Purine nucleoside phosphorylase-catalyzed, phosphate-independent hydrolysis of 2-amino-6-mercapto-7-methylpurine ribonucleoside

Article abstract of DOI:10.1006/bioo.1999.1137

In the presence of 1 mM phosphate, 2-amino-6-mercapto-7-methylpurine ribonucleoside (MESG) is a well-behaved substrate for calf spleen purine nucleoside phosphorylase (PNP). In the absence of phosphate, calf spleen PNP catalyzes a slow hydrolysis of MESG, which is accompanied by inactivation of the enzyme, analogous to the previously observed PNP-catalyzed hydrolysis of inosine and guanosine with formation, in the former case, of a stable PNP·hypoxanthine complex (P.C. Kline and V.L. Schramm (1992) Biochemistry 31, 5964-5973). Qualitative and semiquantitative features of calf spleen PNP- catalyzed hydrolysis of MESG are accounted for by the following model. First, in the absence of phosphate and at pH 7.4, the enzyme exists as an equilibrium mixture of monomer and trimer with a dissociation constant for the trimer of 3 x 10¹⁴ M². Second, a stoichiometric reaction between three molecules of MESG and the PNP trimer results in the formation of a stable PNP·purinethiol complex. Third, the PNP·purinethiol complex initially formed with the monomeric enzyme partitions between product release and formation of a stable complex with 55 turnovers per inactivation event. Fourth, the stable PNP·purinethiol complexes are rapidly dissociated by phosphate to regenerate active enzyme. This dissociation is accompanied by an increase in absorbance at 356 nm consistent with a pK(a) for the purinethiol base on the enzyme of 8.1, compared to a corresponding value of 8.8 in aqueous solution.

Full text of DOI:10.1006/bioo.1999.1137

Bioorganic Chemistry 27, 307–325 (1999)
Article ID bioo.1999.1137, available online at http://www.idealibrary.com on
Purine Nucleoside Phosphorylase-Catalyzed, Phosphate-
Independent Hydrolysis of 2-Amino-6-mercapto-
7-methylpurine Ribonucleoside
Jianming Cheng, Victor Farutin,1 Zhijun Wu,2 Gayatry Jacob-Mosier,3
Brad Riley, Ryan Hakimi, and Eugene H. Cordes4
College of Pharmacy and Department of Chemistry, University of Michigan,
Ann Arbor, Michigan 48109
Received September 18, 1998
In the presence of 1 mM phosphate, 2-amino-6-mercapto-7-methylpurine ribonucleoside
(MESG) is a well-behaved substrate for calf spleen purine nucleoside phosphorylase (PNP).
In the absence of phosphate, calf spleen PNP catalyzes a slow hydrolysis of MESG, which is
accompanied by inactivation of the enzyme, analogous to the previously observed PNP-catalyzed
hydrolysis of inosine and guanosine with formation, in the former case, of a stable PNPиhypoxan-
thine complex (P. C. Kline and V. L. Schramm (1992) Biochemistry 31, 5964–5973). Qualitative
and semiquantitative features of calf spleen PNP-catalyzed hydrolysis of MESG are accounted
for by the following model. First, in the absence of phosphate and at pH 7.4, the enzyme exists
as an equilibrium mixture of monomer and trimer with a dissociation constant for the trimer
of 3 ϫ 1014 M2. Second, a stoichiometric reaction between three molecules of MESG and
the PNP trimer results in the formation of a stable PNPиpurinethiol complex. Third, the
PNPиpurinethiol complex initially formed with the monomeric enzyme partitions between
product release and formation of a stable complex with 55 turnovers per inactivation event.
Fourth, the stable PNPиpurinethiol complexes are rapidly dissociated by phosphate to regenerate
active enzyme. This dissociation is accompanied by an increase in absorbance at 356 nm
consistent with a pKa for the purinethiol base on the enzyme of 8.1, compared to a corresponding
value of 8.8 in aqueous solution.
᭧ 1999 Academic Press
INTRODUCTION
Potent and specific inhibitors of purine-nucleoside phosphorylase (purine-nucleo-
side:orthophosphate ribosyltransferase, EC 2.4.2.1; PNP) have multiple plausible utilit-
ies in clinical medicine (2–4). These derive from the demonstration that inherited
deficiency of this enzyme is associated with profound immunodeficiency in T-cell
function (5,6). It follows that a clinical use might be found for such inhibitors for T-
cell lymphomas and leukemias, autoimmune diseases, psoriasis, and prevention of
organ graft rejection, among other possibilities.
1 Present address: Millennium Pharmaceuticals, Inc., 238 Main Street, Cambridge, MA 02142.
2 Present address: Helios Pharmaceuticals, Inc., Louisville, KY.
3 Present address: 617 Elmwood Road, Bay Village, OH 44140.
4 To whom correspondence should be addressed. Fax: (734) 936-9463. E-mail: cordeseh@umichedu.
307
0045-2068/99 $30.00
Copyright ᭧ 1999 by Academic Press
All rights of reproduction in any form reserved.

308
CHENG ET AL.
Discovery of inhibitors has been facilitated by knowledge of the three-dimensional
structure of human PNP (7) and a series of careful and detailed, if not always mutually
consistent, studies of kinetics and mechanism for this enzyme (1,8–17). The structure-
based search for useful inhibitors of PNP has resulted in discovery of several compound
classes (18–21), which was built on earlier studies (22–26). More recently, detailed
structures of calf spleen PNP and its complexes with a number of substrates, substrate
analogs, and inhibitors (27–29), as well as the structure of the Escherichia coli enzyme
(30), have clarified several aspects of mechanism and catalysis for PNP and provided
an expanded structural basis for inhibitor discovery.
The discovery of novel enzyme inhibitors having properties appropriate for clinical
use is of general interest to this laboratory. Specifically, in collaboration with BioCryst
Pharmaceuticals, Inc., we have initiated a program of research designed to discover
structurally novel, and clinically useful, second-generation inhibitors of human PNP.
The first phase of this work is the creation of a standard database of PNP inhibitor
data and the development of predictive QSAR models based upon it.
At the outset of this work, we considered that 2-amino-6-mercapto-7-methylpurine
ribonucleoside (MESG) might be a suitable substrate to employ in PNP inhibitor
discovery. It has been established that this molecule is a substrate for PNP and the
kinetics of its phosphorolysis (or hydrolysis) can be conveniently followed spectropho-
tometrically in the range 355–360 nm (31,32) (Scheme I).
SCHEME I
This substrate has been successfully employed to assay activities of enzymes in
several classes (31,33,34). The assay derives from the change in pKa for the conjugate
acid of the purinethiol moiety of the ribonucleoside, 6.5, and for the conjugate acid
of the free purinethiol, 8.8. Thus, at the value of pH employed in our studies, 7.4,
about 89% of MESG is present in solution as the zwitterionic conjugate base, ␭
max
330 nm, but about 96% of the purinethiol product is present as the neutral conjugate
acid, ␭ 355 nm (31).
max
In the course of this work, we discovered a phosphate-independent hydrolysis of
MESG catalyzed by calf spleen PNP and have established several features for the
kinetics of this reaction. The results of this work are presented herein.

