Inverse acyl phosph(on)ates: Substrates or inhibitors of β-lactam-recognizing enzymes?

Article abstract of DOI:10.1006/bioo.2001.1218

Acyl phosph(on)ates represent a new class of inhibitors of β-lactam - recognizing enzymes. Previously described members of this class were aroyl phosph(on)ates. These compounds have been shown to acylate and/or phosphylate the active site serine residue, leading to either transient or essentially irreversible inhibition [Li, N., and Pratt, R. F. (1998) J. Am. Chem. Soc. 120, 4264-4268]. The present paper describes the synthesis and evaluation as inhibitors of an inverse pair of acyl phosph(on)ates that incorporate the amido side chain that represents a major substrate specificity determinant of these enzymes. Thus, N-(phenylacetyl)glycyl phenyl phosphate and benzoyl N-(benzyloxycarbonyl)aminomethyl phosphonate were prepared. The former of these compounds was found to be a substrate of typical class A and C β-lactamases and of the DD-peptidase of Streptomyces R61; it thus acylates the active site serine. In contrast, the latter compound was an irreversible inhibitor of the above enzymes, probably by phosphonylation of the active site serine. With each of these enzymes therefore, the amido side chain rather than the acyl group dictates the orientation of the bound phosph(on)ate and thus the mode of reaction.

Full text of DOI:10.1006/bioo.2001.1218

Bioorganic Chemistry 29, 271–281 (2001)
doi:10.1006/bioo.2001.1218, available online at http://www.idealibrary.com on
Inverse Acyl Phosph(on)ates: Substrates or Inhibitors of
␤-Lactam-Recognizing Enzymes?
Michael J. Morrison, Naixin Li, and R. F. Pratt1
Department of Chemistry, Wesleyan University, Middletown, Connecticut 06459
Received February 13, 2001
Acyl phosph(on)ates represent a new class of inhibitors of ␤-lactam—recognizing enzymes.
Previously described members of this class were aroyl phosph(on)ates. These compounds have
been shown to acylate and/or phosphylate the active site serine residue, leading to either
transient or essentially irreversible inhibition [Li, N., and Pratt, R. F. (1998) J. Am. Chem. Soc.
120, 4264–4268]. The present paper describes the synthesis and evaluation as inhibitors of an
inverse pair of acyl phosph(on)ates that incorporate the amido side chain that represents a
major substrate specificity determinant of these enzymes. Thus, N-(phenylacetyl)glycyl phenyl
phosphate and benzoyl N-(benzyloxycarbonyl)aminomethyl phosphonate were prepared. The
former of these compounds was found to be a substrate of typical class A and C ␤-lactamases
and of the DD-peptidase of Streptomyces R61; it thus acylates the active site serine. In
contrast, the latter compound was an irreversible inhibitor of the above enzymes, probably by
phosphonylation of the active site serine. With each of these enzymes therefore, the amido
side chain rather than the acyl group dictates the orientation of the bound phosph(on)ate and
thus the mode of reaction.
᭧ 2001 Academic Press
Key Words: ␤-lactamase; DD-peptidase; acyl phosph(on)ate; enzyme inhibition; enzyme
substrates.
INTRODUCTION
Acyl phosph(on)ates 1 have been shown in this laboratory to be novel substrates/
covalent inhibitors of ␤-lactamases of classes A and C (1). The nucleophilic hydroxyl
group of the active site serine can, in principle, attack either the carbonyl or the
phosphyl center. The former yields an acyl-enzyme and the latter a phosphyl-enzyme
species. Precedent would suggest that the enzyme could catalyze hydrolysis of the
1 To whom correspondence should be addressed. Fax: (860)-685-2211. E-mail: rpratt@wesleyan.edu.
271
0045-2068/01 $35.00
Copyright ᭧ 2001 by Academic Press
All rights of reproduction in any form reserved.

