The case for intrinsically disordered proteins playing contributory roles in molecular recognition without a stable 3D structure.

1.1 MB·English

by  Uversky, Vladimir N.

Published:11January2013
©2013Facultyof1000Ltd
The case for intrinsically disordered proteins playing contributory
roles in molecular recognition without a stable 3D structure
Vladimir N. Uversky1,2 and A. Keith Dunker1*
Addresses:1DepartmentofMolecularMedicine,USFHealthByrdAlzheimer’sResearchInstitute,UniversityofSouthFlorida,Tampa,FL33612,
USA;2InstituteforBiologicalInstrumentation,RussianAcademyofSciences,142290Pushchino,MoscowRegion,Russia;3Centerfor
ComputationalBiologyandBioinformatics,IndianaUniversitySchoolofMedicine,Indianapolis,IN46202,USA
*Correspondingauthor:A.KeithDunker([email protected])
F1000BiologyReports2013,5:1(doi:10.3410/B5-1)
Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttribution-NonCommercialLicense
(http://creativecommons.org/licenses/by-nc/3.0/legalcode),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,
providedtheoriginalworkisproperlycited.Youmaynotusethisworkforcommercialpurposes.
Theelectronicversionofthisarticleisthecompleteoneandcanbefoundat:http://f1000.com/prime/reports/b/5/1
Abstract
The classical ‘lock-and-key’ and ‘induced-fit’ mechanisms for binding both originated in attempts to
explainfeaturesofenzymecatalysis.Forbothofthesemechanismsandfortheirrecentrefinements,
enzymecatalysisrequiresexquisitespatialandelectroniccomplementaritybetweenthesubstrateand
the catalyst. Thus, binding models derived from models originally based on catalysis will be highly
biasedtowardsmechanismsthatutilizestructuralcomplementarity.Ifmerebindingwithoutcatalysis
is the endpoint, then the structural requirements for the interaction become much more relaxed.
Recent observations on specific examples suggest that this relaxation can reach an extreme lack of
specific3Dstructure,leadingtomolecularrecognitionwithbiologicalconsequencesthatdependnot
onlyuponstructural andelectrostaticcomplementarity between thebindingpartnersbutalsoupon
kinetic,entropic,andgeneralizedelectrostaticeffects.Inadditiontothisdiscussionofbindingwithout
fixedstructure,examplesinwhichunstructuredregionscarryoutimportantbiologicalfunctionsnot
involving molecular recognition will also be discussed. Finally, we discuss whether ‘intrinsically
disordered protein’ (IDP) represents a useful new concept.
Introduction However, the lock-and-key model gives rise to some
Preparationsoftheenzymeemulsincleaveb-glycosides, interesting questions. For enzymatic transfer of a
but not a-glycosides, while preparations of invertase phosphate from ATP to an –OH acceptor via a lock-
cleave a-glycosides, but not b-glycosides. From these and-key mechanism, why doesn’t an –OH from water
observations, Emil Fischer suggested in 1894 that the simply outcompete the –OH from the acceptor, thus
enzymeandsubstrateexertamutualeffectoneachother leadingtoATPhydrolysisinsteadofphosphatetransfer?
likealockandkey[1-2].Thus,thelock-and-keyhypothesis Daniel Koshland suggested that sequestering the reac-
was originally derived from studies on catalysis, not tants from water via small conformational changes that
molecularrecognition. hecalled“inducedfit”wouldsolvetheproblem,especially
sincewateritselfwouldbetoosmalltoinducetheneeded
Ashorttimelater,in1897,PaulEhrlichappliedthelock- conformationalchanges[5-6].Inducedfitistheformation
and-key hypothesis to the problem of how an antibody ofanencountercomplexbetweenmoleculesinconforma-
binds specifically to a particular antigen [3-4]. In this tionsdistinctfromtheirfinalconformations,followedby
example, the interaction results in specific binding, not mutual structural adjustment until the intimate fit
catalysis. Thus, Ehrlich converted the lock-and-key betweenthetwopartnersisrealized.Subsequentstructural
hypothesisfromexplainingenzymespecificitytoexplain- studies on a number of enzymes revealed substantial
ingprotein-basedmolecularrecognition. conformational changes upon substrate binding, which
Page1of12
(pagenumbernotforcitationpurposes)
F1000Biology Reports2013, 5:1 http://f1000.com/prime/reports/b/5/1
areconsistentwithinducedfitbutcouldalsobeexplained While enzyme catalysis very often occurs without
byothermechanisms[7]. significant disorder, for several enzymes disorder-to-
order backbone rearrangements is a feature. In these
Besides induced fit, an alternative model called ‘con- cases,thesubstratebindstopartoftheactivesitethatis
formational selection’hasbeenproposed toexplain the structured and then a disordered region folds onto the
association between a macromolecule and a flexible substrate typically including residues involved in the
ligand [8]. In conformational selection, the ligand catalysis(andoftentoexcludewater)andthenthesame
assumesanensembleofconformations,andtheprotein region unfolds againto releasethe product [15-16].
binds to the conformation that gives the best fit to the
binding site. Conformational selection and induced fit Recent studies on enzyme reaction mechanisms using
areoftendiscussedasthetwoextremepossiblemechan- faster methods than previously available have revealed
ismsforthebindingtoaflexibleligandsuchasanIDPto multiple steps in which conformational changes in the
astructured partner. substrates are complementary to multiple conforma-
tional changes in the enzymes. These multiple substeps
For enzymes, the lock-and-key and the induced-fit involve small amounts of energy and only slight
mechanisms lead to different kinetic equations, which conformational changes, resulting in an overall better
can be considered as parts of a unified reaction cycle fit and larger energy decrease in the transition state
[9]. Analysis of enzyme kinetic data in terms of this energy than could likely be achieved by one large
unified reaction cycle suggests that substrate and conformational change[7].
