β-Ketophosphonates as β-lactamase inhibitors: Intramolecular cooperativity between the hydrophobic subsites of a class D β-lactamase

Article abstract of DOI:10.1016/j.bmc.2008.05.045

A series of aryl and arylmethyl β-aryl-β-ketophosphonates have been prepared as potential β-lactamase inhibitors. These compounds, as fast, reversible, competitive inhibitors, were most effective (micromolar K_i values) against the class D OXA-1 β-lactamase but had less activity against the OXA-10 enzyme. They were also quite effective against the class C β-lactamase of Enterobacter cloacae P99 but less so against the class A TEM-2 enzyme. Reduction of the keto group to form the corresponding β-hydroxyphosphonates led to reduced inhibitory activity. Molecular modeling, based on the OXA-1 crystal structure, suggested interaction of the aryl groups with the hydrophobic elements of the enzyme's active site and polar interaction of the keto and phosphonate groups with the active site residues Ser 115, Lys 212 and Thr 213 and with the non-conserved Ser 258. Analysis of binding free energies showed that the β-aryl and phosphonate ester aryl groups interacted cooperatively within the OXA-1 active site. Overall, the results suggest that quite effective inhibitors of class C and some class D β-lactamases could be designed, based on the β-ketophosphonate platform.

Full text of DOI:10.1016/j.bmc.2008.05.045

Bioorganic & Medicinal Chemistry 16 (2008) 6987–6994
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
b-Ketophosphonates as b-lactamase inhibitors: Intramolecular cooperativity
between the hydrophobic subsites of a class D b-lactamase
*
Senthil K. Perumal, S. A. Adediran, R. F. Pratt
Department of Chemistry, Wesleyan University, Middletown, CT 06459, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of aryl and arylmethyl b-aryl-b-ketophosphonates have been prepared as potential b-lactamase
inhibitors. These compounds, as fast, reversible, competitive inhibitors, were most effective (micromolar
Ki values) against the class D OXA-1 b-lactamase but had less activity against the OXA-10 enzyme. They
were also quite effective against the class C b-lactamase of Enterobacter cloacae P99 but less so against the
class A TEM-2 enzyme. Reduction of the keto group to form the corresponding b-hydroxyphosphonates
led to reduced inhibitory activity. Molecular modeling, based on the OXA-1 crystal structure, suggested
interaction of the aryl groups with the hydrophobic elements of the enzyme’s active site and polar inter-
action of the keto and phosphonate groups with the active site residues Ser 115, Lys 212 and Thr 213 and
with the non-conserved Ser 258. Analysis of binding free energies showed that the b-aryl and phospho-
nate ester aryl groups interacted cooperatively within the OXA-1 active site. Overall, the results suggest
that quite effective inhibitors of class C and some class D b-lactamases could be designed, based on the b-
ketophosphonate platform.
Received 31 March 2008
Revised 20 May 2008
Accepted 21 May 2008
Available online 27 May 2008
Keywords:
b-Lactamase
Inhibitor
Phosphonate
Bacterial enzymes
Steady state enzyme kinetics
Cooperativity
Ó 2008 Published by Elsevier Ltd.
1. Introduction
of these enzymes, and particularly of class A, havearisen asthreats to
the effectiveness of b-lactam antibiotics. New variants of classes B
The effectiveness of b-lactam antibiotics was compromised
even at the beginning of their clinical application by the existence
of b-lactamases.1 These enzymes catalyze hydrolysis of b-lactams
and thence the loss of their antibiotic activity. Over the last
60 years, the importance of b-lactamases to the resistance of bac-
teria to b-lactams has only increased through worldwide popula-
tions of pathogenic bacteria.2,3 b-Lactamase inhibitors have been
introduced into clinical practice to oppose the effect of b-lactamases,
but those employed to date are of limited spectrum of activity and
need to be supplemented by broader spectrum alternatives. The
search for, design, synthesis, and testing of b-lactamase inhibitors
therefore continues.4,5
and D have also appeared,7,8 further threatening the future of b-lac-
tam chemotherapy. The b-lactamase inhibitors employed commer-
cially to date are, to a large degree, specific to the class A enzymes
and their effectiveness is continually eroded by the appearance of
b-lactamase variants capable of resisting or hydrolyzing them.9,10
The research described below represents part of a project to explore
newclassesofnon-b-lactamb-lactamase inhibitorsand, primarilyin
this report, inhibitors of serine b-lactamases.
A series of aroyl phosphates (1) and diaroyl phosphates (2) rep-
resent a new lead to non-b-lactam inhibitors of b-lactamases.11–14
These compounds inhibit serine b-lactamases (classes A, C, and D)
by acylation of the active site, presumably of the nucleophilic ser-
ine; deacylation is then generally slow (Scheme 1).
There are four well-defined classes of b-lactamase.6 One of these,
class B, contains metallo-enzymesthat have an essential metal ion at
the active site as the central catalytic entity. The other three classes
(A, C, and D) are serine enzymes where the mechanism of catalysis
involves an active site serine residue whose side chain hydroxyl
group initiates catalysis by acting as a nucleophile against the sub-
strate. A covalentacyl-enzymeintermediateis thus generated which
is subsequently hydrolyzed to regenerate the free enzyme. Of the
four classes of b-lactamase, the most significant medically, for some
time now, have been the classes A and C enzymes.2,3 Many mutants
O
O
P
O
O
P
O
Ar
O
(O)R/Ar'
Ar
O
O
Ar
O-
O-
1
2
Class D b-lactamases, the oxacillinases, appear to be particularly
susceptible to inhibition by the diaroyl phosphates, 2.13,14 The ac-
tive site of these enzymes is known to be hydrophobic15,16 and,
consequently, inhibition was observed to be greatly enhanced by
hydrophobic substituents. Specifically, enzyme acylation was
strongly accelerated by hydrophobic substituents while deacyla-
* Corresponding author. Tel.: +1 860 685 2629; fax: +1 860 685 2211.
E-mail address: rpratt@wesleyan.edu (R.F. Pratt).
0968-0896/$ - see front matter Ó 2008 Published by Elsevier Ltd.
doi:10.1016/j.bmc.2008.05.045

6988
S. K. Perumal et al. / Bioorg. Med. Chem. 16 (2008) 6987–6994
(en route to 10), the reaction between an aryl carboxylic acid
O
O
P
O
fast
+
+
XPO3H-
methyl ester and dibenzyl methylphosphonate (Scheme 2, route
b) was employed.18 Phenyl monoesters (3–5) were then accessible
by alkaline hydrolysis (Scheme 2). Phenyl diesters were converted
to benzyl and biphenylmethyl diesters by a literature procedure19
and the latter compounds mono-dealkylated by iodide in a Finkel-
stein reaction to obtain the benzyl and biphenylmethyl monoesters
6–12. The b-hydroxyphosphates 13 and 14 were obtained by addi-
tion of the dibenzyl methylphosphonate anion to aryl aldehydes to
form the diester intermediates 18, followed by mono-dealkylation
as above (Scheme 2).