272
MORRISON ET AL.
acyl-enzyme to regenerate free enzyme but is less likely to be able to do this to a
phosphyl-enzyme (2,3). Acting in the former mode, 1 would appear to be a substrate,
while in the latter, it would be seen as an irreversible covalent inhibitor. Depending
on the choice of R/Ar, deacylation of the acyl-enzyme may be slow, and thus transiently
inhibited enzymes can be generated by the acyl transfer route (1). To date, our work
with these compounds has involved only simple alky/aryl substituents on 1 (1).
In the present paper, we describe the synthesis and evaluation of compounds 2 and
3 as substrates/inhibitors of ␤-lactamases and a bacterial DD-peptidase. The latter is
a representative of a group of enzymes that catalyze the hydrolysis and transpeptidation
of D-Ala-D-Ala peptides, essential reactions in bacterial cell wall biosynthesis. These
enzymes are also the targets of the ␤-lactam antibiotics. The ␤-lactamases, which
catalyze the hydrolysis of ␤-lactams (Scheme 1), are a major source of bacterial
resistance to ␤-lactams. In these new phosph(on)ates, the specific amido side chain
found in good ␤-lactamase substrates has been incorporated either in the acyl (2) or
in the phosphyl (3) moiety. We were interested in the issue of whether, assuming that
these molecules interacted covalently with the ␤-lactamase active site, the amido side
chain or the acyl group dictated the nature of the reaction observed. Studies of other,
chemically very different, ligands have shown that covalent adducts can be formed
with a variety of chemical moieties in either the amido side chain or leaving group
areas of the active site (4,5). The results of the present research show that, in this
case, the side chain has a dominant effect on the reactivity of 2 and 3 with the enzymes.
MATERIALS AND METHODS
Materials
The ␤-lactamases were obtained from the Centre for Applied Microbiology and
Research (Porton Down, Wiltshire, U.K.) and used as received. Typical specific
activities for these preparations were as previously reported (6). The Streptomyces
`
R61 DD-peptidase was kindly provided by Professor J.-M. Frere (University of
`
`
Liege, Liege, Belgium). Benzylpenicillin was purchased from Sigma Chemical Co.,
SCHEME 1.

INVERSE ACYL PHOSPH(ON)ATES
273
cephalothin was a gift from Eli Lilly and Co., and m-carboxyphenyl phenaceturate
was available from previous studies in this laboratory (7).
Tetraethylammonium N-(phenylacetyl)glycyl phenyl phosphate (2). This compound
was prepared by means of a method described by Kluger et al. (8). First, bis(tetraethy-
lammonium) phenyl phosphate was prepared by the addition, with stirring over 10
min, of phenyl dichlorophosphate (7.86 g, 50 mmol; Aldrich Chemical Co.) to 20
ml of water in an ice bath. The mixture was stirred for a further 1 h and the water
removed by rotary evaporation. Two equivalents of tetraethylammonium hydroxide
(35% in water, 41.1 ml; Aldrich) were then added, the pH adjusted to 7.0 with
hydrochloric acid, and the resulting solution feeeze–dried. The residue (16 g) was
˚
dissolved in 80 ml dichloromethane and dried overnight over 4 A molecular sieves.
Dicyclohexylcarbodiimide (1.2 g, 5.8 mmol) was added to a stirred solution of
phenylacetylglycine (1.16 g, 6 mmol) in dry dichloromethane (125 ml). The resulting
mixture was stirred for 10 min and then bis(tetraethylammonium) phenyl phosphate
(5 mmol) in dichloromethane added. After a further hour, dicyclohexyl urea was
removed by filtration and the dichloromethane solution extracted twice with 50-ml
portions of water. The aqueous solution was freeze–dried to yield the product which
was then purified by Sephadex LH-20 chromatography in dichloromethane. The final
oily product was characterized as the tetraethylammonium salt by its 1H NMR spectrum
[(D2O)␦ 1.24 (t, 12H, NEt4), 3.23 (q, 8H, NEt4), 3.65 (s, 2H, PhCH2), 4.08 (s, 2H,
NHCH2), 7.3 (m, 10H, ArH)] and ESMS [m/e 608.7; bis(tetraethylammonium cation)].
N,NЈ-Dibenzylethylenediammonium bis[benzoyl N-(benzyloxycarbonyl)amino-
methylphosphonate] (3). N-(Benzyloxycarbonyl)aminomethylphosphonic acid was
prepared as previously described (9). The required acyl phosphate was then obtained
by the method of Jencks and Carriuolo (10), where the ratio of starting materials, the
time of mixing, and the mode of removal of pyridine, which catalyzes hydrolysis of
the product as well as its formation, were varied to optimize the yield. Thus, a solution
of benzoic anhydride (0.94 g, 4.2 mmol; Acros Organics) in pyridine (0.5 ml) was
added dropwise with stirring to an ice-cooled solution of N-(benzyloxycarbonyl)ami-
nomethylphosphonic acid (100 mg, 0.41 mmol) and sodium hydroxide (0.85 mmol)
in water (1.8 ml). After 15 min, the reaction mixture was rapidly extracted three
times with diethyl ether, where suction by an aspirator was used to remove the ether
phase. The pH of the aqueous phase was lowered to 3.5 by addition of 1 M HCl and
the resulting solution freeze–dried. This procedure yielded a mixture of product and
starting phosphonate in a ratio of ca. 5:1. The product could be isolated by precipitation
as a N,NЈ-dibenzylethylenediamine salt. Thus, the above product was dissolved in a
minimum volume of water and to it was added an equal volume of a 14 mM aqueous
solution of N,NЈ-dibenzylethylenediamine diacetate. The resulting mixture was stirred
in an ice bath for 10 min and the precipitated product removed by filtration, washed
with water, and dried in vacuo. The product, a colorless solid was characterized by
1
its melting point (133–135ЊC), H NMR spectrum [(2H6-DMSO) ␦ 3.41 (s, 4H,
CH2CH2), 3.60 (dd, J ϭ 6, 12 Hz, CH2P), 4.15 (s, 4H, CH2Ph), 5.11 (s, 2H, CH2O),
7.10 (brt, 1H, NH), 7.5 (m, 20H, ArH)], 31P NMR spectrum [(2H6-DMSO) ␦ 10.5],
IR spectrum [(KBr) ␥Cϭ0 ϭ 1709 cmϪ1], and ESMS (m/e 348.1; M ϩ H+).

