Synthesis and reactivity with β-lactamases of 'penicillin-like' cyclic depsipeptides

Article abstract of DOI:10.1021/jo980564+

Several 7-carboxy-3-amido-3,4-dihydro-2H-1-benzopyran-2-ones have been synthesized as potential β-lactamase substrates and/or mechanism-based inhibitors. Substituted o-tyrosine precursors were prepared by the Sorensen method and then heated in vacuo to give the lactones. These compounds are cyclic analogues of aryl phenaceturates which are known to be β-lactamase substrates. The goal of incorporating the scissile ester group into a lactone was to retain the leaving group tethered to the acyl moiety at the acyl- enzyme stage of turnover by serine β-lactamases, in a manner similar to that during penicillin turnover. Further, in two cases, a functionalized methylene group para to the leaving group phenoxide oxygen was incorporated. These molecules possess a latent p-quinone methide electrophile which could, in principle, be unmasked during enzymic turnover and react with an active site nucleophile. All of these compounds were found to be substrates of class A and C β-lactamases, the first δ-lactones with such activity. Generally, k(cat) values were smaller than for the analogous acyclic depsipeptides, which suggests that the tethered leaving group may obstruct the attack of water on the acyl-enzymes. Further exploration of this structural theme might lead to quite inert acyl-enzymes and thus to significant inhibitors. Despite the apparent advantage offered by the longer-lived acyl-enzymes, the functionalized compounds were no better as irreversible inhibitors than comparable acyclic compounds [Cabaret, D.; Liu, J.; Wakselman, M.; Pratt, R. F.; Xu, Y. Bioorg. Med. Chem. 1994, 2, 757-771]. Thus, even tethered quinone methides, at least when placed as dictated by the structures of the present compounds, were unable to efficiently trap a nucleophile at serine β- lactamase active sites.

Full text of DOI:10.1021/jo980564+

J . Org. Chem. 1999, 64, 713-720
713
Syn th esis a n d Rea ctivity w ith â-La cta m a ses of “P en icillin -lik e”
Cyclic Dep sip ep tid es
D. Cabaret,† S. A. Adediran,‡,§ M. J . Garcia Gonzalez,† R. F. Pratt,*,‡ and M. Wakselman*,†
SIRCOB, Universite´ de Versailles - Saint Quentin-en-Yvelines, Baˆtiment Lavoisier,
45 Avenue des Etats Unis, F-7800, Versailles, France, and Department of Chemistry, Wesleyan University,
Middletown, Connecticut 06459
Received March 26, 1998
Several 7-carboxy-3-amido-3,4-dihydro-2H-1-benzopyran-2-ones have been synthesized as potential
â-lactamase substrates and/or mechanism-based inhibitors. Substituted o-tyrosine precursors were
prepared by the So¨rensen method and then heated in vacuo to give the lactones. These compounds
are cyclic analogues of aryl phenaceturates which are known to be â-lactamase substrates. The
goal of incorporating the scissile ester group into a lactone was to retain the leaving group tethered
to the acyl moiety at the acyl-enzyme stage of turnover by serine â-lactamases, in a manner similar
to that during penicillin turnover. Further, in two cases, a functionalized methylene group para to
the leaving group phenoxide oxygen was incorporated. These molecules possess a latent p-quinone
methide electrophile which could, in principle, be unmasked during enzymic turnover and react
with an active site nucleophile. All of these compounds were found to be substrates of class A and
C â-lactamases, the first δ-lactones with such activity. Generally, kcat values were smaller than for
the analogous acyclic depsipeptides, which suggests that the tethered leaving group may obstruct
the attack of water on the acyl-enzymes. Further exploration of this structural theme might lead
to quite inert acyl-enzymes and thus to significant inhibitors. Despite the apparent advantage offered
by the longer-lived acyl-enzymes, the functionalized compounds were no better as irreversible
inhibitors than comparable acyclic compounds [Cabaret, D.; Liu, J .; Wakselman, M.; Pratt, R. F.;
Xu, Y. Bioorg. Med. Chem. 1994, 2, 757-771]. Thus, even tethered quinone methides, at least when
placed as dictated by the structures of the present compounds, were unable to efficiently trap a
nucleophile at serine â-lactamase active sites.
In tr od u ction
substrate in mechanistic studies.5,6,10,11 One offshoot of
these studies has been the design of functionalized
analogues of 1, molecules such as 2,12 as potential
mechanism-based inhibitors of â-lactam-recognizing en-
zymes. Acylation of the active site serine residue by 2
would yield a p-(bromomethyl) phenol. This should
The â-lactamase enzymes and their evolutionary par-
ents, the bacterial DD-peptidases, remain at the center
of â-lactam research.1,2 The search for new classes of
molecules that interact with these enzymes is therefore
an important part of the quest for new antibiotics.
Demand for the latter has increased in response to the
growing resistance of known pathogens and the emer-
gence of new ones.3
Acyclic depsipeptides have been shown to be substrates
of both â-lactamases and DD-peptidases and much has
been learned about these enzymes from studies with
these substrates.4-11 In particular, the aryl phenaceturate
1 has been extensively employed as a â-lactamase
eliminate bromide in a facile fashion to yield an electro-
philic quinone methide (Scheme 1). Reaction of the latter
with an active site nucleophile (Nu) could then potentially
inactivate the enzyme. This theme has been effective with
other serine hydrolases.13-17
* Corresponding author: Dr. M. Wakselman at the above address.
