Phosphorylation and dephosphorylation of polyhydroxy compounds by class A bacterial acid phosphatases

Article abstract of DOI:10.1039/b304012g

Nonspecific acid phosphatases share a conserved active site with mammalian glucose-6-phosphatases (G6Pase). In this work we examined the kinetics of the phosphorylation of glucose and dephosphorylation of glucose-6-phosphate (G6P) catalysed by the acid phosphatases from Shigella flexneri (PhoN-Sf) and Salmonella enterica (PhoN-Se). PhoN-Sf is able to phosphorylate glucose regiospecifically to G6P, glucose-1-phosphate is not formed. The K_m for glucose using pyrophosphate (PPi) as a phosphate donor is 5.3 mM at pH 6.0. This value is not significantly affected by pH in the pH region 4-6. The K_m value for G6P by contrast is much lower (0.02 mM). Our experiments show these bacterial acid phosphatases form a good model for G6Pase. We also studied the phosphorylation of inosine to inosine monophosphate (IMP) using PPi as the phosphate donor. PhoN-Sf regiospecifically phosphorylates inosine to inosine-5′-monophosphate whereas PhoN-Se produces both 5′IMP and 3′IMP. The data show that during catalysis an activated phospho-enzyme intermediate is formed that is able to transfer its phosphate group to water, glucose or inosine. A general mechanism is presented of the phosphorylation and dephosphorylation reaction catalysed by the acid phosphatases. Considering the nature of the substrates that are phosphorylated it is likely that this class of enzyme is able to phosphorylate a wide range of hydroxy compounds.

Full text of DOI:10.1039/b304012g

View Article Online / Journal Homepage / Table of Contents for this issue
Phosphorylation and dephosphorylation of polyhydroxy compounds
by class A bacterial acid phosphatases
Naoko Tanaka, Zulfiqar Hasan, Aloysius F. Hartog, Teunie van Herk and Ron Wever*
Institute for Molecular Chemistry, University of Amsterdam, Nieuwe Achtergracht 129,
The Netherlands. E-mail: rwever@science.uva.nl; Fax: ϩ31 20 525 5670; Tel: ϩ31 20 525 5110
Received 14th April 2003, Accepted 16th June 2003
First published as an Advance Article on the web 9th July 2003
Nonspecific acid phosphatases share a conserved active site with mammalian glucose-6-phosphatases (G6Pase).
In this work we examined the kinetics of the phosphorylation of glucose and dephosphorylation of glucose-6-
phosphate (G6P) catalysed by the acid phosphatases from Shigella flexneri (PhoN-Sf ) and Salmonella enterica
(PhoN-Se). PhoN-Sf is able to phosphorylate glucose regiospecifically to G6P, glucose-1-phosphate is not formed.
The Km for glucose using pyrophosphate (PPi) as a phosphate donor is 5.3 mM at pH 6.0. This value is not
significantly affected by pH in the pH region 4–6. The Km value for G6P by contrast is much lower (0.02 mM).
Our experiments show these bacterial acid phosphatases form a good model for G6Pase. We also studied the
phosphorylation of inosine to inosine monophosphate (IMP) using PPi as the phosphate donor. PhoN-Sf
regiospecifically phosphorylates inosine to inosine-5Ј-monophosphate whereas PhoN-Se produces both 5ЈIMP
and 3ЈIMP. The data show that during catalysis an activated phospho-enzyme intermediate is formed that is
able to transfer its phosphate group to water, glucose or inosine. A general mechanism is presented of the
phosphorylation and dephosphorylation reaction catalysed by the acid phosphatases. Considering the nature of
the substrates that are phosphorylated it is likely that this class of enzyme is able to phosphorylate a wide range of
hydroxy compounds.
available, showing the striking similarity of the active sites of
Introduction
those enzymes.9,14 A similar architecture of the active site of
the acid phosphatases, the lipid phosphatases, the vanadium
haloperoxidases and the mammalian G6Pases15 suggests that in
principle this super family of enzymes is able to catalyse essen-
tially the same basic reactions with variations in specificity and
catalytic efficiency. This has already been demonstrated by us
showing that vanadate substituted acid phosphatases have
bromoperoxidase activity9 and apo CPO has phosphatase
activity.5
Nonspecific acid phosphatases (NSAPs) are bacterial enzymes
which are able to catalyse the hydrolysis of phosphate
monoesters. On the basis of amino acid sequences, NSAPs
are categorized into three classes, designated as class A, B and
C.1,2 Three domains are present in the amino acid sequences
of class A acid phosphatases which contain their active
site residues. These are also conserved in mammalian glucose-
6-phosphatases (G6Pase), lipid phosphatases, vanadium-
containing chloroperoxidases (CPO) and vanadium-containing
bromoperoxidases (BPO).2–9 The amino acid residues in
sequence motifs of domain I: KX6RP, domain II: PSGH and
domain III: SRX5HX3D play very important roles in catalysis.
The active site residues participate in the binding of vanadate
or phosphate, act as nucleophiles, stabilize the penta-
coordinated transition state and play a role in leaving group
protonation.10
G6Pase, a key enzyme in glucose homeostasis, catalyses the
hydrolysis of G6P to glucose and phosphate, the terminal steps
in gluconeogenesis and glycogenolysis. Mutations have been
identified11 in G6Pase that cause glycogen storage disease type
Ia which is an autosomal recessive metabolic disorder. It was
shown that these mutations alter the active site residues and
abolish G6Pase activity demonstrating the importance of these
residues in phosphatase action. The importance of these resi-
dues was also demonstrated in mutagenesis studies in the con-
served domain of two lipid phosphatases.8,12 Similarly, Renirie
et al. investigated the effect of 6 mutations of putative catalytic
residues on the phosphatase activity of apo CPO.13 Based on
these mutagenesis studies, a model was derived that may serve
as a template for G6Pase and other related phosphatases. The
kinetic data on the phosphatase reaction catalysed by apo CPO
point to a mechanism in which during turnover a phosphoryl-
ated enzyme intermediate is present.
