The transition state of the phosphoryl-transfer reaction catalyzed by the lambda Ser/Thr protein phosphatase

Article abstract of DOI:10.1021/ja990667p

The catalytic reaction of the Mn²⁺ form of the native bacteriophage λ phosphatase and the H76N mutant was studied with the substrate p-nitrophenyl phosphate using heavy atom isotope effects and pH-dependent rate studies. The kinetic isotope eff

Full text of DOI:10.1021/ja990667p

6382
J. Am. Chem. Soc. 1999, 121, 6382-6390
The Transition State of the Phosphoryl-Transfer Reaction Catalyzed
by the Lambda Ser/Thr Protein Phosphatase
†
‡
‡
,†
Richard H. Hoff, Pamela Mertz, Frank Rusnak, and Alvan C. Hengge*
Department of Chemistry and Biochemistry, Utah State UniVersity, Logan, Utah, 84322-0300, Section of
Hematology Research and the Department of Biochemistry and Molecular Biology, Mayo Clinic and
Foundation, Rochester, Minnesota 55905
ReceiVed March 2, 1999
Abstract: The catalytic reaction of the Mn2 form of the native bacteriophage λ phosphatase and the H76N
mutant was studied with the substrate p-nitrophenyl phosphate using heavy atom isotope effects and pH-
dependent rate studies. The kinetic isotope effects in the substrate were measured at the nonbridging oxygen
+
1
8
18
atoms [ (V/K)nonbridge], at the bridging oxygen atom undergoing bond cleavage [ (V/K)bridge], and at the nitrogen
15
atom in the nitrophenol leaving group [ (V/K)]. The isotope effects with native enzyme at the pH optimum
of 7.8 were 1.0133 ( 0.0006 for (V/K)bridge, 1.0006 ( 0.0003 for (V/K), and 0.9976 ( 0.0003 for
V/K)nonbridge. These values were constant within experimental error across the pH range from 6.0 to 9.0 and
were also unchanged for the slower catalytic reaction resulting when Ca was substituted for Mn . The
results indicate that the chemical step of P-O bond cleavage is rate-limiting, the first metallophosphatase for
which this has been shown to be the case. The isotope effects are very similar to those measured for reactions
of protein-tyrosine phosphatases, indicating that the two families of enzymes share similar dissociative transition
states. The 18(V/K)bridge and (V/K) isotope effects for the H76N mutant were slightly increased in magnitude
relative to the native enzyme but were much smaller than the values expected if the leaving group were departing
with a full negative charge. The pH vs kcat profile for the native enzyme is bell-shaped with pKa values of 7.7
1
8
15
1
8
(
2
+
2+
15
(
0.3 and 8.6 ( 0.4. Km values for substrate increased with pH approximately 70-fold across the pH range
5
.8-9.1. The Km for the H76N mutant was similar to that observed for native enzyme at high pH and was
relatively constant across this pH range. The basic limb of the pH-rate profile is reduced but not abolished
in the H76N mutant reaction. The results are discussed in terms of the possible role of His-76 and the nature
of the transition state for catalysis in the native enzyme and mutant.
Introduction
Thr phosphatase families distinguished primarily by substrate
specifities and susceptibility to specific inhibitors; these are
designated types 1, 2A, 2B (also called calcineurin), and 2C.
Among the Ser/Thr phosphatase family crystal structures have
The regulation of metabolism in organisms from bacteria to
higher eukaryotes is accomplished by the reversible phospho-
rylation of proteins. Protein phosphorylation (by kinases) or
dephosphorylation (by phosphatases) occurs primarily on ty-
rosine, serine, or threonine residues. Interestingly, phosphatase
families have evolved which utilize completely different cata-
lytic machinery to accomplish the hydrolysis of phosphate
monoester bonds. The protein-tyrosine phosphatases (PTPases)
have no metal ions and utilize a conserved arginine for substrate
binding and transition-state stabilization, a cysteine nucleophile
to form a phosphoenzyme intermediate, and general acid
catalysis is accomplished by a conserved aspartic acid residue
8
been published for the catalytic subunit of rabbit muscle and
9
10
11
human PP-1, bovine brain, and human calcineurin, for
12
PP2C, and for the related metalloenzyme purple acid phos-
13,14
phatase (PAP).
A variety of experimental evidence indicates
that the binuclear metal center in bovine brain calcineurin is an
Fe-Zn center. The protein phosphatase from bacteriophage λ,
(
5) Barford, D.; Das, A. K.; Egloff, M.-P. Annu. ReV. Biophys. Biomol.
Struct. 1998, 27, 133-164.
(
(
96.
6) Lohse, D. L.; Denu, J. M.; Dixon, J. E. Structure 1995, 3, 987-990.
7) Rusnak, F.; Yu, L.; Mertz, P. J. Biol. Inorg. Chem. 1996, 1, 388-
1-4
which protonates the leaving group.
Less is known about
3
the mechanistic details of catalysis by the Ser/Thr phosphatases,
but these phosphatases are distinguished from the protein-
tyrosine phosphatases by their use of a binuclear metal center
as a key component of catalysis.5 There are four major Ser/
(8) Goldberg, J.; Huang, H. B.; Kwon, Y. G.; Greengard, P.; Nairn, A.
C.; Kuriyan, J. Nature 1995, 376, 745-753.
9) Egloff, M.; Cohen, P. T.; Reinemer, P.; Barford, D. J. Mol. Biol.
(
1
995, 254, 942-959.
-7
(10) Griffith, J. P.; Kim, J. L.; Kim, E. E.; Sintchak, M. D.; Thomson,
J. A.; Fitzgibbon, M. J.; Fleming, M. A.; Caron, P. R.; Hsiao, K.; Navia,
M. A. Cell 1995, 82, 507-522.
(11) Kissinger, C. R.; Parge, H. E.; Knighton, D. R.; Lewis, C. T.;
Pelletier, L. A.; Tempczyk, A.; Kalish, V. J.; Tucker, K. D.; Showalter, R.
E.; Moomaw, E. W. Nature 1995, 378, 641-644.
(12) Das, A. K.; Helps, N. R.; Cohen, P. T.; Barford, D. EMBO J. 1996,
15, 6798-6809.
(13) Klabunde, T.; Strater, N.; Frohlich, R.; Witzel, H.; Krebs, B. J. Mol.
Biol. 1996, 259, 737-748.
(14) Strater, N.; Klabunde, T.; Tucker, P.; Witzel, H.; Krebs, B. Science
1995, 268, 1489-1492.
*
To whom correspondence should be addressed.
Utah State University.
Mayo Clinic and Foundation.
†
‡
(
1) Zhang, Z.-Y.; Wang, Y.; Dixon, J. E. Proc. Natl. Acad. Sci. U.S.A.
994, 91, 1624-1627.
2) Lohse, D. L.; Denu, J. M.; Santoro, N.; Dixon, J. E. Biochemistry
1
1
(
997, 36, 4568-4575.
(
3) Guan, K. L.; Dixon, J. E. J. Biol. Chem. 1991, 266, 17026-17030.
(4) Cho, H.; Krishnaraj, R.; Kitas, E.; Bannwarth, W.; Walsh, C. T.;
Anderson, K. S. J. Am. Chem. Soc. 1992, 114, 7296-7298.
1
0.1021/ja990667p CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/25/1999

Mechanism of λ Protein Phosphatase
J. Am. Chem. Soc., Vol. 121, No. 27, 1999 6383
studies each of the isotope effects revealed alterations in the
transition state when protonation of the leaving group in the
transition state was lost due to mutation of the general acid.
