How do disordered regions achieve comparable functions to structured domains?

Abstract

The traditional structure to function paradigm conceives of a protein's function as emerging from its structure. In recent years, it has been established that unstructured, intrinsically disordered regions (IDRs) in proteins are equally crucial elements for protein function, regulation and homeostasis. In this review, we provide a brief overview of how IDRs can perform similar functions to structured proteins, focusing especially on the formation of protein complexes and assemblies and the mediation of regulated conformational changes. In addition to highlighting instances of such functional equivalence, we explain how differences in the biological and physicochemical properties of IDRs allow them to expand the functional and regulatory repertoire of proteins. We also discuss studies that provide insights into how mutations within functional regions of IDRs can lead to human diseases.

Introduction

The folding of polypeptides into structured proteins has traditionally been seen as a prerequisite for protein function. In recent years, segments called intrinsically disordered regions that do not fold autonomously into stable secondary and tertiary structures have also been shown to be functional in a diverse array of fundamental molecular processes.^1–7 The propensity of a protein segment to be structured spans a continuous spectrum (Fig. 1): at the extreme ends, a protein with no intrinsically disordered regions is deemed a structured protein and conversely, a protein with a disordered sequence entirely devoid of structure is referred to as an intrinsically disordered protein (IDP). Over one third of proteins in higher eukaryotes have large disordered segments (>30 residues), referred to as intrinsically disordered regions (IDRs).^8,9

Figure 1.

The level of disorder in proteins can vary greatly, even within what is traditionally considered a “structured domain” and a “disordered region.” Both structured domains and disordered regions are fundamental units of protein function, and most eukaryotic proteins are composed of both types of region. Reprinted with permission from Babu et al., Science, 2012, 337, 1460–1461, © American Association for the Advancement of Science.

The first molecular structure of a protein was determined for myoglobin¹⁰ and the ensuing expansion of protein structure investigations was driven by a focus on highly structured, relatively rigid proteins. Indeed, many important principles of protein sequence, structure, function, folding and stability have been derived from studies of these structured proteins.^11–18 A number of recent studies have demonstrated that IDRs can perform functions and biological tasks that are comparable to structured proteins, and that in certain cases IDRs provide specialized functional advantages. Structured and disordered regions can also act synergistically, with IDRs enhancing a protein's versatility by increasing the number of possible functional states it can adopt.¹⁹ For instance, alternative splicing of disordered segments can modify protein function and rewire regulatory and signaling networks through the differential inclusion of IDRs that contain linear interaction motifs and/or post-translational modification sites.^20–23 The differential inclusion of such disordered segments can mediate new protein interactions, and hence change the context in which the biochemical or molecular functions are carried out by the protein.²¹ We and others recently showed that the alternative splicing of disordered segments can rewire or fine-tune protein interactions in a tissue-specific manner, for example by including/excluding segments that directly interact with other proteins, DNA, RNA, or ligands^20,21 (Fig. 2). Another mechanism by which functional versatility can be increased is by regulated unfolding or disorder-to-order transitions, which permits some proteins to functionally switch between the structured and disordered states. This provides another mechanism for synergy between disorder and order.²⁴ In addition to enhancing a protein's versatility, IDRs can also directly regulate a protein's activity by influencing its half-life in cells.^25,26 We and others have recently demonstrated that the length, composition, and position of disordered segments in proteins can affect the protein's half-life, and hence its function in cells.^25,27,28 Yeast, mouse, and human proteins with terminal or internal IDRs have significantly shorter half-lives, likely because these features can promote the initiation of degradation by the proteasome (Fig. 3).

Figure 2.

The alternative inclusion or exclusion of tissue-specific exons can rewire interaction networks and modulate protein interactions. (A) When tissue-specific splicing gives rise to isoforms that differ in the presence of disordered protein segments, this can result in the rewiring of an interaction network in the respective tissues. This is illustrated with the PIP5K1C kinase gene that has an exon with different inclusion levels in cerebellum and lymph node. (B) Molecular mechanisms by which a tissue-specific segment can rewire or fine-tune protein interactions. Segments encoded by TS exons can include a region that directly interacts with other proteins, DNA, RNA, or ligands, or indirectly affect the protein's binding properties (e.g., affinity, kinetics, and selectivity). TS segments that are involved in interactions are less frequently domains and more often disordered regions that embed peptide-binding motifs. Reprinted with permission from Buljan et al., Mol Cell, 2012, 46, 871–883, © Cell Press.