Ar
O
X
E-OH
E-O
Ar
O-
slow
H2O
-
E-OH
+
ArCO2
Scheme 1.
tion was more strongly influenced by electronic effects; the lack of
significant hydrophobic influence on the latter presumably reflects
the similarity of hydrophobic interactions in the acyl-enzyme
intermediate and the deacylation transition state.
The phosphonates 3–14 inhibited class A, C, and D b-lactamases
in a fast reversible fashion. 1H NMR spectra of the b-ketophospho-
O
O
CH2P
O-
O
O
CH2P
O-
O
O
3
4
F
F
F
O
O
O
O
CH2P
O-
CH2P OCH2
F
O
O-
F
6
5
O
O
O
O
CH2P OCH2
CH2P OCH2
O-
O-
7
8
O
O
O
O
O
N
CH2P OCH2
CH2P OCH2
O-
O-
Cl
9
10
O
O
O
O
O
O
N
HN
CH2P OCH2
CH2P OCH2
O-
O-
S
11
12
OH
O
OH
O
C
CH2P OCH2
C
CH2P OCH2
H
H
O-
O-
14
13
PhCH2OCONH
R
O
P
O
CH2
O
15
O-
nates in 2H2O showed no evidence of carbonyl hydration. These
carbonyl groups are not, therefore, strongly electrophilic, even in
5. It is unlikely, therefore, that covalent tetrahedral carbonyl ad-
ducts were formed between these ketophosphonates and the b-
lactamase active sites. Thus, fluorination of the aryl ketone in 5
did not lead to strong inhibition—the small improvement of 5 over
3 as an inhibitor is probably due to the greater hydrophobicity of
the pentafluorophenyl group20 (see below). It might be noted here,
in passing, that no such adducts were observed with even more
electrophilic ketophosphonates.17 In one case, that of 4 and the
OXA-1 b-lactamase, more detailed steady state experiments re-
vealed competitive inhibition with a weaker non-competitive com-
ponent (Ksi P 0.5 mM). The results reported below assume simple
competitive inhibition at concentrations comparable to the Ki
values.
In order to better assess the influence of the hydrophobic effect
on both the acyl donor (ArCO–) site and the leaving group
ð—OPO2ꢀXÞ site, a series of aryl and arylmethyl b-ketophospho-
nates 3–12 have been prepared and tested as b-lactamase inhibi-
tors. Of these, 3–5 represent non-hydrolyzable analogues of 1,
and 6–12, of 2. To assess the contribution of the planar carbonyl
group, the reduced species 13 and 14 were also prepared. Few spe-
cific non-hydrolyzable inhibitors of b-lactamases are known. b-
Ketophosphonates such as 15, bearing amido side chains, have pre-
viously been established as b-lactamase inhibitors.17
2. Results and discussion
Diphenyl arylacylphosphonates were, in general, synthesized
by the Michaelis–Arbuzov reaction (Scheme 2, route a). In one case,

S. K. Perumal et al. / Bioorg. Med. Chem. 16 (2008) 6987–6994
6989
O
O
ArCOCH2P OPh
O-
MeOP(OPh)2
NaOH
ArCOCH2Br
ArCO2Me
ArCHO
ArCOCH2P OPh
OPh
140 ºC
16: a Ar = Ph
b Ar = Ph-Ph
c Ar = C6F6
Route a
d Ar = 2-benzoxazolon-5-yl
e Ar = 1.,3-benzothiazol-2-yl
NaH Ar'CH2OH
O
O
nBuLi/THF
MeP(OCH2Ar')2
Route b
NaI
ArCOCH2P OCH2Ar'
ArCOCH2P OCH2Ar'
MEK
OCH2Ar'
O-
17: a Ar = Ph, Ar' = Ph
b Ar = Ph-Ph, Ar' = Ph
c Ar = Ph, Ar' = Ph-Ph
d Ar = Ph-Ph, Ar' = Ph-Ph
e Ar = chlorazol, Ar' = Ph
f Ar = 2-benzoxazolon-5-yl
g Ar = 1.,3-benzothiazol-2-yl
OH
ArCHCH2P
OCH2Ph
18: a Ar = Ph
b Ar = Ph-Ph
O
OH
O
nBuLi/THF
NaI
OCH2Ph
ArCHCH2P OCH2Ph
MEK
O-
MeP(OCH2Ph)2
Scheme 2.
Table 1
Inhibition of b-lactamases by b-ketophosphonates
the class D OXA-1 enzyme and the class C P99 enzyme. A survey of
the OXA-1 data for the ketophosphonates 3–9 suggests an increase
in effectiveness with increase in hydrophobicity of the substituents
on either side of the phosphonyl moiety. This result is in accord
with observations made with the diacyl phosphates 2 where it
was shown that two separate hydrophobic groups could advanta-
geously bind to the hydrophobic active site of the OXA-1 b-
lactamase.14
The present data for compounds 6–9 can be analyzed to assess
the degree of intramolecular cooperativity between the two hydro-
phobic sites. Scheme 3 shows free energy of binding differences
(calculated from the data of Table 1) on changing phenyl to biphe-
nyl substituents on each side of the phosphonate. It appears that
there is approximately the same contribution to the free energy
of binding on addition of one phenyl group on either side of the
phosphonyl moiety of 6. Addition of the second phenyl, however,
and again to either side, increases the binding free energy consid-
erably more than addition of the first (ca. 2.1 kcal/mol cf. ca.
0.6 kcal/mol). There must, therefore, be cooperativity between
the second pair of phenyl groups on binding to the protein. The
source of this cooperativity cannot be definitively identified at this
b-Lactamase, Ki (l
M)a
Phosphonate
OXA-1b
OXA-10
P99b
TEMb
3
4
5
6
7
1050
140
380
640
230
200
6
430
270
170
440
300
>1000
n
>1000
>1000
>1000
>1000
200
n
n
200
100
350
670
190
n
74
>1000
70
n
920
475
150
>1000
n
8
9
n
10
11
12
13
14
>1000
1500
3600
>1000
>1000
n
n
n
24
1080
260
n, not determined.
a
Uncertainties 610%.
P99, class C b-lactamase of Enterobacter cloacae P99; TEM, class A b-lactamase
b
from the TEM-2 plasmid; OXA-1 and OXA-10, class D b-lactamases from E. coli.
The b-lactamase inhibition data collected on 3–14 are reported
in Table 1. The compounds appear to be quite effective inhibitors of
O
O
CH2 P OCH2
O-
6
- 0.6 kcal/mole
- 0.7 kcal/mole
- 2.8 kcal/mole
O
O
O
O
CH2 P OCH2
O-
CH2 P OCH2
O-
7
8
- 2.2 kcal/mole
- 2.1 kcal/mole
O
O
CH2 P OCH2
O-
9
Scheme 3.

