The
Saccharomyces cerevisiae
type 2A protein phosphatase Pph22p
is biochemically different from mammalian PP2A
Piotr Zabrocki
1
, Wojciech Swiatek
1
, Ewa Sugajska
1
, Johan M. Thevelein
2
, Stefaan Wera
2
and Stanislaw Zolnierowicz
1,
*
1
Cell and Molecular Signaling Laboratory, Intercollegiate Faculty of Biotechnology UG-MUG, Gdansk, Poland;
2
Laboratorium
voor Moleculaire Celbiologie, K.U. Leuven, Leuven-Heverlee, Flanders, Belgium
The Saccharomyces cerevisiae type 2A protein phosphatase
(PP2A) Pph22p differs from the catalytic subunits of PP2A
(PP2Ac) present in mammals, plants and Schizosaccharom-
yces pombe by a unique N-terminal extension of approxi-
mately 70 amino acids. We have overexpressed S. cerevisiae
Pph22p and its N-terminal deletion mutant DN-Pph22p in
the GS115 strain of Pichia pastoris and purified these
enzymes to apparent homogeneity. Similar to other
heterologous systems used to overexpress PP2Ac, a low yield
of an active enzyme was obtained. The recombinant
enzymes designed with an 8 ·His-tag at their N-terminus
were purified by ion-exchange chromatography on DEAE-
Sephacel and affinity chromatography on Ni
2+
-nitrilotri-
acetic acid agarose. Comparison of biochemical properties
of purified Pph22p and DN-Pph22p with purified human
8·His PP2Ac identified similarities and differences
between these two enzymes. Both enzymes displayed similar
specific activities with
32
P-labelled phosphorylase aas
substrate. Furthermore, selected inhibitors and metal ions
affected their activities to the same extend. In contrast to
the mammalian catalytic subunit PP2Ac, but similar to the
dimeric form of mammalian PP2A, Pph22p, but not
DN-Pph22p, interacted strongly with protamine. Also with
regard to the effects of protamine and polylysine on phos-
phatase activity Pph22p, but not DN-Pph22p, behaved
similarly to the PP2Ac–PR65 dimer, indicating a regulatory
role for the N-terminal extension of Pph22p. The N-terminal
extension appears also responsible for interactions with
phospholipids. Additionally Pph22p has different redox
properties than PP2Ac; in contrast to human PP2Ac it
cannot be reactivated by reducing agents. These properties
make the S. cerevisiae Pph22p phosphatase a unique enzyme
among all type 2A protein phosphatases studied so far.
Keywords:Saccharomyces cerevisiae; protein phosphatase
Pph22p; protein phosphatase 2A; heterologous expression,
Pichia pastoris.
Reversible protein phosphorylation catalysed by protein
kinases and phosphoprotein phosphatases is a major
mechanism utilized by eukaryotic organisms to regulate
various cellular processes [1]. Protein kinases are apparently
derived from one primordial gene. In contrast, protein
phosphatases are encoded by at least three unrelated gene
families. Based on primary and tertiary structure similarit-
ies, protein phosphatases are currently classified into PPP,
Mg
2+
-dependent PPM (both PPP and PPM are specific
against phosphoserine/phosphothreonine residues) and
PTP (phosphotyrosine residues-specific) families [2,3]. The
PTP family comprises also dual-specificity phosphatases
that are able to dephosphorylate all three phospho-residues
[4]. Mammalian type 2A protein phosphatase (PP2A), a
member of the PPP family, displays a broad substrate
specificity in vitro. However, its in vivo substrate selectivity,
enzymatic activity and subcellular localization are regulated
by the association with regulatory subunits [5,6]. Thus, two
different dimeric forms of PP2A are formed by the
association of the catalytic subunit (PP2Ac) with PR65/A
scaffolding subunit or a4 protein. In addition, association of
a third variable subunit derived from the unrelated protein
families PR55/B, PR61/B¢or PR72/B¢¢ to the PR65/A–
PP2Ac dimer results in the formation of trimeric PP2A [6].
