membranous spindle matrix orchestrates cell division, Nat. Rev. Mol. Cell Biol. 11 (2010) 529–535. [40] D. St Johnston, Moving messages: the intracellular localization of mRNAs, Nat. Rev. Mol. Cell Biol. 6 (2005) 363–375. [41] J. Chen, J. Lippincott-Schwartz, J. Liu, Intracellular spatial localization regulated by the microtubule network, PLoS One 7 (2012) e34919. [42] J. Birkenfeld, P. Nalbant, S.H. Yoon, G.M. Bokoch, Cellular functions of GEF-H1, a microtubule-regulated Rho-GEF: is altered GEF-H1 activity a crucial determinant of disease pathogenesis? Trends Cell Biol. 18 (2008) 210–219. [43] D. Boettiger, Mechanical control of integrin-mediated adhesion and signaling, Curr. Opin. Cell Biol. 24 (2012) 592–599. [44] T.A. Springer, M.L. Dustin, Integrin inside-out signaling and the immunological synapse, Curr. Opin. Cell Biol. 24 (2012) 107–115. [45] R. Nussinov, C.J. Tsai, F. Xin, P. Radivojac, Allosteric post-translational modification codes, Trends Biochem. Sci. 37 (2012) 447–455. [46] J. Monod, J. Wyman, J.P. Changeux, On the nature of allosteric transitions: a plausible model, J. Mol. Biol. 12 (1965) 88–118. [47] H. Frauenfelder, S.G. Sligar, P.G. Wolynes, The energy landscapes and motions of proteins, Science 254 (1991) 1598–1603. [48] K.A. Dill, H.S. Chan, From Levinthal to pathways to funnels, Nat. Struct. Biol. 4 (1997) 10–19. [49] H.S. Chan, K.A. Dill, Protein folding in the landscape perspective: chevron plots and non-Arrhenius kinetics, Proteins 30 (1998) 2–33. [50] C.J. Tsai, S. Kumar, B. Ma, R. Nussinov, Folding funnels, binding funnels, and protein function, Protein Sci. 8 (1999) 1181–1190. [51] C.J. Tsai, B. Ma, R. Nussinov, Folding and binding cascades: shifts in energy landscapes, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 9970–9972. [52] S. Kumar, B. Ma, C.J. Tsai, N. Sinha, R. Nussinov, Folding and binding cascades: dynamic landscapes and population shifts, Protein Sci. 9 (2000) 10–19. [53] A.T. Vaidya, C.H. Chen, J.C. Dunlap, J.J. Loros, B.R. Crane, Structure of a light-activated LOV protein dimer that regulates transcription, Sci. Signal. 4 (2011) ra50. [54] P.S. Brzovic, A. Lissounov, D.E. Christensen, D.W. Hoyt, R.E. Klevit, A UbcH5/ubiquitin noncovalent complex is required for processive BRCA1-directed ubiquitination, Mol. Cell 21 (2006) 873–880. [55] A.D. Capili, C.D. Lima, Taking it step by step: mechanistic insights from structural studies of ubiquitin/ubiquitin-like protein modification pathways, Curr. Opin. Struct. Biol. 17 (2007) 726–735. [56] V. Bernier-Villamor, D.A. Sampson, M.J. Matunis, C.D. Lima, Structural basis for E2-mediated SUMO conjugation revealed by a complex between ubiquitinconjugating enzyme Ubc9 and RanGAP1, Cell 108 (2002) 345–356. [57] D.M. Duda, L.A. Borg, D.C. Scott, H.W. Hunt, M. Hammel, B.A. Schulman, Structural insights into NEDD8 activation of cullin-RING ligases: conformational control of conjugation, Cell 134 (2008) 995–1006. [58] D.M. Duda, D.C. Scott, M.F. Calabrese, E.S. Zimmerman, N. Zheng, B.A. Schulman, Structural regulation of cullin-RING ubiquitin ligase complexes, Curr. Opin. Struct. Biol. 21 (2011) 257–264. [59] S. Jackson, Y. Xiong, CRL4s: the CUL4-RING E3 ubiquitin ligases, Trends Biochem. Sci. 34 (2009) 562–570. [60] A.A. Ogunjimi, D.J. Briant, N. Pece-Barbara, C. Le Roy, G.M. Di Guglielmo, P. Kavsak, R.K. Rasmussen, B.T. Seet, F. Sicheri, J.L. Wrana, Regulation of Smurf2 ubiquitin ligase activity by anchoring the E2 to the HECT domain, Mol. Cell 19 (2005) 297–308. [61] E. Ozkan, H. Yu, J. Deisenhofer, Mechanistic insight into the allosteric activation of a ubiquitin-conjugating enzyme by RING-type ubiquitin ligases, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 18890–18895.