PNP-CATALYZED HYDROLYSIS OF MESG
309
RESULTS
Nonenzymatic decomposition of MESG. MESG is unstable in aqueous solution
(31). In dilute aqueous acid, pH 2–3, we observed a half-life in the range of 20–30
days at room temperature. MESG is reasonably stable in dimethyl sulfoxide solution.
The structure of the decomposition products is not known although the reaction
presumably involves base-catalyzed attack on the cationic purine ring (35). The rate
of nonenzymatic decomposition of MESG under the conditions of our measurements,
pH 7.4 and 25ЊC, is sufficiently rapid to require correction of all kinetic measurements
reported herein.
Since our kinetic measurements were made under zero-order conditions, we deter-
mined the zero-order rate of decomposition of MESG as a function of MESG concen-
tration under our experimental conditions. Manually or employing the robotic work-
bench, zero-order rates for MESG decomposition were determined by following the
increase in absorbance at 314 nm or the decrease in absorbance at 356 nm as a
function of time over the concentration range 10–150 nM. Zero-order rates were
found to be linearly dependent on MESG concentration:
Rate314nm /initial absorbance330nm ϭ 2.4 Ϯ 0.2 ϫ 10Ϫ6 AU/s/AU,
Rate356nm /initial absorbance330nm ϭ Ϫ4.0 Ϯ 1.0 ϫ 10Ϫ6 AU/s/AU.
From time to time, control measurements were carried out to ensure that nonenzymatic
decomposition rates were consistent with these values.
First-order rate constants for MESG decomposition were also determined at 25ЊC
as a function of pH in the presence of 0.05 M tris-hydroxymethylaminomethane (Tris)
buffer: pH 7.0, kobs ϭ 3 ϫ 10Ϫ5 sϪ1; pH 7.4, kobs ϭ 4.3 ϫ 10Ϫ5 sϪ1; pH 8.0, kobs
ϭ
8 ϫ 10Ϫ5 sϪ1. At values of pH greater than 9, decomposition of MESG is rapid.
These observations are in general accord with those of Webb (31) and Broom and
Milne (36).
Basic kinetic behavior of calf spleen PNP-catalyzed phosphorolysis of MESG. At
pH 7.4 (0.1 M N-(2-hydroxyethyl)piperazine-NЈ-(2-ethanesulfonic acid), sodium salt;
Hepes), 25ЊC, and in the presence of 10 mM phosphate, zero-order rates for calf
spleen PNP-catalyzed phosphorolysis of MESG were found to increase linearly with
increasing PNP concentration over the concentration range 30–120 nM.
Zero-order rates for PNP-catalyzed phosphorolysis of MESG at pH 7.4 were deter-
mined as a function of MESG concentration (concentration range 35–280 ␮M) for
two ranges of phosphate concentration: (a) 0.012–0.1 mM and (b) 0.3–4 mM. In
each case, rates were determined at five concentrations of MESG at each of five
concentrations of phosphate. These data were employed to calculate values for the
apparent Km for MESG and phosphate as a function of the concentration of the fixed
varying substrate.
Values of apparent Km for MESG showed modest dependence on phosphate concen-
tration: The apparent Km is near 100 ␮M at 1 mM phosphate and near 400 ␮M at
very low phosphate concentrations. These values are consistent with one of 70 ␮M
determined by Webb (31) for a bacterial PNP at pH 7.6. Values of apparent Km for

310
CHENG ET AL.
phosphate are slightly dependent on the concentration of MESG but are markedly
dependent on the range of phosphate concentration employed for their determination.
At low phosphate and low MESG concentrations, the apparent Km for phosphate is
in the range 50–60 ␮M; at high phosphate and low MESG concentrations, the apparent
Km is in the range 700–800 ␮M. Webb (31) has reported an apparent Km for phosphate
of 25 ␮M for a bacterial PNP at pH 7.6 and Porter has found a value of 140 ␮M for the
calf spleen enzyme (12), both consistent with our value determined at low phosphate
concentrations. Ropp and Traut (11) found a higher value, 3.1 mM, for the Km for
phosphate for the calf spleen enzyme. Two apparent binding sites for phosphate on
PNP have been previously described, based both on structural data for the human
enzyme (7) and on kinetic data (11,12). However, a recent, detailed structure for the
calf spleen enzyme reveals no evidence for a second phosphate binding site (27).
Kinetics of calf spleen PNP-catalyzed hydrolysis of MESG in the absence of phos-
phate. At concentrations of phosphate below about 1 mM, calf spleen PNP catalyzes
the slow decomposition of MESG at a rate that is independent of phosphate concentra-
tion. This is not particularly surprising since PNP-catalyzed hydrolysis of inosine and
guanosine in the absence of phosphate has been previously observed by Kline and
Schramm (1,14). In addition, crystalline calf spleen PNP has been demonstrated to
catalyze the hydrolysis of inosine (27). It is thus highly likely that the phosphate-
independent reaction with MESG is also hydrolysis (a) by analogy with results of
Kline and Schramm (1,14) and Mao et al. (27) and (b) through the identification of
the purinethiol base derived from MESG as a reaction product based on an exact
mass determination employing FAB mass spectrometry. However, we have not identi-
fied ribose as the other reaction product. As is detailed below, the phosphate-indepen-
dent PNP-catalyzed decomposition of MESG terminates after a few catalytic turnovers,
generating very limited amounts of products and making detection of ribose problem-
atic. In what follows, we assume that the phosphate-independent decomposition reac-
tion is, in fact, hydrolysis.
Zero-order rates for PNP-catalyzed hydrolysis of MESG were determined as a
function of the concentration of substrate and enzyme. The ratio of zero-order rates
and substrate concentration is collected for five concentrations of MESG varying
from 1.07 to 6.03 ϫ 10Ϫ5 M and five concentrations of PNP varying from 0.06 to
1.60 ␮M (Table 1). Values of this ratio in each column of Table 1 are sensibly
constant, indicating that the rates are linear in substrate concentration. This finding
is consistent with the value of Km for MESG at low phosphate of 400 ␮M cited
above. In the last row of Table 1, average values for these ratios have been divided
by the concentration of PNP. Here too sensibly constant values are obtained, establish-
ing that the rate of this reaction is linear in enzyme concentration.
In the course of these studies, it was observed that the reactions terminated well
before all MESG substrate initially present was consumed. Data for the extent of
hydrolysis as a function of PNP concentration are collected in Table 2. The extent
of MESG hydrolysis increases with increasing enzyme concentration but less rapidly
than linearly. In other words, the number of catalytic events per enzyme monomer
decreases with increasing enzyme concentration (Table 2). Under the conditions of
these experiments, the hydrolytic reaction terminates following 8–45 catalytic events