Phone: 1 39 25 43 65, Fax: 1 39 25 44 52, e-mail: michel.wakselman@
chimie.uvsq.fr.
Competition between reaction with an active site
nucleophile and diffusion of the quinone methide from
† Universite´ de Versailles - Saint Quentin-en-Yvelines.
‡ Wesleyan University.
§ On leave from the Department of Chemistry, University of Ilorin,
Ilorin, Nigeria.
(8) Adam, M.; Damblon, C.; J amin, M.; Zorzi, W.; Dusart, V.; Galleni,
M.; El Kharroubi, A.; Piras, G.; Spratt, B. G.; Keck, W.; Coyette, J .;
Ghuysen, J .-M.; Nguyen-Diste`che, M.; Fre`re, J .-M. Biochem. J . 1991,
279, 601-604.
(1) Neuhaus, F. C.; Georgopapadakou, N. H. In Emerging Targets
in Antibacterial and Antifungal Chemotherapy; Sutcliffe, J ., Georgo-
papadakou, N. H., Eds.; Routledge, Chapman & Hall: New York, 1992;
pp 205-273.
(2) Medeiros, A. A. Clin. Infect. Dis. 1997, 24, Suppl. 1, S19-S45.
(3) Chadwick, D. J .; Goode, J ., Eds. Antibiotic Resistance: Origins,
Evolution and Spread; Ciba Symposium, Wiley: New York, 1997; Vol.
207.
(4) Pratt, R. F.; Govardhan, C. P. Proc. Natl. Acad. Sci. U.S.A. 1984,
81, 1302-1306.
(5) Govardhan, C. P.; Pratt, R. F. Biochemistry 1987, 26, 3385-3395.
(6) Pazhanisamy, S.; Pratt, R. F. Biochemistry 1989, 28, 6875-6882.
(7) Adam, M.; Damblon, C.; Plaitin, B.; Christaens, L.; Fre`re, J .-M.
Biochem. J . 1990, 270, 525-529.
(9) Xu, Y.; Soto, G.; Adachi, H.; Van der Linden, M. P. G.; Keck, W.;
Pratt, R. F. Biochem. J . 1994, 302, 851-856.
(10) Xu, Y.; Soto, G.; Hirsch, K. R.; Pratt, R. F. Biochemistry 1996,
35, 3595-3603.
(11) Adediran, S. A.; Deraniyagala, S. A.; Xu, Y.; Pratt, R. F.
Biochemistry 1996, 35, 3604-3613.
(12) Cabaret, D.; Liu, J .; Wakselman, M.; Pratt, R. F.; Xu, Y. Bioorg.
Med. Chem. 1994, 2, 757-771.
(13) Vilkas, M. In Enzyme-Activated Irreversible Inhibitors; Seiler,
N.; J ung, M. J .; Koch-Weser, J ., Eds.; Elsevier/North-Holland: Am-
sterdam, 1978; pp 323-335.
10.1021/jo980564+ CCC: $18.00 © 1999 American Chemical Society
Published on Web 01/09/1999

714 J . Org. Chem., Vol. 64, No. 3, 1999
Cabaret et al.
Sch em e 1
Sch em e 2
the active site and subsequent hydrolysis (Nu ) H2O)
would be expected to limit the effectiveness of 2 as an
inhibitor, and, in fact, molecules such as 2 were not
efficient inhibitors of â-lactamases.12 If the quinone
methide were tethered to the acyl group, however, one
might expect greater efficiency, and, indeed, molecules
incorporating this feature have effectively inhibited other
serine hydrolases.13-17 Consequently, we have con-
structed molecules analogous to 2, but where the acyl
group is tethered to the leaving group, and describe in
this paper the interactions of one group of these mol-
ecules, dihydrobenzopyranones, with representative â-lac-
tamases and a DD-peptidase. The specific molecules
involved are the parent compounds 3a and 3b and two
functionalized variants 4 and 5. These compounds were
found to be the first examples of δ-lactone substrates of
â-lactamases.
Finally, the action of heat on these compounds in vacuo
for a few minutes gave the substituted dihydrocoumarins
3a or 3b, also, in each case, in racemic form.
2,4-Dimethyl-5-hydroxybenzoic acid20 was prepared
from 2,4-dimethylbenzoic acid by a more convenient route
than the literature method: nitration, catalytic transfer
hydrogenation, diazotization, and hydroxy-de-diazotiza-
tion21,22 (see the Experimental Section). Esterification of
the carboxyl group and acetylation of the hydroxyl
functional group of this acid gave acetate 11 (Scheme 3).