Recently, Asano et al.16–20 reported a new enzymatic method
of phosphorylation of inosine to produce inosine-5Ј-
monophosphate (5ЈIMP) using pyrophosphate (PPi) as a phos-
phate donor by the recombinant PhoC from Morganella
morganii.17,18,20 They investigated the phosphotransferase activ-
ity of a number of enterobacteria,16,19 and using PPi as the
phosphate donor they showed that especially class A1 acid
phosphatases, which are not inhibited by fluoride, exhibit high
regioselective transphosphorylation activity.19 M. morganii
PhoC belongs to class A1 acid phosphatase in which the non-
specific acid phosphatase from Shigella flexneri (PhoN-Sf ) is
also classified.
G6Pase catalyses not only the hydrolysis of G6P but it has
phosphotransferase activity, being able to synthesize G6P from
glucose and various phosphate donors including PPi.21 These
studies prompted us to study the phosphorylation and dephos-
phorylation of glucose and G6P, respectively, by acid phos-
phatases, and compare the catalytic mechanism and parameters
with those of G6Pase. We also compare the phosphorylation of
glucose by the acid phosphatase with that of inosine. Although
this reaction has been studied by Asano et al.,16–20 some kinetic
details for this enzymatic phosphorylation are still lacking. Our
data point to the evidence of an activated phosphorylated
enzyme intermediate that is able to regiospecifically transfer its
phosphate group to an appropriate substrate acceptor. It is
concluded that the enzymatic formation of phosphate esters
catalysed by the acid phosphatases using pyrophosphate as a
cheap phosphate donor may present a viable alternative for the
classical chemical synthetic methods using POCl3.
Although a crystal structure of the G6Pase is not available
yet, the crystal structures of the nonspecific acid phosphatase
from Escherichia blattae14 and the vanadium-containing
chloroperoxidase from the fungus Curvularia inaequalis3 are
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 8 3 3 – 2 8 3 9
2833

View Article Online
sequence similarity of PhoN-Se and PhoN-Sf is about 40%,
although the active site domains of those enzymes are identical.
Mutagenesis studies of PhoC from Morganella morganii17–19
and non-specific acid phosphatase from Escherichia blattae20
have shown considerable enhancement of phosphorylation
activity of these enzymes by changing a few amino acid residues
near a potential inosine-binding site. Thus the high regio-
selectivity in 5ЈIMP synthesis is determined by amino acid
residues that are not involved in binding the phosphate moiety
of the substrate.
The amounts of 5ЈIMP (Fig. 2; panel A) and 3ЈIMP (Fig. 2;
panel B) produced by PhoN-Se were also dependent on PPi
concentrations. It is clear that the total yields of IMPs differ at
different concentrations of PPi and that they are very low at low
PPi concentrations. Not only was the yield of 5ЈIMP during
phosphorylation higher than 3ЈIMP, but also the dephos-
phorylation rate of 3ЈIMP was faster than that of 5ЈIMP. A
major problem in the study of the phosphorylation by non-
specific acid phosphatases is the hydrolysis of synthesized
5ЈIMP and 3ЈIMP that occurs at the same time, which makes
analysis of the events difficult.
Results and discussion
Expression and purification of recombinant PhoN-Sf
Sh. flexneri phoN was cloned in pET3a and expressed under
control of T7 promoter in pET3a.9 Due to instability of the
expression system using pET3a, it was decided to amplify and
subclone the mature sequence of phoN-sf and fuse it to the gene
III signal sequence in pBAD/gIIIB. PhoN-Sf was inducibly
expressed in the periplasmic space of the E. coli TOP10 host
cell. On average, the yield was approximately 30 mg per litre of
culture. The final specific activity of acid phosphatase in the
strain was about 40 U mgϪ1 with over 95% purity. This expres-
sion system was a significant improvement over the previous
system.9
Phosphorylation of inosine to inosine monophosphate by
nonspecific acid phosphatases
The phosphorylation of inosine to inosine monophosphate
using pyrophosphate (PPi) by recombinant nonspecific acid
phosphatases from Shigella flexneri (PhoN-Sf ) and Salmonella
enterica (PhoN-Se) was tested by using different enzyme con-
centrations. The amount of inosine phosphorylated is linearly
dependent upon the enzyme concentration (results not shown).
This indicates that this phosphotransferase activity is catalysed
by both PhoN-Sf and PhoN-Se. PhoN-Sf catalyses the phos-
phorylation of inosine to inosine-5Ј-monophosphate (5ЈIMP),
whereas PhoN-Se synthesizes both 5ЈIMP and inosine-3Ј-
monophosphate (3ЈIMP). Fig. 1 shows the time course of
5ЈIMP synthesis by PhoN-Sf. The phosphorylating activity
of PhoN-Sf is highly dependent on pH, while the highest yield
of 5ЈIMP was obtained at pH 5.0. However, the activity of
PhoN-Se was hardly affected by pH in the range of 4.5 to 6.0,
only at pH 3.5 the activity was much lower. The maximal
amounts formed were 6.8 mM 5ЈIMP for PhoN-Sf (Fig. 1) and
approximately 4.3 mM 5ЈIMP and 1.7 mM 3ЈIMP for PhoN-Se
(results not shown), respectively. The turnover of PhoN-Sf is
76 minϪ1 (2.8 U mgϪ1) and that of PhoN-Se 5ЈIMP synthesis is
18 minϪ1 (0.67 U mgϪ1). These values are comparable to those
reported by Mihara et al. for various other acid phosphatases.
