Download The Structure and Topology of Protein Serine/Threonine

Chapter 86
The Structure and Topology of Protein
Serine/Threonine Phosphatases
David Barford
Section of Structural Biology, Institute of Cancer Research, Chester Beatty Laboratories, London, England, UK
Introduction
Structural studies of the two families of protein phosphatases responsible for dephosphorylating serine and threonine residues have revealed that, although these families
are unrelated in sequence, the architecture of their catalytic
domains is remarkably similar and distinct from the protein
tyrosine phosphatases. The diversity of structure within the
PPP and PPM families is generated by regulatory subunits
and domains that function to modulate protein specificity
and to localize the phosphatase to particular subcellular
locations [1].
Protein serine/threonine
phosphatases of the PPP
family
The protein Ser/Thr phosphatases PP1, PP2A, and PP2B
of the PPP family, together with PP2C of the PPM ­family,
account for the majority of the protein serine/threonine
phosphatase activity in vivo. While PP1, PP2A, and PP2B
share a common catalytic domain of 280 residues, these
enzymes are divergent within their non-catalytic N and C
termini and are distinguished by their associated regulatory
subunits to form a diverse variety of holoenzymes. Major
members of the PPP family are encoded by numerous
­isoforms that share a high degree of sequence similarity,
especially within their catalytic domains. Greater sequence
diversity occurs within the extreme N and C termini of the
proteins. Although these isoforms have similar substrate
specificities and interact with the same regulatory ­subunits
in vitro, the phenotype of a functional loss is isoform
­specific, indicating they perform distinct functions in vivo.
Handbook of Cell Signaling, Three-Volume Set 2 ed.
Copyright © 2010 Elsevier Inc. All rights reserved.
Overall Structure and Catalytic
Mechanism
Structural analyses of members of the PPP family have
begun to reveal the molecular basis for catalysis, inhibition by toxins, and aspects of their regulatory mechanisms.
Crystal structures are available for (1) various isoforms of
PP1 in complex with natural toxins [2, 3], the phosphate
mimic tungstate [4], and a peptide of the RVxF targeting
motif [5]; (2) PP2B in the auto-inhibited state [6] and as a
complex with FK506/FKBP [6, 7]; (3) the PR65/A-subunit
of PP2A [8]; and (4) a regulatory TPR domain of PP5 [9].
The catalytic domains of the PP1 catalytic subunit (PP1c)
and PP2B share a common architecture consisting of a central -sandwich of two mixed -sheets surrounded on one
side by seven -helices and on the other by a subdomain
comprised of three -helices and a three-stranded -sheet
(Figure 86.1a). The interface of the three -sheets at the top
of the -sandwich creates a shallow catalytic site channel.
Conserved amino acid residues present on loops emanating
from the -strands of this central -sandwich are responsible for coordinating a pair of metal ions to form a binuclear
metal centre (Figure 86.2a). Crystallographic data on PP1c
and PP2B provided the first compelling insight regarding the role of metal ions in the catalytic reaction of the
PPP family. The identity of the two metal ions is slightly
controversial. Proton-induced X-ray emission spectroscopy performed on PP1c crystals produced from the protein expressed in Escherichia coli indicated that the metal
ions were Fe2 (or Fe3) and Mn2 [4], whereas atomic
absorption spectroscopy of bovine brain PP2B indicated
a stoichiometric ratio of Zn2 and iron [10]. There have
also been conflicting reports concerning the iron oxidation
states, with both Fe2 and Fe3 being observed. Native
677
678
Figure 86.1 Ribbon diagrams showing an overview of the structures of protein Ser/Thr phosphatase catalytic domains.
(a) The catalytic subunit of human protein phosphatase 1 in complex
with the RVxF motif PP1 binding peptide. The catalytic site is indicated
by the two metal ions and bound sulfate ion. (b) Human PP2C the
catalytic site is indicated by the two metal ions and bound phosphate ion.
Figure 86.1a reproduced from [5] (Egloff et al., 1997), with permission;
Figure 86.2 is reproduced from [19] (Das et al., 1996), with permission.