Figure 3.

Long terminal or internal disordered segments can influence protein half-life by permitting efficient initiation of degradation by the proteasome. Within a protein, short disordered segments tend to correspond to long protein half-lives, while long disordered segments (i.e. at a terminal end or within the internal segment) are linked to shorter half-lives. Reprinted with permission from Van der Lee et al., Cell Rep, 2014, 8, 1832–1844, © Cell Press.

The importance of IDRs is further demonstrated by the finding that both the mis-regulation²⁹ of IDRs and mutations within IDRs often detrimentally affect molecular function;^30,31 indeed, 20% of all annotated missense disease mutations are within IDRs.^32,33 In this review, we discuss recent research that highlights how disordered regions achieve functions that are comparable in scope and importance to structured domains, focusing on classical structural biology concepts such as protein complex and assembly formation, mediation of controlled conformational changes and allostery.

Formation of Protein Complexes and Assemblies

Protein interactions, ranging from stable obligate permanent complexes to transient interactions, are common in signaling and regulation pathways and serve critical roles in nearly all aspects of biology. Classical work in the field of protein complex formation has largely focused on the structural classification of protein-protein interactions^34–36 (PPIs) and the assembly of protein complexes.^37,38 These studies have provided important insights and have been successfully used to design new protein interfaces and assemblies, an important goal in nanomaterial engineering.^39–41

Properties of protein complexes formed by IDRs

Disordered protein regions can also form functional complexes. Relative to interactions between structured protein regions, disordered regions tend to interact with lower affinity and therefore tend to be involved in more transient interactions (Fig. 4). Interactions between IDRs are frequently mediated by linear motifs^42,43 (3–10 residues) and molecular recognition features^44,45 (MoRFs, 10–70 residues), which both serve as sites of interaction.⁴⁶ These features are positioned within disordered regions, but upon interaction with their partner, a defined conformation is often induced. These linear motifs and MoRFs tend to interact with low affinity, making their interactions more reversible and transient, and hence they tend to be common mediators of dynamic cell processes.^47,48 However, high avidity interactions, which are relatively high affinity interactions that result from the accumulated strength of multiple low affinity interactions, can be achieved by combining linear motifs or MoRFs. An example is the endocytotic protein Eps15, which contains over a dozen linear motifs that help cluster AP2 complexes at endocytotic assembly zones.^49,50 In some instances, IDRs can form more extensive (and consequently, high affinity) interfaces than ordered proteins, of a similar size, via the induced folding that can occur upon IDRs binding their target.² It has been speculated that in these cases, fully structured proteins would need to be much larger to achieve similar interface sizes and high binding affinities, which would be unfavorable to the cell due to an increase in either cellular crowding or cell size.^51,52

Figure 4.

The role of intrinsically disordered proteins (IDPs) and regions (IDRs) in mediating interactions and forming complexes. Compared with ordered proteins, disordered proteins perform these roles differently, for instance by binding more partners and having lower affinity interactions (Box 1). The E. coli protein RNase E, in which a disordered tail utilises molecular recognition feature (MoRF)-mediated binding to organize a functional degradasome complex, exemplifies the unique roles that IDRs can play in interactions and protein complex formation (Box 2). Mutations that disrupt the normal complex-forming behaviour of IDPs can lead to disease, for instance by altering linear motifs or conserved regulatory elements (Box 3).

IDRs can regulate complex formation in diverse ways. Although many IDRs fold upon complex formation,⁵³ others maintain their disordered nature when in a complex.⁵⁴ This facilitates complex formation by decreasing the associated entropic cost.^54,55 IDRs contribute several other features that can further modulate entropic costs of interactions, which in turn can promote or inhibit the formation of interactions. For example, IDRs often possess numerous linear motifs which serve as binding sites, and this multivalency limits the conformational entropic cost of binding as only the motifs must become rigid upon binding and not the regions linking them (see Fig. 2 in Flock et al.⁵⁶). Many important protein families contain large disordered elements in their functional regions that serve essential roles for complex formation and regulation. An example is the family of G protein-coupled receptors (GPCRs), where specific disordered regions are critical for modulating and fine-tuning GPCR signaling due to their roles of facilitating interactions with signaling partners, undergoing specific disorder to order transitions, mediating protein complex assembly and tuning interaction specificity and affinity (see Fig. 2 in Venkatakrishnan et al.⁵⁷).