Then, radical bromination of the methyl substituents,
yielded methyl 2,4-bis-(bromomethyl)-5-acetoxybenzoate
12. This dibromide possessed two reactive bromomethyl
groups. To discriminate between these groups, the mol-
ecule was cyclized23,24 to give the bromomethyl phthalide
13 with only one bromomethyl substituent. Moreover, the
formation of the phthalide ring provided mutual protec-
tion of both the latent carboxylic and hydroxyl functions
of the benzyl bromide 13 in the next step of the reaction
sequence. Alkylation of the anion of compound 6a with
monobromide 13 led to the malonic ester 14. Repeated
hydrolysis and decarboxylation, as above for compound
8a , led in this case to a mixture of diacid 10c and
phthalide 15. On being heated in vacuo, this mixture
gave the pure dilactone 5. By analogy with Olah’s method
of cleavage of phthalide to 2-(iodomethyl)benzoic acid,25
treatment of the dilactone 5 with BBr3, sodium bromide,
and tetrabutylammonium bromide in acetonitrile led to
the bromomethylated dihydrobenzopyranone 4. Despite
the presence of a large excess of bromide anion and of
several variations of the experimental procedure, the
reaction did not go to completion, and a small amount of
the parent dilactone 5 was always present in the product.
Crystallization and chromatography were ineffective in
separating these fragile molecules.
Resu lts
Syn th esis. The dihydrobenzopyranones 3a , 3b, and
5 were synthesized by thermal lactonization of the
corresponding substituted o-tyrosine precursors. These
amino acids were prepared by the So¨rensen method.18
Alkylation of the anion of diethyl phenylacetamido- or
19
benzamido-malonates 6a or 6b
with the substituted
benzyl bromides 7a 12 or 7b led to the malonic esters 8a
and 8b, respectively (Scheme 2). Alkaline hydrolysis and
then acidification of these compounds under mild condi-
tions were incomplete (NMR) and therefore were re-
peated twice. Decarboxylation of the triacids 9a or 9b
furnished the racemic N-(phenylacetyl)-4-carboxy-2-hy-
droxyphenylalanine 10a or its benzoyl analogue 10b.
(14) Vilain, A. C.; Okochi, V.; Vergely, I.; Reboud-Ravaux, M.;
Mazaleyrat, J . P.; Wakselman, M. Biochim. Biophys. Acta 1991, 1076,
401-405.
(15) (a) Nguyen, C.; Blanco, J .; Mazaleyrat, J . P.; Krust, B.;
Callebaut, C.; J acotot, E.; Hovanessian, A. G.; Wakselman, M. J . Med.
Chem. 1998, 41, 2100-2110. (b) Wakselman, M.; Xie, J .; Mazaleyrat,
J . P.; Boggetto, N.; Vilain, A. C.; Montagne, J . J .; Reboud-Ravaux, M.
J . Med. Chem. 1993, 36, 1539-1547.
(16) (a) Vergely, I.; Boggetto, N.; Okochi, V.; Golpayegani, S.;
Reboud-Ravaux, M.; Kobaiter, R.; J oyeau, R.; Wakselman, M. Eur. J .
Med. Chem. 1995, 30, 199-208. (b) J oyeau, R.; Felk, A.; Guilaume,
S.; Wakselman, M.; Vergely, I.; Doucet, C.; Boggetto, N.; Reboud-
Ravaux, M. J . Pharm. Pharmacol. 1996, 48, 1218-1230.
(17) Doucet, C.; Vergely, I.; Reboud-Ravaux, M.; Guilhem, J .;
Kobaiter, R.; J oyeau, R.; Wakselman M. Tetrahedron: Asymmetry
1997, 8, 739-751.
(18) (a) Barrett, G. C. In Chemistry and Biochemistry of the Amino
Acids; Chapman & Hall: London, 1985; Ch. 8. (b) O′Donnell, M. J .
Encyclopedia of Reagents for Organic Synthesis; Paquette, L. A., Ed.;
Wiley: New York, 1995; Vol. 3, p 1771.
Sta bility in Aqu eou s Solu tion . Compounds 3a , 3b,
4, and 5 reacted spontaneously in aqueous buffer solu-
1
tions. H NMR spectra of 3 in NaHCO3 buffer showed
clean conversion to a single product with general upfield
movement of resonance peaks, as anticipated for δ-lac-
(20) Newman, M. S.; Chung, H. M. J . Org. Chem. 1974, 39, 1036-
1038.
(21) Tirouflet, J . Bull. Soc. Sc. Bret. 1951, 26, 7-122 (Chem. Abstr.
1953, 47, 8692g).
(22) Vaughan, M. R.; Baird, J r., S. L. J . Am. Chem. Soc. 1946, 68,
1314-1316.
(23) Davies, W.; Perkin, W. H., J r. J . Chem. Soc. 1922, 2202-2215.
(24) Anzalone, L.; Hirsh, J . A. J . Org. Chem. 1985, 50, 2128-2133.
(25) Olah, G. A.; Karpeles, R.; Narang, S. C. Synthesis 1982, 963-
965.
(19) Arct, B.; Kossowska, B.; Lorentz, M.; Prelicz, D.; Sedzik-Hibner,
D.; Osrowski, J . J . Environ. Sci. Health 1983, B18, 559-568.