274
MORRISON ET AL.
Attempts to prepare the N-phenylacetyl analog of 3 by the same method failed,
probably because of the lability of this compound (9,11).
Analytical and Kinetic Methods
Fresh stock solutions of enzymes and phosph(on)ates were prepared in 20 mM
Mops buffer, pH 7.5, and kinetics experiments were conducted in this buffer at 25ЊC.
Rate constants of spontaneous hydrolysis and enzyme steady-state rate parameters
were determined spectrophotometrically. The analytical wavelengths employed were
270 nm (⌬␧ ϭ 1120 MϪ1 cmϪ1) and 275 nm (⌬␧ ϭ 440 MϪ1 cmϪ1) for 2 and 3,
respectively. The steady-state parameters were obtained from initial rate measurements
by the method of Wilkinson (12). Irreversible inactivation rates were determined by
incubation together of suitable concentrations of enzyme and inhibitor. Small aliquots
were taken at appropriate times and the enzyme activity assayed against either cephalo-
thin or benzylpenicillin. Second-order rate constants, ki, for inactivation were then
obtained by employment of Scheme 2 and the program Dynafit (13). In Scheme 2,
ko is the pseudo-first-order rate constant for spontaneous hydrolysis of I under the
conditions employed, and Q the hydrolysis product of I. Phosphorus analyses on
trichloroacetic acid precipitated enzyme samples (1) were carried out by a combination
of the methods of Hess and Derr (14) and Kapoulas et al. (15), where phenylphosphonic
acid was used to construct a standard curve.
RESULTS AND DISCUSSION
1H NMR experiments demonstrated that 2 and 3 hydrolyzed to the expected carbox-
ylate and phosph(on)ate products in aqueous buffer at pH 7.5. The pseudo-first-order
rate constants for this process, obtained spectrophotometrically, were 6.4 ϫ 10Ϫ5 and
1.2 ϫ 10Ϫ4 sϪ1, respectively. The substantial rates of hydrolysis strongly suggested
that C–O bond cleavage had occurred; i.e., the hydrolyses were acyl transfer reactions,
as is generally found with acyl phosphates (16).
The hydrolysis of 2 was strongly catalyzed by the P99 ␤-lactamase. Steady-state
parameters are given in Table 1. The value of kcat/Km identifies 2 as the most specific
acyclic substrate of this ␤-lactamase yet discovered (7). It seems likely that the
negatively charged phosphate leaving group interacts favorably with the electropositive
active site (17) during reaction of 2 with the enzyme (18). The kcat value is very
similar to that obtained for various aryl phenaceturates such as 4 (7), where the same
SCHEME 2.