Finally, because interactions of IDRs tend to be mediated by short stretches of residues, they are easily modifiable, for instance by post-translational modifications (PTMs).⁵⁸ This allows the creation of simple biological switches,⁵⁹ and several examples of this phenomenon are found in regulatory and signaling proteins⁶⁰ such as in the cell cycle regulatory proteins p21 and p27, which are two well-characterized IDPs that function as cyclin-dependent kinase (Cdk) regulators.⁶¹ In this system, p21 and p27 promiscuously bind Cdk/cyclin complexes. The extended, Cdk-bound version of p27 presents various phosphorylation sites, which act to modulate its Cdk regulatory activity. Strikingly, through a specific phosphorylation, p27 can switch from inhibiting the Cdks to being a partial activator,⁶¹ demonstrating how IDPs can be dramatically repurposed by relatively minor changes via post-translational modifications.

Consequences of IDRs in protein complexes

Intrinsic disorder is often associated with binding promiscuity, or the ability to interact with diverse and often numerous binding partners.^46,62,63 Interactions involving IDRs also often tend to be more transient.⁴⁸ In accord with these two features, IDRs often serve as a platform or docking station for the assembly of multiprotein complexes, which help to mediate signal computation and propagation within the cell.⁶⁴ These large, multisite docking proteins frequently function as central hub proteins within protein-protein interaction networks. Some of the most prominent PPI hubs, for example, p53 and BRCA1, are highly intrinsically disordered,⁶⁵ and indeed it has been estimated that two thirds of all signaling proteins are highly disordered.⁶⁶ As another example, the largely unstructured C-terminal half of the scaffolding protein RNase E organizes the Escherichia coli RNA degradasome complex.⁶⁷ The long disordered region is punctuated by short 15 to 40 residue segments (MoRFs) that can fold upon recognition, and these regions have been identified as interaction sites for specific proteins and RNAs that are crucial to its degradasome function.⁶⁸

While ordered proteins can achieve higher binding partner numbers by linking together multiple ordered domains,⁶² disordered proteins achieve this through a combination of short linear motifs and low complexity (highly repetitive) repeats. Due to their short length, many linear motifs may be linked together on a protein to form numerous binding sites,⁴⁸ while the flexibility of low complexity regions may permit the binding of different targets.⁶⁹ Binding promiscuity can also be achieved by the IDR adopting different conformations, as in the case of calmodulin. An IDR within the protein MBP binds to calmodulin, but due to the large number of conformations MBP-calmodulin can exist in, the same IDR can promiscuously bind other biomolecules such as membranes, cytoskeletal proteins and modifying enzymes.⁷⁰

Because of the binding promiscuity of many IDRs, proteins with IDRs are tightly regulated and undergo strict control of their synthesis rates, abundances and lifetimes.^25,29,71 This is because altered abundance of IDPs can lead to promiscuous interactions in cells due to an increased likelihood of off-target interactions, which may result in disease.^29,72 The ability of IDPs to form numerous complexes also has significant consequences for proteome interconnectivity. ID segments tend to connect functional modules, thereby interweaving disparate parts of PPI networks.^48,73 Within proteomes, the ID segments of proteins frequently contain the actual interacting regions,⁷⁴ and interactions between wholly disordered proteins are enriched in the human PPI network.⁷⁵

Importantly, due to the small number of residues that frequently make an interaction site in IDRs, functional sites can be more easily disrupted by mutations. For instance, the conserved proteosomal degradation motif in the highly disordered N-terminal region of the regulatory protein β-catenin (DEG_SCF_TRCP1_1, 32DpSGIHpS37, that mediates binding with WD40 repeat domain of the beta-TRCP subunit of the SCF-betaTRCP E3 ubiquitin ligase complex) is recurrently mutated in different types of cancers.^31,33 Mutations in this motif alter the abundance of β-catenin, leading to gene expression changes that can result in cancers. It seems likely that mutations in linear motifs within IDRs have a greater impact on disease than is currently appreciated³³ (Fig. 4).

