Hydroxamate inhibitors of Aeromonas hydrophila AE036 metallo-β- lactamase

Article abstract of DOI:10.1006/bioo.1998.1111

Amino acid derived hydroxamates were synthesized and tested for inhibitory activity against different metallo-β-lactamases. Several compounds inhibited the clinically relevant enzyme from Aeromonas hydrophila AE036.

Full text of DOI:10.1006/bioo.1998.1111

Bioorganic Chemistry 27, 35–40 (1999)
Article ID bioo.1998.1111, available online at http://www.idealibrary.com on
Hydroxamate Inhibitors of Aeromonas hydrophila AE036
Metallo-␤-lactamase
Magnus W. Walter,* Maria Hernandez Valladares,† Robert M. Adlington,*
`
Gianfranco Amicosante,‡ Jack E. Baldwin,* Jean-Marie Frere,†,1
Moreno Galleni,† Gian Maria Rossolini,§ and Christopher J. Schofield*,1
´
`
`
†Centre d’Ingenierie des Proteines (B6), Universite de Liege, Sart Tilman, Liege 4000, Belgium,
*The Dyson Perrins Laboratory, South Parks Road, Oxford OX1 3QY, United Kingdom;
´
‡Dipartimento di Scienze e Tecnologie Biomediche, Universita degli studi dell’Aquila,
L’Aquila 67100, Italy; and §Dipartimento di Biologia Moleculare, Sezione di Microbioligia,
´
Universita di Siena, Siena 53100, Italy
Received August 3, 1998
Amino acid derived hydroxamates were synthesized and tested for inhibitory activity against
different metallo-␤-lactamases. Several compounds inhibited the clinically relevant enzyme
from Aeromonas hydrophila AE036.
᭧ 1999 Academic Press
INTRODUCTION
Enzyme-mediated resistance to ␤-lactam antibiotics poses a major challenge for
the efficient treatment of bacterial infections. Metallo-␤-lactamases (class B) form a
group of enzymes which contain zinc ions in their active site and whose relevance
for the development of bacterial resistance has only recently been widely recognized
(1–3). Previously known and clinically used inhibitors of “serine” ␤-lactamases, such
as clavulanic acid, are ineffective against the class B metalloenzymes and few effective
inhibitors of metallo-␤-lactamases have been reported (4,5). Recently, we have demon-
strated the activity of amino acid-derived trifluoromethyl alcohols and ketones as the
first synthetic inhibitors of different metallo-␤-lactamases (6,7). Herein, we describe
the synthesis and the inhibitory activity of amino acid-derived hydroxamates against
the metalloenzyme of Aeromonas hydrophila AE036. This pathogen is of particular
interest as it causes a variety of infections in humans.
EXPERIMENTAL METHODS
Kinetic Essays
The enzymes from A. hydrophila and Bacillus cereus were prepared according to
literature procedures (12,13). Analyses were performed at 25ЊC in 15 mM sodium
1 To whom correspondence should be addressed. Fax: 44 1865 275674. E-mail: christopher.schofield@
chem.ox.ac.uk or jmfrere@ulg.ac.be.
35
0045-2068/99 $30.00
Copyright ᭧ 1999 by Academic Press
All rights of reproduction in any form reserved.

36
WALTER ET AL.
cacodylate buffer at pH 6.5, using 200 ␮M Imipenem as reporter substrate for the
enzyme from A. hydrophila and with 150 ␮M nitrocefin as reporter substrate for the
B. cereus enzyme. The pseudo-first-order rate constants were determined by the
reporter substrate method with 200 ␮M imipenem (14). The hydrolysis of the antibiot-
ics was monitored by following the absorbance variation resulting from the opening
of the ␤-lactam ring, using an Uvikon 860 spectrophotometer equipped with thermo-
statically controlled cells and connected to a Copam PC88C microcomputer via a
RS232C serial interface.