â-Lactamases and “Penicillin-like” Despipeptides
J . Org. Chem., Vol. 64, No. 3, 1999 715
Sch em e 3
F igu r e 1. Upper panel: The hydrolysis of 3a (0.07 mM)
followed spectrophotometrically at 300 nm in the absence
(dashed line) and presence (solid line) of the P99 â-lactamase
(0.2 µM). Lower panel: The hydrolysis of 3a (0.07 mM) in the
presence of the P99 â-lactamase (0.2 µM), followed (arrow) by
addition of the TEM-2 â-lactamase (2.4 µM).
Sch em e 4
tone hydrolysis at neutral pH. The final NMR spectrum
from 3a [(2H2O, HCO3-), 2.86 (dd, J ) 13.2, 10.3 Hz, 1H),
3.20 (dd, J ) 13.2, 4.8 Hz, 1H), 3.37, 3.48 (ABq, J ) 14.4
Hz, 2H), 4.53 (dd, J ) 10.3, 4.8 Hz, 1H), 6.7-7.3 (m, 8H)]
was consistent with that expected of the hydrolysis
product 16. Absorption spectral changes leading from 3a
(λmax 278 nm, log ꢀ ) 3.11) to 16 (λmax 292 nm, log ꢀ )
as spontaneous hydrolysis, was not. These data were
fitted to Scheme 4, where S1 and S2 represent the
enantiomers of 3a , and P1 and P2 their corresponding
hydrolysis products. The scheme assumes that the enan-
tiomers of 3a , â-3a , and R-3a , were present in the
reaction mixture in equal amounts and hydrolyzed
3.54) were consistent with production of a m-hydroxy-
benzoate (see also Figure 1, upper panel). First-order rate
constants for the spontaneous hydrolysis of 3a , 3b, 4, and
5 in 100 mM MOPS buffer (5% acetonitrile) at pH 7.5
and 25 °C were 5.6 × 10-4 s-1, 5.0 × 10-4 s-1, 1.4 × 10-3
s-1, and 1.3 × 10-3 s-1, respectively. These values can be
compared with those of 1 and 2, the acyclic analogues of
3a and 4, respectively, and of benzylpenicillin under
spontaneously with the same rate constant, ko. One of
the enantiomers is also a â-lactamase substrate. It was
also assumed that the nonreactive enantiomer was not
a significant competitive inhibitor. The data for reaction
of all â-lactamases with 3 fitted well, quantitatively, to
Scheme 4, by methods described in the Experimental
Section, leading to values of kcat and Km (Table 1) for the
reactive enantiomer.
The TEM â-lactamase was shown to catalyze the
hydrolysis of the same enantiomer of 3a as did the P99
enzyme. Figure 1, lower panel, shows the effect of
addition of the TEM enzyme after consumption by the
P99 â-lactamase of its preferred enantiomer. The rate of
comparable conditions, viz. 1.0 × 10-5 s-1, 8.0 × 10-5 s-1
,
and 1.5 × 10-5 s-1, respectively.
â-La ct a m a se-Ca t a lyzed Hyd r olysis. Addition of
â-lactamase to solutions of 3a , 3b, 4, and 5 led, as
monitored by product absorption, to a two-phased reac-
tion. Figure 1, upper panel, shows, for example, such data
for 3a (0.07 mM) and the P99 â-lactamase (0.2 µM). The
first phase of reaction was enzyme-catalyzed; the second,
of equal amplitude and having the same rate constant

716 J . Org. Chem., Vol. 64, No. 3, 1999
Ta ble 1. Stea d y Sta te Kin etic P a r a m eter s for â-La cta m a se-Ca ta lyzed La cton e Hyd r olysis
Cabaret et al.
enzyme
substrate
P99
TEM
PCl
3a
kcat (s-1
)
5.6 ( 0.4e
0.29 ( 0.01
1.93 × 104
0.40 ( 0.01
0.031 ( 0.001
1.28 × 104
2.3 ( 0.1
>1.2f
0.024 ( 0.002
0.015 ( 0.004
1.6 × 103
Km (mM)d
>1
k
cat/Km (s-1M-1
3b kcat (s-1
Km (mM)d
cat/Km (s-1M-1
kcat (s-1
Km (mM)d
cat/Km (s-1M-1
kcat (s-1
Km (mM)d
cat/Km (s-1M-1
kcat (s-1
Km (mM)
cat/Km (s-1M-1
BPb kcat (s-1
Km (mM)
cat/Km (s-1M-1
kcat/Km (s-1M-1
)
)
)
)
)
(1.2 ( 0.2) × 103
)
2.0 ( 0.2
>0.32f
0.066 ( 0.002
>1
k
3.0 × 104
(3.2 ( 0.1) × 102
4
)
>0.62f
-
-
-
-
-
-
0.10 ( 0.04
>1
k
2.3 × 104
(6.2 ( 0.1) × 102
5
)
7.3 ( 0.1
>2.7f
0.89 ( 0.01
>1
k
8.2 × 103
(2.7( 0.1) × 103
1a
)
125
25.4
0.030
0.23
2.2
0.19
k
5.4 × 105
1.2 × 104
1.6 × 102
)
50
2000
30
0.015
0.02
e0.01
k
)
)
3.3 × 106
108
g3 × 106
2c
4.2 × 104
1.7 × 103
2.2 × 102
a
b
The P99 and TEM data (20 mM MOPS, pH 7.5, 25 °C) are from ref 11. Benzylpenicillin: The P99 data (0.1 M phosphate, pH 7.5,
25 °C) are from ref 4, the TEM data (0.1 M phosphate, pH 7.0, 30 °C) from ref 50, and the PC1 data (0.1 M phosphate, pH 7.5, 25 °C) from
d
ref 5. c The data for this compound are from ref 12. Km values are calculated on the basis of the concentration of the reactive enantiomer,
i.e., one-half of the total concentration. e Experimentally determined parameters from the present work are given as means with standard
deviations. f Lower limits on Km were taken to be the highest substrate concentrations employed.