INVERSE ACYL PHOSPH(ON)ATES
275
TABLE 1
Kinetic Parameters for Interactions of 2 and 3 with the Enzymes
Compound
Enzyme/Parameter
2
3
P99
TEM
PCl
kcat (sϪ1
)
90 Ϯ 7
0.09 Ϯ 0.01
1.0 ϫ 106
b
30 Ϯ 6
16 Ϯ 4
1880
a
Km (mM)
kcat/Km (sϪ1 MϪ1
)
)
)
)
ki (sϪ1 MϪ1
)
(5.9 Ϯ 0.3) ϫ 104
kcat (sϪ1
Km (mM)
kcat/Km (sϪ1 MϪ1
ki (sϪ1 MϪ1
kcat (sϪ1
Km (mM)
cat/Km (sϪ1 MϪ1
ki (sϪ1 MϪ1
kcat (sϪ1
Km (mM)
cat/Km (sϪ1 MϪ1
ki (sϪ1 MϪ1
)
a
)
b
130 Ϯ 8
)
0.105 Ϯ 0.002
0.26 Ϯ 0.02
400
a
k
)
b
14.7 Ϯ 0.7
(3.5 Ϯ 0.2) ϫ 10Ϫ4
0.21
R61
)
4.1 Ϯ 0.01
0.20 Ϯ 0.02
2.05 ϫ 104
b
k
1.65 Ϯ 0.13
)
5 ϫ 10Ϫ3
Note. P99, ␤-lactamase of Enterobacter cloacae P99; TEM, TEM-2 plasmid ␤-lactamase; PCl,
␤-lactamase of Staphylococcus aureus PCl; R61, DD-peptidase of Streptomyces R61.
a No turnover observed.
b No inhibition observed.
acyl-enzyme would be generated and where the hydrolysis of this intermediate is
rate-determining (6,7). It seems likely then that turnover of 2 by this enzyme involves
an intermediate acyl-enzyme (Scheme 3A). Further support for this proposition was
obtained from experiments with the alternative nucleophiles, methanol and D-␣-amino-
butyrate. Methanol (0–2.5 M) increases the rate of reaction of 2 (0.98 mM) with the
P99 ␤-lactamase (Fig. 1) in a fashion observed for most other substrates of this
1
enzyme (6,7,19). A H NMR experiment where 2 was reacted with the enzyme in a
2
2
solution of 2M H4-methanol in H2O, also containing 50 mM NaHCO3 as buffer,
showed that the methanolysis product, methyl phenaceturate, characterized by its
glycyl methylene peak at ␦4.00 (6), accompanied the hydrolysis product after the
reaction was complete. The conclusion is that deacylation is rate-determining but can
be accelerated by addition of an alternative nucleophile to intercept the acyl-enzyme
(Scheme 3A). The NMR experiment showed that 50% of 1 methanolyzed under the
conditions employed which corresponds to a k4/k3 ratio (Scheme 3A) of 25.5. Treatment
of the kinetic data with the appropriate equation (7) yielded a value for the partition
ratio k4/k3 of 29. These estimates of k4/k3 agree very well with the value of 28 obtained
by Xu et al. (7) for what is presumably the same acyl-enzyme generated from 4.
The behavior of 2 with the P99 enzyme in the presence of
strongly resembles that previously observed with the above-mentioned depsipeptides.
Typical behavior is shown in Fig. 1 where the amino acid employed was -2-aminobu-
tyric acid. The substrate inhibition observed was interpreted, as previously, in terms
D-amino acids also
D