PP2B is most likely to contain Fe2, explaining the timedependent inactivation of PP2B that results from the oxidation of the Fe–Zn center [11]. Oxidation of the binuclear
metal center and phosphatase inactivation may represent
a mechanism for PP2B regulation by redox potential during oxidative stress or as a result of reactive oxygen species
generation following receptor tyrosine kinase activation, in
a process reminiscent of the inactivation of protein tyrosine
phosphatases by oxidation of the catalytic site Cys residue
by hydrogen peroxide and perhaps analogous to a possible
PP2C regulatory mechanism, discussed later [10].
The structure of PP1c with tungstate and PP2B with
phosphate indicated that two oxygen atoms of the oxyanion-substrate coordinate the metal ions (Figure 86.2a)
[4, 7]. Two water molecules, one of which is a metalbridging water molecule, contribute to the octahedral
hexa-­coordination of the metal ions. The metal coordinating residues (aspartates, histidines, and asparagines) are
PART | II Transmission: Effectors and Cytosolic Events
Figure 86.2 Comparison of the catalytic sites of the protein Ser/Thr
phosphatases, showing the common binuclear metal center, coordinating
the phosphate of the phosphorylated protein substrate: (a) human PP1
(for clarity, Arg 96 is not shown) and (b) human PP2C.
invariant among all PPP family members. These residues,
together with Arg and His residues that interact with the
phosphate group of the phosphorylated residue, occur
within five conserved sequence motifs found in other
enzymes, including the purple acid phosphatase, whose
common function is to catalyze phosphoryl transfer reactions to water [12]. These observations suggest that PPP
and purple acid phosphatases evolved by divergent evolution from an ancestral metallophosphoesterase. Consistent
with roles in catalysis, mutation of these residues either
eliminates or profoundly reduces catalytic activity.
PPPs catalyze dephosphorylation in a single step with a
metal-activated water molecule or hydroxide ion. The most
convincing evidence for this notion is that the purple acid
phosphatase, which is generally related in structure to the
PPPs both at the tertiary level and at the catalytic site [13],
promote dephosphorylation with inversion of configuration
of the oxygen geometry of the phosphate ion [14]. This
indicates that a phosphoryl-enzyme intermediate would not
occur. The two metal-bound water molecules are within
van der Waals distance of the phosphorous atom of the
phosphate bound to the catalytic site, and one of them is
likely to be metal-activated nucleophile.
Chapter | 86 The Structure and Topology of Protein Serine/Threonine Phosphatases
Interactions with Regulatory Subunits
PP1 and PP2A are responsible for regulating diverse cellular functions by dephosphorylating multiple and ­ varied
protein substrates. This seemingly paradoxical situation
was resolved by the discovery that distinct forms of PP1
and PP2A holoenzymes occur in vivo, where essentially
the same catalytic subunit is complexed to different targeting and regulatory subunits. For PP1 it has been shown that
targeting subunits confer substrate specificity by directing particular PP1 holoenzymes to a subcellular location
and by enhancing or suppressing activity toward different
substrates. The control of PP1 holoenzyme structure and
activity by the combinatorial selection of different targeting/regulatory subunits has recently been the subject of
numerous reviews [15, 16] and will not be discussed at
length here, although the contrasting structural mechanism
by which the conserved PP1 and PP2A catalytic subunits
are able to form diverse holoenzyme structures will be
discussed.