Sch em e 5
Sch em e 6
the subsequent reaction is that of the spontaneous
hydrolysis, indicating that the TEM â-lactamase also did
not catalyze the hydrolysis of the remaining enantiomer
to any significant extent. Qualitatively, 3b behaved
identically to 3a in its interaction with the above
enzymes.
kcat represents the rate constant for acylation (k2) or
deacylation (k3) of the enzyme. This can be conveniently
determined for the P99 and PCl â-lactamases from the
effect of alternative nucleophiles on kcat. If deacylation
were rate-determining at saturation, kcat would be in-
creased by the presence of alternative nucleophiles,
provided, as is true for the above-mentioned enzymes,
that the enzyme can generally accommodate these nu-
cleophiles.5,6 In the present case, methanol was observed
to increase kcat for the hydrolysis of 3a by the P99 enzyme
(data not shown) to an extent expected from a k4/k3 ratio
(Scheme 6) of 29. This is very similar to the value of this
parameter for 1, viz. 27.10 Methanol has no effect however
on the value of kcat/Km, as expected for the mechanism of
Scheme 6, and as is observed for 1.10 1H NMR spectra of
a reaction mixture containing 3.0 M d4-methanol showed
approximately equal amounts of two products, one the
hydrolysis product, the other, with a similar spectrum,
presumably the methyl ester.
Methanol also accelerated turnover of 3a by the PCl
â-lactamase (data not shown). The effect on kcat reflected
a k4/k3 ratio of 6.7. The comparable value for 1 was
determined from previous data5 (when eq 1 was applied,
see Experimental Section) to be 13.3. Methanol had no
effect on turnover of 3a by the TEM â-lactamase. That
enzyme, however, has not previously been found to
catalyze methanolysis of substrates.5
The R61 DD-peptidase did not catalyze the hydrolysis
of 3a or 3b, nor did these compounds inhibit the enzyme.
Rea ction s of â-La cta m a ses w ith th e P oten tia l
In h ibitor s, 4 a n d 5. Absorption spectra showed that the
hydrolysis of one enantiomer of both 4 and 5 was also
catalyzed by the P99 and TEM â-lactamases. Quantita-
tive rate measurements led to the steady-state param-
eters of Table 1. Although the final sample of 4 contained
ca. 15% of 5, it is clear that both â-lactamases catalyzed
the hydrolysis of 4 rather than only the contaminating 5
since the absorption amplitude of the enzyme-catalyzed
reaction in each case corresponded to that expected for
the hydrolysis of both components. These compounds did
not appear to be particularly effective inhibitors however.
For example, 1 mM 4 produced about 40% inhibition of
the P99 â-lactamase (1.0 µM) in 3 h and a similar degree
of inhibition of the TEM enzyme (2.0 µM) in 2 h;
essentially all 4 would have disappeared from solution
through a combination of enzyme-catalyzed and sponta-
neous hydrolysis in these time periods. The dilactone 5
was no more effective than 4 as an inhibitor of these
enzymes.
â-La cta m a se-Ca ta lyzed Meth a n olysis a n d Am i-
n olysis. It is assumed that the hydrolysis of these
δ-lactones by serine â-lactamases proceeds by way of a
covalent acyl-enzyme intermediate (Scheme 5) in the
same way as does the hydrolysis of â-lactams and acyclic
depsipeptides.5 Analysis of the steady-state parameters
of Table 1 is therefore advanced by knowledge of whether
The rate of reaction of 3a in the presence of the P99
â-lactamase was not increased by D-phenylalanine (up
to 30 mM). No product other than that of hydrolysis was
observed in the 1H NMR spectra of reaction mixtures.
This result contrasts sharply with the results of similar
experiments with 1, where the aminolysis by D-amino
acids was observed to be efficiently catalyzed by this

â-Lactamases and “Penicillin-like” Despipeptides
J . Org. Chem., Vol. 64, No. 3, 1999 717
position most closely resembles the penicillin. Only the
carboxyl group seems out of place. It has been previously
observed, however, that the position of carboxyl substitu-
tion on acyclic carboxyphenyl phenaceturates does not
have a very strong influence on the ability of these
compounds to be â-lactamase substrates;10 the position
of the carboxyl group of 3 and 4, when bound to the
enzyme active site, would, however, be much more
restricted than that of 1.