In vivo substrates of PP2A in mammalian cells comprise
protein kinases and transcription factors [7]. However, the
identity of many physiological substrates of PP2A still
remains elusive.
In budding yeast Saccharomyces cerevisiae protein kin-
ases and protein phosphatases regulate cell growth, cell cycle
progression, bud formation and morphogenesis as well as
nutrient- and pheromone-induced signalling [8]. The num-
ber of protein kinases in yeast (119) is approximately four
times higher than the number of protein phosphatases (31)
[9]. However, by association of a single catalytic subunit
with different regulatory subunits, protein phosphatases can
form several functional holoenzymes and thus match the
complexity of protein kinases [2,3,5–7]. All above listed
families of protein phosphatases are encoded by the
S. cerevisiae genome and represented by 12 (PPP),
Correspondence to S. Wera, Laboratorium voor Moleculaire
Celbiologie, K.U. Leuven, Kasteelpark Arenberg 31,
B-3001 Leuven-Heverlee, Flanders-Belgium.
Fax: + 32 16 32 19 79, Tel.: + 32 16 32 15 00,
E-mail: stefaan.wera@bio.kuleuven.ac.be
Abbreviations: PP2A, protein phosphatase type 2A; PP2Ac, the
catalytic subunit of PP2A; Pph21/22p, PP2Ac from Saccharomyces
cerevisiae; PR65/A, the structural subunit of PP2A; KM71
and GS115, strains of Pichia pastoris; GSSG, glutathione disulfide;
GSH, reduced glutathione.
Enzyme: protein phosphatase 2A (EC 3.1.3.16).
*Note: deceased on 13 February 2001.
(Received 31 January 2002, revised 15 April 2002,
accepted 29 April 2002)
Eur. J. Biochem. 269, 3372–3382 (2002) FEBS 2002 doi:10.1046/j.1432-1033.2002.02965.x
6 (PPM) and 13 (PTP) members [8,9]. In budding yeast
Schizosaccharomyes pombe, PP2A is encoded by PPH21
and PPH22 [10]. Both Pph21p and Pph22p are involved in
actin cytoskeleton reorganization, bud morphogenesis and
cell cycle progression from G
2
to M [11–13]. Pph21p and
Pph22p are highly similar (87%) and apparently perform
overlapping functions. Deletion of both PP2A catalytic
subunit genes in budding yeast results in very slow growth.
Additional deletion of the PP2A-related PPH3 gene is lethal
[11,14]. Four polypeptides, encoded by CDC55,TPD3,
RTS1 and TAP42, form complexes with PP2A catalytic
subunits in yeast [12,15–17]. Cdc55p, Tpd3p, Rts1p and
Tap42p correspond, respectively, to mammalian PR55/B,
PR65/A, PR61/B¢and a4. The corresponding genes are not
essential but their mutation results in specific phenotypes.
Moreover, two genes (RRD1 and RRD2) encoding homo-
logues of mammalian phosphotyrosine phosphatase activa-
tor (PTPA), a protein isolated from mammalian tissue
based on its ability to stimulate PP2A activity against
phosphotyrosine residues, are present in the budding yeast
genome [18,19].
All catalytic subunits of PP2A from various species are
subject to diverse regulatory control mechanisms. Carbo-
xymethylation of Leu309 (Leu377 of S. cerevisiae) influen-
ces PP2A activity of PP2A and is a signal for exchanging
variable regulatory B family subunits [20–22] (reviewed in
[23]). Phosphorylation of Tyr307 is dependent on insulin,
epidermal growth factor, interleukin-1, tumour necrosis
factor a(and some other pathways) and inactivation of
phosphatase activity (reviewed in [24]). However no data
are available concerning phosphorylation of Tyr375 in
S. cerevisiae. PP2Ac is also phosphorylated on a threonine
residue, but the role and site(s) of phosphorylation is
unknown [25]. PP2A interacts with second messenger
C
2
-ceramide and phospholipids, which stimulate its activity
(reviewed in [23,24,26]). Moreover, PP2A can potentially be
regulated by changes in the redox state of the catalytic
subunit [27,28].