enzyme concentrations determine whether conforma-
tional selection, induced fit, or a mixture of the two The importance of this background for the current
underlies the reaction [10]. The key point here is that discussion is that binding for the catalytic event is
notonlyisthestructureofthefinalcompleximportant necessarily highly structured: the enzyme has to bind
but also the mechanism used to achieve the final more tightly to a molecular intermediate form than to
complex. the undistorted ground state or to the product(s) [7,9],
and so the physiochemical requirements for catalysis
As for non-catalytic molecular recognition by flexible dependoninteractionswithexquisitestericandelectro-
proteins, many DNA-binding proteins contain regions static complementarity between the enzyme and the
that undergo disorder-to-order transitions upon bind- ligand. However, we argue that if the catalytic require-
ing to their DNA partners [11]. Georg Schulz collected mentfortheinteractionisdroppedandmerebindingis
several of these examples and suggested in 1979 that the biological function, then the molecular recognition
such a disorder-to-order binding mechanism could be eventitselfoughttobecomemorerelaxed.Thequestion
helpful for some biological interactions by enabling here is whether molecular recognition in the absence of
the combination of relatively high specificity and low catalysiscanbecomesorelaxedthatfixed3Dstructureis
affinity [12]. Thermodynamic studies of several no longer arequirementfor binding.
protein-DNA interactions gave further support for
such disorder-to-order transitions, in which the Effects on binding without specific
authors, Ruth Spolar and Tom Record, described the structure formation
binding process as “coupled-binding and folding,” or Suppose an unstructured protein binds to a partner
in one place “extreme induced fit” [13]. Just as for without forming structure. How would one evaluate
enzyme catalysis, it has been suggested that conforma- such a complex in the absence of structure? How does
tional selection and coupled binding and folding one determine the biological significance of such an
(sometimes called induced fit) represent the extreme interaction if one were found? To address these issues,
possibilitiesforbinding,andjustasforenzymes,ithas several examples are presented below in a progression
beensuggestedthateithermechanismorevenmixtures from the most tothe leastamount ofstructure.
of the two mechanisms may be involved for any
particular interaction [14]. Inthefirstexample,aregionofintrinsicproteindisorder
bindstoapartnerviaadisorder-to-ordertransitionwith
As with the lock-and-key hypothesis, induced fit was partofthestructureremainingunstructured.Suchevents
originally developed to explain specific features of are very common. In one study of short segments that
enzyme catalyzed reactions, and this concept was later bind to protein partners, out of 372 binding segments
converted to a molecular recognition mechanism when containing 10,434 residues, 13% of the residues
the underlying ideas were applied to non-catalytic remained unstructured after binding as determined
bindinginteractions. fromthelackofelectrondensityinthecrystalstructures.
Page2of12
(pagenumbernotforcitationpurposes)
F1000Biology Reports2013, 5:1 http://f1000.com/prime/reports/b/5/1
This is substantially higher than the 7% figure for argued that such a mechanism still uses structure for
disordered residues observed for a set of 848 structured binding,justthattherearemultiplestructuresofsimilar
protein monomers [17]. Are these disordered regions energy that can interconvert over timescales required to
merelypresentordotheycontributetothebindingfree collecteithertheX-rayorNMRdata.Wewouldgivethe
energy? counter argument that adopting multiple, different,
interconverting structures is distinctly different from
In a few of the interactions containing IDP regions, becomingstructureduponbinding.
studieshavebeencarriedoutontheeffectsofremoving
all, or fractions, of the structurally unobserved regions A thirdpossibility is thatelectrostatic interactions could
(reviewedin[18]).Removalofsuchflankingdisordered occur over long range without the formation of any
regions has been shown to result in both positive and specific structure at all. Such a result would truly be
negativechangesinthefreeenergyofbinding(Fuxreiter, interaction without a specific complementary structure.
personalcommunication). Considerationofelectrostaticinteractionsleadsustothe
secondexample.
Inonecase,thedisorderedsplicingfactor1(SF1)bindsto
the large subunit of the U2 small nuclear RNA auxiliary The second example involves the interaction between
actor (U2AF), the SF1 segment binds by a motif of Sic1 and CDC4. In this example, the Sic1 protein is
10 residues with K of 23.8nM, and removal of residues unstructured but contains nine similar motifs that are
d
not in physical contact with U2AF reduces the K to well separated along the sequence. Each of these motifs
d
55.6nM.Onthe other hand, the full-lengthSF1 binding contains a central serine or threonine that can become
withU2AFhasaK of11.8nM,indicatingthattheflanking phosphorylated[22].EachphosphorylatedSic1motifis
d
but unstructured regions contribute significantly to bind- recognizedbytheCDC4protein,whichrecruitsSic1for
ing free energy [19]. These data demonstrate binding ubiquitination, thereby targeting it for degradation in
energywithoutspecificstructureformation.Evidently,the late G1 phase, an event necessary for the onset of DNA
steric complementarity required for enzyme catalysis has replication [23]. Replacing different numbers of serines
becomeconsiderablyrelaxedformereassociation. or threonines with alanines leads to loss of viability if
fewerthansixphosphorylationsitesremain[22].Invitro
In the case above, the disordered region exhibits a studiesshowthatindividualphosphorylatedmotifsbind
measurablefreeenergyofassociationwiththeremainder weakly,buttheaffinityandthesteepnessofthebinding
oftheproteinwithouttheformationofspecificstructure. isotherm increases as more sites are phosphorylated.
There are at least three alternative mechanisms as When any six sites are phosphorylated, the affinity
discussedbelow. sharplyincreasesandthebindingisothermissosharpit
resembles an on-off switch[22].