While a number of PTPases have been mechanistically
characterized using isotope effects, similar experiments with
phosphatases utilizing binuclear metal ion catalysis have been
hindered by the fact that the chemical step is less often rate-
limiting with pNPP as the substrate. Isotope effects were
measured for the reaction of pNPP with calcineurin.28 Although
calcineurin is very similar to λPP the chemical step of
phosphoryl transfer was found to be only partially rate-limiting,
Figure 1. The p-nitrophenyl phosphate substrate showing the positions
where isotope effects were measured.
designated λ protein phosphatase (λPP), is considered a member
of the serine/threonine protein phosphatase family on the basis
6
of sequence comparisons and kinetic and spectroscopic
leaving the interpretation of the isotope effects in terms of
1
5-17
28,29
characterizations.
transition-state structure somewhat uncertain.
Alkaline phos-
The X-ray structures of the members of this family indicate
that the binuclear metal center has a ligand environment which
is very similar to that of purple acid phosphatases. In PAP the
stereochemical course of the reaction occurs with inversion of
configuration at phosphorus, supporting a mechanism involving
phatase, another phosphatase utilizing binuclear catalysis, was
found to have isotope effects of unity, consistent with other
data indicating that a nonchemical step is completely rate-
3
0,31
limiting for kcat/KM.
In the present study we report strong evidence that the
chemical step of phosphoryl transfer is fully rate-limiting for
k /K for the reaction of pNPP with λPP, allowing the full
18
direct transfer of the phosphoryl group to water. The nucleo-
3+
phile has been proposed to be an Fe -bound hydroxide ion on
cat
M
19
the basis of rapid kinetic measurements with anions and pH-
intrinsic isotope effects on the transition state to be observed.
In addition the isotope effects for the H76N mutant have been
measured, as well as for the native enzyme in which the supplied
2
0
rate studies. Kinetic studies, solvent isotope effect data, and
redox studies with calcineurin are also indicative of a phos-
phoryl-transfer mechanism which most likely proceeds by direct
transfer to a metal-bound water molecule.6
2
+
2+
metal is Ca in place of Mn , a substitution which results in
,21-24
16
a reduction in rate of about 17-fold. The results yield
In addition to the ligands of the binuclear metal center, there
are several other conserved amino acids within the region of
the active site which could participate in catalysis. One of these
is a histidine residue which in calcineurin (H151) is within 5 Å
of the two metal ions. A His residue in this region is conserved
information about the transition state of the λPP reaction, the
role of His-76, the identity of the substrate as the monoanion
or the dianion of pNPP, and the effect of changing the metal
2
+
2+
ion from Mn to Ca on the transition state of the reaction.
The results also allow an evaluation of the proposal that the
normally loose transition state for solution pNPP hydrolysis is
altered by coordination of the substrate to the metal ions at the
active site and becomes tighter, with greater bond formation to
the nucleophile and less advanced bond cleavage to the leaving
group. It has been noted that coordination to divalent metal ions
in aqueous solution does not alter the transition state of the
6,7
in other Ser/Thr phosphatases such as PP1 (H125), λPP (H76),
and purple acid phosphatase.1 It has been proposed that this
residue could function as a catalytic general acid in the
phosphoryl-transfer reactions catalyzed by this family of
enzymes. Mutation of this residue in λPP results in substantial
3,14
1
5,17
15
kinetic effects
and spectroscopic differences.
32
In this study we report the kinetic isotope effects and pH-
rate studies on the reaction of p-nitrophenyl phosphate (pNPP)
with native λPP and with the H76N mutant. The substrate is
shown in Figure 1 with the positions indicated at which isotope
effects have been measured. Prior studies of protein-tyrosine
phosphatases using this substrate have shown that isotope effects
hydrolysis reaction. However the nature of the transition state
for enzymatic phosphoryl-transfer reactions remains controver-
sial.
Experimental Section
Synthesis of Compounds. The bis(cyclohexylammonium) salts of
reveal the presence or lack of general acid catalysis in the
14
natural abundance p-nitrophenyl phosphate, [ N]-p-nitrophenyl phos-
transition state of the catalytic reaction.2
5-27
In the PTPase
15
18
3
phate, and [ N, nonbridge- O ]-p-nitrophenyl phosphate were syn-
30
14
15
18
thesized as previously described. [ N]-p-nitrophenol and [ N, O]-
(
15) Mertz, P.; Yu, L.; Sikkink, R.; Rusnak, F. J. Biol. Chem. 1997,
72, 21296-21302.
16) Zhuo, S.; Clemens, J. C.; Hakes, D. J.; Barford, D.; Dixon, J. E. J.
Biol. Chem. 1993, 268, 17754-17761.
17) Zhuo, S.; Clemens, J. C.; Stone, R. L.; Dixon, J. E. J. Biol. Chem.
994, 269, 26234-26238.
18) Mueller, E. G.; Crowder, M. W.; Averill, B. A.; Knowles, J. R. J.
Am. Chem. Soc. 1993, 115, 2974-2975.
19) Aquino, M. A. S.; Lim, J.-S.; Sykes, A. G. J. Chem. Soc., Dalton
Trans. 1994, 429-436.
20) Dietrich, M.; Munstermann, D.; Sauerbaum, H.; Witzel, H. Eur. J.
Biochem. 1991, 199, 105-113.
21) Martin, B. L.; D. J., G. Biochim. Biophys. Acta 1994, 1206, 136-
42.
30
p-nitrophenol were synthesized and then mixed to closely reconstitute
the 0.365% natural abundance of 15N. This mixture was phosphorylated
to produce p-nitrophenyl phosphate as the mixture of isotopomers used
for the determination of the (V/K)bridge isotope effect. The [ N]-p-
nitrophenyl phosphate and [ N, nonbridge- O
2
(
(
18
14
1
15
18
3
]-p-nitrophenyl phos-
(
phate isotopic isomers were mixed to reconstitute the natural abundance
15
18
(
of N, and this mixture was used for measurement of the (V/K)nonbridge
isotope effect. Unlabeled pNPP used for pH versus rate experiments
was purchased from Sigma.
(
Kinetic Isotope Effect Determinations. Isotope effect experiments
were run at 100 mM buffer and 1 mM DTT, at 30 °C. The buffers
used were MES at pH 6.0 and TRIS at pH 7.8 and 9.0. DTT was omitted
at pH 9.0. Reactions were begun with 100 µmol of substrate and
(
1
1
1
1
3
1
7
(
22) Martin, B. L.; Graves, D. J. J. Biol. Chem. 1986, 261, 14545-
4550.
(23) Yu, L.; Haddy, A.; Rusnak, F. J. Am. Chem. Soc. 1995, 117, 10147-
0148.
(28) Hengge, A. C.; Martin, B. L. Biochemistry 1997, 36, 10185-10191.
(29) Martin, B. L.; Jurado, L. A.; Hengge, A. C. Biochemistry 1999, 38,
(24) Yu, L.; Golbeck, J.; Yao, J.; Rusnak, F. Biochemistry 1997, 36,
0727-10734.
3386-3392.
(25) Hengge, A. C.; Zhao, Y.; Wu, L.; Zhang, Z.-Y. Biochemistry 1997,
(30) Hengge, A. C.; Edens, W. A.; Elsing, H. J. Am. Chem. Soc. 1994,
6, 7928-7936.
26) Hengge, A. C.; Sowa, G. A.; Wu, L.; Zhang, Z.-Y. Biochemistry
995, 34, 13982-13987.
27) Hengge, A. C.; Denu, J. M.; Dixon, J. E. Biochemistry 1996, 35,
084-7092.
116, 5045-5049.