PNP-CATALYZED HYDROLYSIS OF MESG
TABLE 1
311
The Ratio of Zero-Order Rates and Substrate Concentration for Calf Spleen PNP-Catalyzed
Decomposition of MESG at pH 7.4 and 25ЊCa
[PNP]b (␮M)
[MESG] ϫ 10Ϫ5M
0.06
0.18
0.30
0.78
1.60
1.07
2.15
3.81
2.72
7.42
7.86
13.6
14.8
36.2
31.3
77.1
65.1
2.68
3.95
7.95
12.5
30.4
60.4
3.75
3.28
8.21
13.1
29.3
61.9
6.03
4.20
9.02
13.7
30.5
59.8
Average
3.59
8.09
13.5
31.5
64.9
Average\[PNP]c
6 ϫ 102
4.5 ϫ 102
4.5 ϫ102
4.0 ϫ 102
4.1 ϫ 102
a Ratios have units of sϪ1 and have been multiplied by 105.
b As the monomer.
c V/[MESG]/[PNP]; values have units of sϪ1 MϪ1
.
per enzyme monomer. As is detailed below, reaction termination is not the consequence
of product inhibition.
In an effort to understand the central aspects of calf spleen PNP-catalyzed hydrolysis
of MESG in the absence of phosphate, four complex experiments were designed and
carried out.
The first experiment is characterized by an excess of enzyme over substrate.
Specifically, four reactions were carried out employing 4 ␮M MESG and 12 ␮M
PNP (as the monomer, 4 ␮M as the trimer). One reaction originally contained 1 mM
TABLE 2
Extent of Hydrolysis of MESG Catalyzed by PNP as a Function of the Concentration of Enzyme
at pH 7.4 and 25ЊCa
Percentage
MESG
MESG
consumed
(␮M)
initially
present
hydrolyzed
Moles MESG hydrolyzed
Moles PNP
Moles MESG hydrolyzed
Moles PNP
΂
΃ ΂
΃
[PNP]b (␮M)
obs
calc
150 ␮M MESG
0.06
0.18
0.30
0.78
1.60
2.70
5.70
6.86
9.12
13.5
1.8
3.8
4.6
6.1
9.0
45
32
23
12
8.4
45
29
23
14
9
60 ␮M MESG
0.78
1.60
7.84
12.3
13.1
20.5
10
7.7
14
9
a As determined by the change in absorbance at 361 nm at reaction termination.
b As the monomer.

312
CHENG ET AL.
phosphate; the other three contained no added phosphate. Plots of absorbance at 356
nm as a function of time for these reactions, as well as subsequent manipulations of
them, are depicted graphically in Fig. 1.
The reaction containing 1 mM phosphate (trace A, Fig. 1) proceeded rapidly to
completion (kobs Ͼ 0.1 sϪ1) with a net change in absorbance of 0.032 AU, consistent
with the absorbance calculated at this wavelength for a 4 ␮M solution of the purinethiol
product. Thus, all the MESG initially present was converted to products in the presence
of 1 mM phosphate and excess enzyme, as expected. The three reactions without
added phosphate (traces B; Fig. 1) proceeded more slowly (values of kobs are
0.0126 Ϯ 0.0006, 0.0117 Ϯ 0.0005, and 0.0129 Ϯ 0.0006 sϪ1 for the three runs) and
with a slightly smaller net change in absorbance of 0.027 AU.
At the 1100-s time point, sufficient concentrated phosphate buffer was added to
the first of these reactions to acheive a final phosphate concentration of 1 mM (trace
C; Fig. 1); a small increase in absorbance to 0.031 AU which was too fast to follow
by conventional means was observed. This value is not different from that observed
for the reaction initially containing phosphate (0.032 AU) when correction is made
for sample dilution and nonenzymatic decomposition of MESG. The phosphate-
induced increase in absorbance may reflect: (a) displacement of product purinethiol
from an enzymeиproduct complex by phosphate or (b) completion of an incomplete
reaction driven by phosphate. The latter possibility seems unlikely in view of the
excess of enzyme over substrate and the essential irreversibility of the hydrolytic
reaction. It is definitively eliminated by results of the following two experiments.
First, to the second of the three non-phosphate-containing reactions was added a
second aliquot of 12 ␮M PNP (trace D, Fig. 1). Except for sample dilution, no change
in absorbance was observed, consistent with the absence of unreacted MESG in the
reaction mixture. Subsequent addition of 4 ␮M MESG (trace E, Fig. 1) resulted in
a reaction which occurred with an increase in absorbance of 0.024 AU. Second, to
the last of the three non-phosphate reactions was added a second aliquot of 4 ␮M
MESG (trace F, Fig. 1). A subsequent reaction was observed with a first-order rate
constant of 0.0074 Ϯ 0.0006 sϪ1 and a net absorbance increase of 0.025 AU, similar
to that observed for the two reactions described above. Addition of 12 ␮M PNP to
this reaction resulted in no change in absorbance other than that consequent to sample
dilution (trace G, Fig. 1), confirming again that no unreacted MESG remained in the
reaction mixture. Finally, addition of sufficient concentrated phosphate to this reaction
mixture to yield 1 mM phosphate resulted in an increase in absorbance of 0.07 AU
(trace H, Fig. 1); roughly twice that observed for the reaction containing half as much
FIG. 1. Absorbance at 356 nm (left ordinate) and product concentrations (right ordinate) plotted as
a function of time for four reaction mixtures initially containing 4 ␮M MESG and 12 ␮M PNP, pH 7.4,
25ЊC. The first reaction contained 1 mM phosphate (trace A). Three other reactions initially contained
no phosphate (traces B). To the first of these reactions was added 1 mM phosphate at the indicated time
point (trace C). To the second of these three were added successively 12 ␮M PNP, 4 ␮M MESG, and 1
mM phosphate at the indicated time points (traces D, E, and H). To the last of these three reactions were
added successivly 4 ␮M MESG, 12 ␮M PNP, and 1 mM phosphate at the indicated time points (traces
F, G, and H).