829

[62] B.A. Schulman, J.W. Harper, Ubiquitin-like protein activation by E1 enzymes: the apex for downstream signalling pathways, Nat. Rev. Mol. Cell Biol. 10 (2009) 319–331. [63] E. Papaleo, N. Casiraghi, A. Arrigoni, M. Vanoni, P. Coccetti, L. De Gioia, Loop 7 of E2 enzymes: an ancestral conserved functional motif involved in the E2-mediated steps of the ubiquitination cascade, PLoS One 7 (2012) e40786. [64] E. Papaleo, V. Ranzani, F. Tripodi, A. Vitriolo, C. Cirulli, P. Fantucci, L. Alberghina, M. Vanoni, L. De Gioia, P. Coccetti, An acidic loop and cognate phosphorylation sites define a molecular switch that modulates ubiquitin charging activity in Cdc34-like enzymes, PLoS Comput. Biol. 7 (2011) e1002056. [65] J. Liu, R. Nussinov, The mechanism of ubiquitination in the cullin-RING E3 ligase machinery: conformational control of substrate orientation, PLoS Comput. Biol. 5 (2009) e1000527. [66] J. Liu, R. Nussinov, Rbx1 flexible linker facilitates cullin-RING ligase function before neddylation and after deneddylation, Biophys. J. 99 (2010) 736–744. [67] N. Zheng, B.A. Schulman, L. Song, J.J. Miller, P.D. Jeffrey, P. Wang, C. Chu, D.M. Koepp, S.J. Elledge, M. Pagano, R.C. Conaway, J.W. Conaway, J.W. Harper, N.P. Pavletich, Structure of the Cul1–Rbx1–Skp1–F boxSkp2 SCF ubiquitin ligase complex, Nature 416 (2002) 703–709. [68] B.A. Schulman, A.C. Carrano, P.D. Jeffrey, Z. Bowen, E.R. Kinnucan, M.S. Finnin, S.J. Elledge, J.W. Harper, M. Pagano, N.P. Pavletich, Insights into SCF ubiquitin ligases from the structure of the Skp1–Skp2 complex, Nature 408 (2000) 381–386. [69] N. Zheng, P. Wang, P.D. Jeffrey, N.P. Pavletich, Structure of a c-Cbl–UbcH7 complex: RING domain function in ubiquitin-protein ligases, Cell 102 (2000) 533–539. [70] T. Cardozo, M. Pagano, The SCF ubiquitin ligase: insights into a molecular machine, Nat. Rev. Mol. Cell Biol. 5 (2004) 739–751. [71] G. Wu, G. Xu, B.A. Schulman, P.D. Jeffrey, J.W. Harper, N.P. Pavletich, Structure of a beta-TrCP1–Skp1–beta-catenin complex: destruction motif binding and lysine specificity of the SCF(beta-TrCP1) ubiquitin ligase, Mol. Cell 11 (2003) 1445–1456. [72] H. Dou, L. Buetow, A. Hock, G.J. Sibbet, K.H. Vousden, D.T. Huang, Structural basis for autoinhibition and phosphorylation-dependent activation of c-Cbl, Nat. Struct. Mol. Biol. 19 (2012) 184–192. [73] J. Song, J. Wang, A.A. Jozwiak, W. Hu, P.M. Swiderski, Y. Chen, Stability of thioester intermediates in ubiquitin-like modifications, Protein Sci. 18 (2009) 2492–2499. [74] D.M. Wenzel, A. Lissounov, P.S. Brzovic, R.E. Klevit, UBCH7 reactivity profile reveals