The phenomenon, that class A1 acid phosphatase PhoN-Sf has
a much higher regioselectivity for the formation of 5ЈIMP, has
also been observed by Mihara et al.18,19 They imply that class
A1 nonspecific acid phosphatases exhibit regioselective PPi-
nucleoside phosphotransferase activity.19 On the other hand
PhoN-Se class A2 acid phosphatase catalysed the synthesis
of 5ЈIMP and 3ЈIMP with a 4-fold lower turnover. The overall
Fig. 2 Time course of 5ЈIMP (panel A) and 3ЈIMP (panel B) synthesis
from inosine and by PhoN-Se using (᭛) 5 mM, (ꢀ) 10 mM, (᭺) 40 mM
and (᭿) 100 mM PPi at pH 6.0. The reaction mixture contains 500 nM
of PhoN-Se, 40 mM inosine, and 100 mM sodium acetate buffer
(pH 6.0).
Basically five reactions are thought to occur at the same time
(Scheme 1). First the enzyme reacts with PPi to produce a
binary pyrophosphate–enzyme complex (step 1). This complex
dissociates (step 2) to yield an activated phosphorylated enzyme
intermediate (“EؒP”). In the next step, a reaction (step 3) with
water may occur resulting in dissociation of the activated phos-
phorylated enzyme as well as the hydrolysis of pyrophosphate.
Fig. 1 Time course of 5ЈIMP synthesis from inosine and PPi by
PhoN-Sf at pH 3.5–6.0. The reaction mixture contains 1 µM of
phosphatase, 40 mM inosine, 100 mM disodium pyrophosphate and
100 mM sodium acetate buffer pH 3.5 (᭜), 4.0 (᭿), 4.5 (ꢀ), 5.0 (᭹),
5.5 (᭺) and 6.0 (ᮀ).
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 8 3 3 – 2 8 3 9
2834

View Article Online
Scheme 1 Overall mechanism of phosphorylation and dephosphorylation catalysed by acid phosphatases. Steps (1) and (2), formation of
phosphorylated enzyme intermediate. Reaction (3), hydrolysis of the intermediate “EؒP”. Equilibrium (4), forward reaction; reaction of “EؒP” to
yield a binary enzyme-phosphorylated substrate (EؒR–O–Pi), backward; dissociation of the dephosphorylated substrate. Equilibrium (5), forward;
dissociation into the phosphorylated substrate and free enzyme, backward; hydrolysis of the phosphorylated substrate.
The intermediate (“EؒP”) may also transfer (step 4) the
phosphate to a bound phosphate acceptor (R–OH, in this case
inosine), which dissociates to form E ϩ ROPi, resulting in the
production of IMP. Thus competition with water occurs,
and a high inosine concentration is required for an effective
transphosphorylation reaction. At low PPi concentration, the
maximal yields of 5ЈIMP and 3ЈIMP are very low because the
intermediate concentration of the phosphorylated enzyme
intermediate is very low and also rapid hydrolysis by water
occurs. Further as the reaction proceeds in time PPi becomes
exhausted and dephosphorylation of IMP occurs via formation
of the activated phosphorylated enzyme intermediate and again
hydrolysis by water occurs. At higher PPi concentration, the
difference in yields between 5ЈIMP and 3ЈIMP became clearer,
and 3ЈIMP was hydrolysed earlier than 5ЈIMP.
Fig. 3 shows the substrate dependency of 5ЈIMP and 3ЈIMP
during the dephosphorylation at pH 6.0. There is a striking
difference in the dephosphorylating activity of PhoN-Se and
PhoN-Sf. The latter enzyme has hardly any activity towards
3ЈIMP and is selectively hydrolysing 5ЈIMP. The Vmax is 24.4 U
mgϪ1 with a Km of 300 µM. The class A2 acid phosphatase
PhoN-Se hydrolyses both 5Ј and 3ЈIMP with a similar Vmax
(6.2 U mgϪ1 and 5.8 UmgϪ1, respectively), but the Km values
differ considerably (225 µM and 34 µM, respectively). This dif-
ference in Km values is probably the reason why the amount of
3ЈIMP produced by PPi decreases more rapidly during turnover
(Fig. 2). The dependence of the phosphorylation rate on the
inosine concentration was also studied but it was difficult to
obtain an accurate Km value due to the solubility of inosine
which is less than 80 mM under our condition. For the non-
specific acid phosphatase E. blattae a Km value of 202 mM was
reported.20
The effect of vanadate on the phosphorylation of inosine
has been studied. Vanadate, which is structurally homologous
to phosphate, is known as an inhibitor of a number of acid
Fig. 3 Dephosphorylation of 5ЈIMP (᭹, solid line) and 3ЈIMP (᭺,
phosphatases.22,23 The Ki for vanadate of PhoN-Sf in the
broken line) by (A) PhoN-Sf and (B) PhoN-Se at pH 6.0. The reaction
dephosphorylation of pNPP is about 70 nM.9 Furthermore,
mixture contains 100 nM of phosphatase, 100 mM sodium acetate
buffer (pH 6.0) and various concentrations of IMPs. Km and Vmax were
calculated by non-linear-regression plot using EnzymeKinetics
(Trinity Software). The data points are means of triplicate measure-
ments.
we have shown that vanadate is incorporated into the
active sites of PhoN-Sf and PhoN-Se with a high affinity and
these vanadate-substituted enzymes show bromoperoxidase
activity.9 When 100 mM PPi was used as phosphate donor
in the phosphorylation of inosine, the addition of 100 µM
vanadate to the reaction mixture at pH 6.0 unexpectedly did
not show inhibition. However, as shown by Tracy et al.24 PPi
is known to form a complex with vanadate and it is likely
that vanadate is scavenged by PPi, therefore inhibition
may be quenched. During the experiment the yields of 5ЈIMP
and 3ЈIMP (results not shown) were increased somewhat. It is
likely that vanadate inhibits hydrolysis of synthesized 5ЈIMP
and 3ЈIMP. Therefore the effect of 100 µM vanadate was
studied on the hydrolysis of 5ЈIMP. Indeed in the presence of
vanadate, the hydrolysis was completely inhibited (results not
shown).