PNP-CATALYZED HYDROLYSIS OF MESG
313

314
CHENG ET AL.
MESG (see traces B and C, Fig. 1). Thus, this result is also consistent with and supports
displacement by phosphate of product purinethiol from an enzymeиproduct complex.
The second experiment is closely related to the first and was designed to confirm
and expand conclusions based on that experiment. In this case, only two reactions
were run: the first contained 8 ␮M MESG, 0.9 ␮M PNP, and 1 mM phosphate. The
second contained 8 ␮M MESG, 12 ␮M PNP, but no added phosphate. For both
reactions, absorbance at 356 nm is plotted as a function of time in Fig. 2.
In the presence of phosphate, the absorbance at 356 nm increased rapidly, kobs
ϭ
0.0320 Ϯ 0.0006 sϪ1, with a net absorbance increase of 0.057 AU (trace A, Fig. 2).
This absorbance change is consistent with complete, or nearly complete, conversion
of MESG to phosphorolysis products. A second addition of 8 ␮M MESG to this
reaction mixture resulted in a second reaction having a rate and absorbance change,
0.055 AU, very similar to that of the first (trace C, Fig. 2), as did a third addition of
8 ␮M MESG (change in absorbance 0.055 AU) (trace E, Fig. 2) and a fourth such
addition (change in absorbance 0.053 AU) (trace G, Fig. 2). These results establish
that (a) the PNP remains fully active after repeated rounds of phosphorolysis of
MESG in the presence of 1 mM phosphate and (b) product inhibition is not marked
for this reaction (see below).
The behavior of the system is distinctly different in the absence of added phosphate.
The initial reaction is slower, kobs ϭ 0.0102 Ϯ 0.0004 sϪ1 (trace B, Fig. 2) (i.e., about
40 times less rapid than the corresponding reaction in the presence of 1 mM phosphate,
correcting for the difference in PNP concentration) and occurs with a slightly smaller
increase in absorbance (0.050 AU as opposed to 0.057 AU), in accordance with
results of the first experiment (vide supra). Thus, all MESG initially present was
converted to product in this reaction. A second addition of 8 ␮M MESG resulted in
a reaction (trace D, Fig. 2) characterized by a smaller absorbance increase, 0.026
AU. This increase reflects conversion of approximately 25% of the added MESG to
products. A third such addition resulted in an absorbance increase of only 0.018 AU
(trace F, Fig. 2), as did a fourth addition of MESG (trace H, Fig. 2). The last two
increases are those expected on the basis of the addition of MESG alone, without
conversion to product. Finally, sufficient phosphate buffer was added to the reaction
mixture to yield a final phosphate concentration of 1 mM. A rapid increase in
absorbance in 0.087 AU resulted (trace I, Fig. 2). This increase can be accounted for
by the sum of two processes: the displacement of bound purinethiol from the enzyme
by phosphate and the phosphorolysis of unreacted MESG remaining in solution.
Quantitative agreement requires correction for nonenzymatic decomposition of MESG.
Thus, results of these two experiments are consistent with a model in which, in
FIG. 2. Absorbance at 356 nm (left ordinate) and product concentrations (right ordinate) plotted as
a function of time for two reaction mixtures, pH 7.4, 25ЊC. The first reaction mixture initially contained
8 ␮M MESG, 0.9 ␮M PNP, and 1 mM phosphate (trace A). The second reaction mixture initially contained
only 8 ␮M MESG and 12 ␮M PNP (trace B). To the first reaction mixture were added successively three
aliquots of 8 ␮M MESG (traces C, E, and G) at the indicated time points. To the seond reaction mixture
were added successively three aliquots of 8 ␮M MESG and, finally, 1 mM phosphate (traces D, F, H,
and I) at the indicated time points.

PNP-CATALYZED HYDROLYSIS OF MESG
315

316
CHENG ET AL.
the absence of phosphate and in the presence of concentrated enzyme, there is a
stoichiometric reaction between MESG and the trimeric enzyme to yield an enzymati-
cally inactive PNPиpurinethiol complex which dissociates very slowly. At the PNP
concentrations employed in these experiments, we expect the enzyme to be present
predominantly (ca. 96%, see below) as the trimer. Addition of phosphate causes rapid
dissociation of the complex with a small increase in absorbance at 356 nm to release
free, active enzyme.
The third experiment is closely related in experimental design to the second except
that a more dilute solution of PNP, 0.9 ␮M, was employed with 5 ␮M MESG. At
this enzyme concentration, we expect a significant fraction of the enzyme to be present
as monomer (ca. 22%, see below), in contrast to the situation in the two experiments
just described. Absorbance at 356 nm as a function of time is plotted in Fig. 3 for
two reaction mixtures, one containing 1 mM phosphate and the other phosphate-free.
In the presence of 1 mM phosphate, the initial reaction plus a series of two
consecutive additions of 5 ␮M MESG resulted in three rapid increases in absorbance
at 356 nm of 0.035, 0.036, and 0.033 AU (traces A,C, and D, Fig. 3), in each
case consistent with complete phosphorolysis of MESG and absence of detectable
product inhibition.
In the absence of phosphate, we again saw slower increases in absorbance with
three consecutive additions of 5 ␮M MESG and these were accompanied by smaller
and decreasing net absorption increases: 0.029, 0.022, and 0.09 AU (traces B, E, and
F, Fig. 3). The initial increase in absorbance reflects a minimum conversion of 4.2
␮M MESG to products (and probably reflects 5.0 ␮M MESG converted to products).
Hence, each enzyme monomer turns over an average of four or five times prior to
cessation of the reaction. The second absorbance increase reflects hydrolysis of
approximately 66% (3.3 ␮M) of the second aliquot of MESG. The last value reflects
an absorbance change due to the addition of MESG alone, without conversion of any
of this material to products. Bringing the final reaction mixture to 1 mM phosphate
elicited an additional absorbance increase of 0.019 AU (trace G, Fig. 3), accounted
for by conversion of unreacted MESG to product as well as the displacement of
enzyme-bound purinethiol by phosphate.
The fourth and final experiment was similar in experimental design to the first but
employed an excess of substrate, 40 ␮M MESG, compared to enzyme, 0.9 ␮M PNP.
As in the first experiment, four reactions were run, one initially containing 1 mM
phosphate and three containing no added phosphate. Plots of absorbance at 356 nm
against time are collected in Fig. 4.
For the phosphate-containing reaction, absorbance increases rapidly with time,
kobs ϭ 0.0240 Ϯ 0.00025 sϪ1, with a net absorbance increase of 0.293 AU (trace A,
Fig. 4) consistent with that expected on the basis of complete conversion of MESG
to product purinethiol. The three reactions without added phosphate proceeded more
slowly (values of kobs are 0.0031 Ϯ 0.0001, 0.0030 Ϯ 0.0001, and 0.0031 Ϯ 0.0001
sϪ1) and with a net absorbance increase of 0.152 AU (traces B, Fig. 4), suggesting
conversion of about 23% of the MESG substrate to product. About 11 molecules of
MESG were processed to product for each monomer of PNP in these reactions. This
value is consistent with that expected at this enzyme concentration on the basis of
the data collected in Table 2.