Synthesis
General Procedure 1: Preparation of O-benzyl hydroxamates. To an ice-cooled
solution of O-benzylhydroxylamine hydrochloride salt (238 mg, 15.0 mmol) in dry
dichloromethane (25 ml) was added triethylamine (0.21 ml, 15.0 mmol). After stirring
for 5–10 min the required N-hydroxysuccinimide ester (10.0 mmol) was added and
stirring continued for 4–5 h. The organic layer was washed with aqueous hydrochloric
acid (25 ml), water (50 ml), and brine (50 ml). After drying over magnesium sulphate
the solvents were evaporated in vacuo and the products crystallised from suitable sol-
vents.
General Procedure 2: Preparation of hydroxamic acids. To a solution of the re-
quired O-benzyl-hydroxamate (2.00 mmol) in methanol (15 ml) was added palladium
acetate (200 mg). The flask was placed under a balloon of hydrogen and stirred until
(usually 1–2 h) t.l.c. analysis indicated completion of the reaction. The product
mixture was diluted with methanol, filtered through a plug of Celite, and the solvents
removed in vacuo. The products were obtained by crystallization from suitable sol-
vents.
(RS)-O-Benzyl-N-phenoxyacetylalanine hydroxamate 4 was prepared from N-
phenoxyacetylalanine-N-hydroxysuccinimide (3.20 g, 10.0 mmol) following General
Procedure 1 and obtained as fine white needles after crystallization from ethyl acetate/
diethyl ether (2.00 g, 61%); mp 95–96ЊC ␯
(KBr): 3165s, 3100-2900br, 1650s,
max
1600m, 1565m, 1495m, 1450m, 1375m, 1175m cmϪ1; ␦H (200 MHz, CDCl3): 1.40
(3H, d, J 7Hz, CH3), 4.37–4.48 (3H, m, 2-H and PhOCH2), 4.92 (2H, s, PhCH2O),
6.89–7.37 (11H, m, 10H of aromatic CH and amide NH), 9.30 (1H, s, hydroxamate
NH); ␦C (50.3 MHz, CDCl3): 18.3 (CH3), 46.2 (C-2), 66.9 (PhOCH2), 78.2 (PhCH2O),
114.9, 122.1, 128.7, 128.9, 129.5, 130.0 (aryl CH), 135.5, 157.3 (ipso aryl), 168.7,
169.6 (CϭO); m/z (Scan AP+): 351 (MϩNa, 10%), 268 (5%), 178 (100%), 149
(15%), 107 (35%). Required for C18H20N2O4: C, 65.84; H, 6.14; N, 8.53%. Found:
C, 65.93; H, 6.10; N, 8.56%.
(RS)-O-Benzyl-N-phenoxyacetylphenylalanine hydroxamate 5 was prepared
from N-phenoxyacetylphenylalanine-N-hydroxysuccinimide (3.96 g, 10.0 mmol) fol-
lowing General Procedure 1 and obtained as fine white plates after crystallization
from diethyl ether (3.59 g, 89%); mp 109–110ЊC ␯ (KBr): 3240s, 1660s, 1600m,
max
1550m, 1495s, 1390s, 1260m cmϪ1; ␦H (200 MHz, CDCl3): 3.04–3.11 (2H, m,
PhCH2), 4.17 (2H, s, PhOCH2), 4.65–4.83 (3H, m, PhCH2O, and 2-H), 6.84 (1H, br,
d, J 8Hz, amide NH), 6.86–7.36 (15H, m, aromatic CH), 9.02 (1H, s, hydroxamate
NH); ␦C (50.3 MHz, CDCl3): 38.3 (PhCH2), 51.6 (C-2), 66.8 (PhOCH2), 78.2
(PhCH2O), 114.7, 122.2, 127.1, 128.5, 128.7, 129.3, 129.4, 129.8 (aryl CH), 135.1,

METALLO BETA-LACTMASE INHIBITORS
37
135.9, 156.9 (ipso aryl), 167.8, 168.6 (CϭO); m/z (NH3): 405 (MH+, 75%), 387
(5%), 299 (70%), 282 (90%), 254 (95%), 151 (35%), 120 (100%), 91 (65%). Required
for C24H24N2O4: C, 71.27; H, 5.98; N, 6.93%. Found: C, 71.39; H, 5.90; N, 6.90%.
