In one of the most intensively studied cases, the interaction between CypA and the Gag protein derived from the HIV virus, the Keywords cell cycle; CKS subunit; dynamics; enzyme; interac
Trang 1Molecular mechanisms of the phospho-dependent prolyl
G Lippens1, I Landrieu1and C Smet2
1 CNRS UMR 8576 Unite´ de Glycobiologie Structurale et Fonctionnelle, Universite´ des Sciences et Technologies de Lille 1-59655,
Villeneuve d’Ascq, France
2 Institut de Recherche Interdisciplinaire, CNRS, Lille, France
Introduction
Prolyl cis⁄ trans isomerases form a special class of
enzymes in many respects Most other enzymes change
the covalent chemistry of their substrates (be it through
cleavage, or addition⁄ removal of a phosphate, acetyl
or methyl group), leading to a product that is
distin-guishable by mass spectrometry or immunochemistry
from the incoming molecular entity The modification
imposed by prolyl cis⁄ trans isomerases is not covalent,
but merely conformational One therefore encounters
many technical difficulties when attempting to observe
their molecular action in vivo and even in vitro
Experi-ments with isolated proteins have indicated the involve-ment of two major classes of isomerases, the cyclophilins (Cyps) and the FK506-binding proteins (FKBPs), in the protein-folding process, where the con-formational state of a prolyl peptide bond can be a rate-limiting step [1] The importance of ‘binding ver-sus catalysis’ remains an open issue in many cases [2], partly because tinkering with the enzymatic activity through mutations also leads to changes in interaction parameters [3] Assessing the role of the isomerization function in vivo is even harder In one of the most intensively studied cases, the interaction between CypA and the Gag protein derived from the HIV virus, the
Keywords
cell cycle; CKS subunit; dynamics; enzyme;
interaction module; phosphorylation; Pin1;
prolyl cis ⁄ trans isomerase; protein
degradation; protein structure
Correspondence
G Lippens, CNRS UMR 8576 Unite´ de
Glycobiologie Structurale et Fonctionnelle,
Universite´ des Sciences et Technologies de
Lille 1-59655, Villeneuve d’Ascq Cedex,
France
Fax: +33 3 20 43 65 55
Tel: +33 3 20 33 42 71
E-mail: Guy.Lippens@univ-lille1.fr
(Received 5 June 2007, revised 3 August
2007, accepted 17 August 2007)
doi:10.1111/j.1742-4658.2007.06057.x
Since its discovery 10 years ago, Pin1, a prolyl cis⁄ trans isomerase essential for cell cycle progression, has been implicated in a large number of molecu-lar processes related to human diseases, including cancer and Alzheimer’s disease Pin1 is made up of a WW interaction domain and a C-terminal catalytic subunit, and several high-resolution structures are available that have helped define its function The enzymatic activity of Pin1 towards short peptides containing the pSer⁄ Thr-Pro motif has been well docu-mented, and we discuss the available evidence for the molecular mecha-nisms of its isomerase activity We further focus on those studies that examine its cis⁄ trans isomerase function using full-length protein substrates The interpretation of this research has been further complicated by the observation that many of its pSer⁄ Thr-Pro substrate motifs are located in natively unstructured regions of polypeptides, and are characterized by minor populations of the cis conformer Finally, we review the data on the possibility of alternative modes of substrate binding and the complex role that Pin1 plays in the degradation of its substrates After considering the available work, it seems that further analysis is required to determine whether binding or catalysis is the primary mechanism through which Pin1 affects cell cycle progression
Abbreviations
APP, amyloid precursor protein; CDK, cyclin-dependent kinase; CKS, cyclin-dependent kinase subunit; CTD, C-terminal domain; Cyp, cyclophilin; FKBP, FK506-binding protein; IRF, interferon regulatory factor; Pol II, polymerase II.
