Determinants for substrate specificity of the bacterial PP2C protein phosphatase tPphA from Thermosynechococcus elongatus

Article abstract of DOI:10.1111/j.1742-4658.2011.08466.x

Members of the Mg²⁺- or Mn²⁺-dependent protein phosphatases/PP2C-like serine/threonine phosphatases (PPM/PP2C) are abundant and widely distributed in prokaryotes and eukaryotes, where they regulate diverse signal transduction pathway

Full text of DOI:10.1111/j.1742-4658.2011.08466.x

Determinants for substrate specificity of the bacterial
PP2C protein phosphatase tPphA from
Thermosynechococcus elongatus
Jiyong Su and Karl Forchhammer
Interfaculty Institute for Microbiology and Infection Medicine, Department of Organismic Interactions, University of Tubingen, Germany
¨
Members of the Mg2+- or Mn2+-dependent protein phosphatases ⁄PP2C-
like serine ⁄ threonine phosphatases (PPM⁄ PP2C) are abundant and widely
distributed in prokaryotes and eukaryotes, where they regulate diverse
Keywords
cross-linking; PII signalling; PPM
phosphatase; protein phosphorylation; signal
transduction
signal transduction pathways. Despite low sequence conservation, the struc-
ture of their catalytic core is highly conserved except for a flexible loop
Correspondence
termed the flap subdomain. Bacterial PPM⁄ PP2C members without C- or
K. Forchhammer, Interfaculty Institute for
N-terminal regulatory domains still recognize their substrates. Based on the
crystal structure of tPphA (a PPM ⁄ PP2C member from the cyanobacte-
rium Thermosynechococcus elongatus), variants of tPphA were generated by
site-directed mutagenesis to identify substrate specificity determinants. Fur-
thermore, a PPM ⁄ PP2C chimera containing the tPphA catalytic core and
the flap subdomain of human PP2Ca was also generated. tPphA variants
and the chimera were tested towards different artificial substrates and
native phosphorylated PII. A binding assay combining chemical crosslink-
ing and pull-down was designed to analyze the binding of the various phos-
phatase variants to phosphoprotein PII. Together, these data showed that
the metal 1–metal 2 cluster in the catalytic center, but not the catalytically
active metal 3, is required for the binding of phosphorylated substrate.
Residues outside the catalytic center are pivotal for the recognition and
turnover of phosphorylated protein substrate. In particular, a histidine resi-
due (His39) of tPphA was identified to play a specific role in protein sub-
strate dephosphorylation. Furthermore, mutations in the variable flap
subdomain can affect enzyme activity as well as substrate specificity.
Microbiology and Infection Medicine,
Department of Organismic Interactions,
University of Tubingen, Auf der
¨
Morgenstelle 28, D-72076 Tubingen,
¨
Germany
Fax: 49 7071 295843
Tel: 49 7071 2972096
E-mail: karl.forchhammer@uni-tuebingen.de
(Received 12 September 2011, revised 7
November 2011, accepted 19 December
2011)
doi:10.1111/j.1742-4658.2011.08466.x
Structured digital abstract
l
tPphA binds to P-II by cross-linking study (View Interaction: 1, 2, 3, 4, 5, 6, 7, 8)
Introduction
Reversible protein phosphorylation affects the struc-
ture and function of proteins that are responsible for
the regulation of nearly all biological processes in liv-
ing organisms. Protein kinases and protein phosphata-
ses control the state of protein phosphorylation.
Protein phosphatases can be divided into three main
subclasses based on their substrate specificity and
protein sequence: protein tyrosine phosphatase [1];
protein serine ⁄ threonine phosphatase [2]; and histidine
phosphatase [3]. Protein serine ⁄threonine phosphatase
can be further classified into three major subfamilies:
phosphoprotein phosphatases, the aspartate-based
phosphatases, and metal-dependent protein phosphata-
ses (PPM ⁄PP2C) [2]. The human PPM member PP2Ca
Abbreviations
IPTG, isopropyl thio-b-D-galactoside; pNPP, p-nitrophenyl phosphate; PII-P, phosphorylated PII protein; PP2C, PP2C-like serine ⁄ threonine
phosphatase; PPM, Mg2+- or Mn2+-dependent protein phosphatase.
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS
1

Determinants for protein phosphatase substrate specificity
J. Su and K. Forchhammer
[4] has been the defining representative of the PPM
family, which is therefore also referred to as the PP2C
family. The catalytic domain of these family members
comprises eight absolutely conserved residues within
11 conserved motifs [5–7]. PP2C homologs are widely
present in eukaryotes and prokaryotes where they reg-
ulate diverse signaling pathways [8]. In the human gen-
ome, there are 16 distinct PP2C genes that give rise to
at least 22 different isoforms [9]. They participate in
many biological processes including regulation of mito-
gen-activated protein kinase cascades, cell cycle
progression, ubiquitination and degradation of pro-
teins, and mechanisms for cell death and survival [8].
In plants, there are even more PP2C genes with Ara-
bidopsis thaliana harbouring 82 PP2C genes [10].
Except for six, all the others can be classified into 10
groups (groups A–J) based on sequence similarity [10].
Group A contains genes whose products have been
shown to act as negative regulators in the abscisic acid
(ABA) signaling pathway, which takes part in the
response to different environmental stresses such as
drought, cold and salinity [8,11]. In bacteria, many
PP2Cs lack potential regulatory domains but neverthe-
less they still specifically dephosphorylate a variety of
substrates [12–15].
The PP2C phosphatases require Mg2+ or Mn2+
ions as ligands, which are coordinated by a universally
conserved core of aspartate residues. Structural analy-
sis of bacterial and plant PP2C members revealed that
these enzymes contain three metal ions (M1, M2 and
M3) embedded in a catalytic core with almost invari-
ant structure [11,16–20] except for the variable region
termed the flap subdomain [17]. The flap, which pro-
trudes near the M3 binding site, consists of a flexible
loop and a-helical segments (Fig. 1A). In the catalytic
center, M1 and M2 coordinate a catalytically active
nucleophilic water molecule and, according to novel
structures, they could interact with two oxygen atoms
of the phosphate group [18]. Recently, we could show
by mutational and structural analysis of a cyanobacte-
rial PP2C family member (tPphA from Thermosyn-
echococcus elongatus) that the loosely bound M3,
which was occasionally observed in several bacterial
and plant PP2C homologs, is required for catalysis,
presumably by activating a water molecule as proton
donor [21]. tPphA is the homolog of PphA from
Fig. 1. (A) The crystal structure of tPphA (PDB 2J86). The amino acids that were analyzed by site-directed mutagenesis are indicated. Metal
1, Metal 2 and Metal 3 (M1, M2 and M3) constitute the catalytic center of tPphA. The flap subdomain is close to M3. (B) The crystal struc-
ture of PII from Synechococcus elongatus PCC 7942 (PDB 1QY7). The T-loop is indicated. (C) The crystal structure of SaSTP (PDB 2PK0).
The His41 residue from monomer C could make a salt bridge with the Glu152 residue from monomer A. The homolog residue of SaSTP
His41 in tPphA is His39. (D) Overlay of the crystal structures of tPphA (colored cyan, PDB 2J86) and human PP2Ca (colored magenta, PDB
1A6Q). The human PP2Ca flap (termed hflap in the figure) and tPphA flap (termed tflap in the figure) show different conformations.
2
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS

J. Su and K. Forchhammer
Determinants for protein phosphatase substrate specificity
Synechocystis sp. PCC 6803, which is the phosphatase
that dephosphorylates the signaling protein PII. The
PII protein has been shown to be phosphorylated ⁄
dephosphorylated on a seryl residue (Ser49) located at
the apex of a large solvent-exposed loop termed T-loop
(Fig. 1B) in response to the binding of signaling
metabolites [22,23]. The genome of Synechocystis
contains eight putative PP2C type homologs [24,25].
However, only one out these eight genes turned out to
be responsible for phospho-PII (PII-P) dephosphoryla-
tion [11,26], raising the question of the mechanistic
basis for the specificity of PII towards PphA. The
alignment of these eight PP2Cs showed the greatest
sequence variability in the flap subdomain. The tPphA
observation that enzymatically active recombinant
human PP2Ca (amino acids 1–297, without the C-reg-
ulatory domain) could not dephosphorylate PII-P
although a synthetic T-loop phosphopeptide could be
readily dephosphorylated [21]. The structure of the
catalytic core of tPphA and human PP2Ca is highly
similar, but their flap subdomains differ substantially
in structure (Fig. 1D) and primary sequence [16].
Since the structures of tPphA and of PII are solved
(Fig. 1A, B) [16,30] and the conformational changes
involved in PII signaling are well studied [31], these
proteins represent a well suited example for studying
the fundamental properties of PP2C substrate interac-
tion. In order to find key residues ⁄ domains in tPphA
for substrate recognition, residues surrounding the
catalytic center including the flap segment were ana-
lyzed by site-directed mutagenesis (Fig. 1A) and a new
PP2C chimera protein (tPphA catalytic cen-
ter + human PP2Ca flap) was generated (Fig. 2B).
Substrate-specific activities of the variants were assayed
and a phosphatase binding assay, based on chemical
crosslinking under enzymatically arrested conditions,
was developed to assay the direct interaction between
the enzymes and the PII protein substrate.
protein displays
a highly similar flap subdomain
sequence to PphA and it could specifically dephos-
phorylate PII-P in an effector-molecule-dependent
manner [16].
A clue to understanding the specific interaction
between PP2C phosphatases and their phosphoprotein
substrates came from the crystallographic analysis of
the Streptococcus agalactiae PP2C phosphatase SaSTP
[19]. An SaSTP crystal with four monomers (A, B, C
and D) in the asymmetric unit revealed monomer C
in a conformation which interacted with the flap
subdomain of the adjacent monomer A (Fig. 1C).
The interactions include two salt bridges (Arg12(C)–
Glu151(A) and His41(C)–Glu152(A)) and three direct
hydrogen bonds (Ser14(C)–Glu152(A), Arg13(C)–
Ser155(A) and Ile162(C)–Pro157(A)). Whether this
interaction mimicked an enzyme–product complex or
was a serendipitous crystallographic artifact was an
open question [19]. However, SaSTP His41 is highly
conserved among bacterial PP2C phosphatases. The
corresponding residue in tPphA is His39, which is
located at the tip of a small loop (Fig. 1A). His39
seems to be flexible, since this residue is invisible in
one space group (C2221) of the wild-type tPphA crys-
tal [16] and in the crystal of the D193A variant [21].
Whether the residue is indeed involved in substrate
recognition or product interaction needs to be
resolved. A further clue for understanding substrate
recognition and regulation was provided by the
atomic structures of A. thaliana PP2Cs in complex
with ABA-regulated PP2C inhibitor proteins [11,27].
The structures revealed a highly conserved tryptophan
residue in the flap subdomain of ABI1 (ABA-insensi-
tive1) and HAB1 (hypersensitive to ABA1), which
directly docked into a cavity of the inhibitor proteins
leading to inhibition of ABI1⁄ HAB1 activity, a fact
that highlights an important regulatory role of the
flap subdomain [11,27–29]. In support of a role of
the flap subdomain for substrate recognition is the
Results
In this study, several new tPphA variants (R13K, S15A,
Q17E, Q17K, M36A, H39A, R160A, T138A and
T138E) were created. The previously created variants
D18A, D34A, D119A, D193A, D231A [21], H161A and
R169A [16] were also included in this research. Fig-
ure 1A shows the location of the mutated residues in the
3D structure of tPphA. According to the classic PP2C
motif nomenclature [5], the primary sequence of tPphA
contains all 11 conserved motifs and eight highly con-
served residues (shown in Fig. 2A). Motif 5b and motif
6 constitute the flap subdomain. The positions of
mutated residues in tPphA primary sequence are indi-
cated (Fig. 2A). Furthermore, the flap subdomains of
tPphA and of human PP2Ca were mutually exchanged
to generate two chimeric proteins (Fig. 2B). In chimera
A, the human PP2Ca flap subdomain was fused to the
tPphA catalytic core. In chimera B, the tPphA flap sub-
domain was fused to the human PP2Ca catalytic core.
Whereas construction of chimera A resulted in soluble
and active protein (see below), chimera B was insoluble
and could not be re-solubilized under various conditions
(details on the manipulation of chimera B are given in
the section entitled ‘Cloning, overexpression and purifi-
cation of human PP2Ca and the chimeras’). Further
experiments were therefore only performed with
chimera A.
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS
3

Determinants for protein phosphatase substrate specificity
J. Su and K. Forchhammer
Fig. 2. (A) The motif nomenclature of tPphA according to Bork et al. 1996 [5]. tPphA contains all 11 conserved motifs and eight highly con-
served residues (colored red). Motif 5b and motif 6 constitute the flap subdomain. The amino acids, which were analyzed by site-directed
mutagenesis, are indicated by blue numbers. (B) The strategy of exchanging the flap subdomains of tPphA and human PP2Ca to generate
two chimeric proteins. The tPphA flap subdomain (tflap) is labeled in blue, the human PP2Ca flap subdomain (hflap) is labeled in green.
Only the catalytic domain (1–297) of human PP2Ca was used, whereas the C-terminal segment (298–382, grey a helices in the figure) was
truncated.
2J82 [16]) reveals that Gln17 indirectly coordinates M2
Enzymatic assays of tPphA variants, human
by a bridging water molecule. In many other PP2C
PP2Ca and chimera A towards artificial
homologs, the corresponding position is occupied by a
substrates and PII-P
glutamyl residue (see Supplementary file S1 in [20]).
To test the effect of amino acid substitutions on tPphA
activity, the various phosphatase variants were assayed
with the following substrates: the artificial substrate
p-nitrophenyl phosphate (pNPP) (Table 1), three dif-
ferent phosphopeptides, of which one corresponds to
the phosphorylated T-loop of PII (Table 2), and the
physiological substrate PII-P (Fig. 3 and Fig. S1A–C).
As shown previously, the alanine variants of the aspar-
tate residues D18, D34, D119, D193 and D231, which
coordinate the metals in the catalytic center, are cata-
lytically inactive [21]. The newly generated variants
displayed a large heterogeneity with respect to their
catalytic properties, as described in more detail in the
following. Figure 4A shows a comprehensive compari-
son of the variants, which were differentially affected
in substrate specificity. To facilitate the comparison,
the enzyme activities towards the five different
substrates were normalized to the respective activity of
wild-type tPphA, which was set to 100%. Among all
newly created variants, variant Q17K was most
strongly affected in catalytic activity, with only about
10% residual activity towards pNPP and the threonyl
peptide and no detectable activity towards seryl pep-
tides and PII-P. The tPphA crystal structure (PDB file
Interestingly, the Gln17 to Glu mutation (Q17E)
resulted in increased catalytic efficiency towards an
artificial phosphoseryl peptide (pS-peptide), whereas
the activity towards the other substrates was very simi-
lar to wild-type tPphA. Of the other novel variants,
variant S15A was not significantly affected in catalytic
properties. Several variants (R13K, M36A, T138A,
R160A, H161A and R169A) showed a general reduc-
tion in catalytic activity, dephosphorylating all tested
substrates more slowly than wild-type tPphA. Appar-
ently, these point mutations affected the catalytic activ-
ity of the corresponding tPphA variants to different
degrees, showing effects on Kcat and Km as presented
in Table 1. Residues Arg13 and Met36 are close to the
catalytic center of tPphA. The conservative replace-
ment of Arg13 by Lys resulted in a moderate reduction
of activity. By contrast, the M36A variant showed
strongly reduced catalytic activity towards all
substrates, suggesting that the bulky methionyl side
chain could be required to stabilize the structure of the
catalytic center or to exclude water from the catalytic
center. Alanine replacement of residues Thr138,
Arg160, His161 and Arg169, all belonging to the flap
subdomain (Figs 1A and 2A), moderately reduced
4
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS

J. Su and K. Forchhammer
Determinants for protein phosphatase substrate specificity
Table 1. Kinetic parameters of tPphA and its variants towards pNPP. pNPP assays were carried out as described in Experimental proce-
dures. From the apparent reaction velocities of three independent repetitions, the kinetic parameters were calculated by linear fitting using
the program GRAPHPAD PRISM 4. Standard errors are given. The results were obtained from single preparations of tPphA variants, human
PP2Ca and chimera A purifications and the data come from single experiments.
)1
Km (mM) pNPP
Kcat (s)1) pNPP
Kcat ⁄ Km (s)1ÆM
)
Km (mM) Mn2+
WT
0.47 0.070
0.88 0.075
0.42 0.038
0.66 0.055
1.02 0.090
0.81 0.059
0.62 0.044
0.45 0.052
0.74 0.065
0.83 0.071
0.57 0.041
0.61 0.062
0.45 0.192
0.51 0.170
0.85 0.055
0.60 0.026
0.83 0.020
1.28 0.050
0.17 0.008
0.26 0.009
1.10 0.036
0.52 0.025
0.26 0.011
0.26 0.011
0.57 0.018
0.49 0.023
1.37 0.259
0.16 0.024
1809 117
682 30
0.57 0.088
0.96 0.052
0.57 0.042
0.74 0.062
2.03 0.187
0.78 0.049
0.68 0.071
0.90 0.072
4.10 0.33
1.34 0.120
0.82 0.053
1.29 0.118
0.48 0.15
0.63 0.093
R13K
S15A
1976 48
1939 76
Q17E
Q17K
167
8
M36A
321 11
1774 58
1156 56
351 15
313 13
1000 32
803 38
3044 576
314 47
H39A
T138A
T138E
R160A
H161A
R169A
Human PP2Ca
Chimera A
Table 2. The activity of tPphA and its variants towards three phosphopeptides. Reactions were performed in the buffer as described in
Experimental procedures. Triplicate assays were used. Standard errors are given. The results were obtained from single preparations of
tPphA variants, human PP2Ca and chimera A purifications and the data come from single experiments. ), activity is below 0.01 nmolÆmi-
n
)1Ælg)1
.
pT-peptide (nmolÆmin)1Ælg)1
)
pS-peptide (nmolÆmin)1Ælg)1
)
T-loop peptide (nmolÆmin)1Ælg)1
)
WT
7.73 1.10
4.61 0.03
8.96 0.06
7.50 0.06
0.68 0.01
3.77 0.01
2.65 0.17
4.13 0.15
0.59 0.05
3.51 0.06
0.87 0.01
4.26 0.09
16.2 0.96
–
3.06 0.21
0.61 0.13
3.11 0.09
5.22 0.29
–
4.22 0.36
1.02 0.10
4.34 0.21
4.50 0.25
–
R13K
S15A
Q17E
Q17K
M36A
H39A
T138A
T138E
R160A
H161A
R169A
0.52 0.01
0.47 0.07
1.90 0.09
0.098 0.08
1.72 0.21
0.21 0.04
0.70 0.04
5.81 0.08
–
0.61 0.02
0.89 0.02
4.20 0.37
1.83 0.03
2.33 0.12
1.61 0.05
2.17 0.13
10.55 0.53
–
Human PP2Ca
Chimera A
enzyme activity towards all tested substrates. This
effect on catalysis could be due to subtle flap subdo-
main interactions with the catalytically active third
metal (M3) as shown in previous structures [20,21].
Finally, two variants, H39A and T138E, were differen-
tially affected in substrate reactivity (Fig. 4A). These
two variants could dephosphorylate artificial substrates
but showed almost no activity towards PII-P. Alanine
replacement of residue His39, which is highly con-
served in bacterial PP2C family members (H39A), did
not at all impair the reactivity towards pNPP, but,
intriguingly, this variant displayed reduced activity
towards artificial phosphopeptides and was almost
completely unable to dephosphorylate PII-P. The
undisturbed activity towards pNPP clearly indicates
that this peripheral part of the enzyme does not
directly take part in catalysis. The impaired reactivity
towards phosphopeptides and phosphoprotein implies
a role of His39 in substrate specificity. Mutation of
Thr138 to Glu (T138E) resulted in a surprisingly high
disturbance of activity. This was unexpected, since in
numerous bacterial PP2C members the corresponding
position is in fact occupied by Asp or Glu residues
(see Supplementary file S1 in [20]). The methyl group
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS
5

Determinants for protein phosphatase substrate specificity
J. Su and K. Forchhammer
Fig. 4. Enzymatic assays of tPphA variants, human PP2Ca and chi-
mera A. Relative activities of tPphA variants, human PP2Ca and chi-
mera A towards different substrates are shown in (A) and (B). The
activities of wild-type tPphA towards five substrates were set as
100% and the activities of other enzymes towards these five sub-
strates were adjusted accordingly. pNPP indicates the value for
Kcat ⁄ Km from the pNPP assay (Table 1). pT-peptide indicates the
relative activity with RRA(pT)VA as substrate, pS-peptide indicates
the activity with RRA(pS)VA peptide as substrate and T-loop pep-
tide indicates the activity towards the CRYRG(pS)EYTV peptide
(Table 2). PII-P indicates the initial activity of PII-P dephosphorylation
(retrieved from the initial slope as shown in the curves of Fig. 3A,
B and Fig. S1C).
Fig. 3. (A) Graphical representation of PII-P dephosphorylation
assays over a period of 45 min analyzed by non-denaturing gels,
shown in Fig. S1A, for tPphA wild-type (WT) and variants S15A,
Q17E, T138A, R13K, R169A and H161A. (B) Graphical representa-
tion of PII-P dephosphorylation assays over a period of 90 min ana-
lyzed by non-denaturing gels, shown in Fig. S1B, for tPphA variants
H161A, R160A, H39A, M36A, T138E, Q17K and D193A.
of Thr138 projects into the basis of the flap subdomain
and introduction of a negative charge at this place
may distort the conformation of the entire flap subdo-
main region. A closer look at the data shows that the
enzyme lost about 80% of activity when assayed with
pNPP; it was even less active towards the artificial pS-
and pT-phosphopeptides and was almost completely
inactive towards phospho-PII. Strikingly, approxi-
mately 50% activity was regained with the T-loop pep-
tide as substrate. Therefore, in addition to general
damage of the catalytic activity, the T138E mutation
seems to affect substrate specificity.
PP2Ca is responsible for the specific exclusion of PII-P
as substrate, the flap was grafted on the catalytic core of
tPphA (chimera A). The resulting hybrid protein
retained indeed some residual activity; however, it could
only dephosphorylate pNPP to some extent (17% resid-
ual activity) but could not dephosphorylate phospho-
peptides and PII-P (Tables 1 and 2 and Fig. 4B),
indicative of an injured enzyme. The catalytic properties
in the pNPP dephosphorylation assay showed that the
hybrid enzyme was strongly affected for Kcat, whereas
the Km values for pNPP and Mn2+ were very similar to
the respective Km values of wild-type tPphA (Table 1).
In agreement with these properties, isothermal titration
calorimetry analysis of metal binding by chimera A
revealed that chimera A bound Mn2+ with a similar
affinity to wild-type tPphA (Fig. S2 and Table S1).
These results imply that chimera A is able to form a
metal cluster which is partly functional, but access of
the substrates to the catalytic site is severely hampered
in the hybrid protein. Only the smallest substrate,
The activities of human PP2Ca and chimera
A (tPphA catalytic core + human PP2Ca flap) were
also assayed towards the above mentioned five different
substrates. Human PP2Ca dephosphorylated artificial
substrates faster than wild-type tPphA but was unable
to dephosphorylate PII-P (Tables 1 and 2 and Fig. S1C).
The inability to dephosphorylate PII-P is not due to an
unfavorable amino acid context around phosphoseryl
residue 49, since the phosphopeptide corresponding to
the phosphorylated T-loop of PII is readily dephospho-
rylated. To prove that the flap subdomain of human
6
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS

J. Su and K. Forchhammer
Determinants for protein phosphatase substrate specificity
pNPP, can be processed in the catalytic site, albeit with
reduced kinetics.
hyde and lysine residues of crosslinked proteins [32].
Finally, conditions could be found under which PII
could be specifically recovered in the tPphA elution
fractions after glutaraldehyde treatment. Controls (PII
with and without crosslinking reaction in the absence
or presence of tPphA) revealed that the recovery of PII
specifically depended on tPphA and proper buffer con-
ditions (Fig. S4A). The fact that almost no crosslink
products between PII and tPphA were found on SDS
gels after heating the samples in SDS sample buffer
can be attributed to the instability of the Schiff base
formed through glutaraldehyde crosslinking [32,33]. A
Ca2+-containing buffer was necessary to efficiently
recover PII-P by this procedure, whereas no pull-down
occurred in the presence of Mg2+ (see the lane labeled
Mg2+ in Fig. 5A) or in the absence of divalent cations
(see the lane labeled ‘)’ in Fig. 5A). Examination of
the phosphorylation status of PII incubated under
these assay conditions revealed that PII-P was dephos-
phorylated in the Mg2+-containing buffer, whereas
An assay to study the interaction between PII-P
and tPphA
To examine how the various mutations of tPphA resi-
dues directly affect the binding of PII-P to the phos-
phatase, we sought to directly analyze the interaction
of the proteins. Initial attempts to pull down PII-P
with His-tagged tPphA on Ni-nitrilotriacetic acid
(Ni-NTA) beads were unsuccessful, indicating that the
interaction is either very weak and ⁄ or transient. Next,
we tried to stabilize the interaction by using conditions
under which the dephosphorylation reaction is
arrested. In the structure of tPphA, obtained by crys-
tallization in a 100 mM CaCl2 containing buffer, Mg2+
was found in the M1 position and Ca2+ in M2 and
M3 sites [16]. Analyzing the effect of Ca2+ on wild-
type tPphA activity showed that Ca2+ did not support
any activity of tPphA and partially inhibited the
Mg2+-supported activity towards phosphopeptides and
PII-P (Fig. S3). Therefore, pull-down assays were per-
formed under inhibitory buffer conditions (Ca2+ or
EDTA, see below) without, however, recovering PII
protein. Another attempt to stabilize the interaction
intermediates was the addition of glutaraldehyde prior
to affinity purification of tPphA and subsequent exam-
ination of PII and tPphA recovery by immunoblot
analysis. The crosslinking reaction by the bis-aldehyde
homobifunctional crosslinker glutaraldehyde proceeds
through the formation of Schiff bases between alde-
almost no dephosphorylation occurred in the Ca2+
-
containing buffer or in the presence of EDTA
(Fig. 5B). Therefore, the low recovery of PII-P in the
presence of Mg2+ was probably due to dephosphoryla-
tion of PII-P and a weaker interaction of non-phos-
phorylated PII with tPphA. On the other hand, the
lack of pull-down in the absence of any divalent metal
may indicate a metal requirement for phosphoprotein
binding (see below). To directly compare the recovery
of phosphorylated and non-phosphorylated PII by this
procedure, the same amount (0.2 lg) of either PII-P or
strep-tagged PII (strep-PII) was used in a pull-down
Fig. 5. (A) PII-P recovered by co-purification with tPphA after glutaraldehyde crosslinking in buffer containing Ca2+, Mg2+ or no divalent cat-
ion ()). 0.5 lg PII-P was used in the binding assay. Afterwards, the samples were analyzed by SDS ⁄ PAGE and immunoblot (anti-PII). CL,
(PII) · 3, (PII) · 2 and (PII) · 1 are crosslinking products, trimeric PII, dimeric PII and monomeric PII, respectively. (B) Dephosphorylation of
PII-P by tPphA under different conditions. 0.5 lg PII-P was incubated with 10 lg tPphA on ice in different buffers for 2 h, corresponding to
the incubation time during the PII-P–tPphA binding assay. The reaction volume was 50 lL containing 110 mM NaCl together with either
2.5 mM Ca2+, 2.5 mM Mg2+and 5 mM EDTA, or without additions, as indicated above. After 2 h incubation, the phosphorylation status of PII
0
was determined by non-denaturing PAGE and immunoblot analysis as described previously [38]. The non-phosphorylated form PII and the
2
phosphorylated forms containing one, two or three phosphorylated subunits (PII1, PII and PII3) are indicated. (C) Comparison of PII-P (0.2 lg)
and strep-PII (0.2 lg) binding to tPphA (10 lg) in Ca2+-containing buffer. PII-P and strep-PII was recovered by co-purification with tPphA after
glutaraldehyde crosslinking in buffer containing Ca2+. Afterwards, the samples were analyzed by SDS ⁄ PAGE and immunoblot (anti-PII).
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS
7

Determinants for protein phosphatase substrate specificity
J. Su and K. Forchhammer
experiment in the presence of Ca2+ (Fig. 5C), and
indeed much more PII-P was recovered than non-phos-
phorylated strep-PII.
Discussion
As deduced from this study, the selection of the phos-
phoprotein substrate, here the PII-P protein, requires a
concerted interaction of the phosphatase with the sub-
strate. The M1–M2 catalytic core is the central region
for binding of PII-P. Structures of the PP2C member
MspP (from Mycobacterium smegmatis) in various
complexes with phosphate, sulfate or the phosphate
analog cacodylate showed that two oxygen atoms of
the substrate phosphate group directly interacted with
M1 and M2 by bidentate coordination [18]. The
MspP–cacodylate complex was suggested to mimic the
competent enzyme–substrate complex [18]. Modeling
of the cacodylate molecule from the MspP structure
into the tPphA structure in such a way that the
distance between cacodylate and M1 and M2 remains
constant showed that the cacodylate oxygen atoms
would replace the oxygen atoms from water molecules
coordinated by M1 and M2 in tPphA [21], which indi-
cates that M1–M2 of tPphA could indeed directly bind
the substrate phosphate. In agreement, the variants,
which cannot correctly assemble the M1–M2 core,
were impaired in recovering PII-P in pull-down experi-
ments. D18A and D34A variants were shown to be
highly impaired in metal binding [21] and the crystal
structure of the D193A showed that the position of
Interaction of tPphA variants, human PP2Ca and
chimera A with PII-P
The interaction of tPphA variants with PII-P was
investigated using the glutaraldehyde crosslinking and
pull-down assay described above. Figure 6A shows the
immunoblot analysis of PII recovered from the pull-
down assays with tPphA variants. The control for
tPphA recovery is shown in Fig. S5. In many cases,
there was
a clear correlation between reactivity
towards and binding of PII-P: variants Q17K and
T138E which displayed very low activity towards PII-P
could not pull-down PII-P, whereas the same position
variants Q17E and T138A, which were only slightly
impaired in reactivity towards PII-P, showed appar-
ently equal or more binding of PII-P than wild-type
tPphA. Furthermore, tPphA variants D18A, D34A
and D193A, coordinating the M1–M2 cluster and
being completely inactive [21], could not bind any
PII-P (Fig. 6A). By contrast, the variants of the loosely
M1-coordinating Asp231 residue (D231A) and of the
M3-coordinating Asp119 residue (D119A), which are
also unable to dephosphorylate PII-P, could still bind
this substrate. The flap subdomain variants (R160A,
H161A as well as R169A), which are partially
impaired in catalytic activity, were also partially
impaired in PII-P binding. Surprisingly, the variant
H39A was able to recover PII-P although the enzyme
was severely impaired in dephosphorylating PII-P. The
human PP2Ca and chimera A, both having an active
metal center (see above), could not recover PII-P
(Fig. 6B), in agreement with the lack of PII-P dephos-
phorylation.
˚
M1 was shifted by 1.8 A in this variant [21]. By con-
trast, the third metal (M3) appears to be less impor-
tant for substrate binding, since the D119A variant
recovered PII-P in the binding assay. The recently
solved crystal structure of the D119A variant [21]
showed that the M1–M2 cluster and the entire cata-
lytic core were not affected compared with the wild-
type structure. However, M3 was specifically missing
and the segment of the flap subdomain near M3 was
˚
moved by 1.8 A towards the center. All these data
Fig. 6. (A) Assay of PII-P binding to wild-type tPphA (WT) and the variants according to the crosslinking and pull-down protocol (see Fig. 5A).
The recovery of PII (in monomeric, dimeric, trimeric and tPphA crosslinked forms, see indications in Fig. 5A) is shown by anti-PII immunoblot-
ting. (B) Assay of PII-P binding to tPphA, chimera A and human PP2Ca. Recovery of PII was visualized by anti-PII immunoblotting. These
assays were performed three times with similar results. All assays were performed as described in Experimental procedures.
8
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS

J. Su and K. Forchhammer
Determinants for protein phosphatase substrate specificity
together with previous investigations [18,21] suggest
that the positively charged M1–M2 core of tPphA
is critical for trapping the substrate phosphate
group, whereas M3 is involved in the subsequent
dephosphorylation reaction.
conclusion, the variant could not recover PII-P in the
pull-down assay (Fig. 6A).
A further region of the enzyme that specifies sub-
strate recognition could be the flap subdomain. Two
major lines of experimental evidence, both with some
limitations, are in favor of this conclusion. The first
line comes from the comparison of human PP2Ca and
tPphA: both enzymes are highly active towards artifi-
cial substrates and human PP2Ca is even very active
towards the peptide corresponding to the phosphory-
lated T-loop of PII; however, it is completely inactive
towards the entire phosphoprotein PII-P. The major
structural difference concerns the flap subdomain
(Fig. 1D), which we therefore suspected plays an
important role in hindering the dephosphorylation of
PII-P. Nevertheless, other differences exist, which could
be important as well. For example, His39, highly con-
served in bacterial PP2C members, is not conserved in
human PP2Ca. The attempt to graft the flap of human
PP2Ca on tPphA and vice versa had only limited suc-
cess: only the combination of tPphA core and human
PP2Ca flap resulted in a soluble protein, but this
enzyme was only partially active. Nevertheless, some
conclusions can be drawn from studying this chimeric
protein. The fact that the enzyme bound metals with
similar affinity to wild-type tPphA and could dephos-
phorylate pNPP, albeit with reduced velocity, indicates
that chimera A can form at least a partially active
catalytic center. The large flap subdomain may col-
lapse on top of the catalytic site and thereby limit its
accessibility, which would explain why only the small-
est substrate, pNPP, can be dephosphorylated. In
future studies, the catalytic core and flap subdomain
of enzymes which are more closely related will be com-
bined, expecting that these chimeras will be catalyti-
cally less impaired. The second experimental approach
to narrow down the function of the flap subdomain
was by introducing point mutations. Almost every
mutation in the flap region (T138E, R160A, H161A
and R169A) reduced the activity of the respective vari-
ant, indicating a general role of the flap subdomain in
catalytic efficiency. Subtle interactions with the nearby
M3 metal could play a role, as suspected earlier
[20,21]. Whereas most flap subdomain variants did not
lead to conclusive results, the T138E mutation is par-
ticularly intriguing. In many other PP2C homologs,
this position is naturally occupied by an Asp or Glu
residue (see Supplementary file S1 in [20]). In tPphA,
the Thr to Glu replacement heavily affected enzyme
activity and strongly reduced binding to PII-P. Activity
could be partially restored (to 50% of wild-type) by
using the T-loop peptide as substrate, whereas the
PII-P substrate is neither turned over nor bound. The
According to the enzymatic properties of the H39A
variant, the His39 residue appears to play an impor-
tant role for the specificity towards substrates, since
reactivity towards pNPP was unaffected but the activ-
ity towards phosphopeptides was substantially
impaired and the reactivity towards PII-P almost van-
ished (Fig. 4A). The catalytic center of the H39A vari-
ant is apparently not impaired, as expected from the
distant location of the His39 residue relative to the
catalytic center. His39 belongs to a small loop (at the
beginning of motif 3, Figs 1A and 2A) which is located
on the distal end of the catalytic face (on the opposite
site of the flap subdomain) flanking a cleft, which
protrudes towards M2 (Fig. 1A). The other site of this
cleft is flanked by Ser15, which can be replaced by Ala
without loss of substrate specificity. His39 is invisible
in one space group (C2221) of the crystal structures of
tPphA [16] and the variant D193A [21], indicating that
this part of the protein is flexible. The present work
provides the first evidence by mutational analysis that
the conserved His39 residue is indeed involved in
substrate recognition, as already suspected from the
crystallographic analysis of the S. agalactiae PP2C
phosphatase SaSTP (see Introduction) [19]. According
to the pull-down data, it appears that His39 is not
essential for the coarse binding of PII-P. Instead, His39
might be involved in fine-tuning the formation of a
productive substrate–enzyme complex that allows the
subsequent dephosphorylation reaction.
In the cleft flanked by His39 and Ser15, near M2,
residue Gln17 plays another crucial role in tPphA
function. Gln17 is the homolog residue of HAB1
Glu203. In the co-crystal structure of PYR1 (pyrabac-
tin resistance1) with HAB1 from A. thaliana, HAB1
Glu203 directly formed a hydrogen bond with PYR1
Ser85. This interaction was suggested to mimic the
interaction of a serine-containing peptide with the cat-
alytic center of a PP2C phosphatase [27]. The study
showed that Gln17 can be replaced by Glu without
loss of function. The Q17E tPphA variant recovered
even higher amounts of PII-P than the wild-type
enzyme. Similar to the interaction of HAB1 Glu203
with PYR1 Ser85, tPphA Glu17 may take part in
substrate binding, explaining the increased recovery
of PII-P by Q17E. By contrast, the Q17K variant was
almost completely inactive. The positive charge of
Lys17 could impair the binding of M2 and thus
compromise enzyme function. In agreement with that
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS
9

Determinants for protein phosphatase substrate specificity
J. Su and K. Forchhammer
side (IPTG) was added to a final concentration of 0.5 mM
to induce protein overproduction. After 4–5 h induction at
25 ꢀC, the cells were harvested by centrifugation and lysed
by sonification (Branson Sonifier 250, Danbury, CT, USA)
on ice in a buffer consisting of 20 mM Tris ⁄ HCl, pH 7.8,
75 mM KCl, 500 mM NaCl, 5 mM MgCl2, 0.5 mM EDTA,
0.2 mM phenylmethylsulfonyl fluoride. The lysate was cen-
trifuged for 30 min at 16 000 g. The supernatant was centri-
fuged again for 30 min at 16 000 g. The clarified cell
extract was used for protein purification. HIS-Select Car-
tridges (Sigma, St Louis, MO, USA) were used to purify
His-tag proteins. After purification, the proteins were dia-
lyzed in a buffer (20 mM Hepes, pH 7.4, 75 mM KCl,
500 mM NaCl, 2 mM MgCl2, 0.5 mM EDTA, 1 mM dith-
iothreitol, 50% glycerol) and analyzed by SDS ⁄ PAGE. The
concentrations of the purified proteins were determined by
the Bradford method (Rotiquant; Carl Roth, Karlsruhe,
Germany) and from UV spectra by calculation with an
absorption coefficient at 280 nm of 27 960 M)1Æcm)1 (calcu-
lated from http://www.expasy.ch/tools/protparam.html).
fact that a single amino acid substitution in the flap
subdomain region is sufficient to completely alter the
substrate specificity agrees with the assumption that
the flap subdomain is involved in substrate selection.
The phosphatase must be able to precisely recognize
its substrate protein, and therefore fine-tuned interac-
tions between phosphatase and substrate are necessary.
Here, we have shown that the His39 residue, M1–M2
core and probably the flap subdomain are involved in
this recognition process. Only when a productive sub-
strate–enzyme complex is formed is the enzyme able to
dephosphorylate its substrate. This mechanism pre-
vents dephosphorylation of non-cognate phosphopro-
teins. For many bacterial PP2Cs, this recognition
process seems to be sufficient, since they lack further
substrate binding domains and solely depend on the
specificity provided by the interaction between sub-
strate and enzyme on the face of the phosphatase. The
flap subdomain is the region with the highest sequence
variability in bacterial PP2C members. Further experi-
ments are required to clearly establish whether the flap
subdomain is indeed the clue to distinguishing between
the different protein substrate specificities of the vari-
ous PP2C members.
Cloning, overexpression and purification of
human PP2Ca and the chimeras
His-human PP2Ca was purified as described previously
with no modification [21]. Artificial genes for chimera A
[tPphA catalytic center (1–137, 174–240) + human PP2Ca
flap (165–219)] and chimera B [human PP2Ca catalytic
center (1–164, 220–297) + tPphA flap (138–173)] were syn-
thesized and cloned into His-tag pET15b vector by Geneart
(Regensburg, Germany). At first, overexpression and purifi-
cation of His-tagged chimeras were performed as in the
previous section. His-tagged chimera A was purified and
stored in a buffer (20 mM Hepes, pH 7.4, 75 mM KCl,
500 mM NaCl, 2 mM MgCl2, 0.5 mM EDTA, 1 mM dith-
iothreitol, 50% glycerol). Unfortunately, most of chimera B
became insoluble inclusion bodies in E. coli BL21 (DE3)
after being induced by IPTG. Only residual soluble chimera
B could be purified by HIS-Select Cartridges (Sigma). After
protein purification, the soluble form chimera B could be
dialyzed against a buffer (20 mM Hepes, pH 7.4, 75 mM
KCl, 500 mM NaCl, 2 mM MgCl2, 0.5 mM EDTA, 1 mM
dithiothreitol, 50% glycerol), and it remained soluble. After
dialysis, the concentration (0.2 lgÆlL)1) of chimera B was
determined by the Bradford method. But it precipitated at
once when it was added into any other buffer.
Experimental procedures
Cloning, overexpression and purification of tPphA
variants
The tPphA gene (tlr2243) encoding tPphA was amplified
from plasmid pET32a + tPphA [16] by using the T7 prim-
ers (Table S2), which amplify genes between the T7 pro-
moter and T7 terminator in the pET (Novagen, Darmstadt,
Germany) series plasmids. The resulting PCR product,
which contains an NdeI site at the tPphA start codon and a
3¢ terminal XhoI restriction site was first cloned using the
pGEM-T easy vector system (Promega, Madison, WI,
USA). Following restriction of the resulting plasmid with
NdeI and XhoI, the tPphA fragment was isolated and
cloned between the NdeI and XhoI sites of His-tag expres-
sion vector pET15b-vector (Novagen), generating an N-ter-
minal fusion to the 6· His-tag sequence from the vector.
The procedure of site-directed mutagenesis of tPphA was
carried out with the QuickChange XL site-directed muta-
genesis kit (Stratagene, La Jolla, CA, USA). The primers
for site-directed mutagenesis are shown in Table S2. All
constructs were checked by DNA sequencing. For overpro-
duction of the recombinant enzymes, the constructs were
transformed into Escherichia coli BL21 (DE3) cells (New
England Biolabs GmbH, Frankfurt am Main, Germany)
and grown in LB medium supplemented with ampicillin
(100 lgÆmL)1). When the optical density (D600) of the
culture reached a value of 1–1.5, isopropyl thio-b-D-galacto-
In order to obtain soluble chimera B, modified methods
for overexpression and purification were applied. In method
1, overexpression of His-tagged chimera B was performed
as in the previous section. When the optical density (A600
)
of the E. coli BL21 (DE3) culture reached a value of 1,
IPTG was added to a final concentration of 0.5 mM to
induce protein overproduction. After 4 h induction at
25 ꢀC, the cells were harvested by centrifugation and lysed
by sonification (Branson Sonifier 250) on ice in a buffer
10
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS

J. Su and K. Forchhammer
Determinants for protein phosphatase substrate specificity
consisting of 20 mM Tris ⁄ HCl, pH 7.8, 75 mM KCl,
500 mM NaCl, 5 mM MgCl2, 0.5 mM EDTA, 0.2 mM phen-
ylmethylsulfonyl fluoride. The lysate was centrifuged for
30 min at 16 000 g. The pellet containing most of chimera
B in the form of inclusion bodies was redissolved and dena-
10 mM NaHCO3 and ferric ammonium citrate was replaced
by ferric citrate. When A750 of the cells reached 1.0, the
cells were treated with 0.5 mM L-methionine sulfoximine for
4 h and were then harvested by centrifugation. All subse-
quent steps were carried out at 0–8 ꢀC. The cells were bro-
ken by sonification (Branson Sonifier 250) in a buffer
consisting of 50 mM Tris ⁄ HCl, pH 7.4, 4 mM EDTA,
0.1 mM phenylmethylsulfonyl fluoride. Cell debris was
removed by centrifugation for 10 min at 10 000 g and the
crude extract was cleared by high speed centrifugation at
100 000 g for 60 min. This crude extract enriched with
phosphorylated PII was used for the experiments described
in the section ‘The activities of tPphA variants, human
PP2Ca and chimera A towards PII-P’ below. The crude
extract was applied to an Econo-Pac heparin cartridge
(Bio-Rad, Hercules, CA, USA) equilibrated in 10 mM He-
pes, pH 7.4. Proteins were eluted with a 100 mL linear gra-
dient of 0–400 mM NaCl in 10 mM Hepes, pH 7.4, at a
flow rate of 1 mLÆmin)1. Fractions (4 mL) were collected in
tubes containing 50 lL of 0.4 M EDTA to prevent dephos-
phorylation of PII-P. The PII-P containing fractions were
detected by immunoblot (dot-blot) analysis using PII-spe-
cific antibodies. Ammonium sulfate was added to a satura-
tion of 80% to precipitate proteins. The precipitated
proteins were collected and resuspended in 10 mM
Tris ⁄ HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA. The solu-
tion was applied to a Superdex 200 10 ⁄ 300 GL column
(Amersham Pharmacia, Freiburg, Germany) equilibrated in
a buffer (10 mM Tris ⁄ HCl, pH 7.4, 150 mM NaCl, 5 mM
EDTA) and developed in the same buffer at a flow rate of
0.5 mLÆmin)1. The PII-P containing fractions were concen-
trated by ultrafiltration (Microcon YM-10, 10 kDa cut-off;
Millipore, Bedford, MA, USA). This purified phospho-PII
was used for the experiments described in the sections ‘PII–
tPphA binding assay’ and ‘Human PP2Ca and chimera
bind to PII-P’ below. Purification of recombinant PII
proteins with a C-terminal-fused strep-tag II peptide was
performed according to Heinrich et al. [35].
tured by
a
buffer containing 0.1 M NaH2PO4 ⁄
Na2HPO4, pH 7.4, 6 M guanidine. The solution was centri-
fuged for 30 min at 16 000 g. After centrifuging, the con-
tent of the supernatant was checked by SDS ⁄ PAGE.
SDS ⁄ PAGE showed a strong band at 29 kDa and proved
that chimera B could be denatured and solubilized by 6 M
guanidine. In order to remove guanidine and refold chimera
B, the supernatant was dialyzed against a series of buffers
containing different concentrations (6, 4, 2, 1 and 0 M) of
guanidine, 40 mM Hepes, pH 7.4, 0.5 M NaCl, 2 mM
MgCl2. After the supernatant containing soluble chimera B
was exchanged from the dialysis buffer containing 2 M gua-
nidine to the dialysis buffer containing 1 M guanidine, all of
chimera B precipitated. Therefore, no chimera B could be
refolded.
In method 2, E. coli BL21 (DE3) cells containing plasmid
pET15b + chimera B were grown in LB medium supple-
mented with ampicillin (100 lgÆmL)1). When the optical
density (A600) of the culture reached 0.4 or 0.6 IPTG was
added to a final concentration of 0.5 mM to induce protein
overproduction. After 3 h induction at 18 ꢀC, the cells were
harvested by centrifugation and lysed by a French pressure
cell homogenizer (Emulsiflex-B15; Avestin, Ottawa, Can-
ada) in a buffer consisting of 20 mM Tris ⁄ HCl, pH 7.8,
75 mM KCl, 500 mM NaCl, 5 mM MgCl2, 0.5 mM EDTA,
0.2 mM phenylmethylsulfonyl fluoride. The lysate was cen-
trifuged for 30 min at 16 000 g. The supernatant was centri-
fuged again for 30 min at 16 000 g. The clarified cell
extract was used for protein purification. HIS-Select Car-
tridges (Sigma) were used to purify His-tag chimera B.
After purification, the proteins were dialyzed in a buffer
(20 mM Hepes, pH 7.4, 75 mM KCl, 500 mM NaCl, 2 mM
MgCl2, 0.5 mM EDTA, 1 mM dithiothreitol, 50% glycerol)
and analyzed by SDS ⁄ PAGE. The concentration
(0.3 lgÆlL)1) of the purified protein was determined by the
Bradford method (Rotiquant; Carl Roth). Chimera B still
precipitated when it was exchanged from the first buffer
(20 mM Hepes, pH 7.4, 75 mM KCl, 500 mM NaCl, 2 mM
MgCl2, 0.5 mM EDTA, 1 mM dithiothreitol, 50% glycerol)
to any other buffer. Therefore, no further experiments
could be performed with chimera B and the characteristics
of chimera B were unknown.
Assay of phosphatase activity with pNPP as
artificial substrate
The activities of tPphA variants, human PP2Ca and chi-
mera A towards pNPP were assayed as described previously
with modifications [16,21]. Standard assays in a volume of
250 lL contained 0.25–10 lg tPphA variants, 0.25 lg
human PP2Ca and 0.25–10 lg chimera A in a buffer con-
sisting of 10 mM Tris ⁄ HCl, pH 8.0, 50 mM NaCl, 1 mM
dithiothreitol and 2 mM MnCl2. Reactions were started by
the addition of 2 mM pNPP at 30 ꢀC and the increase in
absorbance at 400 nm was measured in an ELx808 absor-
bance microplate reader (BioTek, Winooski, VT, USA)
against a blank reaction without enzyme. To determine the
Km for Mn2+, the pNPP concentration was fixed at 1.5 mM
and the Mn2+ concentration was varied from 0.1 to 8 mM.
Preparation of phosphorylated PII and strep-PII
Phosphorylated PII protein (PII-P) was prepared from Syn-
echococcus elongatus PCC 7942 as described previously with
modifications [34]. To achieve a maximum degree of in vivo
PII phosphorylation, the cells were grown in 4.5 L BG-11
medium containing 17.6 mM NaNO3 supplemented with
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS
11