276
MORRISON ET AL.
SCHEME 3.
of Scheme 4. In Scheme 4, D represents the substrate, P and H its hydrolysis products,
and Q, the product of aminolysis by the -amino acid A. Application of the derived
D
steady-state equation (7) to these data yielded a value of the partition ratio in favor
of aminolysis, k3/K2k6, of 1.0 mMϪ1. This is very similar to the value previously
obtained for 4 and D
-phenylalanine, viz. 1.4 Ϯ 0.4 mMϪ1 (7). These results strongly
support the proposition that 2 reacts with the P99 ␤-lactamase by an acyl transfer
mechanism.
Class A ␤-lactamases, from the TEM-2 plasmid and from Staphylococcus aureus
PCl, also catalyze the hydrolysis of 2, presumably also by an acyl-transfer mechanism
since phosphonyl derivatives of these enzymes are known to hydrolyze more slowly
(20,21). Although 2 is not as good a substrate of these enzymes as it is of the class
C enzyme, it is quite similar to the depsipeptide 4 in this regard (7). The kcat value
with the PCl enzyme indicates rate-determining deacylation, as for 4 and for most
other substrates with this enzyme (6,7,22,23).
Finally, 2 is also a quite good substrate of the R61 DD-peptidase, and somewhat
better than 4 in this regard (24). This enzyme is structurally very similar to the P99
␤-lactamase (25) and has been studied as a model DD-peptidase for many years
(26). Deacylation is likely to be rate-determining under saturation conditions here
also (24,27).
In contrast to the behavior of 2, the phosphonate 3 was not a substrate of any of
the above ␤-lactamases, but did irreversibly inactivate them. The relevant inactivation

INVERSE ACYL PHOSPH(ON)ATES
277
FIG. 1. Effect of alternative nucleophiles on turnover of 2 by the P99 ␤-lactamase. (Top) Initial rates
of solvolysis of 2 (0.98 mM) in aqueous methanol as a function of methanol concentration. The
␤-lactamase concentration was 24 nM. (Bottom) Initial rates of total reaction (hydrolysis plus aminolysis)
of 2 in the presence of D-2-aminobutyrate (125 mM) and the P99 ␤-lactamase (10 nM) as a function of
the concentration of 2. The points are experimental and the lines are calculated as described in the text.
SCHEME 4.
constants, ki, are also given in Table 1. The phosphonate 3 can be compared with 5,
which has previously been studied in this laboratory. The phosphonate 5, and various

278
MORRISON ET AL.
analogs of it, inactivate ␤-lactamases by phosphonylation of the active site serine
residue (3,11,17,28). Compounds 3 and 5 both have oxygen leaving groups but the
conjugate acid pKa of the former is significantly lower than that of the former [4.2
vs 7.1 (29)]; the benzoate of 3 would therefore be expected to be the better leaving
group and thus 3 the better phosphonylating agent. Previous studies have suggested
that better leaving groups at phosphorus lead to better inhibitors (3,11). In accord
with that idea, 3 inactivated all three ␤-lactamases more rapidly than did 5. The
inactivated P99 ␤-lactamase contained ca. 0.8 g-atoms of phosphorus per mole of
enzyme. This observation supports the notion that the reaction between 3 and the ␤-
lactamases involves phosphonylation (Scheme 3B).
Thus, a clear generalization can be seen in the reactions of 2 and 3 with the ␤-
lactamases examined. The acyl phosphate 2 is a substrate which is turned over by a
double displacement acyl-transfer reaction while the acyl phosphonate 3 inactivates
the enzyme by phosphonylation. It seems likely that this difference is dictated by the
specific interaction of the amido side chain with the enzyme in each case (Scheme
3). In all crystal structures of class A and C ␤-lactamases with specific substrates
and inhibitors (17,21,30–32), the amido side chain is found hydrogen-bonded to a
backbone carbonyl of the ␤-strand which forms one side of the active site and to the
amido side chain of an Asn residue in the conserved S(Y)XN motif. These interactions
appear to promote both acylation and phosphylation rates—the rates of acylation and
phosphonylation by nonspecific acyl phosphonates are much smaller (1).
The reaction of 3 with the R61 DD-peptidase can then be understood in terms of
the above discussion. This enzyme is very similar in structure to the P99 ␤-lactamase
(25) and 4 is a common substrate (24). A concise overview of the phenomena involved
is seen in Fig. 2, which shows the effect of adding enzyme to a mixture of 3 and the
substrate 4. The hydrolysis of 4 was monitored at 290 nm. A burst of product from
4 is observed, followed by a slow steady-state turnover which slowly accelerates.
Addition of benzylpenicillin (1.0 mM) after the burst led to immediate cessation of
turnover of 3 (not shown), due, presumably, to acylation of the enzyme by the ␤-
lactam. The latter experiment shows that the slow turnover seen in Fig. 2 is in fact
due to the DD-peptidase.
For reasons not understood at present, the R61 DD-peptidase is quite inert to
phosphonates such as 5 (11). Nor is it acylated at a significant rate by non-specific
acyl phosphonates such as benzoyl phenylphosphonate (1). The data of Fig. 2 suggests
rather slow (vs turnover of 4) formation of an inert complex of the enzyme and 3,
followed by even slower turnover. It is also noticeable in Fig. 2, that the steady-state
turnover of 4 slowly increases with time. This is presumably due to depletion of 3
in solution by turnover and by spontaneous hydrolysis. The data of Fig. 2 were
therefore fitted [by Dynafit (13)] to Scheme 5, where S and I represent 4 and 3,
respectively, and P and Q represent the products of hydrolysis of 4 and 3, respectively.