(
(31) Simopoulos, T. T.; Jencks, W. P. Biochemistry 1994, 33, 10375-
10380.
(
(32) Herschlag, D.; Jencks, W. P. J. Am. Chem. Soc. 1987, 109, 4665-
4674.

6384 J. Am. Chem. Soc., Vol. 121, No. 27, 1999
Hoff et al.
sufficient λPP used so that background hydrolysis rates were less than
An estimation of the nonbridge 18O isotope effect for coordination
31
10% of the enzymatic rate. Enzymatic experiments were run in parallel
was performed by monitoring the separation in the P NMR signals
for 1 O- and for O-labeled pNPP as a function of the concentration
of Ca2 . The solution was 5 mM in pNPP at pH 7.0 (100 mM MOPS
buffer). The pNPP consisted of a 1:1 mixture of natural abundance
6
18
with solution experiments to establish background hydrolysis rates under
experimental conditions. The extent of reaction was measured by
transferring an aliquot of the reaction mixture into 0.1 N NaOH and
assaying for p-nitrophenol by measuring the absorbance at 400 nm.
Isotope effect experiments were run in triplicate and were stopped at
extents of substrate turnover ranging from 40% to 60%. Reactions were
stopped by removing the samples from the temperature controller,
placing them on ice, and titrating to pH 5. An aliquot was removed for
the determination of the precise fraction of reaction. This aliquot was
split, with one portion assayed for p-nitrophenol immediately and the
other portion placed in TRIS buffer at pH 9.0 with alkaline phosphatase
for several hours before it was similarly assayed. The ratio of the amount
of p-nitrophenol in these two samples gave the extent of reaction. The
remaining reaction solution was treated to isolate the p-nitrophenol
product by extracting three times with an equivalent volume of diethyl
ether. The aqueous layer, containing the residual substrate, was
evaporated briefly under vacuum to remove dissolved ether, an
equivalent volume of the TRIS pH 9.0 buffer was added, and the pH
was adjusted to 9.0 with NaOH. This sample was treated with alkaline
phosphatase to quantitatively hydrolyze all of the remaining substrate.
This mixture was then titrated back to pH 5.0 and extracted with ether,
the p-nitrophenol in this ether fraction representing the residual substrate
at the point when the λPP enzymatic reaction was stopped. The ether
fractions were dried over magnesium sulfate and filtered, and the solvent
was removed by rotary evaporation. The p-nitrophenol was sublimed
under vacuum at 90 °C, and 1.0-1.5 mg samples were prepared for
isotopic analysis using an ANCA-NT combustion system in tandem
with a Europa 20-20 isotope ratio mass spectrometer.
+
18
3
compound and of [nonbridge- O ]-p-nitrophenyl phosphate. The
chemical shifts were followed as calcium ion concentrations were
increased from 0 to 1 M at a constant ionic strength of 3.0 M maintained
using NaCl. 18O substitution causes a small upfield shift in the
31
P
18
chemical shift of about 0.02 ppm/ O. The separation will change as a
function of metal ion concentration if there is an isotope effect on the
coordination and will reach a maximum at the dissociation constant
D
K followed by a decrease as the titration is concluded. This technique
36
was introduced by Ellison and Robinson to determine the equilibrium
isotope effect for deprotonation of formic acid and has also been used
to measure the isotope effect for deprotonation of phosphate and of
phosphate esters.37
pH-Dependent Kinetic Assays of λ Protein Phosphatase and λ
Protein Phosphatase H76N. The following buffers at 200 mM
concentration (2× the final concentration in the assays) were prepared
for pH studies: MES for pH 5.8, 6.1, and 6.4; MOPS for pH 6.7, 7.0,
and 7.3; Tricine for pH 7.6, 7.9, and 8.2; Bicine for pH 8.5 and 8.8;
and CHES for pH 9.1. Stocks of pNPP (0.5 M) were also prepared at
each pH by titrating a 1.0 M solution of pNPP with HCl or NaOH to
the desired pH followed by dilution with water to the appropriate
volume.
Assays contained 100 mM buffer, 1 mM MnCl , and the appropriate
2
concentration of NaCl to normalize the ionic strength due to varying
concentrations of pNPP used in the assays. pNPP was varied in order
to determine kinetic parameters at each pH. All reagents except enzyme
were incubated for 5 min at 30 °C before initiation of the reaction by
addition of enzyme. Reactions were allowed to proceed for 2 min and
were terminated by the addition of 1 mL of 2 M TRIS (pH 10) and 10
mM EDTA. At this pH (g10), p-nitrophenol is completely deproto-
nated. After mixing, the absorbance at 410 nm was read immediately
Isotope effects were calculated from the isotopic ratio in the
p-nitrophenol product at partial reaction (R
), and in the starting material (R ). Equation 1 was used to calculate
the observed isotope effect from R and R at fraction of reaction f.
p
), in the residual substrate
(R
s
o
p
o
-
1
-1
isotope effect ) log(1 - f)/log(1 - f(R /R ))
(1)
and converted to specific activity using ꢀ410 ) 17 800 M cm .
p
o
The apparent ionization constants of the enzyme-substrate complex
for λPP were determined by fitting the raw data to the appropriate curve
using a nonlinear least-squares analysis method.
Separation of pNPP from Contaminant Inorganic Phosphate. A
preparation of pNPP containing less contaminating inorganic phosphate
was needed in order to carry out assays with λPP and λPP(H76N) at
Equation 2 was used to calculate the observed isotope effect from the
3
3
s o
same reaction using R and R .
isotope effect ) log(1 - f)/log[(1 - f)(R /R )]
(2)
o
was determined from nitrogen isotope ratio mass spectroscopic
s
o
i
high substrate concentrations to avoid product inhibition (K ortho-
phosphate ) 0.71 mM at pH 7.8).16 This was completed by ether
extractions of pNPP. A solution of pNPP was prepared and acidified
with 1 N HCl to a pH of ∼4.5-5 and extracted twice with ether (ratio
of ether to aqueous phase ∼ 3:1). The aqueous phase was retained and
acidified to pH ∼1.0 using 1 N HCl. This was then extracted 5-6
times with ether (ratio of ether to aqueous phase ∼3:1). Anhydrous
R
analysis of the disodium salt of unreacted substrate. As a control this
isotope ratio was compared to that from samples of p-nitrophenol
isolated after complete hydrolysis of the substrate. The agreement of
these two numbers demonstrates that, within experimental error, no
isotopic fractionation occurs during the isolation and purification of
p-nitrophenol.
4
MgSO was added to the ether phase as a drying agent and was removed
The 18O isotope effects were measured by the remote-label method34
using the nitrogen atom in the substrate as a reporter for isotopic changes
at either the bridging oxygen atom or the nonbridging oxygen atoms,
as previously described for solution reactions of pNPP.30 These
experiments yield an observed isotope effect which is the product of
by filtration. The ether was evaporated to dryness using a rotary
evaporator, and the yellow oil which remained was dissolved in a few
milliliters of H
2
O. The pNPP concentration was determined by dilution
-1
of the pNPP stock into 0.5 M MOPS, pH 7.0, using ꢀ310 ) 9500 cm
-1
M
. The inorganic phosphate concentration of this stock was checked
1
5
18
38
the effect due to both the N and the O substitution. The observed
using the Malachite green/ammonium molybdate reagent. Following
ether extraction of pNPP, the contamination from inorganic phosphate
was 0.14-0.20 mol % compared to about 1.2 mol % in the commercial
sample.
15
isotope effects from these experiments were corrected for the N effect
and for incomplete levels of isotopic incorporation in the starting
3
5
material.
Equilibrium Isotope Effects for Coordination of pNPP to Ca+2
.