Trang 2mere interaction of CypA with Gag might protect the
virion from an as yet poorly identified cellular
integra-tion factor [4] Whether the prolyl cis⁄ trans isomerase
activity detected in vitro of CypA on Pro90 in the
full-length Gag protein [5] plays a role in the viral
restriction process in vivo remains an open question In
the case of the Itk kinase, the cis⁄ trans isomerization
catalyzed by CypA was shown to generate novel
inter-action surfaces for two distinct molecular partners [6],
and the conformation of a proline in the linker region
between two SH3 domains of the Crp adaptor protein
determines the autoinhibition of the domains [7] These
latter examples provide a structural basis for the
manner in which prolyl cis⁄ trans isomerization could
act as a conformational switch in biological processes
Beyond the difficulties of characterizing this
confor-mational switch in a cellular context and even more in
a living organism, another problem with the molecular
characterization of the prolyl-isomerase enzymes is
their redundancy in the cell Indeed, a model organism
such as yeast has as many as eight Cyps and four
FKBPs, and even knocking down all of them does not
lead to a clear phenotype under normal conditions [8]
The discovery of Ess1, a novel Saccharomyces cerevisiae
prolyl cis⁄ trans isomerase [9] that proved essential for
cell division, was therefore highly relevant Its human
homolog was identified as a protein interacting with
the NIMA kinase during the G1⁄ S cell cycle stage, and
was called Pin1 for this reason [10] Since its initial
identification, Pin1 has attracted a great deal of
attention, as the enzyme seems to be implicated in
various human diseases, ranging from cancer and
neurodegenerative diseases to inflammation Rather
than adding to the list of excellent general reviews on
Pin1 [11–14] or to those reviews that emphasize its role
in cellular processes related to diseases [15–17], we
focus here on what is known about the molecular
mechanisms of the enzyme action Most importantly,
we want to critically review the evidence that Pin1
would or would not act as a prolyl cis⁄ trans isomerase
Molecular mechanisms of Pin1 action
The initial X-ray structure of the catalytic domain
complexed with an Ala-Pro dipeptide and a sulfate ion
[18] suggested an important role for two structural
ele-ments in catalysis First, the loop between residues 66
and 77 (human Pin1 numbering) is involved in binding
the phosphate moiety of the substrate (Fig 1) This
loop is flexible, as suggested by its different
conforma-tion in a second crystallographic structure, where the
catalytic domain was not complexed to a substrate
peptide [19] Heteronuclear NOE data on human Pin1,
however, indicated only a limited decrease in the flexi-bility of this loop on peptide binding [20] or even a closed conformation in the absence of substrate [21] Our NMR data on the Arabidopsis thaliana analog, Pin1At, showed severe line broadening in the equiva-lent stretch, giving weight to the dynamic character of this loop on the 100 ls to 1 ms time scale [22], but the crystal structure of the Candida albicans Ess1 revealed
a closed loop even in the absence of substrate [23] The dynamic character of this loop was recently shown by NMR relaxation dispersion measurements on the human enzyme during substrate binding [24] and
Fig 1 Top: Molecular structure of Pin1 (Protein Data Bank code: 1Pin [4]), showing the catalytic domain (green), the WW domain (red), and the linker region (yellow) The positively charged residues
of the active site loop (Lys63, Arg68, and Arg69) are in blue, whereas the presumed catalytic Cys113 is in pink The Ala-Pro pep-tide in the active site is indicated by sticks, and is complemented
by a sulfate ion (light blue) A poly(ethylene glycol) molecule (not shown) is sequestered between the catalytic domain and the WW domain Bottom: The complex between Pin1 and a Pol II CTD phospho-peptide (Protein Data Bank code: 1F8A [5]) The active site loop adopts a more extended conformation, and the peptide inter-acts with the WW domain mainly through the phospho-Thr ⁄ Pro moiety.
Trang 3catalysis [25] Interestingly, in the case of CypA, the
dynamic character of its active site was found to be
essential for its catalytic function [3] The fact that
Pin1 can compensate for the lack of Ess1 in yeast was
exploited to further investigate the functional role of
the different Pin1 residues Behrsin et al [26] applied
unigenic evolution, where the capacity of a randomly
mutated pin1 gene to compensate for Ess1 in a yeast
strain devoid of the wild-type ess1 gene is screened
They showed that, in particular, Lys63 is essential for
the anchoring function of the phosphorylated
sub-strate, whereby the K63A mutant affects the catalytic
efficiency because of weaker binding The two
argi-nines in the loop, Arg68 and Arg69, would provide no
more than a single positive charge [26]
A second important structural feature of the initial
X-ray structure of Pin1 was the spatial proximity of
Cys113 to the Ala-Pro bond (Fig 1) [18] Although
this initially suggested a catalytic mechanism through a
nucleophilic attack on the substrate carbonyl carbon
by the Sc of Cys113, the fact that the C113D mutant
remains functional calls this mechanism into question
[26] The Cys113 residue, through its unusually low
pKa value, might indeed maintain an overall
electro-negative environment that is crucial for destabilizing
the double bond character of the pThr-Pro bond [26]
Further evidence is available for the catalytic
mecha-nism of Pin1 resembling more closely that of the other
prolyl cis⁄ trans isomerases, with only its loop region
selecting for Pro residues preceded by a negative
charge In the A thaliana Pin1 homolog, we did not
observe significant chemical shift changes for the
equivalent Cys70 upon saturation of the catalytic
domain with several phosphopeptides [22] In yeast,
the equivalent C120R mutant of Ess1 only prevented
growth at the higher temperature of 37C [27]
Finally, the chemically and structurally similar binding
pockets of Pin1 and FKBP and the structural
resem-blance between their respective high-affinity inhibitors
[28] further underscore the similarity between Pin1 and
the other peptide prolyl isomerases (PPIases)
A second crystallographic structure (Fig 1) with the
Pin1 WW domain complexed to a phospho-peptide
derived from the C-terminal domain of polymerase II
(Pol II CTD) indicated the trans conformation of the