a
Phosphorylation of glucose to glucose-6-phosphate by
nonspecific acid phosphatases
PhoN-Sf and PhoN-Se also mediate the phosphorylation of
glucose to glucose-6-phosphate (G6P) using PPi as the phos-
phate donor. The enzymatic assay system used (see Experi-
mental) allows us to monitor in a convenient way the formation
of G6P. The amount of glucose which was consumed by the
reaction was also determined by HPLC (results not shown) and
is in agreement with that of the enzymatic assay. Glucose-1-
phosphate (G1P) was not formed upon phosphorylation
according to the enzymatic assay using phosphoglucomutase.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 8 3 3 – 2 8 3 9
2835

View Article Online
Table 1 Kinetic constantsa of phoshorylation and dephosphorylation by PhoN-Sf
Reaction
Substrate
pH
Km/mM
8.2 1.9
3.2 0.4
3.2 0.5
3.6 0.2
5.3 1.0
192 22
Vmax/U mgϪ1
Phosphorylation
Glucose
Glucose
Glucose
Glucose
Glucose
Inosine
G6P
4.0
4.5
5.0
5.5
6.0
5.0
6.0
6.0
6.0
26.3 1.3
12.9 0.3
8.2 0.2
4.2 0.1
2.5 0.1
18.5 1.4
19.2 0.4
23.7 1.0
20.3 1.0
Dephosphorylation
0.021 0.004
0.298 0.004
0.087 0.009
5ЈIMP
PPi
a To obtain initial rate constants of glucose phosphorylation, the reaction mixture containing 1 µM of PhoN-Sf, 100 mM disodium pyrophosphate,
100 mM sodium acetate buffer (pH 4.0–6.0) and various concentrations of glucose was incubated for 5 min at 30 ЊC, then the amount of G6P was
assayed using G6P dehydrogenase. For inosine phosphorylation, the reaction mixture containing 1 µM of PhoN-Sf, 100 mM disodium pyro-
phosphate, 100 mM sodium acetate buffer (pH 5.0) and various concentrations of inosine were incubated for 10 min at 30 ЊC, and the amount of
the formed 5ЈIMP was determined using HPLC. Dephosphorylation of G6P, 5ЈIMP and PPi was carried out using the reaction mixture containing
100 nM of PhoN-Sf, various concentrations of each substrate in 100 mM sodium acetate (pH 6.0). After incubating for 1 min, the reaction mixture
was quenched by addition of Biomol Green reagent. The kinetic constants were determined by using EnzymeKinetics (Trinity Software). The data
points are means of triplicate measurements.
Thus this phosphorylation reaction of glucose is also regio-
specific. As shown in Fig. 4, glucose phosphorylation by PhoN-
Sf is dependent upon pH, although the behaviour differs from
the observation of inosine phosphorylation. Unlike inosine
phosphorylation, the hydrolysis of G6P hardly occurred within
8 hours (Fig. 4). However, at low pH after initial formation
of G6P, the phosphorylated sugar was re-hydrolysed as the
reaction continued (results not shown). Since the PhoN-Se
produced only 5 mM G6P at pH 4.0, which is considerably
less than the 40 mM of G6P produced by PhoN-Sf under the
same conditions, we restricted our study to PhoN-Sf. As Fig. 4
shows the amount of G6P produced reaches a maximum
value of approximately 40 mM at pH 4 after 3 hours and then
remains constant. Inspection of Fig. 1 and Fig. 2 show that
under these conditions nearly all PPi is consumed as a con-
sequence of both phosphorylation of glucose and by the
hydrolysis of the activated phosphorylated enzyme intermedi-
ate. It is likely therefore that further conversion of glucose can
be obtained by further addition of PPi. Indeed a second addi-
tion of PPi increased the conversion to about 60% (results not
shown). Fig. 5 illustrates the phosphor nuclear magnetic reson-
ance (31P NMR) spectra taken at several time points during the
incubation. In line with the experiments in Fig. 4, at pH 4.0
approximately 30 mM G6P was produced after 3 hours incu-
bation. Fig. 5 only shows chemical shifts corresponding to G6P,
free phosphate and PPi. Other phosphorylated compounds
were not detected, confirming that phosphorylation of glucose
by PhoN-Sf is very regiospecific.
Fig. 5 Time course of G6P synthesis from glucose and PPi by PhoN-
Sf followed by 31P NMR. The reaction mixture contains 1 µM of
PhoN-Sf, 100 mM glucose, 100 mM disodium pyrophosphate and
100 mM sodium acetate buffer pH 4.0. At zero time a spectrum was
taken from the reaction mixture without PhoN-Sf. The reaction was
initiated by addition of 1 µM PhoN-Sf in the NMR tube. Concen-
trations of products and reactants were determined using dimethyl-
methylphosphonate as an external standard. Peaks at δ Ϫ9.39, 1.21
and 1.84 correspond to pyrophosphate, phosphate and glucose-6-
phosphate, respectively.
Table 1 shows the kinetic constants for the phosphorylation of
glucose by PhoN-Sf that were obtained from initial rate meas-
urements. The Km value for glucose is only slightly dependent
upon pH whereas the Vmax increases more rapidly at low pH
values. The hydrolysis of G6P was also studied, and Table 1
shows that at pH 6.0 the Km value of G6P is considerably
less (21 µM) than for glucose in the phosphorylation reaction
(5.3 mM). Similarly the Km values for inosine in the phosphoryl-
ation reaction is higher (192 mM) than the Km values for 5ЈIMP
(0.3 mM) in the dephosphorylation reaction. The data indicate
that non-phosphorylated substrates have a lower affinity for the
enzyme than the phosphorylated substrates.
Fig. 4 Time course of G6P synthesis from glucose and PPi by PhoN-
Sf. The reaction mixture contains 1 µM of PhoN-Sf, 100 mM glucose,
100 mM disodium pyrophosphate and 100 mM sodium acetate buffer
pH 3.5 (᭜), 4.0 (᭿), 4.5 (ꢀ), 5.0 (᭹), 5.5 (ᮀ) and 6.0 (᭺).