Instrumentation. UV-Vis spectra were recorded on a Cary 1
3
1
The coordination of pNPP with calcium ion was monitored by recording
the shift of the UV-vis spectrum as a function of calcium ion
concentration. A 5 mM solution of the disodium salt of pNPP was
prepared at pH 7.0 with 100 mM MOPS buffer. Its spectrum was
spectrophotometer equipped with a thermostated cell holder. P NMR
data were obtained using a Bruker ARX400 spectrometer operating at
161.976 MHz. The spectra were the sum of 50-2000 (usually <200)
scans and were externally referenced to phosphoric acid (0 ppm) in a
coaxial tube. The data were resolution-enhanced by Gaussian apodiza-
monitored at increasing concentrations of CaCl
strength of 3.0 M (NaCl).
2
at a constant ionic
(
36) Ellison, L. R.; Robinson, M. J. T. J. Chem. Soc., Chem. Commun.
1983, 745.
(37) Knight, W. B.; Weiss, P. M.; Cleland, W. W. J. Am. Chem. Soc.
(
(
33) Bigeleisen, J.; Wolfsberg, M. AdV. Chem. Phys. 1958, 1, 15-76.
34) O’Leary, M. H.; Marlier, J. F. J. Am. Chem. Soc. 1979, 101, 3300-
3
306.
35) Caldwell, S. R.; Raushel, F. M.; Weiss, P. M.; Cleland, W. W.
Biochemistry 1991, 30, 7444-7450.
1986, 108, 2759-2761.
(38) Lanzetta, P. A.; Alvarez, L. J.; Reinach, P. S.; Candia, O. A. Anal.
Biochem. 1979, 100, 95-97.
(

Mechanism of λ Protein Phosphatase
J. Am. Chem. Soc., Vol. 121, No. 27, 1999 6385
Table 1. Isotope Effects for λPP Reactions with PNPP
the separation of the peaks of the isotopic isomers. The
deprotonation of phosphate monoesters results in a maximum
change of 0.015 ppm and an isotope effect of 1.015 ( 0.001 at
enzyme and
conditions
15(V/K)
18(V/K)bridge
18(V/K)nonbridge
37
native, Mn2+ 1.0006 ( 0.0003 1.0133 ( 0.0006 0.9976 ( 0.0003
27 °C
(the isotope effect for protonation will be the reciprocal
2
+
(
pH 7.8)
of this number, 0.985). Since Ca results in a maximal change
in this separation of 0.0024 ppm, the isotope effect for
complexation by two calcium ions is estimated to be about 16%
as large as that for protonation of one of the nonbridge oxygen
2
+
+
native, Mn
1.0006 ( 0.0002 1.0132 ( 0.0004 0.9981 ( 0.0002
1.0007 ( 0.0006 1.0143 ( 0.0005 0.9992 ( 0.0001
1.0007 ( 0.0001 1.0130 ( 0.0004 0.9984 ( 0.0002
1.0016 ( 0.0003 1.0183 ( 0.0009 0.9976 ( 0.0001
(pH 6.0)
2
native, Mn
(pH 9.0)
1
8
2
+
atoms, or approximately 0.997. These Knonbridge values are the
native, Ca
18
(pH 7.8)
isotope effects resulting from O in all three nonbridge oxygen
atoms.
2
+
H76N, Mn
(pH 7.8)
Plots of log(kcat/KM) vs pH for the native and for the H76N
λPP and log kcat vs pH for the native λPP are presented in
Figures4 and 5, respectively. Table 3 contains a listing of kinetic
data with pH. Significant substrate inhibition was observed at
tion prior to Fourier transformation. Peak width at half-height was
typically 0.018-0.030 ppm.
2
+
pH e 7, requiring increased levels of Mn in order to obtain
saturating concentrations of substrate. For this reason the
reaction was not explored below pH 5.8. The theoretical curve
through the data in Figure 5 represents the fit to the equation
Results
The isotope effects for the enzymatic reactions of λPP with
pNPP obtained from the isotopic ratios of product and those
obtained from the isotopic ratios of residual substrate agreed
within experimental error in all cases and were averaged to give
the results shown in Table 1 with their standard errors. Six or
more determinations of each isotope effect were made. Isotope
+
+
kcat ) kcat (max)/(1 + [H ]/K1app + K2app/[H ]), where K1app
and K2app are apparent ionization constants of the enzyme-
3
9
-1
substrate complex. The value for kcat (max) was 500 ( 200 s .
K1app and K2app yield pKa values of 7.7 ( 0.3 and 8.6 ( 0.4,
2
+
effects were measured with the wild-type enzyme with Mn
at the pH optimum of 7.8, at pH 6.0 and 9.0, and separately
respectively.
with Ca2 at pH 7.8. Isotope effects with the H76N mutant were
+
Discussion
2+
measured at pH 7.8 with Mn .
Factors Influencing the Expression of the Intrinsic Isotope
Effects. Experimental data show that the reaction of purple acid
phosphatase, which is structurally very similar to the Ser/Thr
phosphatases, proceeds with inversion of stereochemistry at
The values for the 18O isotope effects have been corrected
for the 1 N effects and for levels of isotopic incorporation. Since
the enzymatic substrate is likely the dianion of pNPP (vide
infra), the values for 18(V/K)nonbridge in Table 1 at pH 6.0 have
5
18
phosphorus, consistent with a single-step mechanism. Kinetic
18
been corrected for the equilibrium O isotope effect of 1.015
21,22
and solvent isotope effect data with calcineurin
also are most
37
30
18
on deprotonation, as previously described. The (V/K)nonbridge
consistent with a mechanism involving direct phosphoryl transfer
to a metal-bound water molecule without a phosphoenzyme
intermediate. This evidence indicates that a model for the
catalytic mechanism for the Ser/Thr phosphatases can be
represented as shown in Scheme 1. In this scheme k2 represents
a hypothetical conformational change or other nonchemical step
after substrate binding. There is no direct evidence for such an
additional step in the mechanism of λPP; however it is
reasonable to assume for the moment that a conformational
change may occur upon substrate binding in order to consider
the possibility that a nonchemical step may be partly or fully
rate-limiting, thereby affecting the observed enzymatic kinetic
isotope effects.
1
8
values shown are the isotope effects resulting from O in all
three nonbridge oxygen atoms. For purposes of comparison the
isotope effects for PTPase reactions with the substrate pNPP
from previous studies are shown in Table 2.
Data from titration experiments of pNPP with Ca2 are shown
in Figures 2 and 3. The UV-vis experiment following the
change in λmax of a 5 mM solution of pNPP as a function of
+
2+
Ca concentration indicates complete complexation at about
00 mM metal ion; λmax gradually changed from 311.1 nm in
the absence of calcium ion to a plateau value of 309.5 at 500
5
2
+
31
mM Ca (Figure 2). The P NMR data, however, indicate
that further complexation occurs at higher calcium ion concen-
trations and that saturation is approached but is not complete
Because the competitive method was used to measure the
isotope effects in this study, they are effects on V/K and thus
are effects on the part of the mechanism up to the first
irreversible step, regardless of which step is rate-limiting in the
overall enzymatic mechanism. The first irreversible step is likely
to be the phosphoryl-transfer step from substrate to the metal-
bound water, which is shown as k3 in Scheme 1. The justifica-
tions for representing this step as irreversible are the poor
nucleophilicity of p-nitrophenol and the observation that p-
2
+
at 1 M Ca (Figure 3). A likely possibility is that the UV-vis
experiment detects formation of an initial 1:1 complex, while
at higher calcium concentrations a 2:1 metal-pNPP complex
forms which causes no further change in the UV-vis spectrum
but is detected by changes in the 31P chemical shift. The NMR
data indicate that full saturation is approached at approximately
twice the Ca levels as indicated by the UV-vis experiment.