In the case of PP1, it is known that the binding of targeting subunits to its catalytic subunit is mutually exclusive, suggesting that there are one or more common or
overlapping binding sites, recognized by all PP1-binding
subunits. It is therefore a little surprising that PP1-binding
subunits are highly diverse structurally and share little to
no overall sequence similarities. The key to understanding this paradox came from the crystal structure of PP1c
in complex with a short, 13-residue peptide derived from a
region of the PP1-glycogen targeting subunit (GM) responsible for PP1c interactions (Figure 86.1a). This structure
showed that the peptide associated with the phosphatase
via two hydrophobic residues (Val and Phe), which engage
a hydrophobic groove on the protein surface formed from
the interface of the two -sheets of the central -sandwich
and remote from the catalytic site [5]. Two basic residues
of the peptide immediately N-terminal to the Val residue
form salt-bridge interactions with Asp and Glu residues at
one end of the peptide binding channel. Alanine substitutions of either the Val or Phe residues of the peptide abolish
PP1–peptide interactions. Analysis of other PP1-binding
subunit sequences revealed the presence of the identical or
related sequence motif RRVxF, found to mediate the interactions between the GM peptide and PP1c [5]. The role of
the degenerative RVxF motif in mediating PP1-regulatory
subunit interactions is now supported by numerous experimental observations. First, for various regulatory subunits, mutation of either hydrophobic residue of the motif
in native proteins weakens or eliminates their association
with PP1c. Second, peptides corresponding to the RVxF
motif competitively disrupt the interactions of regulatory
subunits with PP1c. Third, the use of a common or overlapping PP1c binding site explains why the interactions of
regulatory subunits is mutually exclusive. The number of
PP1 holoenzyme structures that could be generated in this
679
manner is potentially infinite, and to date over a hundred
PP1 regulatory subunits have been characterized.
The residues of PP1 that interact with the RVxF peptide
are conserved in all isoforms of PP1 in all eukaryotic species, although not within PP2A and PP2B, explaining why
PP1-binding subunits are unique to PP1. The mechanism
of combinatorial control of PP2A holoenzyme structure is
different from that of PP1 and is mediated by a scaffolding
subunit termed the PR65/A subunit, which simultaneously
associates with the PP2A catalytic subunit and a variable
regulatory B subunit. The ability to recognize a variety of
regulatory subunits (perhaps over 50) is conferred by the
architecture of the PR65/A subunit, which consists of 15
tandem repeats of a 39 amino-acid sequence termed the
HEAT motif and related in structure to ARM repeats.
These repeats assemble to create an extended molecule ideally suited for mediating protein–protein interactions [8].
Combinatorial generation of variable PP2A holoenzymes is
achieved by the ability of different combinations of HEAT
motifs to select different regulatory B subunits [8].
Interactions of Natural Toxins with PP1
PP1 and PP2A are specifically and potently inhibited by a
variety of naturally occurring toxins such as okadaic acid,
a diarrhetic shellfish poison and powerful tumor promoter,
and microcystin, a liver toxin produced by blue-green algae
[17]. Whereas PP2B is only poorly inhibited by the toxins
that affect PP1 and PP2A, it is known to be the immunosuppressive target of FK506 and cyclosporin in association
with their major cellular binding proteins, the cis-trans peptidyl prolyl isomerase FKBP12 and cyclophilin, respectively
[18]. The mechanism of inhibition of PP1 by okadaic acid
and microcystin LR have been defined by structures of PP1
in complex with these inhibitors. Both inhibitors, although
structurally different, bind to a similar region of the phosphatase, occupying the catalytic channel to directly block
phosphatase–substrate interactions. Regions of PP1 that
contact the toxins include the hydrophobic groove and the
12/13 loop, the latter undergoing conformational changes
to optimize contacts with microcystin [2, 3]. Both toxins
disrupt substrate–phosphatase interactions by competing
for sites on the protein that coordinate the phosphate group
of the substrate. For example, carboxylate and carbonyl
groups of microcystin interact with two of the metal-bound
water molecules [2], whereas okadaic acid contacts the two
phosphate-coordinating arginine residues (Arg-96 and Arg221) [3]. A similar mechanism of phosphatase inhibition
by steric hindrance of substrate binding is observed in the
structure of the full-length PP2B holoenzyme, for which
the autoinhibitory domain lies over the substrate binding channel of the catalytic domain in such a way that a
Glu side chain accepts a hydrogen bond from two of the
metal-bound water molecules [6].