6990
S. K. Perumal et al. / Bioorg. Med. Chem. 16 (2008) 6987–6994
stage but a possible structural basis for it involving the enzyme ac-
tive site structure is described below. It is also possible, however,
that the phenomenon arises from differences between the struc-
tures of the compounds in free solution, whether they exist in an
open or closed (hydrophobically stabilized) conformation, for
example.
(Ar = biphenyl).14 With respect to the latter comparison also, it is
interesting that the carbonyl group of 9 in structure B is well
placed to be attacked by the hydroxyl group of Ser 67, aided by
the Lys 70 carbamate acting as a general base. The carbonyl oxygen
is also well placed to slip into the oxyanion hole (backbone NH
groups of Ser 67 and Ala 215) after such attack. It is true that this
structure is close to that obtained from the original structure build-
ing based on the model of 2 (Ar = Ph–Ph) bound to the enzyme (see
Section 3), but the ligand did move considerably during the
dynamics runs (e.g., structure A) and structure B was the lowest
energy conformation found. It is not unlikely, therefore, that struc-
ture B represents a good model of the non-covalent Michaelis com-
plex formed on binding of 2 to the OXA-1 b-lactamase.
Finally, it is noticeable that in complex A the biphenyl groups of
9 appear to interact with each other (‘stack’) to a greater degree
than in B where the biphenyl ‘arms’ splay apart. The thermody-
namic cooperativity described above on addition of the second
phenyl groups to 7 and 8 might, therefore, more obviously arise
from A. If B were the correct structure, however, the cooperativity
would then likely involve changes in protein structure and/or
dynamics, leading, for example, to a tightening of the complex. A
crystal structure would be helpful in resolving this issue.
Molecular models of 9 bound to the active site of the OXA-1 b-
lactamase were constructed as described in Section 3, and various
possible conformations were explored by molecular dynamics sim-
ulations. In Figure 1, the two stablest structures found (by the cri-
terion of interaction energy14,21) are shown. In the first of these (A),
the carbonyl group of 9 appears hydrogen-bonded to the side chain
functional groups of Lys 212 and Ser 115, both conserved active
site residues, and the phosphonate moiety appears to interact with
the hydroxyl groups of Thr 213 and Ser 258. The latter of these res-
idues is not conserved among class D b-lactamases22 and is re-
placed in OXA-10, for example, with arginine. In structure B,
more favorable than A on the basis of overall interaction energy,
the carbonyl group does not appear to directly interact with any
protein residues, but the phosphonate appears to be hydrogen-
bonded to Lys 212, Thr 213, and Ser 115. In neither A nor B does
the ligand interact directly with the catalytic Lys70 carbamate/
Ser 67 pair, which remains internally hydrogen-bonded.
Addition of the ‘OXA-specific’ chlorazole substituent in 10 did
not lead to strong inhibition. Apparently, the combination of this
substituent and the phosphonate in 10 is not able to take advan-
tage of the position occupied by the side chain of cloxacillin, a good
In both A and B, the biphenyl groups are positioned close to the
hydrophobic residues of the active site, Met 99, Trp 102, Val 117,
Leu 255, and Ile 259, as also observed in the models of 2
Figure 1. Stereoviews of energy-minimized complexes formed between the ketophosphonate 9 and the OXA-1 b-lactamase. The ligand is shown in the A (upper) and B
(lower) conformations (see text); only heavy atoms are displayed.

S. K. Perumal et al. / Bioorg. Med. Chem. 16 (2008) 6987–6994
6991
OXA-1 substrate.23 Replacement of the phenacyl group of 6 with
the more polar heterocycles of 11 and 12 also did not lead to sig-
nificantly stronger inhibition. Modeling had suggested, based on
the above structure with 9, that the heterocycle of 11 may be in
a position to hydrogen bond to backbone amide groups surround-
ing the active site. Any such dual interaction, however, may have
been counterbalanced by the diminished hydrophobic interaction.
Although the diacyl phosphates 2 were comparably effective
inhibitors of the OXA-10 b-lactamase, a representative of another
subclass of class D enzymes, as they were of OXA-1,14 this was
not true of 3–12, which were much poorer inhibitors of the former
enzyme. It may be that generally effective non-covalent inhibitors
will be difficult to find for class D b-lactamases.
Compounds 3–12 were, however, moderately effective inhibi-
tors of the class C P99 b-lactamase (Table 1). Additional hydropho-
bic (phenyl) groups slightly improved the activity over the baseline
of 3 and this correlates with previous observations.12,14,24 The
addition of heterocycles, as seen in 12, particularly, has more dra-
matic effects, again a known property of the class C active site. It is
likely that optimization of heterocyclic substituents could achieve
very potent phosphonate inhibitors of class C b-lactamases, as
demonstrated by Shoichet et al. with boronates.25
Compounds 16b and 16c were prepared as described above for
16a and purified as described below.
3.1.1.2. Diphenyl 4-biphenacylphosphonate (16b). The product
was purified by flash chromatography using ethyl acetate/hexane
(3:7) as eluant (Rf: 0.38). The product was obtained as a colorless
solid in 80% yield. Mp 92–94 °C; 1H NMR (CDCl3) d 8.12 (d,
J = 8.7 Hz, 2H), 7.71 (d, J = 8.1 Hz, 2H), 7.63 (d, J = 6.6 Hz, 2H),
7.41–7.51 (m, 3H), 7.31 (t, J = 6.9 Hz, 4H), 7.15–7.20 (m, 6H), 3.96
(d, J = 23 Hz, 2H), 31P NMR (CDCl3) d 10.26.
3.1.1.3. Diphenyl pentafluorophenacylphosphonate (16c ). The
crude product was subjected to flash chromatography using 30%
ethyl acetate in hexane (3:7) as eluant (Rf: 0.30). The product
was obtained as a colorless viscous oil (25%). 1H NMR (DMSO-d6)
d 7.32 (t, J = 7.8 Hz, 4H), 7.19 (t, J = 7.2 Hz, 2H), 7.13 (d, J = 8.1 Hz,
4H), 3.87 (d, J = 22.5 Hz, 2H). 19F NMR (DMSO-d6) d ꢀ140.46 (d,
J = 8.5 Hz, 2F), ꢀ147.47 (t, J = 9.3 Hz, 1F), ꢀ159.84 (t, J = 15 Hz,
2F). 31P NMR (D2O) d 11.08.