Since full complexation could not be achieved, an exact fit of
the data to determine the nonbridge 18O isotope effect for
formation of this complex is not possible. An estimation of this
isotope effect can be obtained by comparing the maximum
change in the isotopic separation induced by Ca2 complexation
2
+
22
nitrophenol is a poor inhibitor of Ser/Thr phosphatases, which
suggests that its dissociation from the active site is rapid.
When only one step is isotopically sensitive in an enzymatic
+
41
reaction, the isotope effect on V/K is described by eq 3. In
with that which has been reported for analogous experiments
on the protonation of phosphate esters.40 The magnitude of the
*
*
*
(
V/K) ) [ k + c + c ( K )]/(1 + c + c )
(3)
f
r
eq
f
r
isotope effect will be proportional to the maximum change in
this equation (V/K) represents either 18(V/K) or 15(V/K), *k
*
(
39) Taylor, J. R. An Introduction to Error Analysis; University Science
Books: Mill Valley, CA, 1982.
40) Weiss, P. M.; Knight, W. B.; Cleland, W. W. J. Am. Chem. Soc.
986, 108, 2761-2762.
similarly designates the isotope effect on the isotope-sensitive
(
1
(41) Cleland, W. W. Bioorg. Chem. 1987, 15, 283-302.

6386 J. Am. Chem. Soc., Vol. 121, No. 27, 1999
Hoff et al.
Table 2. Isotope Effects for PTPase Reactions with PNPPa
enzyme and conditions
15(V/K)
18(V/K)bridge
18(V/K)nonbridge
native PTP1, YOP, and VHR (range)
native Stp1
general acid (D to N) mutants of PTP1,
0.9999 to 1.0001
1.0007 ( 0.0001
1.0019 to 1.0034
1.0118 to 1.0152
1.0160 ( 0.0005
1.0275 to 1.0294
0.9981 to 1.0003
1.0018 ( 0.0003
1.0018 to 1.0024
YOP, VHR, and Stp1 (range)
a
Data from refs 25-27. Data from the native PTP1, YOP, and VHR were very similar and are reported as a range; data from the native Stp1
differ in small but systematic ways from the other three enzymes and are reported separately. Data for all four enzymes are from reactions at their
optimum pH. Data from the Asp to Asn general acid mutants of all four enzymes were very similar and are reported as a range.
Figure 2. The variation of the λmax of a 5 mM solution of pNPP as a
function of calcium ion concentration at pH 7.0, with ionic strength
maintained at 3.0 M with NaCl.
step, *Keq is the equilibrium isotope effect in the forward
direction, and the constants cf and cr are, respectively, the
forward and reverse commitment factors.42 There will be no
reverse commitment if the phosphoryl-transfer step k3 is
irreversible; the data described above indicates that this is the
2
1
case. With the mechanism in Scheme 1, if the chemical step
is the only isotope-sensitive step, then the isotope effect is given
by eq 4 and the commitment factor cf will equal (k3/k-2)(1 +
k2/k-1).
*
*
(
V/K) ) ( k + c )/(1 + c )
(4)
3
f
f
Figure 3. (A) The 31P chemical shift of a 5 mM solution pNPP is
shown as a function of calcium ion concentration. Filled diamonds
represent the chemical shift of unlabeled pNPP, and open squares
To the extent that a nonchemical step such as substrate
binding or a subsequent conformational change is rate-limiting,
the resulting forward commitment will suppress the magnitude
of the isotope effects on the chemical step and can completely
abolish them, resulting in observed isotope effects of unity. This
is the case with alkaline phosphatase, where the absence of
18
3
represent nonbridge- O -labeled pNPP. (B) The difference between
the chemical shifts of labeled and unlabeled pNPP is shown as a
function of calcium ion concentration. Data from two titrations under
identical experimental conditions are shown. For comparison, when
18
kinetic isotope effects30 as well as other kinetic data indicate
that a nonchemical step is completely rate-limiting except for
substrates having very poor leaving groups. In the calcineurin
reaction commitments partially suppress the kinetic isotope
effects at the pH optimum, although the suppression is reduced
and the isotope effects on the chemical step are fully or nearly
31
3
the isotope effects for the deprotonation of O -labeled phosphate or
18
of
3
O -labeled phosphate esters are measured by the same technique,
the difference in the isotope-induced chemical shift changes from 0.070
40
a
to a maximum of about 0.085 ppm at the pK .
complex following the hypothetical conformational change or
other nonchemical step. If such an additional step is not rapidly
reversible, then the enzyme-substrate complex will partition
completely forward from this step and the large k3/k-2 ratio will
suppress the isotope effects on the chemical step.
2
8
fully expressed if the reaction is studied at higher pH.
The ratio k2/k-1 is a measure of the fate of the initial enzyme-
substrate complex. If the substrate is tightly bound, this ratio
will be large (the substrate will be “sticky”) and the isotope
effects thereby suppressed. With the λPP enzyme pNPP exhibits
a KM value of 14 mM, a value 30-fold higher than the KM of a
phosphoprotein substrate.16 This fact gives us a reasonable
expectation that the ratio k2/k-1 will be small with pNPP. The
ratio k3/k-2 reflects the partitioning of the enzyme-substrate
In cases where a commitment factor is sufficiently large to
suppress but not entirely abolish isotope effects, the magnitudes
of the observed isotope effects will often increase at nonoptimal
pH values due to the slower rate of the chemical step and the
resultant lowering of the commitment factor, just as was found
2
8
with calcineurin. As a test for the degree to which chemistry
is rate-limiting, the isotope effects for the λPP reaction with
pNPP were measured at the pH optimum of 7.8 and also at pH
6.0 and 9.0, where catalysis is significantly slower. The
(42) Northrop, D. B. In Isotope Effects on Enzyme-Catalyzed Reactions;
Cleland, W. W., O’Leary, M. H., Northrop, D. B., Eds.; University Park
Press: Baltimore, MD, 1977; p 122.

Mechanism of λ Protein Phosphatase
J. Am. Chem. Soc., Vol. 121, No. 27, 1999 6387
Scheme 1
measured are the intrinsic ones for the enzymatic phosphoryl-
transfer step with the native enzyme using pNPP as substrate.
Since the chemical step is considerably slower in the H76N
2
+
2+
mutant, and with the native enzyme when Ca replaces Mn ,
chemistry should be rate-limiting in those reactions as well.
Compared with the data from the native enzyme, the values
of 15(V/K) and of (V/K)bridge increase with the H76N mutant.
This could indicate a change in the nature of the transition state
arising from this mutation. An alternative explanation of the
data is that the isotope effects are suppressed in the native
enzyme but become fully expressed in the slower H76N mutant.
This can be ruled out because a commitment factor in the native
enzymatic reaction would suppress all of the isotope effects in
equal proportion. Compared to the values for the native reaction,
18
m
Figure 4. The pH-kcat/K profile for the hydrolysis of pNPP by wild
type λ protein phosphatase (filled squares) and the H76N mutant (filled
triangles).
1
1
5
8
(V/K) for the H76N mutant is increased about 2.5-fold while
V/K)bridge only increases about 1.3-fold. This means that the
(
changes in the isotope effects caused by the mutation are the
result of a change in the transition state for the phosphoryl-
2
+
transfer reaction. In addition, substitution of Ca results in a
significant reduction in catalytic rate but no change in the isotope
effects which would be expected if the isotope effects were being
partially suppressed by commitment factors.