Both Pph21p and Pph22p differ from the catalytic
subunits of PP2A (PP2Ac) of mammals, plants and
Schizosaccharomyces pombe by the presence of a unique
N-terminal extension of approximately 70 amino acids. In
order to assess the impact of this N-terminal extension on
enzymatic properties of PP2A we expressed 8 ·His-tagged
Pph22p and a mutant of Pph22p lacking the N-terminal
extension (DN-Pph22p) in the yeast Pichia pastoris, purified
the phosphatases to apparent homogeneity and compared
their biochemical properties to that of purified 8 ·His-
tagged human PP2Ac expressed in Pichia.
MATERIALS AND METHODS
Host strains, media and buffers
The strain GS115 (his4,AOX1, AOX2)ofP. pastoris was
used for the overexpression of S. cerevisiae Pph22p,
DN-Pph22p and N-terminus of Pph22p (first 77 amino
acids). Human PP2Acaand PR65a/Aawere overexpressed
and purified using KM71 (his4,aox1, AOX2) as described
previously [29]. All strains were grown, transformed, and
analyzed according to the manufacturer’s (Invitrogen)
instructions. Escherichia coli strains DH5aand Top 10F¢
were used for all plasmid constructions and propagations.
The following media were used to grow P. pastoris:RDB-
agar: 1
M
sorbitol, 2% glucose, 1.34% yeast nitrogen base
without amino acids, 4 ·10
)5
% biotin, 2% agar; MD
medium: 1.34% yeast nitrogen base without amino acids,
4·10
)5
% biotin, 2% glucose; MM medium: 1.34% yeast
nitrogen base without amino acids, 4 ·10
)5
%biotin,0.5%
methanol; YPD: 1% yeast extract, 2% peptone, 2% glucose
pH 5.8 adjusted with HCl; MGY medium: 1.34% yeast
nitrogen base without amino acids, 4 ·10
)5
%biotin,1%
glycerol. The following buffers were applied to purify
recombinant Pph22p and DN-Pph22p: SCED buffer: 1
M
sorbitol, 10 m
M
sodium citrate pH 7.5, 10 m
M
EDTA,
10 m
M
dithiothreitol; breaking buffer: 50 m
M
Tris/HCl
pH 7.5, 1 m
M
EDTA, 0.1% 2-mercaptoethanol, 10 m
M
NaCl, 5% glycerol, 10 m
M
phenylmethanesulfonyl fluoride
and 20 m
M
benzamidine; buffer A: 20 m
M
Tris/HCl
pH 7.5, 170 m
M
NaCl (150 m
M
for DN-Pph22p purifica-
tion), 0.1 m
M
EDTA, 0.1% 2-mercaptoethanol, 5% gly-
cerol, 1 m
M
phenylmethanesulfonyl fluoride and 2 m
M
benzamidine; buffer B: 20 m
M
Tris/HCl pH 7.5, 450 m
M
NaCl, 30 m
M
imidazole, 5% glycerol and 0.01% Triton
X-100; buffer C: 20 m
M
Tris/HCl pH 7.5, 20% glycerol
(± 0.5 m
M
dithiothreitol).