3.1.2. General procedure for preparation of phenyl b-ketophos-
phonates
The class A TEM b-lactamase is not effectively inhibited by
these ketophosphonates, reflecting, most likely, the much more po-
lar environment of its active site where Glu 166, Asn 170, and the
hydrolytic water molecule are incorporated. It is possible that po-
lar replacements for the aryl groups may help, although neither the
heterocycles of 11 and 12 nor the phenylacetylamido side chain of
previously examined ketophosph(on)ates17 support this idea.
Finally, the results (Table 1) of reduction of the (planar) keto
group of 6 and 7 to the (tetrahedral) alcohols of 13 and 14 does not
suggest a route to more effective b-lactamase inhibition. It may be,
as was suggested by our exploration of models of complexes of 13
(both enantiomers) with the OXA-1 b-lactamase (not shown), that
alcohols simply make the inhibitors more flexible without supplying
additional interactions to counteract this negative feature.
Thus, our survey of the b-ketophosphonates 3–12 suggests that
quite effective inhibitors of the class D OXA-1 and class C P99 b-
lactamases could be developed, based on this platform. It seems
unlikely, however, on the basis of results now in hand, that a com-
pletely general non-covalent serine b-lactamase inhibitor of this
class could be developed. On the other hand, since no other such
molecules are currently known, further investigation of the b-keto-
phosphonates may be warranted.
3.1.2.1. Sodium phenyl phenacylphosphonate (3). The title com-
pound was obtained by alkaline hydrolysis of the diphenyl ester
16a. Thus, to a solution of 16a (0.2 g, 0.59 mmol) in dioxane (2 ml)
was added an aqueous solution of NaOH (94 mg, 2.3 mmol) in
1 ml of water. The above mixture was stirred at room temperature
for 1 h. The solvent was then reduced to one-third of its original vol-
ume by rotary evaporation and the pH of the solution adjusted to 8
with HCl. This solution was extracted with ethyl acetate. The aque-
ous layer was then acidified to pH 1 by addition of HCl and extracted
with ether (in this case) or methylene chloride. The combined or-
ganic extracts were washed with water and dried over anhydrous
MgSO4. The solvent was then removed under reduced pressure
and the acid thus obtained was converted into its sodium salt by
the addition of an aqueous solution of NaHCO3 (50 mg, 0.59 mmol).
The solution was then freeze–dried and the crude product purified
by aqueous Sephadex G10 chromatography. The appropriate frac-
tions (characterized by UV absorption) were freeze–dried to yield
the title compound in 58% yield. 1H NMR (D2O) d 7.89 (d,
J = 7.8 Hz, 2H), 7.55 (t, J = 7.5 Hz, 1H), 7.41 (t, J = 7.8 Hz, 2H), 7.20
(t, J = 7.8 Hz, 2H), 7.03 (t, J = 7.5 Hz, 1H), 6.91 (d, J = 8.7 Hz, 2H),
3.61 (d, J = 22.2 Hz, 2H). 31P NMR (D2O) d 13.12; HRMS (ES+) calcu-
lated for C14H13O4PNa: 299.0449; found: 299.0443.
3.1.2.2. Sodium phenyl 4-biphenacylphosphonate (4). The title
compound was obtained from 16b in exactly the same way as 3
from 16a in 50% yield. 1H NMR (D2O) d 7.92 (d, J = 8.1 Hz, 2H),
7.63 (t, J = 8.4 Hz, 4H), 7.40 (t, J = 7.2 Hz, 2H), 7.33 (d, J = 7.5 Hz,
1H), 7.16 (t, J = 7.8 Hz, 2H), 6.99 (t, J = 7.2 Hz, 1H), 6.88 (d,
J = 8.4 Hz, 2H), 3.60 (d, J = 22.2 Hz, 2H). 31P NMR (D2O) d 9.37;
HRMS (ES+) calculated for C20H17O4PNa: 375.0762; found:
375.0766.
3. Experimental
3.1. Synthesis
All chemical reagents were purchased from Aldrich, unless
otherwise noted. Methyl diphenyl phosphite was purchased from
Organometallics Inc. and 1-(1,3-benzothiazol-2-yl)-2-bromo-1-
ethanone from Maybridge Chem. Co.
3.1.2.3. Sodium phenyl pentafluorophenacylphosphonate (5). The
title compound was obtained in 36% yield by employing the same
procedure as for 3. 1H NMR (DMSO-d6) d 7.35 (t, J = 7.2 Hz, 2H),
7.09–7.19 (m, 3H), 3.72 (d, J = 21.6 Hz, 2H). 31P NMR (D2O) d 7.90.
19F NMR (D2O) d ꢀ141.73 (d, J = 20 Hz, 2F), ꢀ149.20 (t, J = 16 Hz,
1F), ꢀ161.63 (t, J = 16 Hz, 2F); HRMS (ES+) calculated for C14H7F5O4P-
Na2: 410.9798; found: 410.9796.
3.1.1. General procedure for synthesis of diphenyl b-ketophos-
phonates
3.1.1.1. Diphenyl phenacylphosphonate (16a). Bromoacetophenone
(1.0 g, 5.02 mmol) and methyl diphenyl phosphite (1.25 g, 5.02 mmol)
were mixed and stirred at 140 °C for 2.5 h under a dry N2 atmosphere.
The crude product thus obtained was subjected to flash chromatogra-
phy with ethyl acetate/hexane (1:4) as eluant (Rf: 0.3). The product
was obtained as a colorless solid in 81% yield. 1H NMR (CDCl3) d
8.05 (d, J = 7.2 Hz, 2H), 7.62 (t, J = 7.2 Hz, 1H), 7.50 (t, J = 7.8 Hz, 2H),
7.30 (t, J = 7.8 Hz, 4H), 7.14–7.21 (m, 6H), 3.93 (d, J = 23.1 Hz, 2H).
31P NMR (CDCl3) d 10.21.
3.1.3. General procedure for the synthesis of bis[(bi)phenylme-
thyl] aryl ketophosphonates
3.1.3.1. Dibenzyl phenacylphosphonate (17a). Dry benzyl alco-
hol (0.37 g, 3.42 mmol) was added to a solution of NaH (60% sus-
pension in toluene) (0.23 g, 5.68 mmol) in dry DMSO. This

6992
S. K. Perumal et al. / Bioorg. Med. Chem. 16 (2008) 6987–6994
mixture was stirred at room temperature for 15 min. To the above
solution was added a dry DMSO solution of diphenyl phen-
acylphosphonate 16a (0.4 g, 1.14 mmol). The reaction mixture
was then stirred at room temperature for 4 h and quenched with
saturated NH4Cl solution (20 ml). The mixture was extracted with
methylene chloride (2ꢁ 60 ml) and the combined organic solution
was washed with brine (60 ml), dried over anhydrous MgSO4, and
concentrated to dryness. The crude product was subjected to flash
chromatography using ethyl acetate/hexane (2:3) as the eluant (Rf:
0.38). The product was obtained as a colorless solid in 50% yield. 1H
NMR (CDCl3) d 7.97 (d, J = 7.5 Hz, 2H), 7.59 (t, J = 7.5 Hz, 1H), 7.44
(t, J = 7.8 Hz, 2H), 7.30–7.34 (m, 10H), 4.99–5.12 (m, 4H), 3.67 (d,
J = 22.8 Hz, 2H); 31P NMR (CDCl3) d 18.32.
sively with NaHCO3, water, and brine and dried over anhydrous
MgSO4. The oil obtained after evaporation of the solvent was sub-
jected to flash chromatography with 40% ethyl acetate in hexane
(Rf: 0.4) as the eluant, to afford the product in 48% yield. 1H NMR
(CDCl3)
d 7.18–7.45 (m, 14H), 4.84–5.09 (m, 4H), 3.02 (d,
J = 22.2 Hz, 2H), 2.61 (s, 3H). 31P NMR (CDCl3) d 21.52.