Interpretation of Kinetic Isotope Effects on Phosphoryl-
Transfer Reactions. Kinetic isotope effects can characterize
reactions in detail, in particular yielding information about the
structure of the transition state. The transition state for hydrolysis
of phosphate monoesters in solution can be described as very
loose or “dissociative” in nature, characterized by extensive bond
cleavage to the leaving group and minimal bond formation to
the nucleophile, and in which the transferring phosphoryl group
Figure 5. pH-kcat profile for the wild type λ protein phosphatase-
catalyzed hydrolysis of pNPP. The line represents the fit to the equation
+
+
k
cat ) kcat(max)/(1 + [H ]/K1app + K2app/[H ]), where K1app and K2app
are the apparent ionization constants of the enzyme-substrate complex.
values of 7.7 ( 0.3 and 8.6 ( 0.4, respectively.
K1app and K2app yield pK
a
Table 3. λPP pH-Rate Data
43,44
resembles a metaphosphate ion.
Phosphodiesters and triesters
MnCl
(mM)
2
kcat/K
m
cat (s-1
(s mM-1)
-1
exhibit successively tighter, more associative transition states
characterized by less bond cleavage to the leaving group and
greater bond formation to the nucleophile, where the transferring
phosphoryl group resembles a pentacoordinate phosphorane.43
Linear free-energy relationships indicate that in diesters and
triesters with good leaving groups the reactions are concerted
with no phosphorane intermediate, but that the transition states
become tighter (more associative) than in the dissociative
pH
k
)
K
m
(mM)
Native λPP pH-Rate Data
5.8
6.1
6.4
6.7
7.0
7.3
7.6
7.9
8.5
8.8
9.1
30 ( 1
44 ( 1
1.0 ( 0.1
1.2 ( 0.1
1.1 ( 0.2
2.8 ( 0.3
2.9 ( 0.5
5.0 ( 0.5
14 ( 4
15
30
37
37
36
35
27
13
6.8
5.7
4.2
0.86
10
10
10
10
1
41 ( 2
102 ( 3
101 ( 4
136 ( 4
180 (10
270 ( 40
280 ( 20
210 ( 20
60 ( 20
1
1
1
1
4
5,46
transition state of the monoester reaction.
Isotope effects
40 ( 10
49 ( 7
have been measured for the phosphoryl-transfer reactions of a
30,35,40,47-49
50 ( 10
70 ( 30
number of phosphate esters in solution.
The cumula-
1
tive data indicate that isotope effects can distinguish between
these types of transition states. Calculations predict inverse
H76N λPP pH-Rate Data
5.8
6.4
6.7
7.0
7.6
7.9
8.2
8.5
8.8
1.6 ( 0.1
2.2 ( 0.1
3.9 ( 0.2
4.1 ( 0.5
1.0 ( 0.1
1.8 ( 0.2
0.93 ( 0.08
0.57 ( 0.07
0.56 ( 0.04
22 ( 3
26 ( 4
49 ( 5
40 ( 10
30 ( 10
70 ( 10
35 ( 7
50 ( 10
60 ( 10
15
0.073
0.085
0.080
0.10
0.033
0.026
0.027
0.011
0.0093
(
43) Thatcher, G. R. J.; Kluger, R. AdV. Phys. Org. Chem. 1989, 25,
10
10
10
1
9
9-265.
(
44) Hengge, A. C. In ComprehensiVe Biological Catalysis: A Mecha-
nistic Reference; Sinnott, M., Ed.; Academic Press: San Diego, CA, 1998;
Vol. 1, pp 517-542.
1
1
1
1
(
45) Ba-Saif, S. A.; Waring, M. A.; Williams, A. J. Am. Chem. Soc.
1
990, 112, 8115-8120.
(46) Davis, A. M.; Hall, A. D.; Williams, A. J. Am. Chem. Soc. 1988,
110, 5105-5108.
47) Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc. 1990, 112, 7421-
7422.
(
magnitudes of the isotope effects are nonunity and, within
experimental error, constant over the pH range examined. This
strongly suggests that the chemical step is rate-limiting for V/K
across the pH range 6.0-9.0 and that the isotope effects
(
48) Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc. 1991, 113, 5835-
5
841.
(49) Hengge, A. C.; Tobin, A. E.; Cleland, W. W. J. Am. Chem. Soc.
1995, 117, 5919-5926.

6388 J. Am. Chem. Soc., Vol. 121, No. 27, 1999
Hoff et al.
nonbridge 18O isotope effects for dissociative transition states
Since the isotope effect for protonation of a phosphate monoester
is 0.985 and if protonation of the substrate occurs before or
during catalysis, the observed isotope effect at this position will
be the product of the inverse isotope effect for protonation and
that for the phosphoryl-transfer step. Since the kinetic
4
0
37
and normal values for associative transition states. The
experimental nonbridge 18O isotope effects for monoester
reactions under different conditions are all small and inverse
(
0.9994-0.9997), and the nonbridge 18O isotope effects for
1
8
diesters and triesters which have been measured are (with a
(V/K)nonbridge effects for monoester hydrolysis in all past studies
have been very small (near unity), if the monoanionic form is
single exception, which may be anomalous) normal (1.0040-
1
8
1
.0250).
involved then the observed (V/K)nonbridge effect should be in
the neighborhood of the isotope effect for protonation (0.985).
The isotope effects in the leaving group also distinguish
1
8
The inverse values of (V/K)nonbridge with λPP were all found
to be much smaller (less inverse) than 0.985 (Table 1). Therefore
they represent exactly the magnitude expected for a loose
transition state of the dianionic substrate, as in the uncatalyzed
reaction in solution.
between the loose transition states of monoesters and the tighter
15
ones of diesters and triesters. The isotope effects (V/K) and
18
(V/K)bridge measure charge delocalization in the leaving group
and P-O bond cleavage, respectively. The extensive bond
1
8
cleavage in the monoester dianion reaction shows bridge- O
isotope effects of 1.0202-1.030 and 1 N isotope effects of
.0028-1.0039. The tighter transition states of diesters and
triesters with this leaving group exhibit values for the bridge-
5
The possibility that coordination to the metal ions at the active
site could result in isotope effects at the nonbridging oxygen
atoms also must be considered. Metal analysis shows that the
purified enzyme is devoid of iron, and no EPR signals for Fe3
are observed. Therefore the active form of the enzyme in these
1
+
1
8
15
O isotope effects in the range 1.0039-1.0060 and for the
isotope effect in the range 1.0007-1.0016.
When protonation of the leaving group occurs in the transition
N
2
+
52
2+
experiments is the di-Mn
and presumably the di-Ca
18
enzyme, depending upon which metal ion is supplied in the
state, the normal (V/K)bridge isotope effect arising from P-O
bond cleavage is reduced by the inverse isotope effect arising
from protonation. This is just what is observed in the reaction
of the pNPP monoanion in solution, where proton transfer from
the phosphoryl group to the leaving group occurs during the
reaction, and in the reactions of protein-tyrosine phosphatases
where protonation of the leaving group is accomplished by a
1
8
buffer solution. A few measurements of O isotope effects on
the complexation of metal ions to pNPP have been reported.
Co(III) is capable of forming inert coordination complexes with
phosphate esters such as the Co(ethylenediamine)2-pNPP
complex which can be isolated and crystallized. Cleavage of
the phosphate oxygen-cobalt bond of this complex results in a
5
3
conserved Asp general acid. The maximum 18(V/K)bridge effect
kinetic isotope effect of 1.0135, which is close to the
equilibrium effect for deprotonation of a phosphate ester. The
equilibrium isotope effect for formation of the complex will be
the reciprocal of this value, or 0.9866. Formation of the more
labile complex between Co(cyclen)2 and pNPP results in a
seen in pNPP reactions in which the leaving group is not
protonated is around 1.03, the value observed in mutant PTPases
2
5-27
in which general acid catalysis has been eliminated.