INVERSE ACYL PHOSPH(ON)ATES
279
FIG. 2. Initial portion of the progress curve for turnover of m-carboxyphenyl phenaceturate (4, 0.92
mM) by the R61 DD-peptidase (0.27 ␮M) in the presence of 3 (1.5 mM). Shown here are the experimental
trace and the calculated curve (see text) superimposed.
SCHEME 5.
SCHEME 6.
The thereby derived value for k2 was taken to be kcat for turnover of 3 by this enzyme.
The kcat value of 3 is more suggestive of deacylation than dephosphonylation although
rates of dephosphonylation of this enzyme are not known. Consistent with this interpre-
tation, the rate constant for regeneration of free enzyme from reaction of the R61
DD-peptidase benzoyl phenyl phosphate (1) was determined to be 5.6 ϫ 10Ϫ4 sϪ1,

280
MORRISON ET AL.
which is very similar to the kcat reported here for 3; these rate constants should be
the same if a common benzoyl-enzyme intermediate were generated. The rate constant
for the inactivation of the enzyme by benzoyl phenyl phosphate, viz. 1.7 sϪ1 MϪ1 is
also very similar to the value reported in Table 1 for kcat/Km for 3. The reactions of
Scheme 5 are accompanied by an even slower irreversible inactivation with ki ϭ
5 ϫ 10Ϫ3 sϪ1 MϪ1, detected by incubation of 3 (0–3 mM) with the enzyme for 24
hr; this may represent the phosphonylation rate.
It seems that the R61 DD-peptidase, faced with a choice between two unfavorable
reactions, phosphonylation by a reagent with an amido group in the acyl donor site
(Scheme 3), or acylation by a reagent with a benzoyl group in the acyl donor site
and an amidomethylphosphonate in the leaving group (acyl acceptor) site (Scheme
6), may have selected the acyl transfer reaction as the lesser evil.
The results described in this paper demonstrate how the reactivity of substrates/
inhibitors of these enzymes can be modulated. They further emphasize the importance
of the amido side chain and the charge on the leaving group in the design of substrates
and covalent inhibitors of ␤-lactamases and DD-peptidases. The difference between
the acyl phosph(on)ates 2 and 3 is very clear: 2 is a substrate, 3 is a covalent inhibitor.
Clearly the side-chain effect was dominant in this case.
ACKNOWLEDGMENT
This research was supported by the National Institutes of Health.
REFERENCES
1. Li, N., and Pratt, R. F. (1998) J. Am. Chem. Soc. 120, 4264–4268.
2. Pratt, R. F. (1989) Science 246, 917–919.
3. Rahil, J., and Pratt, R. F. (1992) Biochemistry 31, 5869–5878.
4. Richter, H. G. F., Angehrn, P., Hubschwerlen, C., Kania, M., Page, M. G. P., Specklin, J.-L., and
Winkler, F. K. (1996) J. Med. Chem. 39, 3712–3722.