(RS)-N-Benzoyl-O-benzylalanine hydroxamate 6 was prepared from N-benzoyl-
alanine-N-hydroxysuccinimide (2.90 g, 10.0 mmol) following General Procedure 1
and obtained as fine white needles after crystallization from diethyl ether (2.83, 95%);
mp 147–149ЊC ␯
(KBr): 3285s, 3180m, 3060m, 1690m, 1630s, 1575m, 1490m,
max
1375m, 1245m cmϪ1; ␦H (200 MHz, CDCl3): 1.43 (3H, d, J 7Hz, CH3), 4.66 (1H,
qui, J 7Hz, 2-H), 4.85 (2H, s, PhCH2O), 7.10–7.78 (11H, m, 10H of aromatic CH,
and amide NH), 10.02 (1H, s, br, hydroxamate NH); ␦C (50.3 MHz, CDCl3): 18.1
(CH3), 46.8 (C-2), 78.2 (PhCH2O), 127.5, 127.7, 127.8, 129.1, 129.4, 132.2 (aryl
CH), 133.5, 135.3 (ipso aryl), 168.7, 170.3 (CϭO); m/z (Scan AP+): 299 (MH+, 5%),
225 (100%), 176 (80%), 148 (35%), 105 (25%). Required for C17H18N2O3: C, 68.44;
H, 6.08; N, 9.39%. Found: C, 68.60; H, 6.14; N, 9.44%.
(RS)-N-Phenoxyacetylalanine hydroxamate 7 was prepared from 4 (656 mg, 2.00
mmol) following General Procedure 2 and obtained as fine white needles after
crystallization from methanol/diethyl ether (286 mg, 60%); mp 141–143ЊC; ␯
max
(KBr): 3340s, 3195s, 3400–2800br, 1650s, 1600m, 1540s, 1495m, 1460m, 1375m,
1290m cmϪ1; ␦H (200 MHz, (CD3)2SO): 1.22 (3H, d, J 7Hz, CH3), 4.26 (1H, qui, J
7Hz, 2-H), 4.50 (2H, s, PhOCH2), 6.91–7.37 (5H, m, aromatic CH), 8.15 (1H, d, J
8Hz, NH), 8.90 (1H, br); ␦C (50.3 MHz, (CD3)2SO): 18.9 (CH3), 46.2 (C-2), 67.0
(PhOCH2), 115.2, 121.8, 130.2 (aryl CH), 158.5 (ipso aryl), 168.0, 169.5 (2ϫCϭO);
m/z (Scan AP+): 239 (MH+, 5%), 206 (25%), 178 (100%), 150 (15%), 111 (10%).
HRMS: Calcd for C11H15N2O4 (MH+): 239.1032. Found: 239.1032.
(RS)-N-Phenoxyacetylphenylalanine hydroxamate 8 was prepared from 5 (808
mg, 2.00 mmol) following General Procedure 2 and obtained as fine white needles
(354 mg, 56%) after crystallization from methanol/petroleum ether (30–40); mp
128–130ЊC; ␯ (KBr): 3400–3000br, 3215s, 1665s, 1645s, 1600m, 1540m, 1495s,
max
1375s, 1225m cmϪ1; ␦H (200 MHz, (CD3)2SO): 3.81–3.00 (2H, m, PhCH2), 3.28–3.38
(3H, m, PhOCH2, and 2-H), 6.75–7.27 (11H, m, incl. 10H of aromatic CH), 8.24
(1H, d, J 7Hz, NH), 8.91 (1H, br, NHOH); ␦C (50.3 MHz, (CD3)2SO): 37.9 (PhCH2),
51.7 (C-2), 66.6 (PhOCH2), 114.7, 121.2, 126.5, 128.3, 129.3, 129.6 (aryl CH), 137.7,
157.9 (ipso aryl), 167.5, 167.6 (CϭO); m/z (Scan APϪ): 314 (M+, 20%), 313 (M+ϪH,
100%). Required for C17H18N2O4: C, 64.96; H, 5.77; N, 8.91. Found: C, 64.83; H,
5.42; N, 8.64%.