The
equilibrium isotope effect for protonation of p-nitrophenol is
5
3
5
0
smaller equilibrium effect of 0.9920. Divalent metal ions
should exhibit weaker coordination and therefore yield smaller
isotope effects. A study of the effects of magnesium complex-
0
.985. Thus when P-O bond cleavage and proton transfer
18
are extensive and synchronous, the observed (V/K)bridge isotope
effect should be close to the product of these values, or about
1
8
ation with ATP put an upper limit on this O isotope effect of
1
.015. This is close to the value measured in reactions of pNPP
5
4
0
.999.
We sought to measure, or at least to estimate, the magnitude
with PTPases where protonation of the leaving group occurs in
the transition state.2
5-27
1
8
of the O isotope effect for complexation of the pNPP substrate
Mechanism and Transition-State Structure for the λPP
Reaction. The bell-shaped pH-rate profile for kcat indicates that
maximum activity depends on a species which must be
deprotonated and another which must be protonated. The acidic
limb most likely represents the nucleophilic metal-bound water
molecule. The species responsible for the basic limb results not
just in smaller values for kcat but also large increases in KM
when it is deprotonated (Table 3). The KM values for the H76N
mutant at all pH values are more consistent and are similar to
those seen in the native enzyme at high pH, suggesting that
protonation of this residue assists in substrate binding.
2
+
2+
to Ca . The log of the stability constant for the Ca -pNPP
complex has been reported to be 1.26 ( 0.04, as measured by
5
5
potentiometric pH titration. The determination by UV-vis
titration made in this study gave a value of 1.8, which is in
reasonable agreement given the less sensitive method of
measuring λmax, which changes by only 2 nm from 0 to
saturating metal ion. However the results of the 31P NMR
experiment indicate that higher-order complexes form as the
calcium ion concentration is increased beyond the levels which
give 1:1 complexation. The results from the isotope shift 31
NMR experiment (Figure 3) show that there is a very small
O isotope effect for the coordination of Ca . The data allow
P
If the substrate for catalysis were the monoanion of pNPP,
the rate should increase at low pH as the fraction of the substrate
present in solution as the monoanion increases. The pH-rate
profile suggests that the catalytically active form of the substrate
is the dianion. Due to problems with substrate inhibition below
pH 6, the profile could not be extended to low enough pH values
18
2+
only an estimation to be made for this isotope effect from a
comparison of the maximum separation of the signals of the
isotopically labeled ligands compared to the separation observed
37
in the analogous protonation experiment. The change in
separation reaches a maximum when the ligand is half proto-
nated (or complexed). The maximum change in the isotopic
51
where the pKa of the substrate, which is 4.96 in solution, would
be expected to reveal itself. The values for the 18(V/K)nonbridge
isotope effect give additional evidence that the dianion form of
pNPP is the substrate. At pH 6.0 about 10% of the pNPP
substrate will be present as the monoanion, whereas at pH 7.8
and 9.0 the concentration of the monoanion will be negligible.
37
separation found in the protonation experiment was 0.015 ppm.
At a calcium ion concentration which gives half complexation
(
52) Rusnak, F.; Yu, L.; Todorovic, S.; Mertz, P. Biochemistry 1999,
8, 6943-6952.
53) Rawlings, J.; Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc.
3
(
1
997, 119, 542-549.
(
50) Hengge, A. C.; Hess, R. A. J. Am. Chem. Soc. 1994, 116, 11256-
1263.
51) Bourne, N.; Williams, A. J. Org. Chem. 1984, 49, 1200-1204.
(54) Jones, J. P.; Weiss, P. M.; Cleland, W. W. Biochemistry 1991, 30,
1
3634-3639.
(
(55) Massoud, S. S.; Sigel, H. Inorg. Chem. 1988, 27, 1447-1453.

Mechanism of λ Protein Phosphatase
J. Am. Chem. Soc., Vol. 121, No. 27, 1999 6389
of pNPP by the UV-vis data (indicating the 1:1 complex), this
causing the loss of a hypothetical general acid. The 15(V/K)
isotope effect arises from contributions from a quinonoid
resonance structure of the nitrophenolate anion which involves
change is 0.0014 ppm, suggesting that this isotope effect is at
3
1
most about 10% as large as for protonation. In the P NMR
experiments complete saturation of the pNPP ligand could not
be achieved and so an exact measurement of the isotope effect
4
7
the nitro substituent. If the phenolic oxygen atom is in the
region of a cationic metal ion, then less charge will be
delocalized into the aromatic ring and this isotope effect will
be diminished. This explanation of the data also requires that
2+
for the higher-order Ca -pNPP complex could not be calcu-
lated, although the isotope effect clearly is greater than for the
1
8
1:1 complex. Using the maximum value observed for the change
coordination to the metal ion reduces the (V/K)bridge isotope
effect in the same manner as protonation suppresses the normal
isotope effect arising from P-O bond cleavage. There is no
experimental basis to allow an estimation of the magnitude of
in the isotopic separation gives an estimated value of 0.997,
which is probably an upper limit for the isotope effect for the
formation of a complex between pNPP and two calcium ions.
The value for the observed 18(V/K)nonbridge isotope effect in
the λPP reaction will be the product of the isotope effect for
binding and that for catalysis. The very small value of the
18
an O isotope effect for coordination of nitrophenolate ion to
a divalent metal ion. Fairly strong phenolate-metal complexes
are known in a number of enzymes, both with ligands (purple
2+
57
equilibrium isotope effect for coordination of pNPP to two Ca
acid phosphatase ), with substrate analogues (tyrosine hydroxy-
5
8
59
2+
ions inferred by the data indicates that binding effects are not
likely to be sufficiently large to mask the isotope effect for the
phosphoryl-transfer reaction. Thus the 18(V/K)nonbridge effect for
phosphoryl transfer is essentially unity or slightly inverse, not
a normal value that would be indicative of the tighter transition
states seen in diesters and triesters. This value is also very close
to that observed in the reactions of native PTPases (Table 2),
which indicates that the binuclear metal center does not induce
a change in the transferring phosphoryl group in the transition
state relative to its structure in solution reactions and reactions
of PTPases.
lase ), and with substrates (pyrocatechase, Mn -dependent
6
0
catechol dioxygenase ). Thus the possibility of a significant
isotope effect for such an interaction cannot be dismissed.
The pH-rate profile of the H76N mutant was also determined
in an attempt to ascertain the role of this residue. If His-76
functions as a general acid in catalysis, the basic limb of the
pH-rate profile for the H76N mutant should be lost. The data
for the native enzyme and the H76N mutant are shown in Table
3 and Figure 4. The basic limb which is present in the pH-rate
profile of the native enzyme is smaller in the H76N mutant,
but is not eliminated. Kinetic data were difficult to obtain due
to substrate inhibition at low pH, which required the use of
higher metal ion concentrations and increasing KM values for
substrate at higher pH. It is noteworthy that in the native enzyme
the KM increases at high pH to values which are very similar to
those found across the entire pH range in the H76N mutant.