In addition to mediating protein-protein interactions, disordered regions can also mediate interactions with nucleic acids.^76–78 As a result, IDRs are enriched in DNA-binding and RNA binding proteins.^79,80 Similar to linear motifs, conserved recognition elements (CoREs) in IDRs have been suggested to mediate DNA recognition, and mutations in these elements are often deleterious.⁸¹ For instance, the SOX2 P > R loss of function mutation in the KRPMNAFMVWS CoRE is associated with brain anomalies in humans.^81,82

Formation of macromolecular protein assemblies

Macromolecular protein assemblies are large intracellular structures composed of numerous protein partners, often serving organisational roles. Due to their structural versatility and propensity to form multivalent and transient interactions, disordered proteins play a critical role in the formation and maintenance of macromolecular protein assemblies and spatially segregated structures by forming either amyloid-like polymers or proteinaceous droplets⁸³ (Fig. 5). Low-complexity domains in disordered proteins can facilitate phase transitions, which are implicated in the formation of hydrogels as a result of their multivalency and ability to form low-affinity interactions. The ability of disordered proteins to undergo phase transition from a solution to a gel is likely to accompany the formation of large, dynamic supramacroscopic polymer gels from small protein assemblies.⁸⁴ These hydrogel structures can influence the cell either through organisational roles, or by influencing subcellular activities directly. Whereas macromolecular structures composed of ordered proteins, such as certain cytoskeletal features and structures of muscle filaments, tend to be highly regular and relatively long-lived, disordered macromolecular complexes are especially suited for temporally variable structures undergoing substantial flux. IDPs that can form higher order assemblies and functional aggregates are tightly regulated so that their concentrations are low enough to prevent uncontrolled formation of such assemblies.⁷¹ Furthermore, transcripts that encode such assembly forming IDPs are tightly regulated and asymmetrically localised in cells in order to ensure that such assemblies are not formed at random places but in highly specific locations.⁸⁵

Figure 5.

The role of IDRs in forming macromolecular assemblies. IDPs form assemblies that are different from ordered proteins. For instance, IDRs can undergo phase transitions to form hydrogels (Box 1). An example of this process is RNA granule formation, in which IDRs form hydrogels that allow the membraneless organisation of specific proteins and RNAs (Box 2). Mutations that disrupt the assembly formation of IDRs can cause diseases, for instance by causing aberrant aggregation of proteins, or by repurposing of low complexity domains into oncogenic transactivators (Box 3).

Hydrogels composed of disordered proteins have been implicated in numerous organisational roles. For example, RNA granules are membraneless structures containing specific mRNAs and RNA-binding proteins, and an amyloid-like hydrogel composed of disordered, repetitive regions is thought to be necessary to retain the requisite RNAs and RNA-binding proteins^86,87 (Fig. 5). Hydrogels also play roles in the organization of P granules, or ribonucleoprotein organelles that determine Caenorhabditis elegans germ cells,⁸⁸ as well as in neuronal myelin sheath formation.⁸⁹ Importantly, post-translational modification of disordered proteins, such as (de-) phosphorylation of MEG proteins in C. elegans, regulates dynamics of RNA granules.⁹⁰ Given these findings, it is probable that additional organisational roles of IDP-based hydrogels will continue to emerge. It has been speculated that in addition to serving local organisational roles (as mentioned in the preceding examples), molecules that undergo phase transition may serve more diffuse organisational roles throughout the cytoplasm.^84,91

Hydrogels composed of disordered proteins have also been implicated in the selective transport between the cytoplasm and nucleus through the hydrogel-forming ability of disordered FG repeats in nucleoporins of nuclear pore complexes.⁹² Multivalent, interaction-mediated disordered assemblies can control the cell cycle⁹³ and regulate fungal intercellular connectivity by aggregating at cell-cell channels.⁹⁴ Additionally, higher-order disordered protein assemblies have been implicated in cell signaling processes,⁹⁵ in the proper functioning of stereocilia tip links of inner ear hair cells involved in hearing,⁹⁶ and in T-cell function and differentiation.⁹⁷ Certain neurological disorders and cancers involve aberrant formation of these macromolecular assemblages (Fig. 5). For instance, mutations in RNA binding nuclear proteins FUS and TDP43 which form RNA granules and aid in splicing and mRNA stability, lead to cytoplasmic aggregation resulting in the neurodegenerative disease amyotrophic lateral sclerosis.⁹⁸ On the other hand, the ability of the IDRs to form assemblages is exploited in cancers. For example, the low complexity domains of FUS, EWS, or TAF15 RNA binding proteins fuse with a variety of DNA binding domains, forming polymeric fibers which bind to the RNA polymerase II and function as transcriptional activation domains, driving cancer cell formation.⁹⁹