prolyl bond following the phosphorylated Ser [19]
NMR spectroscopy confirmed that the cis conformer
in a Cdc25-derived peptide could not interact with the
WW domain [29] Despite these findings, recently
pre-sented data hint at other potential binding modes
First, when the role of Pin1 in amyloid precursor
pro-tein (APP) processing and ensuing b-amyloid
produc-tion was studied, the interacproduc-tion between the WW
domain and a phosphorylated peptide (V-pT668-P-E-E) derived from the APP cytoplasmic domain was probed
by NMR spectroscopy [30] For this latter phospho-peptide, the15N-labeled Glu670 was the primary probe
of the interaction Interestingly, the correlation peaks corresponding to both the trans and cis conformers of the pThr668-Pro prolyl bond were found to shift upon interaction with the single WW domain, with the larg-est shift for the cis form In an earlier study, the same group examined in great detail the structure of both cisand trans conformers of the same peptide by NMR spectroscopy [31] The local structure of the cis con-former of this peptide is characterized by a hydrogen bond between the amide proton of Val667 and the Glu671 side chain carboxyl group This particular motif might be recognized by the Pin1 WW domain,
or, alternatively, the presence of two glutamate resi-dues downstream of the proline could lead the WW domain to read the cis form in a reversed manner The initial screen for Pin1 substrates, which identified it as
a phospho-dependent prolyl cis⁄ trans isomerase, indicated that, at least for some peptides, using a glutamate (but not aspartate) rather than a phospho-Ser⁄ Thr group before the critical proline did not decrease Pin1 enzymatic efficiency [32] However, we found that a Tau mutant carrying multiple Glu-Pro motifs did not significantly interact with the Pin1
WW domain (G Lippens, I Landrieu, C Smet and
R Brandt, unpublished results) Further structural characterization of the complex between the Pin1 WW domain and the amyloid peptide will be necessary, and might form a novel starting point for the development
of WW domain inhibitors
Even more surprising is the recent finding that Pin1 could recognize cyclin E via a noncanonical pThr384-Gly385 motif [33] rather than the pThr380-Pro381 motif The main argument was that the latter pThr380-Pro381 motif is buried in the yeast Cdc4 molecular surface that was determined by X-ray crys-tallography [34] In this structure of the peptide–CDC4 complex, however, Pin1 is missing, whereas in the study on cyclin E degradation, Pin1 was brought in by the phospho-cyclin E and not by the CDC4a compo-nent of the final complex (Fig 2) [35] Therefore, the outcome of the molecular competition between Pin1 and CDC4 for the same phosphorylated motif is not clear, and still leaves open the possibility that Pin1 could interfere with the cyclin E–CDC4 interface Pin1 was proposed to isomerize the peptide bond between Pro381 and Pro382 The concomittant structural rear-rangement would cause cyclin E to approach the distant E2 ligase of a different SKP⁄ Cullin ⁄ F-box protein (SCF)Cdc4c complex During our work on the
Trang 4neuronal Tau protein, we studied by NMR spectroscopy
a peptide containing the sequence (LP) pT217 (PPT),
which matches the optimal L⁄ I-L ⁄ I ⁄ P-pT-P CDC4
phospho-degron sequence [36], and moreover strongly
resembles the cyclin E peptide (LL)pT380(PPQ) that
was crystallized at the CDC4 interface [34] In the Tau
peptide, despite significant proportions of cis
conform-ers for the three proline residues preceded by a
phos-pho-Thr [37], the Pro218-Pro219 peptide bond did not
show detectable levels of cis conformer (Fig 3), and this
dipeptide did not interact with Pin1 Moreover, whereas
a catalytic amount of Pin1 greatly enhanced the
isomeri-zation rate for the pThr212-Pro213 bond, we did not
detect any Pin1-catalyzed exchange peak for the
neigh-boring pThr217-Pro218 bond in the same peptide
(Fig 3) The mechanistic details of how CDC4a could
overrule the strict phospho-Ser⁄ Thr dependence of Pin1,
be it for binding or for prolyl cis⁄ trans isomerization,
therefore await further structural elucidation
Structural features of Pin1 substrates
Concerning the function of Pin1, a first intriguing
observation is that many of its substrates meet the
cri-teria for the recently identified class of intrinsically
unfolded proteins Although unstructured regions in
proteins or fully unstructured proteins have been known since the beginning of structural biology, only recently have they been identified as a true class of proteins that challenge the sequence–structure–function paradigm [38] Phosphorylation in such regions is a recurring theme, and transforms them into effective anchoring points for novel components in the multi-protein complexes that govern the fate of the cell [39] Unfortunately, for many of these complexes, we do not yet know how these intrinsically unstructured domains exert their molecular function For Cdc25, for example, one of Pin1’s most extensively studied sub-strates, it is not clear how phosphorylation at the Thr48⁄ Thr67 sites regulates the phosphatase activity; the same is true for Tau, a neuronal protein involved
in tubulin polymerization Tau loses its ability to poly-merize tubulin after phosphorylation at the Thr231 position, and Pin1 can restore this function [40] Understanding the binding of Tau to tubulin and its modulation by phosphorylation will be necessary before we can evaluate the role of Pin1 in this complex process Structural studies on peptides derived from the two proteins, Cdc25 and Tau, have shown that only a low percentage of the prolines downstream of the phospho-Thr⁄ Ser residues adopt the cis confor-mation, typically 3–10% [41,42], and at least for the
Fig 2 Schematic view of the parallel between Pin1 and CKS in protein degrada-tion Top: Model of the SCF CDC4 E3 ligase and the role of Pin1 Pin1 is brought in with the cyclin E substrate, through interaction with the pSer384-Pro motif When the com-plex contains the CDC4a isoform, the iso-merase activity of Pin1 leads to cyclin E dissociation, and allows association with a novel CDC4c complex, where ubiquitin addi-tion would occur (adapted from Brazin et al [6]) Bottom: Model of the SCF Skp2
E3 ligase, where CKS1 associates with the Skp2 protein and hence forms an integral part of the E3 ligase that recognizes its phosphorylated p27 kip1 substrate (adapted from Sarkar et al [7] and Dolinski [8]).