One possibility is that the unobserved residues alter the
polypeptide intheunboundstate.Theremovalofthese The CDC4 molecule contains just one site (!) that
residues would then affect the binding constant in the associateswiththephosphorylatedmotif.Ifthereisonly
boundstate.Forexample,ifthebindingmechanismwere onebindingsite,howdotheadditionalsitesincreasethe
to depend on the conformational selection mechanism, affinity and sharpen the binding isotherm? The main
thentheunobservedresiduescouldreducetheamountof effectseemstobeelectrostaticinteractionsthatkeepthe
time spent in a binding-competent conformation by a phosphorylatedmotifsneartothesinglebindingsite,so
variety of mechanisms, thus reducing the on-rate, and that,assoonasonehopsoff,anotherisnearbytohopon
loweringtheoverallaffinity[8]. [24]. A mathematical description of these effects has
been developed, and the resulting mean field statistical
A second possibility is that there are several different mechanicalmodelfortheelectrostaticinteractionsgives
conformationsthatenablethedisorderedproteintobind reasonably good agreement with the experimental data,
to the surface via short-range contacts. Having several including both estimates of the threshold number of
suchbindingstructurescouldresultinmissingstructure phosphorylation sites for binding, and also including
in X-ray experiments due to incoherent scattering or to experimental affinities of CDC4 for Sic1 fragments with
missing data in NMR experiments due to exchange different total charges [25].
broadening. Indeed, several examples of the same
disorderedregionchangingstructuretobindtodifferent The Sic1-CDC4 and the SF1-U2AF examples discussed
surfaces have been observed [20-21]. All that would be above demonstrate that non-structured protein can
neededistohavethealternativebindingmodescloseto contribute very significantly to binding free energy. As
each other on the same binding surface. It could be reviewed by Tompa and Fuxreiter [18] many additional
Page3of12
(pagenumbernotforcitationpurposes)
F1000Biology Reports2013, 5:1 http://f1000.com/prime/reports/b/5/1
examples have been observed in which unstructured datasupportingtheexistenceofencountercomplexes,the
regions contribute both positively and negatively to the question then becomes whether it is possible for
binding free energy between an IDP and its protein encountercomplexestobelong-lived.
partner. We anticipate that similar results will be found
for IDPs binding to nucleic acids if such examples have The above discussion has an interesting parallel to the
notbeenfoundalready.Theseobservationsclearlymake moltenglobuleconcept.Thecapabilityofapolypeptide
the point that unstructured regions of protein can chain to adopt partially folded intermediates was first
contribute to binding free energy without becoming proposed as a part of the frame-work model of protein
structured in the finalcomplex. folding,withacollapsed, butinternallydynamic,short-
lived transient intermediate protein form in protein
While the CDC4 has a single site for binding one folding [29]. In later studies it became apparent the
phosphate group and the flanking amino acids on Sic1, intermediate with a collapsed, but internally dynamic,
which has multiple sites of phosphorylation along with structurewasshowntobeastableformforsomeproteins,
sequence-similar motifs for the flanking amino acids in followingslightstructuraldestabilizationusingavariety
order to fit onto the binding site on CDC4, the oftreatments[30].Next,thisintermediatewasshownto
mathematical model explaining the increased affinity betransientlypopulatedattheearlystagesoftheglobular
arising from the remainder of the molecule does not protein folding [31]. Finally, certain proteins were
require specific structure, merely electrostatic interac- suggested to form molten globules in their functional
tions between a dynamic IDP and its partner. The states [32]. So, will encounter complexes turn out to
question then arises whether such non-specific interac- exhibit a similar progression, being first recognized as
tions could lead to specific protein association without transients, then as stable forms under some particular
the formation of any long lasting complementary conditions, then as stable forms under physiological
interfaces between an IDP and its partner. In other conditionsforsomeproteinsequences?Timewilltell.
words,isthereamechanismbywhichanIDPcouldbind
to apartner without itself formingspecific structure? Has any protein-protein interaction without
any structure formation ever been observed?
Binding without structure formation: A number of IDPs interact with each other to form
is it possible? dimersthatsometimesexhibitverysimplefoldssuchas
The various experimental observations above could be leucinezippers[33],andthatatothertimesexhibitmore
combined to yield a model with an overall electrostatic complexfoldssuchashelicalbundles[34,35].Forafew
attraction between anIDPanditspartner,coupled with examples, the IDP dimers are highly dynamic with only
several local docking interactions that rapidly convert localizedregionsthatshowevidenceofstructureforma-
from one to another. Such an interaction mechanism tion [36,18]. All of these examples, more or less, fit the
could leadtoaspecific associationbetween anIDP and standard view of coupled binding and folding or
its partner without the formation ofstablestructure. inducedfit.
Let’sconsiderthispossibilityfromamoretraditionalview. However, Sigalov and co-workers have reported homo-
Proteincomplexformationtypicallyinvolvesatleasttwo dimerization of the cytoplasmic region of the T cell
steps. Upon meeting, an encounter complex is formed, receptor zeta subunit that is not accompanied by
either proceeding towards the final complex or towards measurable shifts in the CD spectra [37] nor by measur-
dissociation.Evidencesuggeststhatencountercomplexes able chemical shift changes in the NMR spectra possibly
are dominated by electrostatic forces, but hydrophobic suggesting association without structure formation [38].
interactionscanalsoplayarole[26].Aninterestingvariant Other explanations for this observation include the
is to consider the subsequent events in terms of game possibility that the CD spectra might not be sensitive
theory, according to which the interacting partners enoughtopickupformationofhighlylocalizedstructure
continually affect the conformational landscapes of each or technical problems such as exchange broadening for
otherinsuchawaythatconsecutivestepsdependonprior the same protein regions before and after dimerization
stepsuntilthefinalcomplexisformed[27,28].Whatifthe couldobscuretheformationofstructureduringtheinter-
folding funnel for the overall complex were rather flat, action, so further work needs to be carried out to prove
with no energy minimum corresponding to one specific that the molecular association is truly occurring without
structure? In this case, the moves and counter-moves theformationofproteinstructure.Furthermore,theinitial
would continue endlessly, leading to a long-lived, report was in 2004 and other laboratories have not yet
dynamically fluctuating encounter complex that, of reportedsimilarfindings,andthisabsenceofconfirmatory
course, could dissociate at any moment. Since there are data adds uncertainty to Sigalov’s interpretation of his
Page4of12
(pagenumbernotforcitationpurposes)
F1000Biology Reports2013, 5:1 http://f1000.com/prime/reports/b/5/1
observations.However,giventheresultsforanumberof random-walk search for binding site, but more detailed
complexes in which IDP regions have been shown to studiesareneededtoconfirmthismodel[47].Comparing
contribute to the overall free energy of the protein- theorthologousvoltagechannelsinspermfromdifferent
proteininteraction[24,25,39,40]andgivenevidencefor mammalsshowsthattheexoncorrespondingtothechain
the existence of encounter complexes along with the region has significant length variability that arises from
various models to explain their interaction without insertions and deletions (indels). The number of indel
structure[26-28],itisouropinionthatSigalov’sresults substitutions is 5 to 8 times higher than is generally
cannotbedismissedoutofhand.Whileourownviewis observed in genomic studies and indels within the
that specific protein-protein interactions likely require disordered regions are considerably longer than average
at least some localized structure at a key contact point indels,suggestingthatpositiveselectionisoccurring[48].