Molecular cloning of the Pph22p expression constructs
Genomic DNA of S. cerevisiae strain W303 was obtained
by the ammonium acetate method [30] and used as template
to amplify the PPH22 open reading frame with Pfu DNA
polymerase (Stratagene) using a standard protocol. The
following primers were used: sense (1), 5¢-CGGGATCC
ACCATGCATCATCATCATCATCATCATCATGATA
TGGAAATTGATGACCCTATG-3¢(BamHI site under-
lined, 8 ·His-tag bold) and antisense (2), 5¢-CGGAA
TTCTTATAAGAAATAATCCGGTGTCTTC-3¢(EcoRI
site underlined). For cloning of DN-Pph22p (Pph22p
without first 77 amino acids) and the N-terminus of Pph22p
(only the first 77 amino acids) we used: sense primer:
5¢-CGGGATCCACCATGCATCATCATCATCATCAT
CATCATCTTGACCAATGGATTGAGCATTTG-3¢
(BamHI site underlined, 8 ·His-tag bold) and antisense:
5¢-CGGAATTCTTACTGATTTATATTTGTATTGGT
CAG-3¢(EcoRI site underlined). The PCR products were
digested with EcoRI and BamHI and purified by agarose
gel electrophoresis using the Geneclean III kit (BIO101).
The isolated fragments were first subcloned into pBluescript,
and the resulting plasmid amplified in Escherichia coli.
Subsequently, the Pph22p-encoding fragments were sub-
cloned into the pPIC3.5K vector (Invitrogen). All plasmids
used were sequenced with vector- and cDNA-specific
primers.
Homologous recombination in KM71 and GS115 strains
of
Pichia pastoris
Ten micrograms of plasmid DNA produced in E. coli
DH5astrain, was either used without restriction enzyme
digestion or linearized with either SalI, NotIorBglII in
thecaseofpPIC3.5K-PPH22 and SalI in the case of
pPIC3.5K-DN-PPH22 and pPIC3.5K-Nterm (N-terminus
of Pph22p) or not digested, and transformed by the
spheroplast method into KM71 and GS115 strains of
P. pastoris. Transformed yeast cells were plated on
FEBS 2002 Characterization of protein phosphatase Pph22p (Eur. J. Biochem. 269) 3373
RDB-agar plates and transformants were transferred to
plates with either glucose (MD) or methanol (MM) medium
as a carbon source. Transformants that displayed the ability
to grow on both carbon sources were selected for further
evaluation. The presence of cDNA encoding Pph22p, DN-
Pph22p and N-terminus of Pph22p integrated into the yeast
genome was confirmed by PCR analysis applying sense and
antisense oligonucleotides to amplify the PPH22 gene
(sequences listed above). Transformants obtained using
undigested plasmid DNA in KM71 strain and those
obtained after linearization of plasmid with both SalIand
NotI in GS115 in case Pph22p were used for further
evaluation. In case DN-Pph22p and N-terminus of Pph22p
transformants obtained from both kind of DNA (undigest-
edanddigestedwiththeSalI) were used for further
experiments. In order to select transformants with the
highest copy number of PPH22 genes and mutants genes
inserted into the Pichia genome, yeast colonies were
transferredtoYPD-agarplatesorYPD-agarplatescon-
taining G418 (Calbiochem) added at 2 and 4 mgÆmL
)1
.The
fastest growing colonies were selected from YPD-agar
plates containing 4 mgÆmL
)1
G418 and those were selected
for mini-scale expression studies.
Mini-scale expression of Pph22p and DN-Pph22p
in
P. pastoris
Recombinants obtained in the KM71 strain (His
+
Mut
S
,
slow methanol utilization) or in the GS115 strain (either
His
+
Mut
+
or His
+
Mut
S
, fast and slow methanol utiliza-
tion, respectively) were grown for 24 h in 10 mL of MGY to
reach a D
600
between 2 and 6. Yeast cells were centrifuged
and resuspended in MM medium using 0.2 volume of
starting culture volume for His
+
Mut
S
or adjusting D
600
to 1
for His
+
Mut
+
. Methanol-induced cultures were grown at
30 C in an Aquatron AI 15 incubator (Infors HT) with
shaking set up to 280 r.p.m. Methanol was added to 0.5%
(v/v) every 24 h and the induction was carried out for
9 days. To determine the optimal time for protein expres-
sion, aliquots of the cultures were removed at 24-h intervals
and analyzed for the presence of the heterologous protein
by SDS/PAGE with Coomassie staining, immunodetection
with Tetra-His antibodies (Qiagen) and phosphatase
activity measurements. Recombinant GS115 (His
+
Mut
+
)
obtained after transformation of yeast with the NotI
linearized plasmid displaying the highest level of Pph22p
expression, was used for further experiments. In case of
DN-Pph22p and N-terminus for further experiments trans-
formants GS115 (His
+
Mut
+
) obtained with plasmids
linearized with the SalIwereused.