The difference in Km values for glucose and inosine (5.3 mM
versus 192 mM) may explain that G6P is hardly dephosphoryl-
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 8 3 3 – 2 8 3 9
2836

View Article Online
ated (Fig. 4) and 5ЈIMP is relatively easily dephosphorylated
(Fig. 1) during phosphorylation. Scheme 1 shows that there is a
competition between the phosphate acceptor R–OH and water
for the phosphate–enzyme intermediate “EؒP”. Low values of
Km for the phosphate acceptor R–OH or its high concentration
will drive the equilibrium in the direction of the phosphorylated
intermediate and inhibit the hydrolytic reaction.
ATP as a cofactor. Although recycling of ATP is possible the
system has as a major drawback that for each class of substrate
converted another enzyme has to be used. The kinetic studies
reported here show clearly that these acid phosphatases can be
used to regiospecifically phosphorylate inosine and glucose
under very mild conditions using pyrophosphate as a phos-
phate donor. These substrates are structurally very different
and both polyhydroxy compounds are apparently able to enter
the active site of the acid phosphatases and become phos-
phorylated as well. If so, the enzymatic method may become a
useful alternative to the existing chemical methods.
This also explains why glucose inhibits the dephosphoryl-
ation of G6P. Fig. 6 shows inhibition of PhoN-Sf on G6P
phosphorylation by glucose (pH 6.0). The mechanism of the
inhibition is complex. Panel A indicates competitive inhibition
by glucose, which gives Kic of 50 µM, and panel B indicates
uncompetitive inhibition with Kiu of 150 µM. The inhibition of
G6P dephosphorylation by glucose gives an interesting clue to
the understanding of the scheme of dephosphorylation and
phosphorylation. In this case the glucose reacts with the phos-
phorylated enzyme intermediate resulting in re-formation of
G6P. It is interesting to note that inhibition of mammalian
G6Pase activity by glucose has been observed and this inhib-
ition gave a clue to the mechanism of G6Pase.25 The values
found here for the Km of the acid phosphatase for glucose and
the G6P of 5.3 mM and 0.021 mM at pH 6.0, respectively
(Table 1), are much lower than those for liver microsomal
G6Pase. For this enzyme, values reported are a Km for glucose
of 90–120 mM (pH 5.5–7.0) in the phosphotransferase reaction
and 0.65 mM for the hydrolysis of G6P (pH 6.5), respectively.21
Dephosphorylation and phosphorylation by apo haloperoxidases
It has been shown now that the binding pocket for vanadate
in the haloperoxidases is similar to the phosphate binding site
in the acid phosphatases and G6Pase.5 It has also been
demonstrated that apo CPO had phosphatase activity when
p-nitrophenyl phosphate (pNPP) was used as a substrate.5
Taking this analogy into consideration, we studied the de-
phosphorylation of substrates for acid phosphatases by apo
haloperoxidases. Adenosine- 5Ј-monophosphate (5ЈAMP),
5ЈIMP, guanosine 5Ј-monophosphate (5ЈGMP), adenosine
diphosphate (ADP), inosine triphosphate (ITP), guanosine tri-
phosphate (GTP), xanthine triphosphate (XTP), pNPP, PPi,
G6P, phenyl phosphate, phenolphthalein monophosphate,
phenolphthalein diphosphate, and carbamyl phosphate were
tested, but only pNPP and PPi were hydrolysed by apo CPO
and apo BPO. Since the active site of the vanadium CPO is
similar to that of the vanadium BPO, we also studied the phos-
phatase activity of apo BPO from Ascophyllum nodosum.
Indeed as Fig. 7 (panel A) shows, pNPP was hydrolysed but the
maximal turnover of pNPP hydrolysis by apo BPO is 0.08
minϪ1 at pH 8.7. This turnover is about 15 times lower than that
Fig. 6 Inhibition of PhoN-Sf on G6P dephosphorylation by glucose.
(A) Kic is given by plots of 1/v against glucose (inhibitor) concentrations
at various substrate concentrations. (B) Kiu is given by plots of [S]/v
against glucose (inhibitor) concentrations at various substrate
concentrations. The concentrations of G6P (substrate) were 0.02 mM
(᭛), 0.10 mM (ᮀ), 0.20 mM (᭝), 0.50 mM (᭺) and 2.0 mM (᭿). The
combination of intersecting lines in (A) and (B) indicate a mixed type
of inhibition. Competitive inhibition was seen as in (A) with Kic of
50 µM and uncompetitive inhibition was seen as in (B) with Kiu of
150 µM. The reaction mixture was buffered with 100 mM sodium
acetate (pH 6.0), and the reaction was quenched by addition of Biomol
Green reagent.
It is noteworthy that the Vmax values for the dephosphoryl-
ation catalysed by the PhoN-Sf of G6P, 5ЈIMP and PPi are
about the same (Table 1), despite the difference in size and
charge of these substrates. This points to a step in the mechan-
ism where hydrolysis of the phosphorylated enzyme intermedi-
ate by water and liberation of phosphate is rate-limiting. Such a
rate-limiting step has been suggested for the phosphatase reac-
tion catalysed by apo CPO,5 and it is likely that it corresponds
to step 3 in Scheme 1.
The classical chemical introduction of a phosphate group
into a polyhydroxy compound is tedious since it requires a
number of protection and deprotection steps. Further POCl3
that is being used should be handled with caution and it may
give rise to a number of side products. Employing enzymes may
eliminate many of the disadvantages associated with these
chemical syntheses. In fact kinases have been used to phos-
phorylate a number of compounds.26 A problem associated
with the use of these enzymes is that they require expensive
Fig.