680
Protein serine/threonine
phosphatases of the PPM family
Protein phosphatases of the PPM family are present in both
eukaryotes and prokaryotes for which the defining member
is PP2C. Biochemically, the PPM family was distinguished
from the PPP family by its requirement for divalent metal
ions (Mg2) for catalytic activity, although it is now known
from crystal structures that both PPP and PPM phosphatases catalyze dephosphorylation reactions by means
of a binuclear divalent metal center. Within the PPM family, the PP2C domain occurs in numerous structural contexts that reflect structural diversity [19]. For example, the
PP2C domain of the Arabidopsis ABI1 gene is fused with
EF hand motifs, whereas in KAPP-1, a kinase interaction
domain associated with a phosphorylated receptor precedes
the phosphatase domain. Other less closely related examples include the Ca2 stimulated mitochondrial pyruvate
dehydrogenase phosphatase, which contains a catalytic
subunit sharing 22 percent sequence identity with that of
mammalian PP2C, and the SpoIIE phosphatase of Bacillus
subtilus, which has ten membrane-spanning regions preceding the PP2C-like catalytic domain. A surprising homolog
is a 300-residue region of yeast adenylyl cyclase present
immediately N-terminus to the cyclase catalytic domain
that shares sequence similarity with PP2C. This domain
may function to mediate Ras-GTP activation of adenylyl
cyclase activity and is not known to possess protein phosphatase activity. In eukaryotes, the various isoforms of
PP2C have been implicated in diverse functions such as
regulation of cell-cycle progression mediated by dephosphorylation of CDKs [20], to regulation of RNA splicing
[21], control of p53 activity [22], and regulation of stress
response pathways in yeast [23].
The sequences of protein phosphatases of the PPM family share no similarity with those of the PPP family, and
the natural toxins that inhibit the PPP family have no affect
on PPM family phosphatases. It therefore came as a surprise when the crystal structure of human PP2C revealed
a striking similarity in tertiary structure and catalytic site
architecture to the PPP protein phosphatases (Figure 86.1b)
[19]. Mammalian PP2C consists of two domains; an Nterminal catalytic domain common to all members of the
PP2C family is fused to a 90-residue C-terminal domain,
unique to the mammalian PP2Cs [19]. The catalytic
domain is dominated by a central, buried -sandwich of
11 -strands formed by the association of two antiparallel
-sheets, both of which are flanked by a pair of antiparallel -helices inserted between the two central -strands,
with four additional -helices. The C-terminal domain is
formed from three antiparallel -helices remote from the
catalytic site, suggesting a role in defining substrate specificity rather than catalysis.
At the catalytic site of PP2C, two Mn2 ions, separated
by 4 Å, form a binuclear metal center and are coordinated
PART | II Transmission: Effectors and Cytosolic Events
by four invariant aspartate residues and a non-conserved Glu
residue (Figure 86.2b) [19]. These residues are situated at
the top of the central -sandwich that forms a shallow channel suitable for the dephosphorylation of phosphoserine- and
phosphothreonine-containing proteins. Six water molecules
coordinate the two metal ions. One of these water molecules
bridges the two metal ions and four form hydrogen bonds to
a phosphate ion at the catalytic site. Dephosphorylation is
probably catalyzed by a metal-activated water molecule that
acts as nucleophile in a mechanism similar to that proposed
for the PPP family. A recent kinetic analysis of PP2C by
Denu and colleagues [24] indicated that Mn2 and Fe2 are
the most effective divalent metal ions in promoting dephosphorylation reactions, suggesting that at least one of these
ions must be present at the catalytic site. In contrast, Zn2
and Ca2 competitively inhibit PP2C Mn2-dependent
activity. An Fe2-containing catalytic site would be analogous to the PPP protein phosphatases and possibly explains
the H2O2-mediated inactivation of PP2C [25] consequent on
the redox sensitivity of Fe2 and its oxidation to the inactive
Fe3 valence state [24]. The finding that an ionizable group
with a pKa of 7.0 has to be deprotonated for catalysis was
fully consistent with the notion from the crystal structure
that a metal activated water molecule acts as a nucleophile
[19, 24]. Fluoride has long been used as an inhibitor of both
PPP and PPM phosphatases, and the rationale for this inhibition is now clear from the catalytic mechanism of serine/
threonine phosphatases revealed by the crystal structures.
By substituting for the metal-bound nucleophilic water molecule, fluoride prevents metal-activated nucleophilic attack
on the phospho-protein substrate. Substitution of Ala for the
Asp residues of the catalytic site in yeast and plant PP2C
homologs and in the related phosphatase SpoIIE from B.
subtilus abolishes catalytic activity, supporting a role for the
metal ions in catalysis [26, 27].