It can be safely assumed from all precedent that the
â-lactamase-catalyzed hydrolysis of 3-5 proceeds by way
of the usual double displacement mechanism of serine
â-lactamase catalysis (Scheme 5). Assuming the struc-
tural analogy with benzylpenicillin shown in Figure 2,
it seems likely that the amido substituent of 3, in both
the Michaelis complex 17 and the acyl-enzyme interme-
diate 18, would occupy the usual side-chain specificity
pocket.29,30 In the acyl-enzyme, the tethered leaving group
may have, by virtue of the methylene group, not present
in penicilloates, greater freedom of motion than in the
latter, and therefore a greater opportunity to conform to
the specificity of the active site.
In aqueous methanol solutions, the P99 and PC1
â-lactamases commonly catalyze methanolysis of sub-
strates as well as hydrolysis, as shown in Scheme 6.
Thus, methanol and water compete for acyl-enzyme E-S
in the deacylation step of turnover.
F igu r e 2. AM1 structures of, from the top, the methyl
analogue of â-3a (axial 3-amido conformation), methyl penicil-
lin, and the methyl analogue of â-3a (equatorial 3-amido
conformation).
As described in the Results section, the partition ratio
k4/k3 for 3a with both the above-mentioned â-lactamases
is not greatly different from that for 1 with the same
enzymes. This suggests that methanol experiences no
greater a steric barrier (vs water) to nucleophilic attack
on the acyl-enzyme of either the P99 or the PC1 â-lac-
tamase generated from 3a than from the acyclic analogue
1. It should be noted that these enzymes represent a class
C and a class A â-lactamase, respectively. These enzymes
probably differ in the details of proton transfer during
catalysis by virtue of different active site functional
groups,29,30 and perhaps also in the face of the carbonyl
group of the acyl-enzyme to which the nucleophile ap-
proaches.31,32 It is striking, however, that the P99 â-lac-
tamase does not catalyze the aminolysis of 3a by D-phen-
ylalanine as it does, quite effectively, the aminolysis of
1. On the other hand, the P99 enzyme does not catalyze
the aminolysis of benzylpenicillin by D-phenylalanine.33
The tethered leaving groups derived from 3a and ben-
zylpenicillin in their respective acyl-enzymes may impede
approach of the bulky nucleophile D-phenylalanine.
enzyme.6 D-Phenylalanine also had no effect on the rate
of turnover of 3a by the TEM â-lactamase, a result in
accord with precedent.4
Discu ssion
The cyclic esters 3-5 were significantly more labile in
aqueous solution than their acyclic analogue 1 and the
â-lactam antibiotic benzylpenicillin. This reflects the
identity of the former compounds as δ-lactones, which
are known to be particularly sensitive to nucleophilic
cleavage.26
Despite the lability of 3-5 in aqueous solution, we were
able to demonstrate that they were, in general, â-lacta-
mase substrates. They seem to be the first-reported
δ-lactones whose hydrolysis is catalyzed by these en-
zymes, although monocyclic â-lactones, direct analogues
of monocyclic â-lactams, have been shown to be sub-
strates.27,28 Only one-half of each compound 3-5 was a
â-lactamase substrate. We assume that this means that
only one enantiomerseach compound is racemic at the
3-positionsis a â-lactamase substrate. It would then
seem likely, by analogy to the specificity of â-lactamases
for â-lactams bearing â-amido substituents, that the
reactive enantiomer was â-3a . Figure 2 shows AM1
energy-minimized structures of the acetamido analogues
of â-3a along with that of methylpenicillin. Both axial
and equatorial conformers of â-3a , which have very
similar heats of formation at the AM1 level (the axial
conformer is favored by 0.4 kcal/mol) are shown. Clearly
the conformer with the amido substituent in the axial
These results therefore indicate that the δ-lactones
3-5 react with â-lactamases in a manner that has
features of the reactivity of their acyclic analogue 1 on
one hand, and of â-lactams on the other.
The rate constants of Table 1 merit some comment.
First, there are some interesting differences between 3a
and 3b. Deacylation (kcat) of the P99 enzyme after its
(29) Oefner, C.; D′Arcy, A.; Daly, J . J .; Gubernator, K.; Charnas, R.
L.; Heinze, I.; Hubschwerlen, C.; Winkler, F. K. Nature 1990, 343, 284-
288.
(30) Strynadka, N. C. J .; Adachi, H.; J ensen, S. E.; J ohns, K.;
Sielecki, A.; Betzel, C.; Sutoh, K.; J ames, M. N. G. Nature 1992, 359,
700-705.
(26) Bruice, T. C.; Benkovic, S. A. In Bioorganic Mechanisms; W.
A. Benjamin: New York, 1966; Vol. I, pp 21-22.
(27) Wells, J . S.; Trejo, W. H.; Principe, P. A.; Sykes, R. B. J .
Antibiot. 1984, 37, 802-803.
(28) Tanizawa, K.; Santoh, K.; Kanaoka, Y. Chem. Pharm. Bull.
1989, 37, 824-825.
(31) Pratt, R. F.; Dryjanski, M.; Wun, E. S.; Marathias, V. M. J .
Am. Chem. Soc. 1996, 118, 8207-8212.
(32) Bulychev, A.; Massova, I.; Miyashita, K.; Mobashery, S. J . Am.
Chem. Soc. 1997, 119, 7619-7825.