In addition to organisational roles, macromolecular structure formation can directly influence cellular phenotypes. As an example, the linear motif-mediated formation of Nck–nephrin–N-WASP (neuronal Wiskott–Aldrich syndrome protein) protein droplets correlates with actin polymerising ability.⁸⁴ Using dynamic light scattering and small angle X-ray scattering to monitor polymerisation and phase separation, it was found that the highly multivalent interactions between N-WASP and its partners Nck and nephrin induce a distinct liquid-liquid phase separation and create micron-level liquid droplets in solution. This is thermodynamically coupled to a sol-gel transition (macromolecular polymer gel formation) as increasing intermolecular bonding occurs. Within this system, the droplet formation is functionally important, as it is coupled to a stimulation of Arp2/3-mediated actin assembly.

Mediation of Controlled Conformational Transitions in Proteins

For many proteins, conformational changes are required for function. Despite these different conformations, structured proteins have few highly energetically favorable states. In contrast, in disordered proteins numerous conformations are equally energetically favorable¹⁰⁰ (Fig. 6). Despite this inherent conformational flexibility, disordered regions are capable of undergoing the controlled, defined structural shifts associated with molecular function. Two dramatic (and opposite) types of conformational output for disordered proteins are disorder-to-order transitions and order-to-disorder transition (also known as regulated unfolding). In addition, increasing evidence suggests that disorder can play a central role in mediating allostery.

Figure 6.

The role of IDRs in mediating conformational transitions. Compared with structured proteins, disordered proteins exhibit higher conformational flexibility and can undergo different types of folding-related conformational shifts (Box 1). A unique type of conformational transition in IDRs, induced folding, is a critical mechanism for functional CREB-CBP interactions (Box 2). Mutations disrupting functional conformational shifts of IDRs can lead to diseases, for instance by disrupting the regions of structural disorder that normally undergo coupled binding and folding (Box 3).

Disorder-to-order transitions

Many disordered regions undergo transitions from a disordered to an ordered state upon binding an interaction partner such as another protein, DNA, or ligands. The two most prominent concepts that describe the kinetics of coupled binding and folding are conformational selection and induced folding.^53,101–103 In conformational selection, disordered regions stochastically sample different conformations in their disordered state, and one conformation that is complementary to the binding partner is drawn from equilibrium (as in the binding of p53 to MDM2¹⁰⁴). In the induced folding mechanism, nonspecific contacts are formed first, which induces the disordered segment to progressively fold into the correct structures as it forms more ligand-specific contacts. An example is the association of the KID domain to the KIX domain of CREB (Fig. 6): using NMR titrations and ¹⁵N relaxation dispersion, the dynamic folding upon binding that occurs when the phosphorylated kinase inducible activation domain (pKID) of the transcription factor CREB binds the KIX domain of CBP (CREB binding protein) was elucidated. These experiments have shown that upon binding to the KIX domain, pKID initially forms an ensemble of transient encounter complexes; while still bound, these intermediate states are progressively stabilized into their final state by intermolecular interactions found in the final bound state.¹⁰⁵ This interaction has downstream effects on CREB target genes, including genes involved in gluconeogenesis.

Functional disorder-to-order changes also add an element of adaptability due to their ability to respond to environmental conditions.¹⁰⁶ For example, in the mitochondrial inner membrane a redox-mediated disorder-to-order transition permits mammalian COX17 to transfer copper to cytochrome c oxidase.^107,108 Nearly 20% of disease mutations cause aberrations in disorder-to-order transitions.^30,109 For instance, the frameshift mutation in the GPCR Frizzled4 (Fz4) leads to a rare form of familial exudative vitreoretinopathy (Fig. 6). The mutation results in the loss of disorder, making the tail highly structured. This gain of structure induces the formation of aggregates in the endoplasmic reticulum, blocking the localization of the receptor to plasma membrane.¹¹⁰