suchasobservedfortheSic1-CDC4interaction,forma- Theauthorssuggestthattheobservedchain-regionlength
tionofproteincomplexeswithhighlytransientstructure variability, which can affect sperm motility, may be an
formationoughttobetakenasapossibilityuntilruled important determinant in sperm competition, thus
outbyfurtherstudies. accountingforthepositiveselection.
Flexible linkers, flexibility, and Another very interesting example is provided by the
molecular recognition entirely unstructured ~200 residue kinase inhibitor
Direct involvement in molecular recognition is not the p27kip1 [49], which plays a key role in the control of
only typeof biological function carried out by proteins. eukaryotic cell division by the inhibition of several
In addition, disorder can affect molecular recognition different cyclin-dependent protein kinases (Cdks)
withoutdirectinvolvementinthebindinginterface.Two [50,51]. For one such complex, ~70 residues of p27kip1
aspects will be discussed here: flexible linkers and free wrap around the outside of a dimer of a cyclin and its
energyintheunboundstate.Theseexamplesemphasize cognate Cdk [52]. The flexibility of p27kip1 allows it to
thatthefinal3Dstructureofthecomplexisnottheonly associate and dissociate segmentally [50,53], thereby
biologically important aspect of molecular recognition, providingopportunitiesforregulationandcontrol.More
and that the on-rates, off-rates, and conformational specifically, segmental dissociation enables phosphory-
changes enabling association and dissociation can also lationofp27kip1’sY88byanon-receptortyrosinekinase,
contribute to biological function. leading to the exposure of the Cdk’s active site, which
then phosphorylates p27kip1’s T187. This second phos-
Flexible linkers canenablecombinations ofinteractions phorylation provides a signal for ubiquitination, which
thatwouldnotbestericallyallowedbycompletelyrigid then leads to digestion of p27kip1 via the proteasome,
structures.Forexample,calmodulinusesaflexiblelinker whichinturnpromotescellcycleprogression[50,51].
between two domains to wrap around its target helix
[41]. Formation and dissociation of the calmodulin- Thispathwayofphosphorylationleadingtoubiquitina-
targetcomplexwouldsimplynotbepossiblewithoutthe tion, in turn leading to degradation, is very commonly
flexibility of the linker. Multiple zinc fingers connected observed in eukaryotic cells as a means to remove or
by flexible linkers enable certain transcription factors to deplete the levels of key regulatory proteins [54]. Both
wrap up their target DNA molecules [42]. Again such phosphorylationandubiquitinationcommonlyoccurin
complexes could not form or dissociate without the regions of disorder [55-57], and having a sufficiently
flexibilityof the linkers. longregionofdisorderappearstobeimportantforentry
into the proteasome [58].
Flexible linkers can also affect rates of association and
affinities.Aparticularlyinterestingexampleisprovidedby Asafinalexampleoftheconsequencesofflexiblelinkers
voltage gated ion channels. Such channels exhibit three on molecular recognition, Kuriyan and Eisenberg [59]
states: closed (voltage sensitive), open, and inactive argue that the proximity brought about by flexible
(voltage insensitive). While in the open state, ion flow linkers brings about an amplification of the effects on
through the open channel collapses the cell membrane one domain of random mutations in the colocalized
potential.Thus,theamountoftimespentintheopenstate domain.Throughnaturalselection,thisamplificationof
isimportantforthebiologicalfunctionofthechannel.The effectsbyproximityleadstospecificinteractionsandtoa
mechanismofclosingisviaa‘ballandchain’,wherethe startlingvariety ofcomplex allostericcontrols.
terminus of a disordered region closes the open channel
by a binding event [43-44]. Lengthening of the ‘chain’ Asforeffectsarisingfromthefreeenergyoftheunbound
slows the closure, shortening the chain speeds it up state,theunbound,flexiblestateisthestartingpointfor
[45,46],suggestingthepossibilitythattheballundergoesa themanyexamplesinvolvingdisorder-to-ordertransitions
Page5of12
(pagenumbernotforcitationpurposes)
F1000Biology Reports2013, 5:1 http://f1000.com/prime/reports/b/5/1
upon binding to their partners, which in turn might be IDPs/IDRs over ordered proteins and domains is the
structuredoralsodisordered.Whilestructureaccountsfor ability of one disordered segment to bind to multiple
thefinalrecognition,therateofassociationordissociation partners due to its ability to adopt different conforma-
might also be important for biological function: muta- tions inthe boundstate.
tions that do not involve any of the interacting residues
butthataffectthefreeenergyintheunboundstatewould Recent analysis revealed that the C-terminal recognition
affectthefinalbindingconstant[60].Suchanalterationin domain of p53 is not a unique entity and several other
free energy by mutation could also be viewed as an IDPs can be engaged in the formation of multifacial
alteration in the underlying protein ensemble. This complexes [21]. These examples highlight the transient
ensemble view was recently used to explain mutational nature of the intrinsic disorder-based interactions and
effectsonbindingeventsthatleadtoallostery[61,62].In emphasize the extreme adaptability of IDPs. In general,
this view, rather than affecting a Rube-Goldberg type complexes involving disordered proteins are drastically
pathway underlying allostery, the mutation could be differentfromthecomplexesformedbyorderedproteins.