Conclusions
Although the PPP and PPM phosphatases share a similar
tertiary structure, characterized by a central -sandwich
and flanking -helices, with related catalytic site architectures and mechanisms, two observations suggest that
these protein families have evolved from distinct ancestors.
First, the secondary structure topology of the PPP and PPM
families are unrelated and there is no simple rearrangement
of chain connectivities that would allow the PPP topology to be transformed into the PPM topology. Second, the
conserved sequence motifs of the PPP family required for
metal binding and catalysis, typical of some metallophosphoesterases, are distinct from the conserved sequence
motifs of the PPM family. The comparison of the PPP and
PPM families of Ser/Thr protein phosphatases provides
interesting contrasts with the protein phosphatases of the
C(X)5R-motif superfamily composed of three distinct gene
Chapter | 86 The Structure and Topology of Protein Serine/Threonine Phosphatases
families: (1) the conventional protein tyrosine phosphatases
(PTPs), including dual-specificity phosphatases and PTEN
lipid phosphatases; (2) low-molecular-weight protein tyrosine phosphatases (lmPTPs); and (3) Cdc25 dual-specificity
phosphatases [28]. These proteins share a similar overall
tertiary structure composed of a central -sheet surrounded
on both sides by -helices which results in an identical
catalytic site architecture in all three families characterized
by the nucleophilic Cys-residue located within a phosphate
binding cradle. What distinguishes the three protein families are differences in their secondary structure topology;
however, the topologies of the PTP and lmPTP families are
related by a simple permutation of secondary structure connectivity, suggesting that these families may have evolved
from the same ancestor and diverged as a result of exon
shuffling events. In recent years, much has been learned of
the structural details of Ser/Thr protein phosphatases relating to their overall folds, catalytic mechanisms, and interactions with regulatory subunits and toxins. However, still
elusive is the structure of a Ser/Thr protein phosphatase
in complex with a phosphoprotein substrate that would
explain the basis for the selectivity for Ser/Thr residues and
reveal the mechanism of substrate selectivity by regulatory
subunits. It is hoped that future studies of Ser/Thr phosphatase will address these questions.
References
1. Hubbard MJ, Cohen P. On target with a new mechanism for the regulation of protein phosphorylation. Trends Biochem Sci 1993;18:172–7.
2. Goldberg J, Huang HB, Kwon YG, Greengard P, Nairn AC, Kuriyan
J. Three-dimensional structure of the catalytic subunit of protein serine/threonine phosphatase-1. Nature 1995;376:745–53.
3. Maynes JT, Bateman KS, Cherney MM, Das AK, Luu HA,
Holmes CF, James MN. Crystal structure of the tumor-promoter okadaic acid bound to protein phosphatase-1. J Biol Chem
2001;276:44,078–44,082.
4. Egloff MP, Cohen PT, Reinemer P, Barford D. Crystal structure of
the catalytic subunit of human protein phosphatase 1 and its complex
with tungstate. J Mol Biol 1995;254:942–59.
5. Egloff MP, Johnson DF, Moorhead G, Cohen PT, Cohen P, Barford D.
Structural basis for the recognition of regulatory subunits by the catalytic subunit of protein phosphatase 1. EMBO J 1997;16:1876–87.
6. Kissinger CR, Parge HE, Knighton DR, Lewis CT, Pelletier LA,
Tempczyk A, Kalish VJ, Tucker KD, Showalter RE, Moomaw EW, et
al. Crystal structures of human calcineurin and the human FKBP12–
FK506–calcineurin complex. Nature 1995;378:641–4.
7. Griffith JP, Kim JL, Kim EE, Sintchak MD, Thomson JA, Fitzgibbon
MJ, Fleming MA, Caron PR, Hsiao K, Navia MA. X-ray structure
of calcineurin inhibited by the immunophilin-immunosuppressant
FKBP12–FK506 complex. Cell 1995;82:507–22.