(RS)-N-Benzoylalanine hydroxamate 9 was prepared from 6 (596 mg, 2.00 mol)
following General Procedure 2 and obtained as fine white needles (187 mg, 45%)
after crystallization from methanol/petroleum ether (30–40); mp 144–146ЊC ␯
max
(KBr): 3400–2800br, 1665m, 1630s, 1605m, 1560s, 1490m, 1455m, 1320s, 1245m
cmϪ1; ␦H (200 MHz, (CD3)2SO): 1.32 (3H, d, J 7Hz, CH3), 4.40 (1H, qui, J 7Hz, 2-
H), 7.41–7.91 (5H, m, aromatic CH), 8.51 (1H, d, J 7Hz, NH), 8.85 (1H, br), 10.69
(1H, br); ␦C (50.3 MHz, (CD3)2SO): 18.1 (CH3), 47.0 (C-2), 127.6, 128.2, 131.3 (aryl
CH), 134.1 (ipso aryl), 166.1, 169.4 (CϭO); m/z (Scan AP+): 225 (MϪHϩNHϩ4 ,
100%), 209 (MH+, 20%), 176 (60%), 148 (50%), 105 (30%). HRMS: Calcd for
C10H13N2O3 (MH+): 209.0927. Found: 209.0938.

38
WALTER ET AL.
RESULTS AND DISCUSSION
Target hydroxamic acids 7–9 (Scheme 1) were chosen because of their structural
similarity to ␣-amino trifluoromethyl ketones which have been shown previously to
inhibit metallo-␤-lactamases. It was expected that the hydroxamate group would bind
to the metal center in the active site of the enzyme. Variation of the N-substituent
would allow the influence of the side-chain on the inhibitory activity to be investigated.
The N-phenoxyacetyl group was chosen because it is part of the commonly used
antibiotic penicillin V. Racemic mixtures of N-acyl amino acids 1–3 were converted
into O-benzyl protected hydroxamic acids 4–6 via the corresponding N-hydroxysuc-
cinimides using a standard coupling procedure, followed by treatment with O-benzyl
hydroxylamine hydrochloride salt and triethylamine. Catalytic hydrogenolysis using
Pd(OAc)2 as catalyst afforded hydroxamates 7–9 in good yields (8). This stepwise
procedure was preferable to the direct coupling of 1–3 with hydroxylamine because
the analytically pure hydroxamate products could be obtained by crystallization, thus
avoiding column chromatography which was found to be necessary when employing
direct coupling using hydroxylamine (9).
Compounds 4–9 were tested as inhibitors of metallo-␤-lactamases from B. cereus
and A. hydrophila. None were active against the B. cereus enzyme, but significant
inhibition was observed in the case of the A. hydrophila enzyme. Hydroxamic acids
7 and 8 and somewhat surprisingly, protected hydroxamates 4 and 6, but not 5 (which
was only soluble up to 150 ␮M) and 9, inhibited the enzyme. The lack of inhibitory
activity in the case of 9 rules out the possibility that the protected hydroxamates are
converted into the free hydroxamic acids by the enzyme.
Compounds 4, 6, 7, and 8 apparently behaved as slow-binding inhibitors of the A.
hydrophila ␤-lactamase. In their presence the rate of hydrolysis of 200 ␮M Imipenem
progressively decreased to reach a steady state after a few minutes. With compounds
4 and 8 this steady state was concentration dependent. The residual activity was ca.