Midi-scale expression of recombinant proteins
in
P. pastoris
For the midi-scale expression of proteins the selected GS115
(His
+
Mut
+
) strain was cultured in a 100-mL baffled flask
in 25 mL of MGY medium at 30 C with shaking at
205 r.p.m. This yeast preculture reached an D
600
5after
24 h; then 5 mL of the preculture was used to inoculate five
portions of 1 L each of MGY medium and grown in 3 L
flasks to D
600
between 2 and 6. Cells were harvested, washed
and resuspended in 1 L of MM medium to induce
overexpression of heterologous proteins. These cultures
were grown for 24 h at 30 C with shaking set at 205 r.p.m.
After centrifugation at 2000 gfor 5 min at room tempera-
ture the cell pellet was washed with ice-cold water and
stored at )80 C.
Purification of recombinant Pph22p and DN-Pph22p
Methanol-stimulated GS115-PPH22 (or GS115-DN-
PPH22) cells (approximately 50 g) were resuspended in
100 mL of SCED buffer supplemented with 150 mg
(84.7 UÆmg
)1
)ofyeastlyticasefromArthrobacter luteus
(ICN) and incubated at 30 Cfor90mintoachieve
spheroplast formation. The spheroplasts were harvested by
centrifugation at 750 gfor 10 min at 25 C and resus-
pended in 50 mL of ice-cold breaking buffer. Acid-washed
glass beads (size 450–500 lm) were added (1 : 1, v/v) and
the mixture was vortexed (MS-1 minishaker, IKA) 10
times for 1 min each with 1-min intervals for cooling on
ice. The lysates were cleared by centrifugation at 30 000 g
for 30 min at 4 C. Cell-free supernatants were combined
and fractionated with ammonium sulfate added to obtain
45% saturation. The precipitated protein was collected by
centrifugation, dissolved in 20 mL of buffer A and dialysed
against buffer A. The dialysate was loaded at a flow rate of
15 mLÆh
)1
onto a DEAE-Sephacel column (2 ·10 cm)
equilibrated previously with buffer A. The column was
washed with 20 column volumes of buffer A and
phosphatase activity was eluted with a linear gradient
from 170 m
M
to 500 m
M
NaCl (in the case of DN-Pph22p
from 150 to 700 m
M
NaCl) in buffer A collecting 3-mL
fractions. The fractions containing phosphatase activity
were combined, dialysed against buffer B and loaded onto
aNi
2+
-nitrilotriacetate agarose (Qiagen) column
(2 ·3 cm). The column was washed with buffer B and
protein eluted with a linear gradient of imidazole from 30
to 200 m
M
, collecting 2-mL fractions. The fractions were
analysed by immunodetection with Tetra-His antibodies
and phosphatase activity assays. Combined fractions
corresponding to the peak of Pph22p activity were dialysed
against buffer C with or without dithiothreitol and stored
in small aliquots at )80 C. Protein concentration was
determined by the Bradford method using bovine serum
albumin as standard.
Antibodies and immunodetection
To detect recombinant His-tagged proteins monoclonal
mice IgG
1
Tetra-His antibodies (Qiagen) were applied as
primary antibodies followed by goat anti-mouse horse
radish peroxidase-coupled secondary antibodies (Santa
Cruz Biotechnology Inc.). The colour reaction was devel-
oped in the presence of reduced form of NAD plus either
nitro blue tetrazolium (Sigma) or 4-chloro-1-naphthol (Sig-
ma). When goat anti-mouse alkaline phosphatase-coupled
secondary antibodies were applied the colour reaction was
developed in the presence of nitro blue tetrazolium and
5-bromo-4-chloro-3-indolyl phosphate (ICN).