Determinants for protein phosphatase substrate specificity
J. Su and K. Forchhammer
For pNPP catalytic constants, the pNPP concentration was
varied from 0.1 to 2 mM. From the linear slope of each reac-
tion, the kinetic parameters Km and Vmax were calculated by
nonlinear hyperbolic fitting using the GRAPHPAD PRISM 4
program (GraphPad Software Inc., La Jolla, CA, USA).
extract His-tagged tPphA, 30 lL Ni-NTA magnetic agarose
beads (Qiagen, Hilden, Germany) were added. The mixture
was gently mixed by rotation for 1 h. Thereafter, the
Ni-NTA beads were removed with a magnet and washed
three times with 0.5 mL washing buffer (10 mM Tris ⁄ HCl,
pH 8.0, 2.5 mM CaCl2, 100 mM NaCl, 20 mM imidazole).
Finally, the bound proteins were eluted by denaturing the
sample with 50 lL SDS sample buffer (75 mM Tris ⁄ HCl,
pH 6.8, 1.5% dithiothreitol, 1% SDS, 10% glycerol, 0.04%
bromophenol blue) at 95 ꢀC for 5 min. Subsequently, the
eluted proteins were visualized by SDS ⁄ PAGE and immu-
noblot analysis as described previously [39]. In order to
reveal the effect of divalent cations on the binding between
PII and tPphA, 2.5 mM CaCl2 was either removed or chan-
ged to 2.5 mM MgCl2 in buffer I, buffer II and the washing
buffer.
Reactivity of tPphA variants, human PP2Ca and
chimera A towards phosphopeptides
Three different phosphopeptides were used to assay the
activities of tPphA variants and chimera A. The sequences
of the three peptides are CRYRG(pS)EYTV, RRA(pT)VA
and RRA(pS)VA. CRYRG(pS)EYTV corresponds to the
T-loop sequence of the PII protein. In a standard assay,
0.25–10 lg tPphA variants, 0.25 lg human PP2Ca and
0.25–10 lg chimera A were reacted with 100 lM phospho-
peptides in a reaction volume of 100 lL containing 50 mM
Tris ⁄ HCl, pH 8.0, 50 mM NaCl, 2 mM MnCl2 and 1 mM
dithiothreitol. Reactions were incubated at 30 ꢀC for
2–10 min, and then stopped by the addition of 50 lL 4.7 M
HCl. The released Pi was quantified colorimetrically by the
malachite green assay [36,37]. The absorbance of the solu-
tion at 630 nm was measured in an ELx808 absorbance
microplate reader (BioTek) against a blank reaction, which
was stopped at the start point by 50 lL 4.7 M HCl. The
activity of all enzymes toward peptides was calculated with
a phosphate standard.
Human PP2Ca and chimera A bind to PII-P
The procedures were same as in the previous section.
Acknowledgements
We thank Christina Herrmann and Oleksandra Fokina
for practical help. This work was supported by DFG
grant Fo195 and by GRK 685 ‘infection biology’.
References
The activities of tPphA variants, human PP2Ca
and chimera A towards PII-P
1 Tonks NK (2006) Protein tyrosine phosphatases: from
genes, to function, to disease. Nat Rev Mol Cell Biol 7,
833–846.
Dephosphorylation assays of PII-P were carried out in
10 mM Tris ⁄ HCl, pH 8.0, 50 mM NaCl, 1 mM dithiothreitol
and 2 mM MgCl2. 30 lL reactions contained 5 lL Synecho-
coccus extract containing 60 ng highly phosphorylated PII
and 100 ng enzymes. Reactions were started by the addition
of purified phosphatase. After incubation at 30 ꢀC the reac-
tions were stopped by adding 2.5 lL 100 mM EDTA on
ice. Subsequently, the phosphorylation state of PII was
determined by non-denaturing PAGE and immunoblot
analysis as described previously [38].
2 Shi Y (2009) Serine ⁄ threonine phosphatases: mechanism
through structure. Cell 139, 468–484.
3 Rigden DJ (2008) The histidine phosphatase
superfamily: structure and function. Biochem J 409,
333–348.
4 Das AK, Helps NR, Cohen PTW & Barford D (1996)
Crystal structure of the protein serine ⁄ threonine phos-
phatase 2C at 2.0 angstrom resolution. EMBO J 15,
6798–6809.
5 Bork P, Brown NP, Hegyi H & Schultz J (1996) The
protein phosphatase 2C (PP2C) superfamily: detection
of bacterial homologues. Protein Sci 5, 1421–1425.
6 Kennelly PJ (2002) Protein kinases and protein phos-
phatases in prokaryotes: a genomic perspective. FEMS
Microbiol Lett 206, 1–8.
PII– tPphA binding assay
Each step of the following binding assay was carried out
on ice or at 4 ꢀC. tPphA (10 lg wild-type or tPphA variant
proteins) was incubated with PII-P (0.2 or 0.5 lg) or strep-
PII (0.2 or 5 lg) in 50 lL of buffer I containing 1 mM
Tris ⁄ HCl, pH 7.4, 2.5 mM CaCl2, 110 mM NaCl, 0.05 mM
EDTA and 0.08% (w ⁄ w) glutaraldehyde (Carl Roth). After
60 min, 2.5 lL of 1 M imidazole (pH 8.0) was added to the
reaction. After further incubating on ice for 15 min, the
reaction was diluted with 420 lL buffer II containing
2.5 mM CaCl2, 100 mM NaCl and 0.05% v ⁄ v Tween-20. To
7 Shi L, Potts M & Kennelly PJ (1998) The serine, threo-
nine, and ⁄ or tyrosine-specific protein kinases and
protein phosphatases of prokaryotic organisms: a family
portrait. FEMS Microbiol Rev 22, 229–253.
8 Arino J, Casamayor A & Gonzalez A (2011)
Type 2C protein phosphatases in fungi. Eukaryot Cell
10, 21–33.
12
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS

J. Su and K. Forchhammer
Determinants for protein phosphatase substrate specificity
9 Lammers T & Lavi S (2007) Role of type 2C protein
phosphatases in growth regulation and in cellular stress
signaling. Crit Rev Biochem Mol 42, 437–461.
10 Schweighofer A, Hirt H & Meskiene L (2004) Plant
PP2C phosphatases: emerging functions in stress signal-
ing. Trends Plant Sci 9, 236–243.
22 Forchhammer K & Demarsac NT (1994) The P-II pro-
tein in the cyanobacterium Synechococcus sp. strain
Pcc-7942 is modified by serine phosphorylation and sig-
nals the cellular N-status. J Bacteriol 176, 84–91.
23 Forchhammer K (2008) P-II signal transducers: novel
functional and structural insights. Trends Microbiol 16,
65–72.
11 Melcher K, Ng LM, Zhou XE, Soon FF, Xu Y, Suino-
Powell KM, Park SY, Weiner JJ, Fujii H, Chinnusamy
V et al. (2009) A gate-latch-lock mechanism for
hormone signalling by abscisic acid receptors. Nature
462, 602–610.
24 Kaneko T, Sato S, Kotani H, Tanaka A, Asamizu E,
Nakamura Y, Miyajima N, Hirosawa M, Sugiura M,
Sasamoto S et al. (1996) Sequence analysis of the
genome of the unicellular cyanobacterium Synechocystis
sp. strain PCC6803. II. Sequence determination of the
entire genome and assignment of potential protein-cod-
ing regions. DNA Res 3, 109–136.
12 Obuchowski M, Madec E, Delattre D, Boel G, Iwanicki
A, Foulger D & Seror SJ (2000) Characterization of
PrpC from Bacillus subtilis, a member of the PPM
phosphatase family. J Bacteriol 182, 5634–5638.
13 Shakir SM, Bryant KM, Larabee JL, Hamm EE,
Lovchik J, Lyons CR & Ballard JD (2010) Regulatory
interactions of a virulence-associated serine ⁄ threonine
phosphatase-kinase pair in Bacillus anthracis. J Bacteriol
192, 400–409.
25 Zhang CC, Gonzalez L & Phalip V (1998) Survey, anal-
ysis and genetic organization of genes encoding eukary-
otic-like signaling proteins on a cyanobacterial genome.
Nucleic Acids Res 26, 3619–3625.
26 Kloft N & Forchhammer K (2005) Signal transduction
protein P-II phosphatase PphA is required for light-
dependent control of nitrate utilization in Synechocystis
sp. strain PCC 6803. J Bacteriol 187, 6683–6690.
27 Dupeux F, Antoni R, Betz K, Santiago J, Gonzalez-
Guzman M, Rodriguez L, Rubio S, Park SY, Cutler
SR, Rodriguez PL et al. (2011) Modulation of abscisic
acid signaling in vivo by an engineered receptor-insensi-
tive protein phosphatase type 2C allele. Plant Physiol
156, 106–116.
14 Treuner-Lange A, Ward MJ & Zusman DR (2001)
Pph1 from Myxococcus xanthus is a protein phospha-
tase involved in vegetative growth and development.
Mol Microbiol 40, 126–140.
15 Irmler A & Forchhammer K (2001) A PP2C-type phos-
phatase dephosphorylates the PII signaling protein in
the cyanobacterium Synechocystis PCC 6803. Proc Natl
Acad Sci USA 98, 12978–12983.
16 Schlicker C, Fokina O, Kioft N, Grune T, Becker S,
Sheldrick GM & Forchhammer K (2008) Structural
analysis of the PP2C phosphatase tPphA from Thermo-
synechococcus elongatus: a flexible flap subdomain
controls access to the catalytic site. J Mol Biol 376,
570–581.
28 Miyazono K, Miyakawa T, Sawano Y, Kubota K,
Kang HJ, Asano A, Miyauchi Y, Takahashi M, Zhi
YH, Fujita Y et al. (2009) Structural basis of abscisic
acid signalling. Nature 462, 609–614.
29 Cutler SR, Rodriguez PL, Finkelstein RR & Abrams
SR (2010) Abscisic acid: emergence of a core signaling
network. Annu Rev Plant Biol 61, 651–679.
17 Pullen KE, Ng HL, Sung PY, Good MC, Smith SM &
Alber T (2004) An alternate conformation and a third
metal in PstP ⁄ Ppp, the M-tuberculosis PP2C-family
Ser ⁄ Thr protein phosphatase. Structure 12, 1947–1954.
18 Bellinzoni M, Welhenkel A, Shepard W & Alzari PM
(2007) Insights into the catalytic mechanism of PPM
Ser ⁄ Thr phosphatases from the atomic resolution struc-
tures of a mycobacterial enzyme. Structure 15, 863–872.
19 Rantanen MK, Lehtio L, Rajagopal L, Rubens CE &
Goldman A (2007) Structure of Streptococcus agalactiae
serine ⁄ threonine phosphatase – the subdomain confor-
mation is coupled to the binding of a third metal ion.
FEBS J 274, 3128–3137.
30 Xu YB, Carr PD, Clancy P, Garcia-Dominguez M,
Forchhammer K, Florencio F, de Marsac NT, Vasud-
evan SG & Ollis DL (2003) The structures of the PII
proteins from the cyanobacteria Synechococcus sp PCC
7942 and Synechocystis sp PCC 6803. Acta Crystallogr
D 59, 2183–2190.
31 Fokina O, Chellamuthu VR, Forchhammer K & Zeth
K (2010) Mechanism of 2-oxoglutarate signaling by the
Synechococcus elongatus P(II) signal transduction pro-
tein. Proc Natl Acad Sci USA 107, 19760–19765.
32 Migneault I, Dartiguenave C, Bertrand MJ & Waldron
KC (2004) Glutaraldehyde: behavior in aqueous solu-
tion, reaction with proteins, and application to enzyme
crosslinking. BioTechniques 37, 790–802.
20 Wehenkel A, Bellinzoni M, Schaeffer F, Villarino A &
Alzari PM (2007) Structural and binding studies of the
three-metal center in two mycobacterial PPM Ser ⁄ Thr
protein phosphatases. J Mol Biol 374, 890–898.
21 Su JY, Schlicker C & Forchhammer K (2011) A third
metal is required for catalytic activity of the signal-
transducing protein phosphatase M tPphA. J Biol Chem
286, 13481–13488.
33 Hermanson G (2008) Glutaraldehyde. In Bioconjugate
Techniques, 2nd edn (Hermanson GT ed), pp 265–268.
Academic Press, New York.
34 Irmler A, Sanner S, Dierks H & Forchhammer K
(1997) Dephosphorylation of the phosphoprotein P(II)
in Synechococcus PCC 7942: identification of an ATP
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS
13

Determinants for protein phosphatase substrate specificity
J. Su and K. Forchhammer
and 2-oxoglutarate-regulated phosphatase activity. Mol
Microbiol 26, 81–90.
of PII-P dephosphorylation by tPphA, human PP2Ca
and chimera A.
Fig. S2. Mn2+ binding to tPphA, chimera A and
human PP2Ca studied by isothermal titration calorim-
etry.
Fig. S3. (A) Inhibitory effect of Ca2+ on the relative
activities of tPphA towards pNPP, pT-peptide, T-loop
peptide and PII-P. (B) Inhibitory effect of Ca2+ on the
dephosphorylation of PII-P by wild-type tPphA.
Fig. S4. Glutaraldehyde (GDA) crosslinking followed
by Ni-NTA affinity purification (pull-down, PD) using
His-tPphA (10 lg) with PII-P (0.5 lg).
35 Heinrich A, Maheswaran M, Ruppert U & Forchham-
mer K (2004) The Synechococcus elongatus P signal
transduction protein controls arginine synthesis by com-
plex formation with N-acetyl-L-glutamate kinase. Mol
Microbiol 52, 1303–1314.
36 Vanveldhoven PP & Mannaerts GP (1987) Inorganic
and organic phosphate measurements in the nanomolar
range. Anal Biochem 161, 45–48.
37 Ekman P & Jager O (1993) Quantification of subn-
anomolar amounts of phosphate bound to seryl and
threonyl residues in phosphoproteins using alkaline-
hydrolysis and malachite green. Anal Biochem 214,
138–141.
Fig. S5. Recovery of tPphA from the pull-down assays
shown in Fig. 6(A).
Table S1. The affinity and thermodynamic parameters
of tPphA, chimera A and human PP2Ca from ITC
assay.
38 Ruppert U, Irmler A, Kloft N & Forchhammer K
(2002) The novel protein phosphatase PphA from
Synechocystis PCC 6803 controls dephosphorylation of
the signalling protein PII. Mol Microbiol 44, 855–864.
39 Forchhammer K & Hedler A (1997) Phosphoprotein
PII from cyanobacteria – analysis of functional conser-
vation with the PII signal-transduction protein from
Escherichia coli. Eur J Biochem 244, 869–875.
Table S2. Primers used for PCR amplification of
tPphA and for site-directed mutagenesis.
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be reorganized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
Supporting information
The following supplementary material is available:
Fig. S1. (A), (B) Time course assay of PII-P dephos-
phorylation by tPphA variants. (C) Time course assay
14
FEBS Journal (2012) ª 2011 The Authors Journal compilation ª 2011 FEBS