3.1.4. General procedure for the synthesis of benzyl and (4-
phenyl)benzyl aryl ketophosphonates (6–10)
The phosphonate monoesters 6–10 were synthesized as shown
in Scheme 2. Thus, for example, to a solution of dibenzyl phen-
acylphosphonate (0.214 g, 0.56 mmol) in MEK (10 ml), NaI
(84 mg, 0.56 mmol) was added. The above solution was refluxed
for 2.5 h. The solvent was removed under reduced pressure and
the resulting solid was collected and washed with ice-cold acetone
to remove starting materials. Trace solvents were removed from
the filtered solid under reduced pressure.
3.1.3.2. Dibenzyl 40-biphenacylphosphonate (17b)
. This com-
pound was obtained from 16b by essentially the same procedure
as was used for 17a. The crude product was recrystallized from
50% benzene/cyclohexane in 50% yield. Mp 79–80 °C; 1H NMR
(CDCl3) d 8.01 (d, J = 8.0 Hz, 2H), 7.64 (d, J = 8.0 Hz, 2H), 7.61 (d,
J = 8.0 Hz, 2H), 7.48 (t, J = 8.0 Hz, 2H), 7.43 (t, J = 8.0 Hz, 1H), 7.29
(br, 5H), 5.07 (dAB q, J = 9.6, 13.6 Hz, 4H), 3.69 (d, J = 24.5 Hz,
2H). 31P NMR (CDCl3) d 18.40.
3.1.4.1. Sodium benzyl phenacylphosphonate (6 ). The crude
product from above was recrystallized from ethanol (95%) to yield
a colorless solid in 57% yield. 1H NMR (D2O) d 7.78 (d, J = 7.2 Hz,
2H), 7.47 (t, J = 7.2 Hz, 1H), 7.31 (t, J = 7.8 Hz, 2H), 7.12–7.21 (m,
3H), 7.03–7.10 (m, 2H), 4.63 (d, J = 10.2 Hz, 2H), 3.43 (d,
J = 21.6 Hz, 2H). 31P NMR (D2O) d 12.01; HRMS (ES+) calculated
for C15H15O4PNa: 313.0606; found: 313.0594.
3.1.3.3. Bis[(4-biphenyl)methyl] phenacylphosphonate (17c). The
title compound was obtained from 16a by essentially the same pro-
cedure as was used for the synthesis of 17a. The biphenylmethanol
that was found to be the only contaminant of the crude product
was removed by flash chromatography using 40% ethyl acetate in
hexane as the eluant (Rf: 0.4), followed by a methanol wash to obtain
the pure product. The product was thus obtained as a colorless solid
in 57% yield. Mp 84–86°C; 1H NMR (CDCl3) d 7.98 (d, J = 7.2 Hz, 2H),
7.53–7.57 (m, 9H), 7.44 (t, J = 7.9 Hz, 6H), 7.35–7.37 (m, 6H), 5.06–
5.18 (m, 4H), 3.71 (d, J = 22.5 Hz, 2H). 31P NMR (CDCl3) d 18.57.
3.1.4.2. Sodium benzyl 4-biphenacylphosphonate (7). The crude
product was recrystallized from 50% aqueous ethanol as a colorless
solid in 65% yield. 1H NMR (DMSO-d6) 8.13 (d, J = 7.5 Hz, 2H), 7.71
(2d, J = 7.5 Hz, 4H), 7.48 (t, J = 7.5 Hz, 2H), 7.39 (t, J = 7.5 Hz, 1H),
7.28 (m, 5H), 4.69 (d, J = 6.9 Hz, 2H), 3.26 (d, J = 20.9 Hz, 2H). 31P
NMR (DMSO-d6) d 5.40; HRMS (ES+) calculated for C21H19O4PNa:
389.0919; found: 389.0919.
3.1.3.4. Bis(4-biphenylmethyl) 40-biphenacylphosphonate (17d).
The title compound was obtained from 16b by essentially the same
procedure as used for the synthesis of 17b. The crude product, how-
ever, was purified by recrystallization from ethyl acetate. The product
was obtained as a colorless solid in 38% yield. Mp 200–202 °C; 1H
NMR (CDCl3) d 8.03 (d, J = 8.7 Hz, 2H), 7.52–7.65 (m, 12H), 7.33–
7.47 (m, 13H), 5.07–5.20 (m, 4H), 3.74 (d, J = 22.5 Hz, 2H). 31P NMR
(CDCl3) d 18.45.
3.1.4.3. Sodium 4-biphenylmethyl phenacylphosphonate (8). The
crude product was recrystallized from water/methanol (4:1) to yield
a colorless solid in 77% yield. 1H NMR (D2O) d 7.78 (d, J = 8.4 Hz, 2H),
7.54 (d, J = 8.1 Hz, 2H), 7.45 (d, J = 7.8 Hz, 3H), 7.38 (t, J = 7.8 Hz, 2H),
7.31 (t, J = 8.4 Hz, 3H), 7.17 (d, J = 7.5 Hz, 2H), 4.73 (d, J = 7.2 Hz, 2H),
3.47 (d, J = 21 Hz, 2H). 31P NMR (D2O) d 12.12; HRMS (ES+) calculated
for C21H19O4PNa: 389.0919; found: 389.0921.
3.1.4.4. Sodium 4-biphenylmethyl 40-biphenacylphosphonate
3.1.3.5. Methyl 3-[20-chlorophenyl]-5-methylisoxazole-4-carbo-
xylate. Triethylamine (0.8 g, 7.8 mmol) was added to a solution of
chlorazol chloride (2 g, 7.8 mmol) in methanol (150 ml). This mix-
ture was stirred at room temperature for 5 h. The precipitated
amine hydrochloride was removed by filtration, the filtrate con-
centrated under reduced pressure, and the residue taken up into
ethyl acetate. The ethyl acetate solution was washed with aqueous
sodium bicarbonate and brine, and dried over anhydrous MgSO4.