Phosphatase activity assays
Protein phosphatase activity was measured with
32
P-labelled
phosphorylase a(10 l
M
) as substrate as described pre-
viously [31]. When indicated, protamine (33 lgÆmL
)1
)and
3374 P. Zabrocki et al. (Eur. J. Biochem. 269)FEBS 2002
ammonium sulfate (16 m
M
), were included in the assay
buffer. The recombinant PR65/A subunit was preincubated
with PP2Ac, Pph22p or DN-Pph22p in 20 m
M
Tris/HCl
pH 7.4, 50 m
M
NaCl, 0.1 m
M
EDTA and 0.1% 2-mercapto-
ethanol for 10 min at 30 C before the reaction was
started with
32
P-labelled phosphorylase a. To measure the
effect of pH on PP2A activity the buffer containing 20 m
M
sodium acetate/acetic acid, 20 m
M
imidazole/HCl and
20 m
M
Tris/HCl covering pH from 5.0 to 10.0 was applied.
One unit of phosphatase activity corresponds to 1 lmol of
32
P
i
released from
32
P-labelled phosphorylase aper min at
30 C.
For activity assays with lipids, reactions were carried out
described previously [26] with minor changes. Lipids and
phosphatases were incubated on ice for 30 min, prior to the
phosphatase activity assay. Reactions were carried out for
15–30 min at 30 C. The assay was terminated by addition
of 0.1 mL 1 m
M
KH
2
PO
4
in 1
M
H
2
SO
4
and 0.3 mL 2%
ammonium molybdate. After 10 min of incubation a
toluene/isobutyl alcohol mixture (1 : 1) was added, vortexed
for 10 s and centrifuged for 10 min. Free
32
P
i
was
determined from the radioactivity recovered in the organic
phase.
32
P-Labelled phosphorylase ahydrolysis did not
exceed 20% of total phosphorylase ainallsamples.All
illustrated data represent the mean of at least two
independent experiments.
Preparation of liposomes
Phospholipids were solubilized in chloroform and after
evaporation of chloroform were resuspended in 50 m
M
Tris
pH 7.4, 0.1 m
M
EDTA, 0.1% 2-mercaptoethanol buffer by
sonication under argon. Sonication was carried out in an
ice-bath for 10 min with breaks (24 kHz) (BioMetra
Ultrasonicator). Before use, liposomes were kept for 2 h
on ice to allow association of lipids.
Determination of the influence of disulfides on yeast
and mammalian recombinant PP2Ac
Pph22p, PP2Ac and DN-Pph22p were incubated with
20 m
M
dithiothreitol overnight at 4 C. Mixtures were
dialyzed extensively against buffer containing 50 m
M
Tris,
pH 7.4, 0.1 m
M
EDTA and 20% glycerol. Determination
of the influence of glutathione disulfide (GSSG) and
reduced glutathione (GSH) was carried out by mixing this
redox agent with the purified phosphatase and incubation at
30 C for 30 min. The phosphatase assay was initialized by
adding
32
P-labelled phosphorylase a. The reaction buffer
contained 20 m
M
Tris, pH 7.4, 0.1 m
M
EDTA and 10%
glycerol.
Reactivation of PP2Ac and Pph22p activity
Reactivation was carried out as described previously [27]
with minor changes. PP2Ac, Pph22p and DN-Pph22p were
inactivated by incubation with 20 m
M
GSSG overnight at
4C. Mixtures were dialyzed extensively against buffer
contained 50 m
M
Tris, pH 7.4, 0.1 m
M
EDTA and 20%
glycerol.