8. Groves MR, Hanlon N, Turowski P, Hemmings BA, Barford D. The
structure of the protein phosphatase 2 A PR65/A subunit reveals
681
the conformation of its 15 tandemly repeated HEAT motifs. Cell
1999;96:99–110.
9. Das AK, Cohen PW, Barford D. The structure of the tetratricopeptide
repeats of protein phosphatase 5: implications for TPR-mediated protein–protein interactions. EMBO J 1998;17:1192–9.
10. Wang X, Culotta VC, Klee CB. Superoxide dismutase protects calcineurin from inactivation. Nature 1996;383:434–7.
11. Yu L, Haddy A, Rusnak F. Evidence that calcineurin accommodates an active site binuclear metal center. J Am Chem Soc
1995;117:10,147–10,148.
12. Lohse DL, Denu JM, Dixon JE. Insights derived from the structures
of the Ser/Thr phosphatases calcineurin and protein phosphatase 1.
Structure 1995;3:987–90.
13. Klabunde T, Strater N, Frohlich R, Witzel H, Krebs B. Mechanism of
Fe(III)–Zn(II) purple acid phosphatase based on crystal structures. J
Mol Biol 1996;259:737–48.
14. Mueller EG, Crowder MW, Averill BA, Knowles JR. Purple acid
phosphatase: a diron enzyme that catalyses a direct phosphogroup
transfer to water. J Am Chem Soc 1993;115:2974–5.
15. Bollen M. Combinatorial control of protein phosphatase-1. Trends
Biochem Sci 2001;26:426–31.
16. Cohen PT. Protein phosphatase 1: targeted in many directions. J Cell
Sci 2002;115:241–56.
17. MacKintosh C, MacKintosh RW. Inhibitors of protein kinases and
phosphatases. Trends Biochem Sci 1994;19:444–8.
18. Liu Jr. J, Farmer JD, Lane WS, Friedman J, Weissman I, Schreiber
SL Calcineurin is a common target of cyclophilin-cyclosporin A and
FKBP–FK506 complexes. Cell 1991;66:807–15.
19. Das AK, Helps NR, Cohen PTW, Barford D. Crystal structure of
human protein serine/threonine phosphatase 2C at 2.0-Å resolution.
EMBO J 1996;15:6798–809.
20. Cheng A, Ross KE, Kaldis P, Solomon MJ. Dephosphorylation of
cyclin-dependent kinases by type 2C protein phosphatases. Genes
Dev 1999;13:2946–57.
21. Murray MV, Kobayashi R, Krainer AR. The type 2C Ser/Thr
phosphatase PP2C is a pre-mRNA splicing factor. Genes Dev
1999;13:87–97.
22. Takekawa M, Adachi M, Nakahata A, Nakayama I, Itoh F, Tsukuda
H, Taya Y, Imai K. p53-inducible wip1 phosphatase mediates a negative feedback regulation of p38 MAPK-p53 signaling in response to
UV radiation. EMBO J 2000;19:6517–26.
23. Shiozaki K, Russell P. Counteractive roles of protein phosphatase 2C
(PP2C) and a MAP kinase kinase homolog in the osmoregulation of
fission yeast. EMBO J 1995;14:492–502.
24. Fjeld CC, Denu JM. Kinetic analysis of human serine/threonine protein phosphatase 2C. J Biol Chem 1999;274:20,336–20,343.
25. Meinhard M, Grill E. Hydrogen peroxide is a regulator of ABI1, a protein phosphatase 2C from Arabidopsis. FEBS Letts 2001;508:443–6.
26. Adler E, Donella-Deana A, Arigoni F, Pinna LA, Stragler P. Structural
relationship between a bacterial developmental protein and eukaryotic
PP2C protein phosphatases. Mol Microbiol 1997;23:57–62.
27. Sheen J. Mutational analysis of protein phosphatase 2C involved in
abscisic acid signal transduction in higher plants. Proc Natl Acad Sci
USA 1998;95:975–80.
28. Barford D, Das AK, Egloff M-P. The structure and mechanism of
protein phosphatases. Insights into catalysis and regulation. Annu Rev
Biophys Biomol Struct 1998;27:133–64.