parkin and HHARI to be RING/HECT hybrids, Nature 474 (2011) 105–108. [75] J.N. Pruneda, P.J. Littlefield, S.E. Soss, K.A. Nordquist, W.J. Chazin, P.S. Brzovic, R.E. Klevit, Structure of an E3:E2 approximately Ub complex reveals an allosteric mechanism shared among RING/U-box ligases, Mol. Cell 47 (2012) 933–942. [76] A.A. Yunus, C.D. Lima, Lysine activation and functional analysis of E2-mediated conjugation in the SUMO pathway, Nat. Struct. Mol. Biol. 13 (2006) 491–499. [77] D. Reverter, C.D. Lima, Insights into E3 ligase activity revealed by a SUMO– RanGAP1–Ubc9–Nup358 complex, Nature 435 (2005) 687–692. [78] M. Tozluoglu, E. Karaca, R. Nussinov, T. Haliloglu, A mechanistic view of the role of E3 in sumoylation, PLoS Comput. Biol. 6 (2010). [79] E. Karaca, M. Tozluoglu, R. Nussinov, T. Haliloglu, Alternative allosteric mechanisms can regulate the substrate and E2 in SUMO conjugation, J. Mol. Biol. 406 (2011) 620–630. [80] S. Maleri, Q. Ge, E.A. Hackett, Y. Wang, H.G. Dohlman, B. Errede, Persistent activation by constitutive Ste7 promotes Kss1-mediated invasive growth but fails to support Fus3-dependent mating in yeast, Mol. Cell. Biol. 24 (2004) 9221–9238. [81] M. Good, G. Tang, J. Singleton, A. Remenyi, W.A. Lim, The Ste5 scaffold directs mating signaling by catalytically unlocking the Fus3 MAP kinase for activation, Cell 136 (2009) 1085–1097. [82] M.A. Seeliger, J. Kuriyan, A MAPK scaffold lends a helping hand, Cell 136 (2009) 994–996. [83] J.G. Zalatan, S.M. Coyle, S. Rajan, S.S. Sidhu, W.A. Lim, Conformational control of the Ste5 scaffold protein insulates against MAP kinase misactivation, Science 337 (2012) 1218–1222. [84] S.M. DeWire, S. Ahn, R.J. Lefkowitz, S.K. Shenoy, Beta-arrestins and cell signaling, Annu. Rev. Physiol. 69 (2007) 483–510. [85] K. Xiao, S.K. Shenoy, K. Nobles, R.J. Lefkowitz, Activation-dependent conformational changes in {beta}-arrestin 2, J. Biol. Chem. 279 (2004) 55744–55753. [86] R. Nussinov, How do dynamic cellular signals travel long distances? Mol. Biosyst. 8 (2012) 22–26. [87] G. Kar, O. Keskin, A. Gursoy, R. Nussinov, Allostery and population shift in drug discovery, Curr. Opin. Pharmacol. 10 (2010) 715–722. [88] Y. Pan, C.J. Tsai, B. Ma, R. Nussinov, Mechanisms of transcription factor selectivity, Trends Genet. 26 (2010) 75–83. [89] W. Shao, M. Brown, Advances in estrogen receptor biology: prospects for improvements in targeted breast cancer therapy, Breast Cancer Res. 6 (2004) 39–52. [90] J.M. Hall, D.P. McDonnell, Coregulators in nuclear estrogen receptor action: from concept to therapeutic targeting, Mol. Interv. 5 (2005) 343–357.