Trang 5full-length Tau protein, our preliminary NMR data
indicate that the same low population of cis
conform-ers is found in the full-length protein For the cis
con-former to be the predominant one, it needs to be
stabilized by the surrounding (folded) structure It was
suggested that multiple phosphorylations might create
a locally strained conformation [43], favoring the
cis conformation of one or more prolines, but NMR
studies have as yet not detected a prevalent cis
con-former in peptides carrying two or more of these
motifs [37,44] Because many of the Pin1-recognized
phosphorylation motifs are in unstructured regions, we
thus can reasonably expect conformational
heterogene-ity at the level of its substrate pSer⁄ Thr-Pro prolyl
bonds, with the cis form being the less common one The predominant trans conformation at the pThr⁄ Ser-Pro bonds combined with the Pin1-mediated increase in immunoreactivity of the MPM-2 antibody [45] towards its substrates suggests that the antibody could recognize the cis conformer of the pThr-Pro prolyl bond in substrates such as phospho-Cdc25 or phospho-Tau Indeed, Pin1 as an isolated enzyme would merely lower the energetic barrier separating both conformations without changing their relative populations However, when coupled to another molecular process that is conformer dependent (such
as protease sensitivity or antibody recognition), isomerization could catalyze changes in the relative
tPro213
Fig 3 NOESY spectrum of the triply phosphorylated SRSRpT212PpS214LPpT217PPTR peptide of Tau The cis conformation for the Pro219 is below the limit of detection Upon addition of a catalytic amount of Pin1, enhanced cis ⁄ trans isomerization of the pThr212-Pro213 peptide bond leads to an additional red peak (red box, peak connecting the trans Pro213 Ha resonance at 4.94 p.p.m and the cis Pro213 Ha at 4.42 p.p.m.), whereas in the same spectrum, the equivalent exchange peak for the pThr217-Pro218 bond (which should be in the green box
at 4.72 p.p.m., trans Pro218 Ha, and 5.20 p.p.m., cis Pro218 Ha) was not detected.
Trang 6populations, as one of the forms would continuously
disappear from the pool of free peptides Structural
characterization of the MPM-2 antibody with a
phospho-peptide substrate might be very informative
in this aspect, and shed further light on this issue
Another well-studied process for which Pin1’s
cata-lytic activity has been put forward is
dephosphoryla-tion by the PP2A phosphatase [46] The crystal
structure of a cyclin-dependent kinase (CDK)2⁄
cyclin A kinase [47] revealed the existence of the trans
conformation of the Ser-Pro bond in the peptide
sub-strate Similarly, the major proline-directed
phospha-tase PP2A recognizes and dephosphorylates only the
trans conformer of its phosphorylated peptide
sub-strate [46] Pin1 could thus enhance the molecular
function of this phosphatase by speeding up the
cis–trans interconversion rate, as was proposed for
both Tau and Cdc25 [46] Indeed, for the small
phos-pho-peptides that are used in most studies, a pool of
mainly cis conformers can be obtained by lithium⁄
trifluoroethanol stabilization and⁄ or selective
proteo-lytic cleavage of the major pool of trans conformers
[46,48] The dephosphorylation of this pool of mainly
cis conformers takes less time in the presence of Pin1
[46], and this unambiguously involves Pin1’s prolyl
cis⁄ trans isomerase activity For the in vitro
phosphor-ylated full-length Tau and CDC25, however, no effort
was made to prepare a similar large pool of cis
con-formers We found at the peptide level that the
pThr231-Pro bond of Tau is mainly in the trans form
[37] Extending our combined in vitro phosphorylation
and NMR spectroscopy of full-length Tau from
protein kinase A [49] to a CDK kinase, we have
preliminary data that the pThr231-Pro bond in this
CDK-phosphorylated full-length Tau also adopts the
trans conformation to a major extent (I Landrieu,
L Amniai and G Lippens, unpublished results)
Isomerization from cis to trans hence cannot be
invoked any more as the sole mechanism promoting
the accelerated dephosphorylation by PP2A Pin1
might in an as yet unidentified manner favor the
inter-action between PP2A and its phosphorylated
sub-strates, and hence stimulate their dephosphorylation
without necessarily requiring its catalytic prolyl
cis⁄ trans isomerase activity
Role of Pin1 in protein stability
Regulation of protein degradation seems to be an
all-important role for Pin1, and as such, a remarkable
parallel with the CKS (CDK subunit) family can be
established CKS targets the activated CDK complex
towards phosphorylated substrates such as CDC25,