Aliquots of the inactivated enzymes were mixed with
dithiothreitol or 2-mercaptoethanol at various concentra-
tions. Phosphatase activity in samples was determined after
a 10-min incubation at 30 C by addition of
32
P-labelled
phosphorylase a. Reactions were carried out for 30 min
under standard conditions.
RESULTS AND DISCUSSION
Comparison of PP2A catalytic subunits
from
S. cerevisiae
and other species
S. cerevisiae protein phosphatases encoded by PPH21 and
PPH22 are homologues of mammalian PP2Ac. Pph21p
consists of 369 amino acids and Pph22p of 377 amino acids.
Both enzymes are hence larger than PP2Ac from mammals,
plants and S. pombe which are composed of 306–322 amino
acids. This difference in size results from the presence of an
acidic stretch of approximately 70 amino acids (pI 3.78 and
4.07 for Pph21p and Pph22p, respectively) at the N-termini
of Pph21/Pph22p (Fig. 1). The role of this N-terminal
extension present in budding yeast PP2Ac is currently
unknown. One may speculate that these regions are
responsible for targeting Pph21p and Pph22p to intracellu-
lar compartments or to specific substrates, or fulfil a special
regulatory function. Interestingly, the N-terminal regions of
Pph21p and Pph22p are quite divergent showing only
49.4% amino-acid sequence identity (the first N-terminal 42
amino acids of Pph22p display only 33.3% identity to the
corresponding region in Pph21p) whereas the overall
identity between enzymes equals 87%. This might indicate
that the N-termini of Pph21p and Pph22p may have distinct
functions. In order to determine whether the N-terminal
extension present in Pph22p influences its biochemical
properties we decided to overexpress this phosphatase and
its deletion mutant without 77 N-terminal amino acids in
P. pastoris, purify these enzymes to apparent homogeneity
and compare their enzymatic properties to those of human
PP2Ac.
Purification of
S. cerevisiae
Pph22p expressed
in
P. pastoris
We determined the growth curves of control GS115,
recombinant GS115-PPH22, and GS115-DN-PPH22
strains in minimal medium containing methanol (data not
shown). A lag period of approximately 100 h was observed
in the case of the GS115-PPH22 and GS-115-DN-PPH22
strains cultured starting from a D
600
of 0.05, but not in the
wild-type control. After this period the recombinant strains
resumed growth and eventually reached a D
600
similar to
that of the wild-type strain. Protein levels in both strains
were similar, but phosphatase activity in GS115-DN-PPH22
lysates was lower than in lysates of GS115-PPH22; it is
likely that more DN-Pph22p was in the insoluble state and
this might also explain why the yield of purification was
lower in case of DN-Pph22p. Cultures in the stationary
phase (high D
600
) showed less pronounced differences
between the strains, but even under these conditions the
strain overexpressing Pph22p grew somewhat slower.
Phosphorylase phosphatase activity was measured in cell-
free extracts of all strains and its dependence on growth
phase (reflected in D
600
value) was analysed. At stationary
phase (high D
600
), both Pph22p and DN-Pph22p proteins
were maximally overexpressed 24 h after methanol induc-
tion; amounts of active phosphatase decreased after this
FEBS 2002 Characterization of protein phosphatase Pph22p (Eur. J. Biochem. 269) 3375
time, as confirmed also by Western blotting analysis (data
not shown).
The long lag period in the growth of the GS115-PPH22
strain observed after transferring the cells to methanol-
containing medium is similar to that described for the strain
overexpressing human PP2Ac [29] and might reflect effects
of higher phosphatase activity on yeast growth or on the cell
cycle.
Figure 2 illustrates the purification of Pph22p and
DN-Pph22p from P. pastoris cells using ammonium sulfate
fractionation, DEAE-Sephacel and Ni
2+
-nitrilotriacetic
acid agarose, as described in the Materials and methods
section. The final preparation, stained with Coomassie
Brilliant Blue, appeared to be homogeneous. Purity was
confirmed by gel filtration (data not shown). Pph22p and
DN-Pph22p proteins were purified with a yield of active
protein of 80 and 60 lgÆL
)1
of P. pastoris culture, respect-
ively, in intracellular overexpression.