affecting the conformational ensemble and hence the
allostery. In a recent study, the binding constants of The case of the elastomeric proteins
associations between structured proteins were shown to An important example of biologically important yet
have a correlation with the measured off-rates and to be disorderedcomplexesisthatoftheelastomericproteins,
rather independent of on-rates. On the other hand, which have a wide range of crucial functions and are
binding constants of associations involving an initially involved in various unique mechanisms where they
disorderedproteinwereshowntohaveacorrelationwith provide the high efficiency elastic recoil necessary to
the on-rates [63]. A possible mechanism here is that a undergo reversible deformation [64]. These proteins are
reducedfreeenergyintheunboundstate(ascomparedto found in the human arterial wall, the capture spiral of
the bound state) would be expected to both reduce the spiderwebs,thehingeofscallopshells,andareinvolved
final affinity and to slow the on-rate. These disorder- in the jumping mechanism of fleas. Since it has been
dependent effects on binding kinetics and affinity values suggested that the elastic recoil of proteins is due to a
don’tdirectlyinvolvethemoleculardetailsofthebound combination of internal energy and entropy, and since
state, but are nevertheless likely to lead to important the dominant driving force in this recoil process is the
biologicalconsequences. increased entropy of the relaxed state relative to the
stretchedstate, itwaspointedoutthatintrinsicdisorder
One-to-manybindersandmultifacialcomplexes plays a crucial role in the function of these rubber-like
IDPs are known to participate in one-to-many and elastomericproteins.Inagreementwiththishypothesis,
many-to-oneinteractions,whereoneIDPoroneintrinsi- the functional aggregates of these proteins were
cally disordered region (IDR) binds to multiple partners described as intrinsically disordered or fuzzy complexes
potentiallygainingvery different structuresinthe bound with high polypeptide chain entropy [64]. Although
state,orwheremultipleunrelatedIDPs/IDRsbindtoone these disordered elastomers possess a broad range of
partner,potentiallygainingsimilarstructuresinthebound sequence motifs, mechanical properties and biological
state[20,21,60]. functions, all of them are dramatically enriched in
proline and glycine residues [65]. This P and G
The one-to-many binding mechanism is especially enrichment plays a central role in defining the elastin-
interestingsinceitmightgeneratemultifacialcomplexes, like properties of disordered elastomers that form
where the same region of an IDP can be engaged in disordered functional aggregates (including disordered
interactionwithmultipleunrelatedpartnersandbeable fibers), and clearly separates all elastomeric proteins
to fold into very different conformations in the bound from the amyloidogenic proteins/peptides that form
state. One of the illustrative examples of such one-to- insolubleamyloid-like fibrilscharacterizedbythecross-
manybindersisp53,asingleC-terminalregion,whichis b-sheet structure with the b-strands perpendicular to the
known to interact with at least four different partners fiberaxis.Figure1illustratesthisobservationbyshowing
[20].Theaminoacidsinvolvedineachinteractionshow a two-dimensional diagram that relates the P and G
a significant overlap and no two of these interactions contents of natural elastomeric protein domains and
could exist simultaneously. Furthermore, the same proteinsthatwereexperimentallyshowntoformamyloid
residues adopt helix, sheet, and two different irregular fibrils[65].Inthisplot,aclearseparationisseenbetween
structures when associated with the different partners. elastomeric and amyloidogenic sequences. These data
Finally,thesameaminoacidsareburiedtoverydifferent provide a clear explanation of why elastomeric proteins
extent in each of the molecular associations [20]. These are expected to form disordered fibers: their amino acid
results show that one of the functional advantages of sequencesareenrichedinthestructure-breakingGandP
Page6of12
(pagenumbernotforcitationpurposes)
F1000Biology Reports2013, 5:1 http://f1000.com/prime/reports/b/5/1
Figure1. Atwo-dimensionalplotcorrelatingprolineandglycinecontentforawidevarietyofelastomericandamyloidogenicpeptides
ElastomericproteinsarecharacterizedbyhighPGcontentandarelocatedintheupper-rightpartofthisplot.Contrarily,amyloidogenicpeptidesare
characterizedbylowPGcontentandthereforearelocatedintheleftbottomcorneroftheplot.Thecoexistenceregion(shadedingray)containsPandG
compositionsconsistentwithbothamyloidogenicandelastomericproperties.Elastomericproteins,includingthedomainsofelastin,majorampullate
spindroin(MaSp)2,flagelliformsilk,theelasticdomainsofmusselbyssusthread,andabductin,appearaboveacompositionthreshold(upperdashedline).
AmyloidogenicsequencesareprimarilyfoundbelowthePG-threshold,alongwithrigidlizardeggshells,tubulliformsilk(TuSp1),aprotectivesilkforspider
eggs,andaciniformsilk(AcSp),usedforwrappingprey.Thecoexistenceregioncontainsamyloid-likepeptidesaswellastheelastomericadhesiveproducedby
thefrogNotadenbennetti,thePEVKdomainsoftitin,wheatgluteninprotein,andthestrongestspidersilks,namelyMaSp1andminorampullatespindroin
(MiSp).Figurereproducedfrom[65].Abbreviations:AcSp,aciniformsilk;MaSp,majorampullatespindroin;MiSp,minorampullatespindroin;TuSp1,
tubulliformsilk.
residues and therefore are naturally selected not to form Subsequentlymanyregionsofproteinshavebeenfound
lengthyorderedsegments. tolack3Dstructureunderapparentlyphysiologicalcon-
ditionsintheabsenceofabindingpartner,andNMRhas
The ‘IDP’ concept: necessary or not? revealed many additional proteins that appeared to be
To our knowledge, the first report of a significant-sized entirely unstructured. Release 6.0 of DISPROT [68] lists
regionofmissingelectrondensitywasinthestructureof 667 proteins with 1,467 disordered regions that are
the extracellular nuclease from Staphylococcus aureus, as associatedwithbiologicalfunction;thissetincludes112
describedin1971[66].Twosuchregionswereobserved whollydisorderedproteins.