and is as such essential for the entry of Xenopus laevis egg extracts into mitosis [50] Well characterized struc-turally, CKS proteins can be found in different confor-mations with regard to their last b-strand Folded back on itself in a monomeric compact form [51], the C-terminal b-strand in the swapped dimer is locked in
a second monomer [52] The structural differences between compact monomer and swapped dimer are mainly limited to the conformation of the Glu-Pro dipeptide in the hinge region between the last b-strand and the core of CKS The crystal structure of the com-plex of CKS with CDK2⁄ cyclin A clearly showed that only the monomeric, compact CKS could bind to the CDK subunit [53], the swapped dimer giving rise to important steric clashes preventing the interaction Pin1 antagonizes the stimulatory role of CKS in mito-sis entry [50], and we initially assumed that this was through a direct interaction with the Glu-Pro hinge motif and subsequent conformational transition between both structural forms Experiments proved the hypothesis wrong, as we did not obtain any evidence
of an interaction between Pin1 and this Glu-Pro dipep-tide of CKS We did, however, show in vitro competi-tion for the same Cdc25-derived phosphorylated peptide between the Pin1 WW domain and the CKS binding module [54], and showed that the interaction surfaces were quite similar in terms of amino acid composition and structure (Fig 4)
As well as a comparable role in regulating the phos-phorylation state of CDC25 or other substrates, the parallel between the WW and CKS interaction domains can be drawn further when considering their respective implications for the ubiquitination process directing proteins towards degradation Indeed, human Cks1 was identified as an important factor in the SCFSkp2 ubiquitin E3 ligase SCFSkp2 plays a role in the degradation of the CDK inhibitor p27kip1 in late
G1phase after phosphorylation on its Thr187 residue [55,56] CKS1 interacts both with Skp2 and with a p27kip1-derived pThr187 peptide [57], and hence plays the role of an adaptor protein that is an integral part
of the E3 ligase (Fig 2) Similarly, Schizosaccharo-myces pombe p13suc1 binds to the activated anaphase-promoting complex (APC)⁄ cyclosome [58]
Pin1 intervenes in a complex manner with the degra-dation of cyclin E through regulation of the interaction
of cyclin E with the SCFCdc4 complex It stimulates ubiquitin addition to cyclin E by the SCFCdc4c E3 ligase, after releasing the same cyclin E from the complex with SCFCdc4a [33] Pin1 would, however, not
be part of the initial SCFCdc4a⁄ c complex, but would
be brought in by the substrate itself (Fig 2) The enhanced cyclin E degradation by Pin1 contrasts with
Trang 7its role in cyclin D stability Pin1 overexpression was
found to stabilize this cyclin at both the protein and
mRNA levels [59] A similar stabilizing effect was
described for Emi1, an inhibitor of the APC complex
required to induce S-phase and M-phase entry by
driv-ing cyclin A and cyclin B accumulation Pin1 prevents
its association with SCFbTrcp, and hence stabilizes
Emi1 when this latter is phosphorylated on its
Ser10-Pro motif [60] In yet another case, the Drosophila
homolog Dodo was shown to facilitate the degradation
of the transcription factor CF2 [61] Finally, the
neuro-nal Tau protein also contains the aforementioned
phospho-degron sequence (LPT217PPLSP) Although
Pin1 has as yet not been implicated directly in its
deg-radation, phosphorylated Tau is targeted for
proteaso-mal degradation through the E3 ubiquitin ligase
CHIP, complexed to an Hsc70 moiety [62], where
CHIP would selectively ubiquitinate (natively)
unfolded proteins by collaborating with the molecular chaperone [63] If this CDC4 phospho-degron sequence
on Tau is physiologically phosphorylated and recog-nized by the ubiquitin ligase, this would close the circle
of Pin1’s preference for unfolded substrates In any case, the WW domain of Pin1 is an excellent example
of the complex roles played by protein–protein inter-action modules [64], and whether the catalytic domain
is a second interaction domain or rather a genuine enzyme awaits further elucidation
Functional overlap of Pin1 with other prolyl cis ⁄ trans isomerases?
Regulation of the transcription machinery was early described as an essential function of Ess1 [65,66] Its interaction with the numerous YSPTSPS heptapeptide repeats of the Pol II CTD [67,68], although of weak
R12
W29 S13
R99
R30
Q78
W82
Fig 4 Molecular surface (left) and ribbon
diagrams (right) of the Pin1 WW domain
(top) or CKS (p13 Suc1 ) interaction domain.
Color coding is according to the chemical
shift changes observed with a
CDC25-derived phosphopeptide The residues
whose chemical shift is most affected upon
peptide binding (red) are Arg12 and Trp29
for the WW domain, and Arg30, Gln78 and
Trp82 for p13 Suc1 Blue color indicates the
absence of chemical shift changes.