10% of that of the uninhibited enzyme in the presence of 0.7 and 0.8 mM 8 and 4,
SCHEME 1

METALLO BETA-LACTMASE INHIBITORS
39
respectively. With both compounds, the pseudo-first-order rate constants increased
with inhibitor concentration in a linear fashion from 5.9 to 12.2 ϫ 10Ϫ3 sϪ1 with
compound 8 (0.4–1.1 M) and from 3.4 to 5 ϫ 10Ϫ3 sϪ1 with compound 4 (0.4–1
mM). On the basis of a simple model,
k
ϩ1
E ϩ X } EX,
k
Ϫ1
where X is the inhibitor and ki ϭ kϩ1 ϩ kϪ1, these data allow calculation of the
following values for the constants: compound 8, kϩ1 ϭ 18 MϪ1 sϪ1, k
ϭ 2.2 ϫ
Ϫ1
10Ϫ3 sϪ1; compound 4, kϩ1 ϭ 5.4 MϪ1 sϪ1, kϪ1 ϭ 2.2 ϫ 10Ϫ3 sϪ1. However, the
observed residual values at the steady state were 1.5- to 7-fold lower than expected
on the basis of the equation vss/v0 ϭ kϪ1/ki , directly derived from the model. The
reasons for this discrepancy are unclear (see below). In contrast, with compounds 6
and 7, the residual activity at the steady state (ca. 8 and 10%, respectively) did not
change when the inhibitor concentration was increased from 0.4 to 1.1 mM. Similarly,
in the same concentration ranges, the ki values remained unaltered, 5 ϫ 10Ϫ3 and
4 ϫ 10Ϫ3 for compounds 6 and 7, respectively. These results indicated a saturation
phenomenon according to a two-step model:
E ϩ X } EX } EX*.
However, if the substrate and inhibitor interact with the enzyme in a competitive
manner, one would expect, under these conditions, complete inhibition of the enzyme
at all inhibitor concentrations. The observed results are thus indicative of a noncompeti-
tive phenomenon. After dialysis complete recovery of activity was observed in all
cases, which could indicate either reversible inhibition or slow turnover of the com-
pounds by the enzyme. The zinc ion content of the inactivated samples after removal
of excess hydroxamic acid by centrifugation was determined by atomic absorption
spectroscopy. The measurements indicated the presence of one zinc ion per molecule
of enzyme. The unchanged ratio of metal cation to enzyme after inhibition excludes
the possibility that the hydroxamate inhibitors act as simple metal chelators in solution.
The results of the kinetic studies are interpreted as evidence therefore that com-
pounds 4, 6, 7, and 8 bind to the active site of the A. hydrophila enzyme and lead
to inhibition of the normal mode of enzymatic action, i.e., ␤-lactam hydrolysis. It is
interesting to note the difference between the A. hydrophila enzyme and that from
B. cereus, which is not inhibited by these hydroxamates. It has been shown that both
enzymes contain two zinc ion binding sites, but with significantly different degrees
of occupancies for the second binding site and. A study of the metal to enzyme ratios
(M/E) in solution gave a M/E ϭ 0.95 Ϯ 0.08 for the A. hydrophila enzyme and
M/E ϭ 1.5 Ϯ 0.02 for the B. cereus enzyme (10,11). The first X-ray crystallographic
analysis of the structure of the B. cereus enzyme, however, revealed only one metal
cation, although amino acid residues with suitable functional groups for a second
metal-binding site were found in the vicinity of the active site. A later study by
Concha et al. of the closely related metallo-␤-lactamase from Bacteroides fragilis

40
WALTER ET AL.
clearly showed the presence of a second zinc-ion close to the active site. (3) Currently
it is unclear whether metallo ␤-lactamases require one or two zinc ions to be bound
at their active sites and if only one which of the two sites is essential for activity.