and the authors suggested that these were “disordered”
andbothhighlysolvatedandhighlymobile.Theauthors When such disordered proteins and regions were first
also reported the extreme trypsin sensitivity of these discovered, the standard view was that they were some-
regions. Even earlier, optical rotatory dispersion was howlikelytobestructured,exceptthattheyweredenatured
usedtoidentify afewproteinsthatappearedtobe fully during isolation or lacked a critical partner that got lost
unstructuredunderapparentlyphysiologicalconditions, duringisolation.Indeed, someofthescientists who were
andfromsuchstudiesoneauthorsuggestedthatthereis involvedincarryingoutearly,keyworkontheseproteins
acategoryof “disordered proteins” [67]. [69,70] tell us that when they were graduate students
Page7of12
(pagenumbernotforcitationpurposes)
F1000Biology Reports2013, 5:1 http://f1000.com/prime/reports/b/5/1
doing the work in the citations just given, they repeated started studying these proteins, one of us chose the
protein purification multiple times using different proto- descriptor ‘natively unfolded protein’ [72,74] and the
cols because neither they nor their advisors could believe otheroneofuschose‘disorderedprotein’[71].Bothofus
that unstructured proteins could be carrying out the bio- considered but rejected ‘flexibility’ as a descriptor. Our
logical functions being observed (Daughdrill, Kriwacki, views regarding ‘flexibility’ were that this term is applied
personalcommunications). to both structured and unfolded proteins but describes
entirelydifferentprocessesforthetwoproteinforms.That
A key development in the study of these proteins in our is, for structured proteins, flexibility refers to periodic or
view has been the development of disorder predictors slightly aperiodic motions as atoms oscillate about their
that used amino acid sequence or composition as inputs equilibrium positions, with higher flexibility referring to
[71,72]. These predictors give results much better than largeramplitudesfortheoscillations.Duringtheseoscilla-
expected by chance, leading to the conclusion that, to a tions,theoverallshapeofthemoleculechangesverylittle.
considerable degree, lack of structure is encoded by the Ontheotherhand,forunfoldedproteins,flexibilityrefers
aminoacidsequence.Inotherwords,disorderedproteins to massive changes in backbone and side chain dihedral
andregionshaveaminoacidcompositionsthataredistinc- angles, leading to large-scale changes in overall shape.
tlydifferentfromthecompositionsofstructuredproteins. Given that flexibility has entirely different meanings for
Thus,thisobservationlinksdisordertotheDNAsequence, structuredandunstructuredproteins,usingthesameterm
leading to an extension of the standard Central Dogma. for both protein forms tends to blur the very large differ-
Thatis,thestandardCentralDogmaisgivenbythefollow- encesinbehavior.
ingsteps:(1)DNAsequence,(2)RNAsequence,(3)Protein
sequence, (4) structure, and (5) function. The extension, Withregardtoreplacingdisorderwitheitherflexibilityor
however,isgivenbythefollowingsteps:(1)DNAsequence, PWPs, our view on thesesuggestions can be summarized
(2) RNA sequence, (3) Protein sequence, (4) intrinsically byawell-knownphrase:“What’sinaname?Thatwhichwe
disorderedensemble,and(5)function. callarosebyanyothernamewouldsmellassweet”.[75].
Inourview,aportionof‘foldingcode’(andsometimesa The fundamental distinction between folding first and
significant part of it) that defines the ability of ordered then binding as compared to concomitant binding and
proteins to spontaneously gain a unique biologically folding is reflected by the marked differences in the
activestructureismissingforIDPs.Thismissingportion amino acid composition between the two types of
ofthe‘foldingcode’(orapartofit)canbesupplemented proteins. Indeed, early theoretical studies on protein
by binding partner(s). As a result, a key difference folding suggested that whether a protein folds or not
between structured and disordered proteins is that the depends on its amino acid composition, and if it has a
formerfoldfirstandthenbindtotheirpartnerswhilethe composition commensurate with folding, then the
latter remain unfolded until they bind their partners. sequence patternsdetermine which fold is favored [76].
Other researchers suggest that this distinction makes no We[72,77,78]andothers[79]havepointedoutthatthe
difference. To emphasize that this is a distinction aminoacidcompositionsofIDPsareentirelyconsistent
withoutadifference,recentlytheterm“proteinswaiting with their lack of folding. The high polarity of these
for partners” (PWPs) was proposed as an alternative to sequencesisverymuchalongthelinesofthesuggestions
the term “disordered” [73]. Above we describe many from the early theoretical studies for sequences that
examples in which disordered proteins have functions would fail tofold[76].
other thanpartner binding,so this term cannot be used
for all types of disordered protein. Also, it is not clear Researchers who question the existence of IDPs in vivo
howthePWPsconceptwouldapplytoexamplessuchas oftenpointout thatcellshaveelaboratemechanisms to
the C-terminus of p53 described above, in which the deal with misfolded proteins and that disordered
same disordered region assumes four different confor- proteins would be cleared by these mechanisms. Thus,
mations when binding to four different partners. This they conclude that disordered proteins cannot exist in
examplesuggeststhepossibilitythatthesamedisordered cellsexcepttransiently.Inouropinion,suchsuggestions
region canswitch fromonepartnertoanother,withthe are misguided for three reasons. First, the misfolded
disordered region changing its shape as it changes its protein response is confined to the endoplasmic
partner. This sort of behavior seems to be much more reticulum, so it is unclear to what extent other parts of
dynamicthan just ‘waiting for apartner’. the cell are under surveillance for protein misfolding.