7
Phosphatase activity of apo BPO and apo CPO. (A)
Dephosphorylation of various concentrations of pNPP by 500 nM apo
BPO in 100 mM Tris–acetate at pH 8.7. The reaction mixture was
incubated for 21 hours. (B) Dephosphorylation of various
concentrations of PPi by 500 nM apo CPO in 100 mM sodium acetate
buffer at pH 4.5. The reaction mixture was incubated for 30 min and
quenched by Biomol Green reagent.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 8 3 3 – 2 8 3 9
2837

View Article Online
for apo CPO.5 Apo CPO was also able to hydrolyse PPi with a
turnover of about 1.9 minϪ1 at pH 4.5 (Fig. 7 panel B). This
should be compared to the hydrolysis rate of PPi by PhoN-Sf
(Vmax of 550 minϪ1 at pH 6.0, Table 1), which is about 300 times
faster than that of the apo haloperoxidases. The Km for PPi on
hydrolysis by apo CPO was approximately 50 µM, which is
nearly in the same order as that of PhoN-Sf (87 µM at pH 6.0,
Table 1).
We also investigated whether apo BPO or CPO were able to
phosphorylate inosine or glucose using 100 mM PPi as a
phosphate donor. Conversion of inosine into 5ЈIMP or 3ЈIMP
or that of glucose into G6P was not observed even after incu-
bation with these enzymes for 24 hours. Our studies on the acid
phosphatases show that an activated phosphorylated enzyme
intermediate (“EؒP”) must be present catalysing the transfer of
a phosphate group to an acceptor (R–OH). Considering the
very low hydrolysis rate of PPi by apo haloperoxidase, only a
very low intermediate concentration of an activated phospho-
enzyme (“EؒP”) is present. This makes phosphate transfer by
these enzymes to a phosphate acceptor very unlikely.
The obvious question is why the PPi reacts so sluggishly with
the apo haloperoxidases compared to the acid phosphatases
despite the fact that the binding pocket for phosphate is identi-
cal in the two classes of enzymes. However, the nature of the
amino acid residues in the channel giving access to the phos-
phate binding pocket is probably very different. A comparison
of the active site channel of the apo CPO3 and that of the acid
phosphatase from E. blattae14 shows clearly that the active site
channel in the former enzyme is much more restricted. This
may also be the reason why the NSAPs have such a wide range
in substrate specificity.
and HindIII sites, respectively, are underlined). PCR was per-
formed using the Expand High Fidelity PCR System (Roche)
with the following conditions: 10 ng pET3a phoN-sf gene, 1 µg
each primer, 200 µM each dNTP, 1.5 mM MgCl2, 2.6 U high-
fidelity polymerase mix in a final volume of 50 µl. A “hot start”
of 3 min at 95 ЊC was followed by 30 cycles of denaturation (1
min at 95 ЊC), annealing (1 min at 55 ЊC) and extention (1 min
at 72 ЊC), and 7 min at 72 ЊC using a programmable heating
block (Eppendorf Mastercycler 5330). pBAD/gIIIB (Invitro-
gen) was chosen for the phoN-Sf gene, as it holds the gene III
signal sequence for secretion of the recombinant protein into
the periplasmic space.9 The PCR product was restricted with
NcoI and HindIII cloned into the corresponding sites of pBAD/
gIIIB, and the resulting clone was confirmed by double-
stranded DNA sequencing.
E. coli TOP10 harbouring the recombinant plasmid pBAD/
gIIIB phoN-sf was grown at 37 ЊC in LB medium until the
absorbance of the culture suspension reached an A600 of 0.4–
0.6. The expression of recombinant PhoN-Sf was induced by
adding 0.02% -arabinose and the growth was continued at
37 ЊC for 4 h. The bacterial cells were harvested by centri-
fugation, and secreted PhoN-Sf was released from the E. coli
periplasmic space by osmotic shock. The cell pellet was re-
suspended in osmotic shock solution 1 (20 mM Tris–HCl, pH
8.0, 2.5 mM EDTA and 20% sucrose) to A600 = 5, and was
incubated on ice for 10 min. After centrifugation for 15 min at
4 ЊC, the cell pellet was resuspended in osmotic shock solution 2
(20 mM Tris–HCl, pH 8.0 and 2.5 mM EDTA) to A600 = 5, and
was incubated on ice for 10 min. The secreted PhoN-Sf was
obtained in the supernatant (osmotic shock fluid) after centri-
fuging for 15 min at 4 ЊC. The osmotic shock fluid was applied
onto a DEAE sepharose (Pharmacia Biotech) ion-exchange
column which was washed with 20 mM Tris–HCl (pH 8.8), and
the enzyme was eluted with 1 M NaCl in 20 mM Tris–HCl
(pH 8.8). The active fractions were pooled and exhaustively
dialysed against 20 mM sodium acetate (pH 5.0). The dialysate
was centrifuged at 20000 g for 30 min and the supernatant was
passed through a 0.45 µm filter (Millipore) and then applied
onto a SP Sepharose Fast Flow (Pharmacia Biotech) ion-
exchange column. The enzyme was eluted with a linear gradient
of NaCl (0–0.3 M) in 20 mM sodium acetate (pH 5.0). The
active fractions were pooled and dialysed against 50 mM Tris–
acetate (pH 7.5) and concentrated by membrane filter Amicon
PM 10 (Millipore). Sephadex G75 column (Pharmacia Biotech)
was used to perform a gel filtration step. The recombinant
PhoN-Sf was eluted with 50 mM Tris–acetate (pH 7.5) at a flow
rate of 0.1 ml minϪ1, and was concentrated on a membrane
filter Amicon PM 10.
Experimental
Materials
All standard recombinant DNA procedures were performed as
described by Sambrook et al.27 Expression and purification of
Salmonella enterica PhoN (PhoN-Se) was as described else-
where.9 Plasmid pET3a harbouring phoN-sf was a kind gift
from Prof. A. J. Lange.9
The host strains Escherichia coli and TOP10 (Invitrogen)
were used in subcloning and expression experiments. Bacteria
were routinely grown at 37 ЊC in Luria-Bertani (LB) medium
(10 g LϪ1 tryptone, 10 g LϪ1 NaCl and 5 g LϪ1 yeast extract at
pH 7.5)27 containing 100 µg mlϪ1 ampicillin. Expression vectors
pBAD/gIIIA and pBAD/gIIIB (Invitrogen) were used to clone
the phoN gene from S. enterica and Sh. flexneri respectiv-
ely. pBAD/gIIIA and pBAD/gIIIB carry the gene III signal
sequence for secretion of the recombinant protein into the
E. coli periplasmic space.
The purity of the preparations was checked on SDS-PAGE
gels stained with Coomassie Brilliant Blue R-250, and the pro-
tein concentration was determined by using a protein assay kit
(Bio-Rad) with BSA as the standard.