PNP-CATALYZED HYDROLYSIS OF MESG
317
FIG. 3. Absorbance at 356 nm (left ordinate) and product concentrations (right ordinate) plotted as
a function of time for two reaction mixtures, each initially containing 5 ␮M MESG and 0.9 ␮M PNP,
pH 7.4, 25ЊC. The first reaction mixture also contained 1 mM phosphate (trace A). At the indicated time
points, two successive additions of 5 ␮M MESG were added (traces C and D). To the second reaction,
initially containing no phosphate (trace B), were added successively two aliquots of 5 ␮M MESG and,
finally, 1 mM phosphate (traces E, F, and G).

318
CHENG ET AL.

PNP-CATALYZED HYDROLYSIS OF MESG
319
The first of these three reactions was brought to 1 mM phosphate, causing an
absorbance increase to 0.269 AU (trace C, Fig. 4) consistent with that observed for the
original phosphate-containing reaction, when corrected for dilution and nonenzymatic
decomposition of MESG. To the second of these reactions was added an additional
40 ␮M MESG; the absorbance increased immediately by 0.098 AU, consistent with
the value expected on the basis of MESG alone, and then decreased slowly, reflecting
nonenzymatic decomposition of unreacted MESG (trace D, Fig. 4). These results
establish that the enzyme present in this reaction mixture was inactive. Subsequent
addition of 0.9 ␮M PNP resulted in a very modest subsequent reaction (trace F, Fig.
4) and, finally, 1 mM phosphate, which caused unreacted MESG to be converted to
product with a kobs of 0.019 Ϯ 0.007 sϪ1 and the enzymeиproduct complex to be
dissociated (trace H, Fig. 4). To the third of these reactions was added an additional
0.9 ␮M enzyme (trace E, Fig. 4). An immediate drop in absorbance due to dilution
was followed by a slow first-order reaction, kobs ϭ 1.9 ϫ 10Ϫ3 sϪ1, with a net increase
of absorbance of 0.02 AU, indicating that the added enzyme consumed a small fraction
of the remaining unreacted MESG. Subsequent addition of 40 ␮M MESG increased
absorbance by an amount accounted for by MESG itself (trace G, Fig. 4). Finally,
addition of 1 mM phosphate converted unreacted MESG to products with a kobs of
0.027 Ϯ 0.0004 sϪ1 (trace I, Fig. 4).
Search for product inhibition. Inhibition of the phosphate-independent PNP-cata-
lyzed hydrolysis of MESG by the products of the reaction—D-ribose, purinethiol, or
both—is a possible explanation, in part at least, for the observations reported above.
Rates for the reaction at pH 7.4 were independent of purinethiol concentration over
the concentration range 0 to 570 ␮M and increased about 70% with increasing
ribose concentration over the same concentration range. A combination of 1 mM
D
-
-
D
ribose plus 1 mM purinethiol elicited a two- to threefold increase in rate compared
to that in the absence of these agents. It is clear that the anticipated products of the
phosphate-independent reaction do not inhibit the reaction under the conditions re-
ported here and, at the concentrations present in the reaction mixtures, have no effect
on the enzymatic reaction rate.
Ribose 1-phosphate (1 mM) and a combination of 1 mM ribose 1-phosphate plus
1 mM purinethiol decreased the rate of the phosphate-independent reaction two- to
threefold. This presumably reflects modest inhibition by ribose 1-phosphate which
is not potentiated by the purinethiol base. The presence of very small amounts of
ribose 1-phosphate among reaction products would have no effect on the enzymatic
reaction rates under the conditions employed in these studies.
FIG. 4. Absorbance at 356 nm (left ordinate) and product concentrations (right ordinate) plotted as
a function of time for four reaction mixtures initially containing 40 ␮M MESG and 0.9 ␮M PNP, pH
7.4, 25ЊC. The first of these reactions also contained 1 mM phosphate (trace A). The other three were
initially phosphate free (traces B). To the first of the these three reactions was added 1 mM phosphate
at the indicated time point (trace C). To the second of the three reactions were added 40 ␮M MESG, 0.9
␮M PNP, and 1 mM phosphate at the indicated time points (traces D, F, and H). To the last of these
three reactions were added 0.9 ␮M PNP, 40 ␮M MESG, and 1 mM phosphate at the indicated time
points (traces E, G, and I).