This implies that protonation of His-76 electrostatically assists
in substrate binding (though actual proton transfer to the
substrate to form the monoanion is ruled out by the
The isotope effects in the leaving group for the reaction of
2+
2+
the native λPP with both Ca and Mn are similar to those
observed in reactions of the native PTPases, where P-O bond
cleavage is extensive in the transition state but the leaving group
is neutralized by protonation. The small 15(V/K) value of 1.0006
indicates that a small but measurable negative charge develops
on the leaving group in the transition state, suggesting that
charge neutralization of the leaving group does not quite keep
up with P-O bond cleavage. A similar value was previously
found for the native low molecular weight PTPase Stp-1 (Table
1
8
(V/K)nonbridge isotope effects as discussed earlier). Similar
electrostatic assistance in stabilization of the transition state is
a reasonable assumption. These facts, and the perturbation of
the metal center which results from the H76N mutation, make
it difficult to interpret the pH-rate dependency strictly for or
against a role for His-76 as a general acid. In sum, neither the
isotope effects nor the kinetic data allow a definitive answer to
this question. This residue may function more decisively as a
general acid in reactions of protein phosphate substrates, where
the leaving group is a more basic oxyanion of a serine or
threonine residue. If His-76 has a higher pKa than the p-
nitrophenolate leaving group, it may merely hydrogen bond to
the bridge oxygen atom rather than protonate it.
2
).
One of the unanswered questions for the Ser/Thr phosphatases
is whether the histidine residue which is conserved in the region
of the active site in Ser/Thr phosphatases functions as a general
acid. The precise roles of the metal ions are also not known,
and it is possible that one or both of them assists in stabilization
of the leaving group, analogous to the situation in alkaline
5
6
phosphatase. Spectroscopic studies have shown that mutation
of His 76 results in a perturbation of the ligand environment of
the binuclear metal center, possibly by disruption of a hydrogen
bond between the histidine and a metal-coordinated water
1
5
It is noteworthy that the H76N mutation has no effect on the
(V/K)nonbridge isotope effect. In all of the PTPases it has been
molecule. In the present study mutation of His-76 in λPP to
1
8
15
asparagine resulted in an increase in (V/K) from 1.0006 to
.0016 and an increase in 1 (V/K)bridge from 1.0133 to 1.0183
8
found that loss, or even partial interference with the neutraliza-
tion of the leaving group by the general acid, always resulted
1
(
Table 1). These increases are analogous to those observed in
1
8
1
5
18
in normal (V/K)nonbridge isotope effects. This was attributed to
more nucleophilic participation in the transition state, which
(V/K) and (V/K)bridge in reactions of PTPases that arise from
the mutation of their general acids (Table 2), but the increases
in the λPP case are smaller in magnitude. There are two possible
explanations for the smaller increase in these isotope effects.
One possibility is that the H76N λPP reaction proceeds with
an earlier transition state in which P-O bond cleavage is less
advanced. An alternative explanation is that charge neutralization
of the leaving group is accomplished by one of the metal ions
and that the increases in the 15(V/K) and (V/K)bridge isotope
effects in the H76N mutant are due to perturbation of the ligand
environment which interferes with but does not eliminate this
interaction in the transition state, rather than the mutation
can be rationalized by the need for more of a push to expel the
leaving group when charge neutralization is lost.2
5-27
In the
H76N mutant of the λPP the leaving group also departs bearing
more of a negative charge, but there is no change in the
1
8
(V/K)nonbridge isotope effect. This may be a result of the
(57) Vincent, J. B.; Olivier-Lilley, G. L.; Averill, B. A. Chem. ReV. 1990,
18
90, 1447-1467.
58) Michaud-Soret, I.; Andersson, K. K.; Que, L., Jr.; Haavik, J.
Biochemistry 1995, 34, 5504-5510.
59) Que, L., Jr.; Heistand 2d, R. H.; Mayer, R.; Roe, A. L. Biochemistry
1980, 19, 2588-2593.
60) Whiting, A. K.; Boldt, Y. R.; Hendrich, M. P.; Wackett, L. P.; Que,
L., Jr. Biochemistry 1996, 35, 160-170.
(
(
(
(56) Kim, E. E.; Wyckoff, H. W. J. Mol. Biol. 1991, 218, 449-464.

6390 J. Am. Chem. Soc., Vol. 121, No. 27, 1999
Hoff et al.
different nucleophile in the λPP reaction (an oxyanion versus a
thiolate in the PTPases) or may be due to other electrostatic
influences in the catalytic site.
a transition state with associative character and considerable
6
1,62
nucleophilic participation
in contrast to the reaction in
solution which is dissociative or loose in nature. The data
obtained in this study do not lend support to the notion that the
binuclear center in the λPP induces such a mechanistic change.
The isotope effects are most consistent with a reaction similar
to the reaction in solution and in the reactions of native PTPases,
which is noteworthy in light of the vastly different catalytic
machinery utilized by the two classes of phosphatases. The
In summary, the isotope effect data are most consistent with
a transition state for the λPP reaction which is very similar to
those of the PTPase reactions. In the reactions of the native
PTPases, the observed 15(V/K) and (V/K)bridge isotope effects
resulting from bond cleavage and charge development are
reduced by protonation of the leaving group. It is conceivable
that in the λPP reaction no such neutralization occurs and that
the observed 15(V/K) and (V/K)bridge isotope effects only
serendipitously resemble these isotope effects in the PTPase
reactions, and instead reflect an earlier transition state with bond
cleavage much less advanced in a tighter transition state, more
similar to reactions typical of diesters or triesters. However, if
no neutralization of the leaving group occurs in the λPP reaction,
it is difficult to explain why the 15(V/K) and (V/K)bridge isotope
effects increase in the mutant. In addition, phosphoryl-transfer
reactions with tighter transition states are characterized by
normal 1 (V/K)nonbridge isotope effects, while the small inverse
values for 18(V/K)nonbridge seen in this study are characteristic
of the loose transition states of monoester reactions. It may also
be conceivable that the 18(V/K)nonbridge isotope effects for λPP
are altered by binding to the binuclear metal center in such a
way that they also merely serendipitously resemble the isotope
effects seen with monoesters. However the solution results with
18
2+
2+
substitution of Ca for Mn in the binuclear metal center also
does not result in a change in the transition state for phosphoryl
transfer. The H76N mutation results in small changes in the
isotope effects, indicative of greater negative charge borne on
the leaving group in the transition state. Neither the isotope
effect data nor the pH-rate data clearly answer the question of
whether His-76 functions as a general acid in the enzymatic
reaction, and it is still uncertain whether this residue or the metal
ions participate in charge neutralization of the leaving group.
The kinetic data indicate that substrate binding is assisted when
His-76 is protonated.
18
18
8
Acknowledgment. Financial support of this work came from
NIH Grant GM47297 to A.C.H. and Grant GM46865 to F.R.
and from the US Army Advanced Civil Schooling Program for
support of R.H.H. R.H.H. and P.M. contributed equally to this
work.
2+
Ca and the fact that the H76N mutation, which is known to
perturb the metal center, alters the isotope effects in the leaving
group but does not change the 1 (V/K)nonbridge isotope effect,
suggest that this isotope effect is not significantly altered by
substrate binding. Thus the simplest explanation of the data is
that the transition state for the λPP reaction is similar to those
of the reactions of PTPases.
Supporting Information Available: Table of pNPP λmax
8
data as a function of calcium ion concentration shown graphi-
31
cally in Figure 2; table of the P chemical shift data shown in
31
Figure 3; and a sample P spectrum from the calcium ion
complexation experiment. This material is available free of
charge via the Internet at http://pubs.acs.org.
Conclusions
JA990667P
It has been suggested that the presence of positive charge in
phosphatases around the transferring phosphoryl group implies
(
61) Mildvan, A. S. Proteins 1997, 24, 401-416.
(62) Schlichting, I.; Reinstein, J. Biochemistry 1997, 36, 9290-9296.