Regulated unfolding and cryptic disorder

Another key type of controlled conformational transition that is gaining increasing recognition is the regulated unfolding of proteins.^24,111 This concept states that (i) regulated unfolding can permit sampling of the conformational space and hence the functional space, and that (ii) a structured domain may have a function when it is partially or even fully unfolded. This is in contrast to the classical idea of a protein having one optimal, folded conformation that is critical for successful function. For example, the histone chaperone nucleophosmin exists in equilibrium between ordered and disordered forms, which have distinct functions and sub-cellular localizations. Nucleophosmin switches between these two states via the PTM-regulated unfolding of its N-terminal domain into a disordered state.^112,113 Another example of PTM-mediated functional unfolding is KSRP, which is a protein that promotes AU-rich mediated mRNA decay. Phosphorylation within KSRP's N-terminal domain acts as a switch by unfolding its KH1 domain, leading to the protein's increased localization in the nucleus and hence impeding its ability to promote mRNA decay in the cytoplasm.¹¹⁴

Regulated protein unfolding has also been described as an output response in signal transduction pathways. For instance, phosphorylation leads to the partial unfolding of p27, causing selective p27 ubiquitination and degradation, which leads to a cascade that terminates in the cell's entry into the S phase of the cell cycle.¹¹¹ On an even larger scale, cryptic disorder has been described to mediate organ-level effects. The elastic protein titin plays a large role in determining diastolic blood volume in the left ventricle of the heart; mechanical unfolding of titin immunoglobulin (Ig) domains exposes cryptic cysteines, which can then be S-glutathionylated. This S-glutathionylation decreases the stability and folding ability of the Ig domain, promoting the extended forms of titin and ultimately mediating the mechanochemical modulation of the elasticity of human cardiomyocytes.¹¹⁵

Like disorder-to-order transitions, regulated unfolding can also be responsive to environmental conditions. In a number of oxidation sensor proteins, large-scale structural rearrangements of domains into disordered states dramatically affect protein function. In bacterial Hsp33, an elevated oxidation status unfolds and activates the protein, allowing it to fulfil its chaperone role. The IDRs of Hsp33 permit it to preferentially stabilize structured folding intermediates under stress conditions and release them in more folding-competent states upon a return to low stress conditions.¹¹⁶

Mediation of allostery in protein structures

Allostery refers to the ability to influence events at a functional site by interactions at a distal site. A prominent example of allostery in structured proteins is hemoglobin, in which the binding of oxygen to one of the subunits induces conformational changes that are propagated to other subunits, raising their affinity for oxygen.¹¹⁷ Several IDRs have recently been shown to also enable allosteric coupling.^118,119 Theoretical description of allostery indicates that IDRs allow for much higher modularity and tunability than ordered regions, suggesting they could play a crucial role in allosteric regulation^120–124 (Fig. 7). Disordered regions can modulate how closely different domains are allosterically coupled, thereby optimizing allosteric responses^120,125 and in general, proteins with higher intrinsic disorder are more amenable to allosteric modulation.¹²⁶ This principle is illustrated by steroid hormone receptors (SHRs), which are allosterically regulated transcription factors.¹²⁶ IDRs in certain regions of these receptors can optimize allosteric responses. Each major domain of SHR reversibly binds ligands, and IDRs modulate allosteric coupling, or the degree of perturbation propagation, between domains. The effect that IDRs have on coupling between domains depends on the conformational and interaction coupling energies between the domains. Before ligand binding, the SHR exists in an ensemble of conformations. Upon the binding of a ligand (e.g. cofactors binding to the disordered N-terminal/AF1 regions), ensemble probabilities are redistributed according to these coupling energies, but the exact redistribution also depends on the probabilities of each state before its binding. Upon considering these factors, a model emerges in which both the ensemble nature of the SHR and the conformational flexibility of certain IDRs seem to account for a long-standing problem in the field, which is how SHRs can have distinct tissue and cell specific effects.

Figure 7.

The role of IDRs in mediating allostery. Disordered proteins can modulate allosteric mechanisms in proteins in a variety of ways (Box 1), and can reverse signal output by switching between positive and negative cooperativity, as in the case of E1A-CBP-pRB interactions (Box 2). Disruption of disorder-mediated allostery can manifest as disease states (Box 3).