Second, such suggestions are based on the assumption
‘Flexibility’isoftenproposedtodescribemotionsinpro- thatanaminoacidsequencethatiscommensuratewith
teinscoveringbothfoldedandunfoldedforms.Whenwe folding (e.g. a high level of hydrophobic groups and
Page8of12
(pagenumbernotforcitationpurposes)
F1000Biology Reports2013, 5:1 http://f1000.com/prime/reports/b/5/1
aromatics) but that is unfolded or misfolded and that a ordered proteins to maintain their disordered structure
highly polar sequence that has evolved to be unfolded under physiological conditions, which is necessary for
willbothbereadilyrecognizedbythe‘unfoldedprotein theirfunctions.
responsesystems’andclearedbythecell.Wethinkthatit
is equally likely that IDPs and regions have evolved to Until recently, disordered proteins or regions were
avoidtheunfoldedproteinresponsebyhavingsequences largely ignored. However, each week there are now
notrecognizedbythosesystems.Whereistheproofthat about 17 publications (estimated by Caron Morales
these systems recognize all types of sequences? Indeed, fromthelast10weeksusingDisProt’sstandardkeyword
themechanismbywhichtheunfoldedproteinresponse searchofPubMed)thatfocusonthecharacterizationand
recognizes its substrate proteins is currently ambiguous, functionsoftheseproteins.Itcouldbereasonablyargued
andwhether,orwhich,IDPsareclearedbythissystemis that there is nothing new in the IDP concept, that all of
not yet understood (Ron Wek, personal communica- thecurrentviewsoftheseproteinsfollownaturallyfrom
tion).Also,in-cellNMRdatademonstratestheexistence long-heldviewsofproteinstructureandfunction.From
andstabilityofIDPsevenwheninsidebothprokaryotic a chemistry and physics point of view, that is certainly
and eukaryotic cells [80-90]; why doesn’t the misfolded true. However, the fact that these proteins were largely
protein response rapidly remove these proteins? Third, ignored previously and now they are being actively
to use humans as an example organism, some of our studied suggests that developing the IDP concepts has
proteins turn over with half-lives of less than a minute served the useful purpose of bringing attention to these
whileothersexistforthelifeofthehuman,givingalmost proteins and to understanding the biological functions
eightordersofmagnitudedifferenceinproteinstability. with which they are involved. The reader is invited to
The stability of each protein is an important aspect of make up his or her own mind regarding the utility, or
its biology. Studies on the relationship between protein lack thereof, of the IDP concept.
lifetimes and protein disorder suggest that some dis-
ordered proteins have long lifetimes, but, on the other Abbreviations
hand,theshort-lifetimeproteinsarerichindisorder[91]. Cdk, cyclin-dependent protein kinase; IDP, intrinsically
Perhapsbothadisorderedregionandaparticularsignal, disordered protein; IDR, intrinsically disordered region;
suchasthePESTmotif[92,93],areneededforaprotein PWPs,proteinswaitingforpartners;SF1,splicingfactor1;
to exhibit a short half-life. The important point here is U2AF, U2smallnuclearRNAauxiliaryactor.
that disordered regions likely help to promote short
protein life-times as animportantaspect of this biology Disclosure
or to put it another way, life-time modulation is an The authors declare that they haveno disclosures.
importantbiologicalfunctionofdisorderedproteins.
Acknowledgements
Is there any evidence that disorder is a product of ThisworkwassupportedinpartbythegrantEF0849803
evolution?Studiesontheevolutionofstructuredproteins from the National Science Foundation (to A.K.D and
suggestthatregionsofproteinswithahighpackingdensity V.N.U.) and the Program of the Russian Academy of
showfeweraminoacidchangesoverevolutionarytimeas Sciences for the “Molecular and Cellular Biology” (to
comparedtoregionswithlowerpackingdensities.Thatis, V.N.U.). Caron Morales is thanked for estimating the
if positional mutation rates (expressed as Shannon’s numberofpublicationsappearinginPubMedeachweek
entropy)areplottedversustightnessofpacking(expressed on intrinsicallydisordered proteins.
as 1/density), virtually a straight line is observed with
lower packing densities showing higher sequence varia- References
bility(seeFigure3in[94]).Ofcoursemanyyearsearlierit 1. Fischer E: Einfluss der configuration aurf die wirkung der
enzyme.BerDtChemGes1894,27:2985-93.
waspointedoutthattheresiduesinthecoreofaprotein
2. LemieuxRU,SpohrU:HowEmilFischerwasledtothelockand
familyexhibitfewermutationsoverevolutionarytimeas
keyconceptforenzymespecificity.AdvCarbohydrChemBiochem
comparedtoresiduesonthesurfaceofthesameproteins 1994,50:1-20.
[95]. Thus, mutation rates are strongly correlated with 3. Ehrlich P: Die werthbemessugn des diphtherieheilserums und
structuralfeaturesofproteins.Ifthemutationratesofthe derentheoretischegrundlagen.KlimishesJahrbuch1897,6:299-326.
structured and disordered regions of proteins are com- 4. TanfordC,ReynoldsJ:Nature’sRobots.InOxford:OxfordUniversity
Press;2001:176-87.
pared, in general (but not always),the mutation rates of
5. Koshland DE, Jr: Enzyme flexibility and enzyme action. J Cell
thedisorderedregionsaremuchhigherthanthemutation
CompPhysiol1959,54:245-58.
ratesofthestructuredpartsofthesameproteins[96-98].In
ourview,theseobservationsaresimplyexplainedassum-
ing that disordered proteins evolve differently from 6. KoshlandDE,Jr:Crazy,butcorrect.Nature2004,432:447.
Page9of12
(pagenumbernotforcitationpurposes)
F1000Biology Reports2013, 5:1 http://f1000.com/prime/reports/b/5/1
7. HammesGG,BenkovicSJ,Hammes-SchifferS:Flexibility,diversity, polyvalentligandwithasingle-sitereceptor.ProcNatlAcadSci
andcooperativity:pillarsofenzymecatalysis.Biochemistry2011, USA2008,105:17772-7.
50:10422-30.
8. Burgen AS,Roberts GC, Feeney J: Binding of flexibleligands to
macromolecules.Nature1975,253:753-5. 25. Borg M, Mittag T, Pawson T, Tyers M, Forman-Kay JD, Chan HS:
Polyelectrostaticinteractionsofdisorderedligandssuggesta
9. Hammes GG: Enzyme catalysis and regulation. New York:
physicalbasisforultrasensitivity.ProcNatlAcadSciUSA2007,
AcademicPress;1982.