Removal of the solvent under reduced pressure afforded the pro-
duct in 77% yield. Mp 52–54 °C; 1H NMR (CDCl3) d 7.31–7.48 (m,
4H), 3.68 (s, 3H), 2.75 (s, 3H).
(9). The crude product was recrystallized from 95% ethanol to
yield a colorless solid in 66% yield. 1H NMR (DMSO-d6) d 7.95 (d,
J = 8.0 Hz, 2H), 7.63 (d, J = 8.4 Hz, 2H), 7.56 (d, J = 8.0 Hz, 2H), 7.45
(d, J = 7.6 Hz, 5H), 7.17–7.31 (m, 7H), 4.95–4.98 (m, 2H), 3.92 (d,
J = 22.4 Hz, 2H). 31P NMR (DMSO-d6) d 14.21; HRMS (ES+) calcu-
lated for C27H23O4PNa: 465.1232; found 465.1233.
3.1.4.5. Sodium benzyl 3-[20-chlorophenyl]-5-methylisoxazole-
4-acylphosphonate ( 10 ). The crude product was purified by
Sephadex G-10 chromatography. The appropriate fractions (char-
acterized by UV absorption) were then freeze–dried to yield the ti-
tle compound in 30% yield. 1H NMR (D2O) d7.37–7.45 (m, 2H), 7.20–
7.33 (m, 5H), 7.05–7.08 (m, 2H), 4.55 (d, J = 6.9 Hz, 2H), 2.86 (d,
J = 21.6 Hz, 2H), 2.51 (s, 3H). 31P NMR (D2O) d 10.95; HRMS (ES+)
calculated for C19H18ClNO5P: 406.0611; found: 406.0602.
3.1.3.6. Dibenzyl 3-[20-chlorophenyl]-5-methylisoxazole-4-acy-
lphosphonate (17e). To a solution of dibenzyl methylphosphonate
(2.6 g, 9.4 mmol) in dry THF, stirred at ꢀ78 °C under a dry N2 atmo-
sphere, was added 3.8 ml of 2.5 M BuLi in hexane (0.603 g,
9.4 mmol). After an hour of stirring at ꢀ78 °C, a solution of methyl
3.1.5. Heterocyclic ketophosphonates
3-[20-chlorophenyl]-5-methylisoxazole-4-carboxylate
(0.28 g,
3.1.5.1. 2-Benzoxazolone-5-acyl bromide. The title compound
was synthesized by a reported literature procedure.26 Thus, anhy-
drous dimethyl formamide (2.1 ml, 33.5 mmol) was added drop-
wise to finely ground AlCl3 (13.3 g, 0.1 mol), with stirring, under
argon. The mixture was heated to 45 °C and benzoxazolone
(1.35 g, 10.0 mmol) and bromoacetyl bromide (1.33 ml,
1.23 mmol) in 3 ml of dry THF was added dropwise to the above
solution. The reaction mixture was stirred for 1 h at ꢀ78 °C after
which the coolant bath was removed. The reaction mixture was
then hydrolyzed with 10% aqueous acetic acid and extracted with
ethyl acetate (3ꢁ 40 ml). The organic layer was washed succes-

S. K. Perumal et al. / Bioorg. Med. Chem. 16 (2008) 6987–6994
6993
15.0 mmol) were added slowly. After 30 min, the mixture was
heated to 95 °C for 4.5 h, poured into ice (0.5 kg), and stirred for
1 h. The precipitate was collected by filtration and washed with
0.5 l of water, dried, and recrystallized from methanol to give
2.1 g of a light brown solid in 82% yield, mp 204–205 °C (lit26
205 °C). 1H NMR (DMSO-d6) 12.16 (br s, 1H), 7. 91 (s, 1H), 7.89
(d, J = 8.1 Hz, 1H), 7.23 (1H, d, J = 8.1 Hz), 4.90 (s, 2H).
solution of dibenzyl methylphosphonate (1.56 g, 5.66 mmol) in
anhydrous THF at ꢀ78 °C under dry N2 atmosphere was added a
2.5 M hexane solution of nBuLi (2.25 ml, 5.6 mmol). The above
solution was stirred at ꢀ78 °C for 30 min, after which the aldehyde
(4.72 mmol) in anhydrous THF was added slowly to the stirred
solution. The reaction was allowed to continue for another hour
at ꢀ78 °C, and then another 30 min at room temperature. An aque-
ous saturated solution of NH4Cl (10 ml) was added to the reaction
mixture and the mixture was extracted with ethyl acetate. The or-
ganic layer was washed with brine and dried over MgSO4. The 1H
NMR spectrum (below) showed it to contain the required product
as the only organic component. This material was used for the sub-
sequent synthesis without any further purification.
3.1.5.2. Diphenyl 2-Benzoxazolone-5-acylphosphonate (16d ).
The above bromoketone (0.88 g, 3.43 mmol) and methyl diphenyl
phosphite (2.13 g, 8.58 mmol) were mixed together in DMF
(0.5 ml) and stirred at 140 °C for 2.5 h under a dry N2 atmosphere.
The residue obtained was diluted with methylene chloride/ethyl
acetate (1:1) and stored at room temperature for 1 h. The solid thus
obtained was filtered, dried and recrystallized from ethyl acetate to
give the product as a colorless solid in 80% yield. Mp 198–200 °C; 1H
NMR (DMSO-d6) d 12.15 (br s, 1H), 8.00 (s, 1H), 7.97 (d, J = 8.1 Hz,
1H), 7.39 (t, J = 7.8 Hz, 4H), 7.23 (d, J = 8.0 Hz, 2H), 7.13–7.19 (m,
5H), 4.35 (d, J = 23.0 Hz, 2H); 31P NMR (DMSO-d6) d 13.45.
3.1.6.1. Dibenzyl ( ) 2-hydroxy-2-phenylethylphosphonate
(18a). The product was obtained as a colorless solid in 95% yield.
Mp 62–65 °C; 1H NMR (CDCl3) d 7.24–7.38 (m, 15H), 4.93–5.11 (m,
5H), 3.78 (br s, 1H), 2.13–2.32 (m, 2H); 31P NMR (CDCl3) d 27.37.
3.1.6.2. Dibenzyl ( ) 2-hydroxy-2-(40-biphenyl)ethylphospho-
nate (18b). The product was obtained as a colorless solid in 94%
yield. Mp 74–76 °C;1H NMR (CDCl3) d 7.24–7.38 (m, 15H), 7.47–
7.50 (m, 4H), 4.90–5.07 (m, 5H), 3.69 (br s, 1H), 2.11–2.29 (m,
2H); 31P NMR (CDCl3) d 27.32.