At their extremes, IDPs may utilise allostery to reverse certain signals in addition to simply regulating signal strength as structured proteins do¹²² (Fig. 7). For example, using single-molecule Förster/fluorescence resonance energy transfer to study the binding processes of adenovirus early region 1A (E1A) oncoprotein, it was found that E1A–CBP–pRb interactions may exhibit either positive or negative cooperativity depending on which E1A interaction sites are exposed (through a biophysically complex mechanism, see Ferreon et al.¹²⁷ for details). Briefly, the availability of the E1A N-terminal region can change the cooperativity of CBP TAZ2 and pRb binding to E1A. When only the C-terminal end of E1A is available for interaction, there is negative cooperativity between pRb and CBP TAZ2 (possibly due to the partial overlap between the pRb and CBP TAZ2 binding sites in the E1A CR1 region), meaning that the formation of the binary E1A complexes (E1A-pRb and E1A-CBP) is favored over the ternary complex (E1A-CBP-pRb).¹¹⁹ However, when the N-terminal end of E1A is available, there is positive cooperativity for the interactions between CBP TAZ2 and pRb, leading to more ternary complex formation. The positive cooperativity in ternary complex formation could enhance E1A's function to exit the cell cycle and promote cell differentiation, whereas the negative cooperativity could support the role of E1A as a promiscuous molecular hub IDP by increasing the population of binary complexes and thereby enhancing their activities.¹¹⁹ This allostery modulation controls the cell cycle and transcription¹¹⁹ and sheds light on the efficiency with which viral E1A disrupts host cell activities.¹²⁸ Such cases of modulated allostery may be a common mechanism in IDP molecular hubs.²

Finally, disorder in loops and linker regions may help to facilitate signaling among a protein's domains, which is commonly viewed to occur via allosteric propagation, by helping to propagate strain energies.¹²⁹ For instance, in sortase, a disorder-to-order transition of a loop co-operatively affects another loop region.¹³⁰ The linkers in proteins can efficiently transfer information between sites on different domains, not solely because of their flexibility, but due to the encoding of successive conformational states within the linker sequence.¹³¹ For example, the glucocorticoid receptor (GR) fine tunes its target genes via allosteric effects propagated through its linker. Binding of the GR to DNA response elements induces conformational changes in the disordered lever arm, which in turn alters cofactor and coactivator binding sites in order to modulate glucocorticoid activity.^131,132

Such active regulatory roles for linkers may potentially account for why mutations in linker regions that alter sequence or length sometimes disrupt protein function. For example, certain mutations of the regulatory linker in Tec kinases, which are tyrosine kinases implicated in hematopoietic cancers, can lead to constitutively active kinase domains.¹³¹ Mutations in the allosteric architecture of IDRs are known to cause a number of genetic disorders. This is seen in a small disordered region in glucokinase (GCK) that allows the concentration dependent allosteric modulation of the turnover rate, which is important for glucose homeostasis. Mutations in GCK can result in persistent hypoglycaemic hyperinsulinemia in infants (Fig. 7). Small molecule therapeutic agents that affect allosteric communication mediated by disordered regions are being investigated to treat these conditions.¹³³

Conclusion

Recently, intrinsically disordered regions (IDRs) have emerged as fundamental units of protein function and regulation. Furthermore, IDRs have expanded our view of protein function by their involvement in areas once conceived of as the exclusive purview of structured domains. An explosive research effort into disordered proteins in the past decade has uncovered a plethora of functions for disordered regions; in this review, we have highlighted how IDRs can facilitate protein complex and assembly formation, orchestrate defined conformational transitions, and mediate allostery. We have also recently achieved a newfound understanding of the biomedical relevance of IDRs by discovering that IDR mutations can be of substantial consequence to disease causation by, for example, affecting functional features of IDRs (such as PTM sites or linear motifs), altering their disorder-to-order transitions, causing aggregation of IDRs or by changing protein half-lives. Since disordered regions can evolve rapidly, a major challenge is to make sense of disease mutation data within disordered regions from genome sequencing projects to uncover the relevant ones from the neutral mutations. The large number of significant biological discoveries that have emerged recently from the study of disordered proteins underscores their central importance in numerous aspects of cellular function. Given that the study of IDRs is in its infancy, it is indeed an exciting time for researchers to further investigate the properties, functionality and disease involvement of IDRs.