The inclusion of Triton X-100 (0.01%) in the buffers used
for chromatography on Ni
2+
-nitrilotriacetic acid agarose
greatly enhanced recovery of active phosphatase from
this column. Similarly to mammalian PP2Ac, Pph22p and
DN-Pph22p migrated as a doublet of two proteins. Pph22p
migrated on SDS/PAGE with a molecular mass of
52–53 kDa, different from its calculated molecular mass
of 44 kDa. DN-Pph22p migrated on SDS/PAGE at its
theoretical molecular mass of 37 kDa. The specific activity
of the final Pph22p and DN-Pph22p appropriate prepara-
tions was 1.3 and 1.8 lmolÆmin
)1
Æmg protein
)1
using phos-
phorylase aas substrate. The specific activity of DN-Pph22p
is similar to the 1.7 lmolÆmin
)1
Æmg protein
)1
obtained for
recombinant human PP2Ac [29], but the value for Pph22p is
lower indicating an inhibitory effect of the N-terminus.
Characterization of purified Pph22p and DN-Pph22p
PP2Ac was initially described as a metal-ion-independent
protein phosphatase [32]. In agreement with this, none of
metal ions tested increased significantly the activity of
Pph22p, DN-Pph22p or PP2Ac (Table 1). In contrast,
several metal ions applied at a concentration of 1 m
M
(Co
2+
,Ni
2+
,Fe
2+
,Fe
3+
and Zn
2+
) inhibited Pph22p
and PP2Ac activities with phosphorylase aas a substrate. In
order to exclude the latter effects being substrate dependent
we confirmed the data from Table 1 using kemptide as
substrate (not shown). It remains to be determined whether
the inhibitory effect of these high concentrations of metal
ions reflect an interaction with SH groups exposed on the
enzyme surface or formation of complexes with amino-acid
Pph22_S.cerevisiae
Pph21_S.cerevisiae
ppa1_S.pombe
PP2Ac/beta_rabbit
PP2Ac/alfa_H.sapiens
ppa2_pombe
PP2Ac_2_A.thaliana
Pph3_S.cerevisiae
Pph22_S.cerevisiae
Pph21_S.cerevisiae
ppa1_S.pombe
PP2Ac/beta_rabbit
PP2Ac/alfa_H.sapiens
ppa2_pombe
PP2Ac_2_A.thaliana
Pph3_S.cerevisiae
Pph22_S.cerevisiae
Pph21_S.cerevisiae
ppa1_S.pombe
PP2Ac/beta_rabbit
PP2Ac/alfa_H.sapiens
ppa2_pombe
PP2Ac_2_A.thaliana
Pph3_S.cerevisiae
Fig. 1. Alignment of PP2A from S. cerevisiae and other organisms. Sequence alignment of PP2A catalytic subunits from S. cerevisae (Pph21p,
Pph22p, Pph3p), S. pombe (ppa1), rabbit, Homo sapiens and Arabidopsis thaliana. Conserved residues are coloured. The N-terminal extension is
only found in the S. cerevisiae PP2A isoforms. The region deleted in DN-Pph22p is framed.
Fig. 2. Purification of Pph22p and DN-Pph22p from overexpressing
P. pastoris cells. Aliquots of Pph22p and DN-Pph22p overexpressed in
P. pastoris and purified by using three steps of purification (protein
precipitation with ammonium sulfate, ion-exchange chromatography
on DEAE-Sephacel and affinity chromatography on Ni
2+
-nitrilotri-
acetatic acid agarose were taken and analysed by polyacrylamide
(10%) gel electrophoresis and staining with Coomassie Brilliant
Blue. St, molecular mass standard (kDa); lane 1, Pph22p; lane 2,
DN-Pph22p.
3376 P. Zabrocki et al. (Eur. J. Biochem. 269)FEBS 2002