104:9650-5.
10. HammesGG, ChangYC, Oas TG: Conformational selectionor
inducedfit:afluxdescriptionofreactionmechanism.ProcNatl
AcadSciUSA2009,106:13737-41.
26. Ubbink M: The courtship of proteins: understanding the
11. Nadassy K, Wodak SJ, Janin J: Structural features of protein- encountercomplex.FEBSLett2009,583:1060-6.
nucleicacidrecognitionsites.Biochemistry1999,38:1999-2017.
12. SchulzGE:NucleotideBindingProteins.In:BalabanM(ed)Molecular
MechanismsofBiologicalRecognition.Amsterdam:Elsevier/North-Holland
27. AntalMA,BodeC,CsermelyP:Perturbationwavesinproteins
BiomedicalPress;1979:79-94.
andproteinnetworks:applicationsofpercolationandgame
13. Spolar RS, Record MT, Jr: Coupling of local folding to site- theoriesinsignalinganddrugdesign.CurrProteinPeptSci2009,
specificbindingofproteinstoDNA.Science1994,263:777-84. 10:161-72.
14. Espinoza-Fonseca LM: Reconciling binding mechanisms of 28. Kovacs IA, Szalay MS, Csermely P: Water and molecular
intrinsically disordered proteins. Biochem Biophys Res Commun chaperones act as weak links of protein folding networks:
2009,382:479-82. energylandscapeandpunctuatedequilibriumchangespoint
towardsagametheoryofproteins.FEBSLett2005,579:2254-60.
15. FershtAR,Knill-JonesJW,BedouelleH,WinterG:Reconstruction
by site-directed mutagenesis of the transition state for the 29. Ptitsyn OB: [Stages in the mechanism of self-organization of
activation of tyrosine by the tyrosyl-tRNA synthetase: a proteinmolecules].DoklAkadNaukSSSR1973,210:1213-5.
mobile loop envelopes the transition state in an induced-fit
mechanism.Biochemistry1988,27:1581-7.
30. Dolgikh DA, Gilmanshin RI, Brazhnikov EV, Bychkova VE,
Semisotnov GV, Venyaminov S, Ptitsyn OB: Alpha-Lactalbumin:
16. Schnell JR,Dyson HJ, Wright PE:Effect ofcofactor bindingand compact state with fluctuating tertiary structure? FEBS Lett
loop conformation on side chain methyl dynamics in 1981,136:311-5.
dihydrofolatereductase.Biochemistry2004,43:374-83.
31. Gilmanshin RI, Ptitsyn OB: An early intermediate of refolding
alpha-lactalbumin forms within 20 ms. FEBS Lett 223 1987,
223:327-9.
17. MohanA,OldfieldCJ,RadivojacP,VacicV,CorteseMS,DunkerAK, 32. BychkovaVE,PainRH,PtitsynOB:The‘moltenglobule’stateis
Uversky VN: Analysis of molecular recognition features involvedinthetranslocationofproteinsacrossmembranes?
(MoRFs).JMolBiol2006,362:1043-59. FEBSLett1988,238:231-4.
18. Tompa P, Fuxreiter M: Fuzzy complexes: polymorphism and
structural disorder in protein-protein interactions. Trends
BiochemSci2008,33(1):2-8. 33. Bracken C, Carr PA, Cavanagh J, Palmer AG, 3rd: Temperature
19. Selenko P, Gregorovic G, Sprangers R, Stier G, Rhani Z, Kramer A, dependenceofintramoleculardynamicsofthebasicleucine
Sattler M: Structural basis for the molecular recognition zipperofGCN4:implicationsfortheentropyofassociation
betweenhumansplicingfactorsU2AF65andSF1/mBBP.Mol withDNA.JMolBiol1999,285:2133-46.
Cell2003,11:965-76.
34. XuD,TsaiCJ,NussinovR:Mechanismandevolutionofprotein
20. Oldfield CJ, Meng J, Yang JY, Yang MQ, Uversky VN, Dunker AK: dimerization.ProteinSci1998,7:533-44.
Flexiblenets:disorderandinducedfitintheassociationsofp53
and14-3-3withtheirpartners.BMCGenomics2008,9(Suppl1):S1.
21. HsuWL,OldfieldC,MengJ,HuangF,XueB,UverskyVN,RomeroP, 35. Gunasekaran K, Tsai CJ, Nussinov R: Analysis of ordered and
Dunker AK: Intrinsic protein disorder and protein-protein disordered protein complexes reveals structural features
interactions.PacSympBiocomput2012:116-127. discriminatingbetweenstableandunstablemonomers.JMol
Biol2004,341:1327-41.
22. Nash P, Tang X, Orlicky S, Chen Q, Gertler FB, Mendenhall MD,
Sicheri F, Pawson T, Tyers M: Multisite phosphorylation of a 36. Fuxreiter M, Tompa P: Fuzzy complexes: a more stochastic
CDK inhibitor sets a threshold for the onset of DNA viewofproteinfunction.AdvExpMedBiol2012,725:1-14.
replication.Nature2001,414:514-21.
37. Sigalov A, Aivazian D, Stern L: Homooligomerization of the
cytoplasmicdomainoftheTcellreceptorzetachainandof
otherproteinscontainingtheimmunoreceptortyrosine-based
23. VermaR,AnnanRS,HuddlestonMJ,CarrSA,ReynardG,DeshaiesRJ: activationmotif.Biochemistry2004,43:2049-61.
Phosphorylation of Sic1p by G1 Cdk required for its
degradationandentryintoSphase.Science1997,278:455-60.
24. Mittag T, Orlicky S, Choy WY, Tang X, Lin H, Sicheri F, Kay LE, 38. Sigalov AB, Zhuravleva AV, Orekhov VY: Binding of intrinsically
TyersM,Forman-KayJD:Dynamicequilibriumengagementofa disordered proteins is not necessarily accompanied by a
Page10 of12
(pagenumbernotforcitationpurposes)