320
CHENG ET AL.
Finally, to search for possible rate effects elicited by the nonenzymatic degradation
products of MESG, a concentrated sample of MESG was permitted to degrade com-
pletely at pH 7.4 and 25ЊC in the absence of enzyme. Addition of an aliquot of the
degradation product to produce a concentration higher than could have been achieved
in our reaction mixtures caused no change in the enzymatic reaction rate.
DISCUSSION
In confirmation of the results of Webb (31) and others (33,34) we have found
MESG to be a reasonably well-behaved substrate for calf spleen PNP. Its optical
properties are well suited to continuous spectrophotometric assays, manual and robotic.
However, this substrate has one central drawback for careful kinetic work: instability
in aqueous solution, particularly under mildly basic conditions. Perhaps a related
substrate, 7-methylinosine, will prove to be an improved option (37). The instability
of MESG is not a major drawback for rapid reactions but necessitates substantial
corrections to observed rates for several slow reactions in the absence of phosphate
reported herein.
We have not attempted a complete, quantitative treatment of the kinetic data reported
here. In addition to the limitations on accuracy of measured rates imposed by the
nonenzymatic decomposition of MESG, as noted above, any such treatement would
be complicated for a number of reasons: (i) PNP undergoes a facile concentration-
dependent and phosphate-dependent equilibration between monomeric and trimeric
forms (11), the monomer having approximately 50-fold higher specific activity than
the trimer. (ii) The underlying kinetic mechanism for calf spleen PNP is not clear.
Several early kinetic studies suggested an ordered addition of substrates and release
of products (7, 10); a subsequent study suggested a random order of addition of
substrates for the synthetic reaction but an ordered addition of substrates for the
phosphorolysis reaction with phosphate binding first (12). However, more recent
data, including the present work, establish that PNP catalyzes hydrolysis of purine
nucleosides in the absence of phosphate, indicating that these substrates can bind in
the absence of phosphate (1,14,15). This conclusion is supported by detailed structural
studies on the calf spleen enzyme, strongly suggesting that nucleoside and phosphate
add to the enzyme in a random manner but that purine base must add to the enzyme
before ribose 1-phosphate in the synthetic direction (27), exactly the opposite of an
earlier proposal (12). (iii) Release of guanine from a PNPиguanine complex may be
largely or completely rate determining for phosphorolysis of guanosine (12) as is
release of hypoxanthine from a PNPиhypoxanthine complex for the arsenolysis of
inosine (13). Thus, rapid-equilibrium assumptions for treatment of kinetic data for
PNP-catalyzed reactions are not likely to be justified. (iv) Structural data for the
human enzyme (7), but not for the calf spleen enzyme (27), and kinetic data (11,12)
suggest the existence of a second binding site for phosphate on PNP; occupancy of that
site by phosphate influences the rate of the phosphorolysis reaction. The dependence of
Km for phosphate on phosphate concentration (see above) also suggests two phosphate
binding sites for the calf spleen enzyme. (v) Finally, for the human erythrocytic
enzyme at least, a second binding site for nucleoside may be present (9).
Our data for phosphate-independent, PNP-catalyzed hydrolysis of MESG can be

PNP-CATALYZED HYDROLYSIS OF MESG
321
accounted for by a model based, in part, on previous observations of Kline and
Schramm (1,14) (see below).
Our model is defined by the following: (i) equilibration between enzyme monomer
and trimer with a dissociation constant Kdiss ϭ (monomer)3/(trimer) ϭ 3 ϫ 10Ϫ14 M2
(in the absence of substrate) [this value is considerably larger than that calculated
earlier (11)]; (ii) a stoichiometric reaction between the enzyme trimer and three
molecules of MESG (all-of-the-sites reactivity) to yield an enzymatically inactive
enzymeиproduct complex; (iii) partitioning of an enzyme monomerиpurinethiol com-
plex between product release and formation of a tight, inactive enzymeиproduct
complex, with 55 turnovers to one inactivation event; (iv) dissociation of the tight
enzymeиpurinethiol complex by added phosphate; and (v) a monomer–trimer equili-
bration rate which is slow compared to rates for catalytic events. This model is
summarized graphically in Fig. 5.
We emphasize that this model is very likely simpler than the actual molecular
events involved. At the same time, it is in general accord with what is known about
PNP and a model of this type appears to be the simplest that accounts satisfactorily
for our data. The essential features of the model are likely to be correct. More complex
models can certainly be constructed which will also accord with the experimental
findings, but relating these to molecular events at our current level of understanding
seems fruitless.
Using this model, we have calculated the ratio of moles of MESG hydrolyzed to
moles of enzyme monomer present and include the results in Table 2. The calculated
values accord with the experimental values within the error of the measurements.
This model also accounts satisfactorily for experimental observations of the four
complex experiments summarized in Figs. 1–4.
In the first two experiments (Figs. 1 and 2), a high PNP concentration, 12 ␮M,
was employed. Using the equilibrium constant for dissociation of trimer into monomer,
FIG. 5. Basic model for the calf spleen PNP-catalyzed hydrolysis of MESG.

322
CHENG ET AL.
K
diss ϭ 3 ϫ 10Ϫ14 M2, a simple calculation reveals that about 96% of PNP monomer
units are present as trimer at this concentration. In the first experiment, Fig. 1, the
initial reaction is predicted to consume all MESG present with formation of 4 ␮M
PNPиpurinethiol complex (as the trimer), which is dissociated by phosphate, in accord
with results. Since 8 ␮M PNP, largely present as trimer, will remain active, the
subsequent addition of 4 ␮M MESG is predicted to again result in complete consump-
tion of substrate with formation of an additional 4 ␮M PNPиpurinethiol complex,
again in accord with observations.
Similarly, in the second experiment, Fig. 2, the initial reaction between 12 ␮M
PNP and 8 ␮M MESG is predicted to consume all the MESG, as observed, with
formation of 8 ␮M PNPиpurinethiol complex (as the trimer), leaving 4 ␮M active
trimeric PNP. The subsequent addition of 8 ␮M MESG is predicted to result in partial
(ca. 50%) consumption of the substrate (we actually observed about 25% MESG
consumption) with inactivation of the enzyme, as found. The qualitative and semiquan-
titative agreement is satisfactory given the complexity of the system and the experimen-
tal uncertainties.
In the third and fourth experiments, Figs. 3 and 4, a more dilute PNP solution, 0.9
␮M, was employed. Our model predicts that 0.2 ␮M PNP will be present initially
as monomer, the remainder being present as trimer. The third experiment, Fig. 3,
employed 5 ␮M MESG. Our model predicts that 0.7 ␮M MESG will be consumed
by the trimer with formation of inactive PNPиpurinethiol complex and the remainder of
the MESG will be consumed by monomer, with formation of 0.08 ␮M PNPиpurinethiol
complex, leaving 0.12 ␮M PNP active. This accords with the observation of complete
consumption of all MESG initially present. The 0.12 ␮M active PNP is calculated
to include 0.08 ␮M PNP present as the monomer. Since 0.08 ␮M ϫ 55 (catalytic
events per monomer) ϭ 4.4 ␮M, the model predicts that most, but not quite all, of
a second addition of MESG should be consumed, leaving all the PNP initially present
in the form of the inactive PNPиpurinethiol complex. These predictions again accord
with experimental observations.
The final experiment, Fig. 4, employed a high MESG concentration, 40 ␮M. The
0.2 ␮M PNP calculated to be present as the monomer (see above) is expected to
consume 0.2 ϫ 55 ϭ 11 ␮M MESG, about 28% of the substrate initially present,
prior to complete enzyme inactivation. This is in good agreement with the observation
of 23% substrate consumption, with enzyme inactivation. However, the addition of
0.9 ␮M PNP to solutions containing unreacted MESG (traces E and F, Fig. 4) generates
less vigorous reactions than anticipated on the basis of the model.
It remains to explain the small absorbance increases at 356 nm observed when
PNPиpurinethiol complexes are dissociated by addition of phosphate (see Figs. 1–4).
In aqueous solution, the pKa of the purinethiol is 8.8 (12) so that almost all of this
species is present as the conjugate acid at pH 7.4. A reduction in the effective pKa
for the purinethiol when complexed to PNP would reduce the fraction present as the
conjugate acid and, hence, the absorbance at 356 nm, the wavelength of maximum
absorption of the conjugate acid. The anionic form of the purinethiol has a very small
extinction coefficient at 356 nm compared to the neutral form. A simple calculation
suggests that the effective pKa of the purinethiol in the PNPиpurinethiol complex is
near 8.1.