Trang 8affinity when these are tested as isolated peptides
in vitro [69], could lead to the assembly of different
molecular complexes when mRNA is transcribed Ess1
would regulate the phosphorylation state of the CTD,
and the yeast Fcp1 phosphatase plays a similar role as
PP2A in this process However, conflicting results have
been reported, with Ess1 either stimulating CTD
dephosphorylation by Fcp1 [70] or inhibiting this same
process [71,72] Differences in the exact nature of the
substrate (full-length Pol II or the isolated CTD
domain) might explain this discrepancy [72] Fcp1 was
found to more effectively dephosphorylated the pSer5
position [70], whereas the WW domain of Pin1 binds
with slightly better affinity the pSer2 position [69] This
suggests a sequential mechanism, with preliminary
binding of the WW domain to the pSer2-Pro motif
and subsequent isomerization by the catalytic domain
at the Ser5 position The yeast Ess1 protein, however,
prefers both to bind and isomerize pSer5-Pro over
pSer2-Pro [73] A role of Pin1 in transcriptional
regu-lation through the (de)stabilization of complexes
formed at the CTD of Pol II was suggested by the
observation that Ess1 and the yeast Nedd4 ubiquitin
ligase Rsp5 compete for the largest subunit of the
RNA Pol II, possibly through their respective WW
domains [74]
Using a number of ess1 temperature-sensitive
mutants, two groups unexpectedly discovered that
CypA can functionally replace Ess1 [66,75] Both
prolyl cis⁄ trans isomerases could catalyze protein
con-formational changes essential for the assembly and⁄ or
activity of the Sin3–Rpd3 histone deacetylase complex,
but not through binding and⁄ or catalytic action
towards the same peptide motifs [76] Ess1 interacts
directly with the Sin3 component, and downregulates
in this manner the deacetylase activity of Rpd3,
whereas CypA would drive the equilibrium towards
the formation of a Sin3–Rpd3–Sap30 complex
Whereas this provides the first evidence of crosstalk
among different PPIase families, the observation of a
basal enzymatic activity towards phosphorylated
sub-strates in cell lysates from Pin1–⁄ – knockout mice [77]
hints at a direct functional overlap An intriguing
com-plementarity is further found in the inflammatory
response towards antiviral double-stranded RNA Pin1
interacts with the pSer339-Pro340 motif on interferon
regulatory factor (IRF)-3, leading ultimately to its
deg-radation and ensuing impaired production of
inter-feron-b [78] In the homologous IRF-4, the Ser-Pro
motif of IRF-3 is interrupted by a Leucine, despite
being in one of the best conserved regions between
both transcription factors Pin1 no longer recognizes
this motif, and no IRF-4 regulation by Pin1 has been
reported However, IRF-4 is regulated by FKBP52, a member of the FK5060-binding prolyl cis⁄ trans isome-rases [79] The tetratricopeptide repeats of FKBP52 mediate the interaction with IRF-4 and hence might be the equivalent of Pin1’s WW domain, whereas its catalytic domain could induce structural changes in the N-terminal proline-rich domain of IRF-4 In the same field of immunology, Pin1 also regulates the production of such proinflammatory cytokines as granulocyte–macrophage colony-stimulating factor [80], interleukin-2 and interferon-c [81] The ARE-con-taining cytokine mRNAs interact with AUF1 factors, and this interaction targets them for degradation Pin1 interferes with this interaction, and hence stimulates cytokine production In the resting eosinophils, how-ever, Pin1’s activity is suppressed through phosphory-lation on one or more of its own Ser⁄ Thr residues Regulation of Pin1 function through phosphoryla-tion is indeed an important topic that has not been extensively explored Phosphorylation at the Ser16 resi-due in the WW domain prevents its interaction with phosphorylated substrates [82], and thereby partially inactivates the function of Pin1 Polo-like kinase-1-mediated phosphorylation, on the other hand, stabi-lizes Pin1 by inhibiting its ubiquitination [83] Pin1 stability and regulated activity itself hence intervene in its complex relationship with phosphorylation
Conclusions and perspectives The list of potential substrates of Pin1 seems never-ending, and one wonders how one single protein could
be involved in such a variety of cellular processes We can only propose some possibilities First, the WW domain is clearly not very selective with regard to its molecular targets Its binding pocket mostly sequesters the phosphate moiety and the proline side chain (Figs 1 and 4), whereas other amino acids around this motif only marginally contribute to the binding affin-ity When studying phosphorylated peptides derived from Tau, we found that the best binder was actually the dipeptide pThr-Pro, with a KDof 100 lm [84] Par-allel studies with Pol II CTD-derived peptides have shown similar results, with only a two-fold better affin-ity for the pSer5-Pro motive over the pSer2-Pro motif [66] We thus believe that the WW domain will recog-nize in vitro basically any pThr⁄ pSer-Pro pattern, as long as it is in a rather unstructured region Second, the weak affinity precludes the formation of stable complexes, and leaves room for the Pin1 molecule to sample a large number of potential substrates during its half-life Finally, the group of S Hanes, who was the first to describe the Sacch cerevisiae parvulin Ess1