3.1.5.3. Diphenyl 2-(10,30-benzothiazol-20-yl)-2-oxo-ethylphos-
phonate (16e ). 1-(10,30-Benzothiazol-20-yl)-2-bromo-1-ethanone
(1 g, 3.9 mmol) and methyl diphenyl phosphite (0.97 g, 3.9 mmol)
were mixed together and stirred at 140 °C for 2.5 h under dry N2
atmosphere. The residue obtained was then subjected to flash
chromatography using ethyl acetate/hexane (2:3) as eluant (Rf:
0.25). The product was thus obtained as a colorless solid in 35%
yield. 1H NMR (CDCl3) d 8.09 (d, J = 7.8 Hz, 1H), 7.91 (d, J = 7.8 Hz,
1H), 7.45–7.54 (m, 2H), 7.06–7.27 (m, 10H), 4.24 (d, J = 23.0 Hz,
2H); 31P NMR (CDCl3) d 9.34.
The phosphonate monoesters 13 and 14 were prepared as de-
scribed above for 6.
3.1.6.3. Sodium benzyl ( ) 2-hydroxy-2-phenylethylphospho
nate (13). The crude material was recrystallized from ethanol/
diisopropyl ether (9:1) to afford the required product as a colorless
solid in 60% yield. 1H NMR (DMSO-d6) d 7.14–7.38 (m, 10H), 7.00
(br s, 1H), 4.77 (d, J = 8.0 Hz, 2H), 4.63–4.71 (m, 1H), 1.61–1.71 (m,
1H), 1.34–1.47 (q, J = 13.0 Hz, 1H); 31P NMR (CDCl3) d 16.18; HRMS
(ES+) calculated for C15H17O4PNa: 315.0762; found: 315.0763.
Dibenzyl phosphonate esters 17f and 17g were prepared as de-
scribed above for 17a.
3.1.5.4. Dibenzyl 2-benzoxazolone-5-acylphosphonate (17f).
The product was obtained as a pale yellow solid in 59% yield. Mp
120–122°C; 1H NMR (CDCl3) d 9.55 (br s, 1H), 7.74 (d, J = 8.7 Hz, 1H),
7.71 (s, 1H), 7.34–7.38 (m, 10H), 6.75 (d, J = 7.8 Hz, 1H), 5.03–5.17
(m, 4H), 3.62 (d, J = 23.0 Hz, 2H); 31P NMR (CDCl3) d 17.93.
3.1.6.4. Sodium benzyl ( ) 2-hydroxy-2-(40-biphenyl)ethylphos
phonate (14). The crude product was recrystallized from ethanol
to afford the required material as a colorless solid in 72% yield.
1H NMR (DMSO-d6) d 7.63 (d, J = 7.8 Hz, 2H), 7.55 (d, J = 8.1 Hz,
2H), 7.44 (t, J = 7.8 Hz, 2H), 7.25–7.39 (m, 8H), 7.09 (br s, 1H),
4.78 (d, J = 7.8 Hz, 2H), 4.71 (t, J = 9.1 Hz, 1H), 1.71 (t, J = 14.3 Hz,
1H), 1.44 (q, J = 12.8 Hz, 1H); 31P NMR (CDCl3) d 16.18; HRMS
(ES+) calculated for C21H21O4PNa: 391.1075; found: 391.1079.
3.1.5.5. Dibenzyl 2-(10,30-benzothiazol-20-yl)-2-oxo-ethylphos-
phonate (17g ). The product was obtained as a pale brown oil in
82% yield. 1H NMR (CDCl3) d 8.13 (d, J = 7.8 Hz, 1H), 7.95 (d,
J = 7.8 Hz, 1H), 7.51–7.60 (m, 2H), 7.20–7.30 (m, 10H), 5.02–5.16
(m, 4H), 4.05 (d, J = 22 .0 Hz, 2H); 31P NMR (CDCl3) d 17.25.
Benzyl phosphonate monoesters 11 and 12 were prepared as
described above for 6.
3.2. Enzyme kinetics
The class C b-lactamase of Enterobacter cloacae P99 and the
TEM-2 class A b-lactamase from Escherichia coli were purchased
from the Centre for Applied Microbiology and Research (Porton
Down, Wiltshire, UK) and used as received. The class D E. coli
OXA-1 b-lactamase was generously provided by Dr. Michiyoshi
Nukaga, Jyosai International University, Japan. The class D E. coli
OXA-10 b-lactamase was a generous gift from Dr. J.-M. Frère of
3.1.5.6. Sodium benzyl 2-benzoxazolone-5-acyl phosphonate
(11). The crude material obtained was recrystallized from etha-
nol/methanol (2:1) to yield the product in 55% yield. 1H NMR
(DMSO-d6) d 12.40 (br s, 1H), 7.90 (s, 1H), 7.82 (d, J = 8.4 Hz, 1H),
7.22–7.34 (m, 5H), 6.97 (d, J = 8.2 Hz, 1H), 4.74 (d, J = 7.2 Hz, 2H),
3.27 (d, J = 21.0 Hz, 2H); 31P NMR (CDCl3) d 5.59; HRMS (ES+) calcu-
lated for C16H15NO6P: 348.0637; found: 348.0646.
the University of Liège, Liège, Belgium. Cephalothin and 7
cephalosporanic acid (7 -ACA) were purchased from Aldrich. Nitr-
ocefin was purchased from Oxoid.
a-amino-
a
3.1.5.7. Sodium benzyl 2-(10,30-benzothiazol-20-yl)-2-oxo-ethyl-
phosphonate (12). The crude material was recrystallized from
ethanol/methanol (4:1) to yield the required product in 67% yield.
1H NMR (DMSO-d6) d 8.16–8.22 (m, 2H), 7.56–7.65 (m, 2H), 7.19–
7.26 (m, 5H), 4.75 (d, J = 6.9 Hz, 2H), 3.55 (d, J = 20.0 Hz, 2H); 31P
NMR (CDCl3) d 4.24; HRMS (ES+) calculated for C16H15NO4PS:
348.0459; found: 348.0467.
All kinetics experiments were carried out at 25 °C in a buffer
containing 20 mM 3-morpholinopropanesulfonic acid (MOPS) at
pH 7.50, and, for the experiments with the OXA-1 and OXA-10
b-lactamases, the buffer also contained 50 mM bicarbonate.
The activities of the b-keto and b-hydroxyphosphonates as
b-lactamase inhibitors were obtained from spectrophotometric
measurements of initial rates of reaction of appropriate substrates
when catalyzed by the various enzymes. Experimental details are
provided in Table 2. The phosphonates 3–7 and 10–14 were suffi-
ciently soluble in water that 10 mM stock solutions were readily
3.1.6. Synthesis of b-hydroxyphosphonates
The dibenzyl esters 18a and 18b were synthesized by modifica-
tion of a procedure reported for a similar synthesis.27 Thus, to a

6994
S. K. Perumal et al. / Bioorg. Med. Chem. 16 (2008) 6987–6994
Table 2
sentative snapshots led to the structures of Figure 1 as the most
stable [lowest ligand–protein interaction energies (Eint21)]
structures.