Further, the possibility exits that if there is only a single zinc present during catalysis
it may move from one binding site to the other. The unusual kinetic results, including
the apparent discrepancies in the data obtained for 8 and 4, may reflect the seemingly
complex nature of zinc binding by the metallo ␤-lactamases.
The lack of inhibitory activity of the hydroxamates against the enzyme from B.
cereus may reflect the level of occupancy of the second metal center by a zinc
ion which may obstruct binding at the first zinc binding site. However, the recent
crystallographic analyses of tetrazole inhibitors bound to the active site of the B.
fragilis enzyme have shown complexation to the second zinc binding site only (5),
rendering this explanation unlikely. Further studies need to be carried out in order to
rationalize the differences in behavior of the hydroxamates with respect to the two
enzymes. The activity of the protected hydroxamates was unexpected and may occur
by a different mechanism. The mechanistic details of the inhibition are currently
under investigation and it is hoped that these studies will lead to an improved under-
standing of the mode of action of metallo-␤-lactamases.
ACKNOWLEDGMENTS
This work was supported by the European Community via HCM (CHRX-CT93-0268) and TMR
(FMRX-CT98-0232) networks (M.W.W. and M.H.V.).
REFERENCES
`
1. Frere, J. M. (1995) Mol. Micobiol. 16, 385–395.
´
`
2. Carfi, A., Pares, S., Duee, E., Galleni, M., Duez, C., Frere, J. M., and Dideberg, O. (1995) EMBO
J. 14, 4914–4921.
3. Concha, N. O., Rasmussen, B. A., Bush, K., and Herzberg, O. (1996) Structure 4, 823–836.
4. Payne, D. J., Bateson, J. H., Gasson, B. C., Proctor, D., Khushi, T., Farmer, T. H., Tolson, D. A.,
Bell, D., Skett, P. W., Marshall A. C., Reid, R., Ghosez, L., and Combret, Y. (1997) Antimicrob.
Agents Chemother. 41, 135–140.
5. Toney, J. H., Fitzgerald, P. M. D., GoverSharma, N., Olson, S. H., May, W. J., Sundelof, J. G.,
Vanderwall, D. E., Clearly, K. A., Grant, S. K., Wu, J. K., Kozarich, J. W., Pompliano, D. L., and
Hammond, G. G. (1988) Chem. Biol. 5, 185–196.
6. Walter, M. W., Felici, A., Galleni, M., Paul Soto, R., Adlington, R. M., Baldwin, J. E., Frere, J. M.,
`
Gololobov, M., and Schofield, C. J. (1996) Bioorg. Med. Chem. Lett. 6, 2455–2458.
7. Walter, M. W., Adlington, R. M., Baldwin, J. E., and Schofield, C. J. (1997) Tetrahedron 53, 7275–
7290.
8. Baldwin, J. E., Adlington, R. M., Elend, A. S., and Smith, M. L. (1995) Tetrahedron 51, 11581–11594.
9. Nikam, S. S., Komberg, B. E., Johnson, D. R., and Doherty, A. M. (1995) Tetrahedron Lett 36,
197–200.
10. Paul Soto, R. per. commun.
11. Carfi, A., Duee, E., Galleni, M., Duez, C., Frere, and J. M. Dideberg, O. (1988) Acta Crystallogr.
D 54, 313–323.
´
`
12. Hernandez Valladares, M., Felici, A., Weber, G., Adolp, H. W., Zeppezauer, M., Rossolini, G. M.,
`
Amicosante, G., and Frere, J. M. (1997) Biochemistry 36, 11534–11541.
13. Lynn Meyers, J., and Shaw, R. W. (1989) Biochim. Biophys. Acta 995, 264–272.
14. Demeester, F., Joris, B., Reckinger, G., Bellefroid-Bourguignon, C., Frere, J. M., and Waley, S. G.
(1987) Biochem. Pharmacol. 36, 2393–2403.