Trang 9[9], has described a large redundancy of protein copy
numbers in the cell, at least under normal growth
con-ditions Indeed, they found that although wild-type
yeast cells contain on the order of 200 000 molecules
of Ess1 per cell, a level lower than 400 molecules per
cell is sufficient for growth, leaving plenty of Ess1
molecules for many substrates [73] Only under certain
conditions of stress does the large pool of Ess1 seem
essential for growth, which probably brings us
proba-bly to the situation existing in human diseases
Whereas the detrimental role of Pin1 in human
dis-eases, and especially cancer, seemed at first to be
evi-dent, recent findings suggest that the picture in certain
cases might be more complex [85] Certainly, Pin1
overexpression correlates strongly with poor prognosis
in a variety of cancers, and these clinical data cannot
be overlooked [86] Nonetheless, Pin1 also stabilizes
p53 and increases its transcriptional activity, which is
essential to counteract oncogenesis [87,88] At the
cel-lular level, its role in cyclin E and c-Myc degradation
or Emi1 stabilization would equally point to a
protec-tive role as a conditional tumor suppressor Yeh et al
have pointed out that the genetic background of the
mouse lines might lead to different outcomes for the
same mutation, making the construction of a single
coherent framework more problematic [89] As is the
case for p53, where the relative levels of protein and
its inhibitors⁄ activators can lead to subtle but
signifi-cant differences between results in cell and animal
models [90], careful analysis of in vivo models will be
needed to validate all data acquired in vitro or in cell
models before drawing conclusions on Pin1’s role in
cancer Finally, in the context of Alzheimer’s disease,
Pin1 was shown to have a beneficial role, as it restores
the capacity of Cdc2-phosphorylated Tau to
polymer-ize tubulin into microtubules [40] However, the
tan-gles of Tau and other amyloid species, although
characteristic in Alzheimer’s disease and correlating
well with cognitive decline, are now seen in a new
light by the scientific community Over a period of
10 years, they have shifted from being an important
cause of the disease towards consituting a cellular
defense against the toxic oligomeric but soluble
spe-cies, although these latter still await clear
identifica-tion [91] Could Pin1 be intended primarily as a
protective mechanism, recognizing aberrant
phosphor-ylated Ser⁄ Thr-Pro motifs and targeting them through
interaction or conformational change towards
dephos-porylation, degradation, or aggregation? Is prolyl
cis⁄ trans isomerization required for this function? and
could this mechanism go awry in certain diseases such
as cancer? Further research will be necessary to
deter-mine the exact role of Pin1 in human disease, and
thereby its potential as a molecular target for novel drugs
Acknowledgements
We thank two anonymous reviewers and Dr
E Appella (NIH, Bethesda, USA) for careful reading and constructive suggestions, and Dr X Hanoulle (Lille, France) for help with the figures The NMR facility used in this work was sponsored by the Re´gion Nord-Pas de Calais, the CNRS, the Pasteur Institute
of Lille, and the University of Lille I
References
1 Lang K, Schmid FX & Fischer G (1987) Catalysis of protein folding by prolyl isomerase Nature 329, 268– 270
2 Fischer G, Tradler T & Zarnt T (1998) The mode of action of peptidyl prolyl cis⁄ trans isomerases in vivo: binding vs catalysis FEBS Lett 426, 17–20
3 Eisenmesser EZ, Millet O, Labeikovsky W, Korzhnev
DM, Wolf-Watz M, Bosco DA, Skalicky JJ, Kay LE & Kern D (2005) Intrinsic dynamics of an enzyme under-lies catalysis Nature 438, 117–121
4 Sokolskaja E & Luban J (2006) Cyclophilin, TRIM5, and innate immunity to HIV-1 Curr Opin Microbiol 9, 404–408
5 Bosco DA, Eisenmesser EZ, Pochapsky S, Sundquist
WI & Kern D (2002) Catalysis of cis⁄ trans isomeriza-tion in native HIV-1 capsid by human cyclophilin A Proc Natl Acad Sci USA 99, 5247–5252
6 Brazin KN, Mallis RJ, Fulton DB & Andreotti AH (2002) Regulation of the tyrosine kinase Itk by the pept-idyl–prolyl isomerase cyclophilin A Proc Natl Acad Sci USA 99, 1899–1904
7 Sarkar P, Reichman C, Saleh T, Birge RB & Kalodimos
CG (2007) Proline cis–trans isomerization controls autoinhibition of a signaling protein Mol Cell 25, 413–426
8 Dolinski K, Muir S, Cardenas M & Heitman J (1997) All cyclophilins and FK506 binding proteins are, individually and collectively, dispensable for viability in Saccharomyces cerevisiae Proc Natl Acad Sci USA 94, 13093–13098
9 Hanes SD, Shank PR & Bostian KA (1989) Sequence and mutational analysis of ESS1, a gene essential for growth in Saccharomyces cerevisiae Yeast 5, 55–72
10 Lu KP, Hanes SD & Hunter T (1996) A human pept-idyl–prolyl isomerase essential for regulation of mitosis Nature 380, 544–547