Experimental conditions employed for spectrophotometric inhibition kinetics
Enzyme (concn)a
Substrate (concn, KM
)
Wavelength (nm)
P99 (2 M)
TEM (2 nM)
l
ACAb (0.5 mM, 0.29 mM)
Cephalothin (0.1 mM, 0.12 mM)
340
278
482c
482
482
Acknowledgment
Nitrocefin (0.1 mM, 86
l
M)
M)
M)
OXA-1 (20 nM)
OXA-10 (5 nM)
Nitrocefin (25
Nitrocefin (80
l
M, 7.6
l
This research was supported by the National Institutes of
Health.
l
M, 22.6
l
a
P99, class C b-lactamase of Enterobacter cloacae P99; TEM, class A b-lactamase
from the TEM-2 plasmid; OXA-1 and OXA-10, class D b-lactamases from E. coli.
b
ACA, 7-aminocephalosporanic acid.
Employed for 11 and 12.
References and notes
c
1. Abraham, E. P.; Chain, E. Nature 1940, 146, 837.
2. Babic, M.; Hujer, A. M.; Bonomo, R. Drug Resist. Updates 2006, 9, 142.
3. Rossolini, G. M.; Docquier, J. D. Future Microbiol. 2006, 1, 295.
4. Buynak, J. D. Biochem. Pharmacol. 2006, 71, 930.
5. Phillips, O. A. Expert Opin. Ther. Pat. 2006, 16, 319.
6. Matagne, A.; Dubus, A.; Galleni, M.; Frère, J.-M. Nat. Prod. Rep. 1999, 16, 1.
7. Walsh, T. R.; Toleman, M. A.; Poirel, L.; Nordmann, P. Clin. Microbiol. Rev. 2005,
18, 306.
8. Walther-Rasmussen, J.; Hoiby, N. J. Antimicrob. Chemother. 2006, 57,
373.
9. Sader, H. S.; Hsiung, A.; Fritsche, T. R.; Jones, R. N. Diagn. Microbiol. Infect. Dis.
2007, 57, 341.
10. Drieux, L.; Brossier, F.; Sougakoff, W.; Jarlier, V. Clin. Microbiol. Infect. 2008, 14,
90.
11. Li, N.; Pratt, R. F. J.Am. Chem. Soc. 1998, 120, 4264.
12. Kaur, K.; Pratt, R. F. Biochemistry 2001, 40, 4610.
13. Adediran, S. A.; Nukaga, M.; Baurin, S.; Frère, J.-M.; Pratt, R. F. Antimicrob.
Agents Chemother. 2005, 49, 4410.
14. Majumdar, S.; Adediran, S. A.; Nukaga, M.; Pratt, R. F. Biochemistry 2005, 44,
16121.
15. Monaghan, C.; Holland, S.; Dale, J. W. Biochem. J. 1982, 205, 413.
16. Sun, T.; Nukaga, M.; Mayama, K.; Braswell, E. H.; Knox, J. R. Protein Sci. 2003, 12,
82.
17. Perumal, S. K.; Pratt, R. F. J. Org. Chem. 2006, 71, 4778.
18. Maung, J.; Mallari, J. P.; Girtswan, T. A.; Wu, I. Y.; Rowley, J. A.; Santiago, N. M.;
Brunelle, A. N.; Berkman, C. E. Bioorg. Med. Chem. 2004, 12, 4969.
19. Prisbe, E. J.; Martin, J. C.; McGee, D. P. C.; Barker, M. F.; Smee, D. F.; Duke, A. E.;
Matthews, T. R.; Verheyden, J. P. H. J. Med. Chem. 1986, 29, 671.
20. Hansch, C.; Leo, A.; Hoekman, D. Exploring QSAR: Hydrophobic, Electronic and
Steric Constants. ACS Professional Reference Book, ACS: Washington, DC, 1995.
21. Curley, K.; Pratt, R. F. J. Am. Chem. Soc. 1997, 119, 1529.
22. Poirel, L.; Héritier, C.; Tolün, V.; Nordmann, P. Antimicrob. Agents Chemother.
2004, 48, 15.
23. Ledent, P.; Raquet, X.; Joris, B.; Van Beeumen, J.; Frère, J.-M. Biochem. J. 1993,
292, 555.
24. Weston, G. S.; Blazquez, J.; Baquero, F.; Shoichet, B. K. J. Med. Chem. 1998, 41,
4577.
25. Tondi, D.; Powers, R. A.; Caselli, E.; Negri, M.-C.; Blazquez, J.; Costi, M. P.;
Shoichet, B. K. Chem. Biol. 2001, 8, 593.
26. Blum, A.; Gazit, A.; Levitzki, A. J. Biol. Chem. 2003, 278, 40442.
27. Dellaria, J. F., Jr.; Maki, R. G.; Stein, H. H.; Cohen, J.; Whittern, D.; Marsh, K.;
Hoffman, D. J.; Plattner, J. J.; Perun, T. J. J. Med. Chem. 1990, 33, 534.
obtained. The ketophosphonates 8 and 9, however, were not solu-
ble enough in water to provide stock solutions at this concentra-
tion. Hence a 5 mM stock solution of 8 was prepared in DMSO/
water (1:1). The solubility of 9 was even lower than that of 8, both
in water and in DMSO, and thus only a 1 mM solution in DMSO
could be obtained. Inhibitor concentrations were varied in the
range of 0–1.0 mM. The data were fitted to a simple competitive
inhibition equation to obtain the Ki values of Table 1.
3.3. Molecular modeling
The computations were performed by means of an sgi octane2
workstation with the InsightII 2000 suite of molecular modeling
programs (Accelrys, San Diego, CA). The starting point for the
structures was a previously published14 model of 2 (Ar = Ph–Ph)
covalently bound as a tetrahedral intermediate to the active site
of the OXA-1 b-lactamase; this, in turn, was derived from a pub-
lished16 crystal structure of the enzyme [PDB entry 1M6K]. As a
first step toward modeling a complex with the b-ketophosphonate
9, the covalent bond between the active site serine and the car-
bonyl carbon was broken, thereby converting the serine side chain
oxygen to a neutral hydroxyl, and hybridizing the tetrahedral
carbon atom of the tetrahedral intermediate to an sp2 carbonyl
carbon and the attached oxygen to an sp2 carbonyl oxygen. The
remaining structural modification of the ligand was also achieved
by means of the Builder module of InsightII. MNDO charges for
the ligand were calculated from MOPAC and the charges on the
protein were assigned by InsightII.
A 15 Å sphere of water
molecules, centered at Ser 67 O , was added to each structure,
c
supplementing the crystallographic water. Molecular dynamics
simulations (200 ps) followed by energy minimization on repre-