Vanadium-containing bromoperoxidase was purified from
the seaweed Ascophyllum nodosum as described previously.28,29
The Curvularia inaequalis vanadium chloroperoxidase was
purified using the Saccaromyces cerevisiae recombinant expres-
sion system.30
Enzymatic assay of phosphotransferase activity
The enzymatic phosphorylation of inosine was assayed as
described by Asano et al.16–20 A standard reaction mixture con-
tains 40 mM inosine, 100 mM disodium pyrophosphate (PPi)
and 0.1–1 µM of enzyme solution in a total volume of 1 ml. It
was not convenient to use tetrasodium pyrophosphate under
acidic conditions because the pH of the solution was alkaline
(about pH 11), therefore disodium pyrophosphate (pH 4.2) was
used for the detailed studies unless otherwise mentioned. The
pH of all reaction mixtures was adjusted after addition of
sodium acetate buffer and PPi. For the time course study, 100 µl
of sample was taken from the reaction mixture and diluted 2–10
times before injection into the HPLC. The amount of inosine
and phosphorylated products 5ЈIMP and 3ЈIMP were deter-
mined by HPLC using a Nucleosil 100–5 C18 column (0.45 ×
12 cm; Macherey-Nagel) equipped with a Pharmacia LKB-
-glucose, glucose-6-phosphate, inosine, 5ЈIMP, 3ЈIMP, tetra-
sodium pyrophosphate (Na4P2O7) and disodium pyrophos-
phate (Na2H2P2O7) were purchased from Sigma (USA), and
glucose-6-phosphate dehydrogenase, phosphoglucomutase and
NADP were purchased from Roche (Germany).
Expression and purification of recombinant PhoN-Sf
The gene for PhoN-Sf was cloned in the expression vector
pBAD/gIIIB as follows. The mature sequence of phoN-sf
(i.e. phoN gene without the 5Ј-end coding for the native secre-
tion signal) was PCR amplified from the pET3a phoN-sf gene
using the forward primer 5Ј-CATGCCATGGCTCAATTC-
CTCCGGGAAATG-3Ј and the reverse primer 5Ј-CCCAA-
GCTTATTTTTTCTGATTGTTAGCGAATTC-3Ј (the NcoI
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 8 3 3 – 2 8 3 9
2838

View Article Online
HPLC pump 2248 and an LKB Bromma 2140 rapid spectral
detector. The column was eluted with 10 mM CH3COOH–
NH4OH (pH 5.0) containing 200 µM sodium azide at a flow
rate of 0.5 ml minϪ1. The retention time for 5ЈIMP, 3ЈIMP and
inosine were 4, 6 and 15 min, respectively. The HPLC effluent
was monitored at 254 nm. The Borwin software program
(JMBS developments) was used for HPLC data acquisition and
evaluation.
Acknowledgements
This work was supported by the Council of Chemical Science
of the Netherlands Organisation for Scientific Research. We
gratefully acknowledge Dr. W. Hemrika and Dr.V. Dumay
for providing us the plasmid harbouring, the S. enterica ser.
typhimurium PhoN gene, and Prof. A. J. Lange, University
of Minnesota (USA) for providing us the plasmid pET3a
phoN-sf.
During glucose phosphorylation, the formation of G6P was
assayed enzymatically by glucose-6-phosphate dehydrogenase.
This assay is based on the method of Noltman et al.31 A phos-
phorylation reaction mixture contains 1 µM PhoN, 100 mM
glucose and 100 mM disodium pyrophosphate in 100 mM
sodium acetate (pH 3.5–6.0). To determine the amount of
G6P, 10 µl of phosphorylation reaction mixture was added to
1 ml of a G6P assay mixture. This assay mixture contains
0.01 mg mlϪ1 glucose-6-phosphate dehydrogenase, 1 mM
of NADPϩ and 10 mM MgCl2 in 100 mM Tris–acetate
(pH 7.5). The formed NADPH can be monitored at 340 nm
(molar absorption coefficient 6.22 mMϪ1 cmϪ1). The form-
ation of glucose-1-phosphate (G1P) was measured by conver-
sion of G1P to G6P using 0.06 mg mlϪ1 phosphoglucomutase
in the same assay system as mentioned above. After measur-
ing the absorption at 340 nm and a stable reading was
obtained, phosphoglucomutase was added to the same mixture,
and the additional increase of absorption at 340 nm was
monitored.
The quantity of glucose and free phosphate were also deter-
mined by HPLC using an Alltech OA 1000 organic acid column
(0.65 × 30 cm) equipped with a Dionex 580 LPG pump and
Dionex UVD-340D/Shodex RI-101 detector. The column was
eluted with 25 mM H2SO4 at a flow rate of 0.4 ml minϪ1. The
Chromeleon software program (Dionex) was used for HPLC
data acquisition and evaluation.
G6P, PPi and free phosphate were quantified by phosphor
nuclear magnetic resonance (31P NMR). Spectra were deter-
mined in D2O on a Varian Unity Inova at 202 MHz. Chemical
shifts (δ) are expressed in ppm relative to 85% phosphoric
acid.
References
1 G. M. Rossolini, S. Schippa, M. L. Riccio, F. Berlutti, L. E.
Macaskie and M. C. Thaller, Cell. Mol. Life. Sci., 1998, 54, 833–850.
2 J. Stukey and G. M. Carman, Protein Sci., 1997, 6, 469–472.
3 A. Messerschmidt and R. Wever, Proc. Natl. Acad. Sci. USA, 1996,
93, 392–396.