PNP-CATALYZED HYDROLYSIS OF MESG
323
CONCLUSIONS
Kline and Schramm (1,14) have previously demonstrated that, at high concentrations
in which PNP is present largely as the trimer, formation of hypoxanthine and ribose
at the first site causes PNP to adopt a conformation in which the hypoxanthine is
bound with a dissociation constant near 10Ϫ12 M and the adjacent sites are inert. In
contrast, hypoxanthine binds to the empty enzyme about 106-fold less strongly and
binding is uniform at all three sites. The case described herein is related to that of
Kline and Schramm but with significant differences. The purinethiol product is bound
less tightly to the monomeric enzyme, with a dissociation constant of perhaps 10Ϫ8
M, than is hypoxanthine, thus permitting some turnover at all three sites. The pu-
rinethiol also binds weakly to the empty enzyme, evidenced by our ability to detect
significant product formation. In both cases, phosphate causes rapid release of bound
purine from the tight complexes.
Although the behavior of PNP-catalyzed MESG hydrolysis is complex, these obser-
vations do not compromise our ability to employ this substrate, in the presence of
phosphate, for determination of kinetic parameters for PNP inhibitors (results to
be published).
EXPERIMENTAL
Materials
Synthesis of MESG. Synthesis of MESG was first reported by Broom and Milne
(36). An alternative synthesis has been reported by Webb (12). In our hands, and
those of others (37), these syntheses proved problematical, initially yielding substantial
amounts of MESG, much of which was subsequently lost in workup during exposure
to protic solvents. We elected to develop a synthesis of the hydroiodide salt of MESG
by a process which avoids use of protic solvents.
2-Amino-6-chloropurine ribonucleoside (234 mg, 0.77 mmol) was dissolved in dry
dimethylformamide (DMF) and sodium carbonate (200 mg) was suspended in this
solution. Methyl iodide (1.5 mL) was added. The mixture was stirred at room tempera-
ture for 13 h, at which time the reaction was judged to be complete by absorbance
at 321 nm. Excess methyl iodide was removed under reduced pressure. Thiourea (234
mg, 3 mmol) was added and the mixture was stirred at room temperature for 30 min.
The suspended sodium carbonate was removed by filtration, acetone (10 mL) was
added, the resulting precipitate was removed by filtration. The filtrate was taken to
dryness under reduced pressure. Acetone was added to the residue and the insoluble
material collected by centrifugation, washed once with acetone, and dried under
reduced pressure. This material was redissolved in 0.50 mL of dry DMF, 5 mL of
acetone was added, insoluble material was removed by filtration, and the filtrate was
taken to dryness under reduced pressure. The residue was again treated with acetone;
the insoluble material was collected by centrifugation, washed once with acetone,
and dried under reduced pressure to yield MESG hydroiodide as a yellowish powder:
209 mg, 61% yield, mp Ͼ300ЊC, pKa 6.5. UV max pH 6, 348 nm (␧ ϭ 11 ϫ 103
M cmϪ1); pH 7.4, 333 nm (␧ ϭ 11 ϫ 103 M cmϪ1). Anal Calcd for C11H16N5O4S-
1
I(DMF)0.6: C, 31.69; H, 4.20; N, 16.17. Found: C, 31.60; H, 4.26; N, 16.36. H
NMR (DMSO-d6): 3.59 (1H, dd, J ϭ 16 and 3 Hz), 3.74 (1H, dd, J ϭ 16 and 3 Hz),

324
CHENG ET AL.
4.00 (1H, m), 4.19 (3H, s), 4.37 (1H, t, J ϭ 4.2 Hz), 5.82 (1H, d, J ϭ 4.2 Hz), 6.24
(2H, b), 7.20(4H, b), 9.62 (1H, s). MS (FAB): Calcd for C11H16N5O4S 314.0923.
Found: 314.0921.
Synthesis of 2-amino-7-methyl-6-purinethiol. A modification of the method of Pra-
sad and Robbins (38) was used. 7-Methylguanine (1 g, 6.5 mmol) and phosphorous
pentasulfide (7.0 g) were refluxed in pyridine (100 mL) for 10 h. The pyridine was
removed by distillation under reduced pressure and water (100 mL) was added to the
residue; this was refluxed for 6 h, yielding a yellowish solution, which formed a
white precipitate on cooling in an ice bath. The precipitate was collected by filtration
and washed with MQ water. The filter cake readily dissolved in 1% potassium
hydroxide (70 mL, room temperature). The solution was filtered and neutralized with
glacial acetic acid and was maintained in an ice bath for 30 min. The fine white
precipitate which formed was collected by filtration, dissolved in 5% potassium
hydroxide (12 mL), and neutralized with glacial acetic acid. The precipitate was again
collected by filtration, washed with water, and dried overnight under vacuum to
yield 56 mg of product, 57% yield (lit. (38) 46%). MS (FAB): Calcd for C6H7N5S
181.0422. Found:181.0419.
Calf spleen PNP (phosphate free) was the best grade available from Sigma (purity
Ͼ98%) and was used without further purification. Unless noted otherwise, concentra-
tions of PNP reported are in terms of monomeric units present. Other reagents were
obtained commercially from either Sigma or Aldrich and were the highest grade avail-
able.
Kinetic Measurements
Kinetic measurements were carried out spectrophotometrically with the aid of an
HP 8453 diode array spectrophotometer. Most of the measurements were made manu-
ally but several were made with the aid of an HP ORCA laboratory robot equipped
with an HP 8453 spectrophotometer. Concordant results were obtained. All measure-
ments were made at 25ЊC pH 7.4 (maintained with 0.1 M Hepes buffer). Reactions
were typically followed at 356 nm. Kinetic parameters were estimated from the
collected data employing the Leonora statistics package. Values of pH were determined
with a Radiometer PHM240 pH meter. Unless otherwise noted, all enzymatic reactions
were initiated with addition of enzyme.
ACKNOWLEDGMENT
We appreciate the insightful comments of Professor Vernon Schramm on an early draft of the manuscript.
This work has benefited from his advice.
REFERENCES
1. Kline, P. C., and Schramm, V. L. (1992) Biochemistry 31, 5964–5973.
2. Sircar, J. C., and Gilbertsen, R. B. (1988) Drugs Future 13, 653–688.
3. Montgomery, J. A. (1994) Exp. Opin. Invest. Drugs 3, 1303–1313.
4. Stockler, J. D. (1984) in Developments in Cancer Chemotherapy (Glazer, R. I., Ed., pp. 35–60, CRC
Press, Boca Raton, FL.
5. Markert, M. L. (1991) Immunodefic. Rev. 3, 45–81.
6. Gilblett, E. R., Ammann, A. J., Wara, D. W., Sandman, R., and Diamond, L. K. (1975) Lancet
i, 1010–1013.