5. Weston, G. C., Bla´zquez, J., Baquero, F., and Shoichet, B. K. (1998) J. Med. Chem. 41, 4577–4586.
6. Govardhan, C. P., and Pratt, R. F. (1987) Biochemistry 26, 3385–3395.
7. Xu, Y., Soto, G., Hirsch, K. R., and Pratt, R. F. (1996) Biochemistry 35, 3595–3601.
8. Kluger, R., Li, X., and Loo, R. W. (1996) Can. J. Chem. 74, 2395–2400.
9. Rahil, J., and Pratt, R. F. (1991) J. Chem Soc. Perkins Trans. 2, 947–950.
10. Jencks, W. P., and Carriuolo, J. (1959) J. Biol. Chem. 234, 1272–1274.
11. Li, N., Rahil, J., Wright, M. E., and Pratt, R. F. (1997) Bioorg. Med. Chem. 5, 1783–1788.
12. Wilkinson, G. N. (1961) Biochem. J. 80, 324–332.
13. Kuzmic, P. (1996) Analyt. Biochem. 237, 260–273.
14. Hess, H. H., and Derr, J. E. (1975) Analyt. Biochem. 63, 607–613.
15. Kapoulas, V. M., Mastronicolis, S. K., and Nakhel, I. C. (1984) Z. Naturforsch. 39C, 249–251.
16. DiSabato, G., and Jencks, W. P. (1961) J. Am. Chem. Soc. 83, 4400–4405.
17. Lobkovsky, E., Billings, E. M., Moews, P. C., Rahil, J., Pratt, R. F., and Knox, J. R. (1994) Biochemistry
33, 6762–6772.
18. Kaur, K., and Pratt, R. F. (2001) Biochemistry, 40, 4610–4615.
19. Knott-Hunziker, V., Petursson, S., Waley, S. G., Jaurin, B., and Grundstrom, T. (1982) Biochem. J.
¨
207, 315–322.
20. Rahil, J., and Pratt, R. F. (1991) Biochem. J. 275, 793–795.
21. Chen, C. C. H., Rahil, J., Pratt, R. F., and Herzberg, O. (1993) J. Mol. Biol. 233, 165–178.
22. Pratt, R. F., McConnell, T. S., and Murphy, S. J. (1988) Biochem. J. 254, 919–922.
23. Christensen, H., Martin, M. T., and Waley, S. G. (1990) Biochem. J. 266, 853–861.
24. Adediran, S. A., and Pratt, R. F. (1999) Biochemistry 38, 1469–1477.
25. Knox, J. R., Moews, P. C., and Fre`re, J.-M. (1996) Chem. Biol. 3, 937–947.

INVERSE ACYL PHOSPH(ON)ATES
281
26. Ghuysen, J.-M., Fre`re, J.-M., Leyh-Bouille, M., Coyette, J., Dusart, J., and Nguyen-Diste`che, M.
(1979) Annu. Rev. Biochem. 48, 73–101.
27. Jamin, M., Adam, M., Damblon, C., Christiaens, L., and Fre`re, J.-M. (1991) Biochem. J. 280, 499–506.
28. Rahil, J., and Pratt, R. F. (1993) Biochem. J. 246, 389–393.
29. Jencks, W. P., and Regenstein, J. (1976) in Handbook of Biochemistry and Biology (Fasman, G. D.,
Ed.), 3rd ed., Vol. 1, pp. 305–351, CRC Press, Cleveland, OH.
30. Oefner, C., D’Arcy, A., Daley, J. J., Gubernator, K., Charnas, R., Heinze, I., Hubschwerlen, C., and
Winkler, J. K. (1990) Nature 343, 284–288.
31. Strynadka, N. C. J., Adachi, H., Jensen, S. E., Johns, K., Sielecki, A., Betzel, C., Sutoh, K., and
James, M. N. G. (1992) Nature 359, 700–705.
32. Maveyraud, L., Massova, I., Birck, C., Miyashita, K., Samama, J.-P., and Mobashery, S. (1996)
J. Am. Chem. Soc. 118, 7435–7440.