4 M. Weyand, H.-J. Hecht, M. Kiess, M.-F. Liaud, H. Vilter and
D. Schomburg, J. Mol. Biol., 1999, 293, 864–865.
5 W. Hemrika, R. Renirie, H. L. Dekker, P. Barnett and R. Wever,
Proc. Natl. Acad. Sci. USA, 1997, 94, 2145–2149.
6 A. F. Neuwald, Protein Sci., 1997, 6, 1764–1767.
7 W. Plass, Angew. Chem., 1999, 38, 909–912.
8 Q.-X. Zhang, C. S. Pilquil, J. Dewald, L. B. Berthiaume and
D. N. Brindley, Biochem. J., 2000, 345, 181–184.
9 N. Tanaka, V. Dumay, Q. Liao, A. J. Lange and R. Wever, Eur. J.
Biochem., 2002, 269, 2162–2167.
10 W. Hemrika, R. Renirie, S. Machedo-Ribeiro, A. Messerschmidt
and R. Wever, J. Biol. Chem., 1999, 274, 23820–23827.
11 A. Ghosh, J.-J. Shieh, C.-J. Pan, M.-S. Sun and J. Y. Chou, J. Biol.
Chem., 2002, 277, 32837–32842.
12 D. A. Toke, M. L. McClintick and G. M. Carman, Biochemistry,
1999, 38, 14606–14613.
13 R. Renirie, W. Hemrika and R. Wever, J. Biol. Chem., 2000, 275,
11650–11657.
14 K. Ishikawa, Y. Mihara, K. Gondoh, E.-I. Suzuki and Y. Asano,
EMBO J., 2000, 19, 2412–2423.
15 W. Hemrika and R. Wever, FEBS Lett., 1997, 409, 317–319.
16 Y. Asano, Y. Mihara and H. Yamada, J. Mol. Catal. B: Enzymat.,
1998, 6, 271–277.
17 Y. Asano, Y. Mihara and H. Yamada, J. Biosci. Bioeng., 1999, 87,
732–738.
18 Y. Mihara, T. Utagawa, H. Yamada and Y. Asano, Appl. Environ.
Microbiol., 1998, 66, 2811–2816.
At zero time a spectrum was taken from the reaction mixture
containing 100 mM PPi, 100 mM -glucose, 100 mM sodium
acetate buffer (pH 4.0) in a 10 mm NMR tube. The reaction was
initiated by addition of 1 µM PhoN-Sf, and was carried out in
the NMR tube. Concentrations of product and reactants were
determined using dimethylmethylphosphonate as an external
standard. The peak at δ = 1.84 corresponds to that seen in
authentic G6P (80 mM).
19 Y. Mihara, T. Utagawa, H. Yamada and Y. Asano, J. Biosci. Bioeng.,
2001, 92, 50–54.
20 K. Ishikawa, Y. Mihara, N. Shimba, N. Ohtsu, H. Kawasaki,
E.-I. Suzuki and Y. Asano, Protein Eng., 2002, 15, 539–543.
21 W. J. Arion, B. K. Wallin, P. W. Carlson and A. J. Lange, J. Biol.
Chem., 1972, 247, 2558–2565.
22 R. L. VanEtten, P. P. Waymack and D. M. Rehkop, J. Am. Chem.
Soc., 1974, 96, 6482–6785.
23 P. J. Stankiewiecz and M. J. Gresser, Biochemistry, 1988, 27, 206–
212.
24 A. S. Tracey, M. J. Gresser and S. Liu, J. Am. Chem. Soc., 1988, 110,
Enzymatic assay of phosphatase activity
2869–5874.
25 L. F. Hass and W. L. Byrne, J. Am. Chem. Soc., 1960, 82, 947–954.
26 K. Faber, in Biotransformation in organic chemistry, 4th edition,
Springer, Berlin, 2000, p. 123–134.
27 J. Sambrook, E. F. Frischt and T. Maniatis, in Molecular cloning:
A Laboratory manual, Cold Spring Harbor Laboratory Press,
Cold Spring Harbor, NY, 1989.
28 E. de Boer, Y. van Kooyk, M. G. M. Tromp, H. Plat and R. Wever,
Biochim. Biophys. Acta, 1986, 869, 48–53.
29 R. Wever, H. Plat and E. de Boer, Biochim. Biophys. Acta, 1985, 830,
181–186.
30 P. Barnett, D. H. Hondmann, L. H. Simons, P. F. Ter Steeg and
R. Wever, International Patent Application (PCT), 1995, WO95/
27046.
31 E. A. Noltman, C. J. Gubbler and S. A. Kuby, J. Biol. Chem., 1961,
236, 1225–1230.
32 B. Martin, C. J. Pallen, J. H. Wang and D. J. Graves, J. Biol. Chem.,
1985, 260, 14932–14937.
33 K. W. Harder, P. Owen, L. K. Wong, R. Aebersold, I. Clark-Lewis
and F. R. Jirik, Biochem. J., 1994, 298, 395–401.
The phosphatase activity was measured by hydrolysis of 10 mM
p-nitrophenyl phosphate as a substrate in 100 mM sodium
acetate (pH 6.0). The reaction mixtures were quenched with
0.5 M NaOH to change the pH to 12, and the production of
p-nitrophenol was measured at 410 nm (molar absorption
coefficient 16.6 mMϪ1 cmϪ1).
When PPi, 5ЈIMP, 3ЈIMP or G6P were used as substrates,
Biomol Green reagent (Biomol), which is a modification of a
malachite green method,32,33 was used for the detection of free
phosphate. A reaction mixture contained 100 mM sodium
acetate (pH 6.0), various concentrations of substrate and 20–
100 nM of PhoN in a total volume of 100 µl, and was incubated
for 1 min at room temperature. The reaction was quenched by
addition of 1 ml of Biomol Green reagent. Following 20 min
incubation at room temperature, the absorbance values at
620 nm were measured.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 8 3 3 – 2 8 3 9
2839