PNP-CATALYZED HYDROLYSIS OF MESG
325
7. Ealick, S. E., Rule, S. A., Carter, D. C., Greenhough, T. J., Babu, Y. S., Cook, W. J., Habash, J.,
Helliwell, J. R., Stoeckler, J. D, Parks, R. E., Jr., Chen, S-f., and Bugg, C. E. (1990) J. Biol. Chem.
265, 1812–1820.
8. Kim, B. K., Cha, S., and Parks, R. E., Jr. (1968) J. Biol. Chem. 243, 1771–1776.
9. Agarwal, R. P., and Parks, R. E., Jr. (1969) J. Biol. Chem. 244, 644–647.
10. Lewis, A. S., and Glantz, M. D. (1976) Biochemistry 15, 4451–4457.
11. Ropp, P. A., and Traut, T. W. (1991) J. Biol. Chem. 266, 7682–7687.
12. Porter, D. J. T. (1992) J. Biol. Chem. 267, 7342–7351.
13. Kline, P. C., and Schramm, V. L. (1993) Biochemistry 32, 13212–13219.
14. Kline, P. C., and Schramm, V. L. (1995) Biochemistry 34, 1153–1162.
15. Erion, M. D., Stoeckler, J. D., Guida, W. C., Walter, R., and Ealick, S. E. (1997) Biochemistry
36, 11735–11748.
16. Erion, M. D., Takabayashi, K., Smith, H. B., Kessi, J., Wagner, S., Honger, S., Shames, S. L., and
Ealick, S. E. (1997) Biochemistry 36, 11725–11734.
17. Stoeckler, J. D., Poirot, A. F., Smith, R. M., Parks, R. E., Jr., Ealick, S. E., Takabayashi, K., and
Erion, M. D. (1997) Biochemistry 36, 11749–11756.
18. Montgomery, J. A., Niwas, S., Rose, J. D., Secrist, J. A., III, Babu, Y. S., Bugg, C. E., Erion, M.
D., Guida, W. C., and Ealick, S. E. (1993) J. Med. Chem. 36, 55–69.
19. Secrist, J. A., III, Niwas, S., Rose, J. D., Babu, Y. S., Bugg, C. E., Erion, M. D., Guida, W. C.,
Ealick, S. E., and Montgomery, J. A. (1993) J. Med. Chem. 36, 1847–1854.
20. Erion, M. D., Niwas, S., Rose, J. D., Anathan, S., Allen, M., Secrist, J. A., III, Babu, Y. S., Bugg,
C. E., Guida, W. C., Ealick, S. E., and Montgomery, J. A. (1993) J. Med. Chem. 36, 3771–3783.
21. Guida, W. C., Elliott, R. D., Thomas, H. J., Secrist, J. A., III, Babu, Y. S., Buff, C. E., Erion, M.
D., Ealick, S. E., and Montgomery, J. A. (1994) J. Med. Chem. 37, 1109–1114.
22. Stoeckler, J. D., Cambor, C., Kuhns, V., Chu, S.-H., and Parks, R. E., Jr. (1982) Biochem. Pharmacol.
31, 163–171.
23. Stoeckler, J. D., Ryden, J. B., Parks, R. E., Jr., Chu, M.-Y., Lim, M.-L., Ren W.-Y., and Klein, R.
S. (1986) Cancer Res. 46, 1774–1778.
24. Halazy, S., Ehrhard, A., and Danzin, C. (1991) J. Am. Chem. Soc. 113, 315–317.
25. Sircar, J. C., Suto, M. J., Scott, M. E., Dong, M. K., and Gilbertsen, R. B. (1986) J. Med. Chem.
29, 1804–1806.
26. Berry, D. A., Gilbertsen, R. B., and Cook, P. D. (1986) J. Med. Chem. 29, 2034–2037.
27. Mao, C., Cook, W. J., Zhou, M., Federov, A. A., Almo, S. C., and Ealick, S. E. (1998) Biochemistry
37, 7135–7146.
28. Bzowska, A., Luic, M., Schroder, W., Shugar, D., Saenger, W., and Koellner, G. (1995) FEBS Lett.
367, 214–218.
29. Koellner, G., Luic, M., Shugar, D., Saenger, W., and Bzowska, A. (1997) J. Mol. Biol. 265, 202–216.
30. Mao, C., Cook, W. J., Zhou, M., Koszalka, G. W., Krenitsky, T. A., and Ealick, S. E. (1997) Structure
5, 1373–1383.
31. Webb, M. R. (1992) Proc. Natl. Acad. Sci. USA 89, 4884–4887.
32. Kulikowska, E., Bzowska, A., and Wierzchowski, J., and Shugar, D. (1986) Biochim. Biophys. Acta
874, 355–363.
33. Wedler, F. C., Ley, B. W., and Moyer, M. L. (1994) Anal. Biochem. 218, 449–453.
34. Chen, Q., Wang, Z., and Killilea, S. D. (1995) Anal. Biochem. 226, 68–73.
35. Townsend, L. B., and Robins, R. K. (1963) J. Am. Chem. Soc. 85, 242–243.
36. Broom, A. D., and Milne, G. H. (1975) J. Heterocyclic Chem. 12, 171–174.
37. Rieger, C. E., Lee, J., and Turnbull, J. L. (1997) Anal. Biochem. 246, 86–95.
38. Prasad, R. N., and Robins, P. (1957) J. Am. Chem. Soc. 79, 6401–6407.