11 Lu KP (2000) Phosphorylation-dependent prolyl isomer-ization: a novel cell cycle regulatory mechanism Prog Cell Cycle Res 4, 83–96
Trang 1012 Joseph JD, Yeh ES, Swenson KI, Means AR &
Win-kler (2003) The peptidyl–prolyl isomerase Pin1 Prog
Cell Cycle Res 5, 477–487
13 Landrieu I, Smet C, Wieruszeski JM, Sambo AV,
Wintjens R, Buee L & Lippens G (2006) Exploring the
molecular function of PIN1 by nuclear magnetic
reso-nance Curr Protein Pept Sci 7, 179–194
14 Wulf G, Finn G, Suizu F & Lu KP (2005)
Phosphoryla-tion-specific prolyl isomerization: is there an underlying
theme? Nat Cell Biol 7, 435–441
15 Wulf G, Ryo A, Liou YC & Lu KP (2003) The prolyl
isomerase Pin1 in breast development and cancer
Breast Cancer Res 5, 76–82
16 Lu KP (2004) Pinning down cell signaling, cancer
and Alzheimer’s disease Trends Biochem Sci 29,
200–209
17 Butterfield DA, Abdul HM, Opii W, Newman SF, Joshi
G, Ansari MA & Sultana R (2006) Pin1 in Alzheimer’s
disease J Neurochem 98, 1697–1706
18 Ranganathan R, Lu KP, Hunter T & Noel JP (1997)
Structural and functional analysis of the mitotic
rotam-ase Pin1 suggests substrate recognition is
phosphoryla-tion dependent Cell 89, 875–886
19 Verdecia MA, Bowman ME, Lu KP, Hunter T & Noel
JP (2000) Structural basis for phosphoserine–proline
recognition by group IV WW domains Nat Struct Biol
7, 639–643
20 Jacobs DM, Saxena K, Vogtherr M, Bernado P, Pons
M & Fiebig KM (2003) Peptide binding induces large
scale changes in inter-domain mobility in human Pin1
J Biol Chem 278, 26174–26182
21 Bayer E, Goettsch S, Mueller JW, Griewel B,
Guiber-man E, Mayr LM & Bayer P (2003) Structural analysis
of the mitotic regulator hPin1 in solution: insights into
domain architecture and substrate binding J Biol Chem
278, 26183–26193
22 Landrieu I, Wieruszeski JM, Wintjens R, Inze D &
Lip-pens G (2002) Solution structure of the single-domain
prolyl cis⁄ trans isomerase PIN1At from Arabidopsis
thaliana J Mol Biol 320, 321–332
23 Li Z, Li H, Devasahayam G, Gemmill T, Chaturvedi V,
Hanes SD & Van Roey P (2003) Structural analysis of
the mitotic regulator hPin1 in solution: insights into
domain architecture and substrate binding J Biol Chem
278, 26183–26193
24 Namanja AT, Peng T, Zintsmaster JS, Elson AC,
Shakour MG & Peng JW (2007) Substrate recognition
reduces side-chain flexibility for conserved hydrophobic
residues in human Pin1 Structure 15, 313–327
25 Labeikovsky W, Eisenmesser EZ, Bosco DA & Kern D
(2007) Structure and dynamics of Pin1 during catalysis
by NMR J Mol Biol 367, 1370–1381
26 Behrsin CD, Bailey ML, Bateman KS, Hamilton KS,
Wahl LM, Brandl CJ, Shilton BH & Litchfield DW
(2007) Functionally important residues in the peptidyl–
prolyl isomerase Pin1 revealed by unigenic evolution
J Mol Biol 365, 1143–1162
27 Wu X, Wilcox CB, Devasahayam G, Hackett RL, Arevalo-Rodriguez M, Cardenas ME, Heitman J & Hanes SD (2000) The Ess1 prolyl isomerase is linked to chromatin remodeling complexes and the general transcription machinery EMBO J 19, 3727–3738
28 Zhang Y, Daum S, Wildemann D, Zhou XZ, Verdecia
MA, Bowman ME, Lu¨cke C, Hunter T, Lu KP, Fischer
G et al (2007) Structural analysis of the mitotic regula-tor hPin1 in solution: insights into domain architecture and substrate binding ACS Chem Biol 2, 320–328
29 Wintjens R, Wieruszeski JM, Drobecq H, Rousselot-Pailley P, Buee L, Lippens G & Landrieu I (2001) 1H NMR study on the binding of Pin1 Trp-Trp domain with phosphothreonine peptides J Biol Chem 276, 25150–25156
30 Pastorino L, Sun A, Lu PJ, Zhou XZ, Balastik M, Finn
G, Wulf G, Lim J, Li SH, Li X et al (2006) The prolyl isomerase Pin1 regulates amyloid precursor protein pro-cessing and amyloid-beta production Nature 440, 528– 534
31 Ramelot TA & Nicholson LK (2001) Phosphorylation-induced structural changes in the amyloid precursor protein cytoplasmic tail detected by NMR J Mol Biol
307, 871–884
32 Yaffe MB, Schutkowski M, Shen M, Zhou XZ, Stu-kenberg PT, Rahfeld JU, Xu J, Kuang J, Kirschner
MW, Fischer G et al (1997) Sequence specific and phosphorylation-dependent proline isomerization: a potential mitotic regulatory mechanism Science 278, 1957–1960
33 Yeh ES, Lew BO & Means AR (2006) The loss of PIN1 deregulates cyclin E and sensitizes mouse embryo fibroblasts to genomic instability J Biol Chem
281, 241–251
34 Orlicky S, Tang X, Willems A, Tyers M & Sicheri F (2003) Structural basis for phosphodependent substrate selection and orientation by the SCFCdc4 ubiquitin ligase Cell 112, 243–256
35 van Drogen F, Sangfelt O, Malyukova A, Matskova L, Yeh E, Means AR & Reed SI (2006) Ubiquitylation of cyclin E requires the sequential function of SCF com-plexes containing distinct hCdc4 isoforms Mol Cell 23, 37–48
36 Nash P, Tang X, Orlicky S, Chen Q, Gertler FB, Men-denhall MD, Sicheri F, Pawson T & Tyers M (2001) Multisite phosphorylation of a CDK inhibitor sets a threshold for the onset of DNA replication Nature 414, 514–521
37 Smet C, Wieruszeski JM, Buee L, Landrieu I & Lippens
G (2005) Regulation of Pin1 peptidyl–prolyl cis⁄ trans isomerase activity by its WW binding module on a multi-phosphorylated peptide of Tau protein FEBS Lett 579, 4159–4164