(33) Pazhanisamy, S.; Govardhan, C. P.; Pratt, R. F. Biochemistry
1989, 28, 6863-6870.

718 J . Org. Chem., Vol. 64, No. 3, 1999
Cabaret et al.
reaction with 3a is more rapid than with 3b while the
opposite applies to the PC1 enzyme. Also remarkable too
for the PC1 enzyme is the much less stable acyl-enzyme
(Km) from 3b than from 3a and the complementary higher
deacylation rate constants (kcat) for 3b. The comparison
of kinetic data between 3a and its acyclic analogue 1 also
raises interesting contrasts. Both the P99 and TEM
enzymes prefer 1 to 3a , but again PC1 differs. The latter
result arises from a more stable acyl-enzyme formed from
3a than from 1. Apparently 3a fits well into the PC1
active site at the acyl-enzyme stage. The significantly
greater deacylation rate (kcat) of 1 vs 3a from the P99
active site may reflect steric hindrance by the tethered
leaving group on the â-face of the acyl group. The PC1
deacylation rates (kcat), where 1 and 3a react equally well,
indicate no hindrance to the occluded nucleophile on the
R-face of the carbonyl group of the acyl-enzyme in a class
A enzyme.
Acylation of all enzymes (kcat/Km) by benzylpenicillin
is much more rapid than by 3a . This presumably reflects
the differences in structure between these substrates, and
in particular perhaps in the different position of the
carboxylate groups (Figure 2). The P99 and more par-
ticularly the PC1 enzyme also catalyze the hydrolysis of
the acyl-enzyme derived from benzylpenicillin much more
effectively than that from 3a . The combination of a Km
of 15 µM and a kcat of 0.024 s-1 (an acyl-enzyme half-life
of 29 s) almost qualifies 3a as an inhibitor of the PC1
â-lactamase. The tight binding of 3b to the P99 enzyme
is also noticeable.
to the narrower, more rigid, and more specific active site
cleft of the DD-peptidase. It is interesting to contrast this
result with that obtained with N-(phenylacetyl)aziridine-
2-carboxylate, a structurally closer analogue to penicil-
lin.39
The extent of irreversible inhibition of â-lactamases
caused by 4 and 5 was not significantly greater than that
by the acyclic analogue 2.12 The modest inhibitory activity
of the latter compound thus may have arisen as much
from the absence of a suitable active site functional group
in the vicinity of the electrophilic methylene (Scheme 1)
as from rapid diffusion of the product or the quinone
methide into solution. The lifetime of the acyl-enzyme
formed by reaction between 4 and the P99 â-lactamase,
ca. 0.4 s, should be long enough to allow the elimination
reaction of Scheme 1 to occur and create the quinone
methide at the active site.40
We have therefore shown that the δ-lactones 3-5 are
â-lactamase but not DD-peptidase substrates. Further,
in certain instances, quite inert acyl-enzymes were
generated. Despite these acyl-enzyme lifetimes, the func-
tionalized δ-lactones 4 and 5 were not significantly better
irreversible inhibitors than the acyclic analogue 2. A
problem in the development of these compounds as
enzyme inhibitors, either by way of their forming inert
acyl-enzymes, or in more complex ways such as is
indicated in Scheme 1, is their rapid spontaneous hy-
drolysis in neutral solution which competes very ef-
fectively with any reaction with an enzyme. Future
inhibitor design based on these compounds would clearly
require more hydrolytically stable species; more specific
placing of the carboxylate group would also be worth
further exploration.
Turnover rates of 4 and 5 (and 2) by the P99 and TEM
â-lactamases appear to be closely similar to that of 3a .
These compounds are therefore substrates to some
degree, as required of mechanism-based inhibitors.34 The
substrate activity of the dilactone 5, lacking a carboxy-
late, is of interest. A comparable situation would be that
of cephalosporin lactone 19 which is a remarkably
effective â-lactamase substrate.35,36 Again the point is
made that although a well-placed carboxylate can be very
important in a â-lactamase substrate, it is certainly not
essential.
Exp er im en ta l Section
En zym es. The â-lactamases were purchased from the
Centre for Applied Microbiology and Research, Porton Down,
Wilts., U.K. The DD-peptidase of Streptomyces R61 was the
generous gift of Dr. J .-M. Fre`re of the University of Liege,
Liege, Belgium.
An a lytica l a n d Kin etic Meth od s. Stock solutions of the
lactones were prepared in acetonitrile (Aldrich, Sureseal)
rather than aqueous buffer because of the instability of these
compounds in the latter solvent (see Results). All kinetics
measurements were made in 100 mM MOPS buffer, pH 7.5,
25 °C. The total acetonitrile concentration after addition of
the substrate was 5% v/v. â-Lactamase reactions were moni-
tored spectrophotometrically by means of a Hewlett-Packard
HP 8452 spectrophotometer. Reactions of 3a , 3b, 4, and 5 were
monitored at 300 nm. Enzyme concentrations were 0.2 µM
(P99, 3a ), 0.7 µM (P99, 4b and 5), 2 µM (P99, 3b), 0.2 µM
(TEM, 3b), 2.65 µM (TEM, 3a ), 6.0 µM (TEM, 4 and 5), and
2.65 µM (PC1, 3a and 3b). To obtain steady state parameters,
total progress reaction curves were obtained at 0.5 mM and
1.0 mM substrate, and the data quantitatively fitted to Scheme
4 by means of the FITSIM program.46
The absence of any reaction between the R61 DD-
peptidase and the δ-lactones is very striking when
9
compared with the rapid acylation (k2/Ks ) 1070 s-1M-1
)
of this enzyme by benzylpenicillin and by the acyclic
depsipeptide 1 (k2/Ks ) 3520 s-1 M-1),37 particularly when
the strong similarity between the active sites of the DD-
peptidase and the P99 â-lactamase is taken into ac-
count.38 This difference presumably reflects a difficulty
encountered by the more rigid (than 1) and somewhat
differently shaped (than penicillin: Figure 2) 3 in binding
(39) Murphy, B. P.; Pratt, R. F. Biochemistry 1991, 30, 3640-3649.
(40) There is only a limited amount of data in the literature on the
rates of elimination of Br- from p-(bromomethyl)phenols and on the
lifetimes of quinone methides in aqueous solution. The information
available41-45 suggests that quinone methide formation may be faster
than 2.3 s-1 (kcat for 4 with the P99 â-lactamase) while its hydration
is likely to be slower, i.e., the quinone methide should accumulate as
an acyl-enzyme.
(34) Maycock, A.; Abeles, R. Acc. Chem. Res. 1976, 9, 313-319.
(35) Laws, A. P.; Page, M. I. J . Chem. Soc., Perkin Trans. 2 1989,
1577-1581.
(36) Varetto, L.; DeMeester, F.; Monnaie, D.; Marchand-Brynaert,
J .; Dive, G.; J acob, F.; Fre`re, J .-M. Biochem. J . 1991, 278, 801-807.
(37) Xu, Y. Ph.D. Thesis, 1995, Wesleyan University, Middletown,
CT.
(41) Filar, L. J .; Winstein, S. Tetrahedron Lett. 1960, 25, 9-16.
(42) Leary, G.; Miller, I. J .; Thomas, W.; Woolhouse, A. D. J . Chem.
Soc., Perkin Trans. 2 1977, 1737-1739.
(43) de la Mare, P. D. B.; Newman, P. A. J . Chem. Soc., Perkin
Trans. 2 1984, 1797-1802.
(38) Knox, J . R.; Moews, P. C.; Fre`re, J .-M. Chem., Biol. 1996, 3,
937-947.
(44) Richard, J . P. J . Am. Chem. Soc. 1991, 113, 4588-4595.

â-Lactamases and “Penicillin-like” Despipeptides
J . Org. Chem., Vol. 64, No. 3, 1999 719
The inhibitory activity of 4 and 5 against the â-lactamases
was assessed by monitoring the loss of enzyme activity in
incubation mixtures of enzyme and inhibitor; small aliquots
were removed at appropriate times and assayed (initial rates)
against benzylpenicillin. The activity of the R61 DD-peptidase
in the presence of 4 and 5 was determined spectrophotometri-
cally by means of the substrate N-(phenylacetyl)glycyl-D-
thiollactate/4,4′-dipyridyl disulfide.19
To study the effects of methanol on the steady-state
parameters, the latter were determined in the presence of
methanol (0-3.0 M) as described above, and the partition ratio,
k4/k3 (Scheme 6), obtained from eq 1 where kMcat is the observed
kcat in methanol solution, kocat the value in the absence of
67.2, 123.8-136.0, 150.1; 165.1, 166.9, 167.7, 169.7. Anal.
Calcd for C25H27NO9: C, 61.85; H, 5.61; N, 2.89. Found: C,
61.82; H, 5.48; N, 2.71.
7-Ca r boxy-3-(p h en yla ceta m id o)-3,4-d ih yd r o-2H-1-ben -
zop yr a n -2-on e (3a ). Malonate 8a (130 mg, 0.24 mmol) was
dissolved in ethanol (5 mL), 20% sodium hydroxide (5 mL)
added, and the mixture stirred at room temperature for 6 h.
The ethanol was evaporated, and the aqueous solution was
washed with diethyl ether, acidified with 50% hydrochloric
acid, and extracted with ethyl acetate. Evaporation of the
solvent gave a product which was stirred in a solution of TFA
(1 mL) in CH2Cl2 (1 mL) for 2 h. The solution was then
evaporated to dryness, and the residue containing the substi-
tuted malonic acid 9a was heated for a few minutes at 180 °C
for decarboxylation. The saponification reaction was incom-
plete (presence of ethyl groups in the NMR spectrum) and was
therefore repeated. Thus, the product was dissolved in a
mixture of ethanol (5 mL) and 20% sodium hydroxide (5 mL)
and stirred at room temperature for 2 h. Evaporation of the
ethanol, acidification with hydrochloric acid, extraction with
ethyl acetate, and evaporation of the solvent gave the racemic
N-(phenylacetyl)-4-carboxy-2-hydroxyphenylalanine 10a . 1H
NMR (CD3OD) δ 2.91 (dd, 10.0 and 13.5 Hz, 1H), 3.32 (dd, 4.7
and 13.5 Hz, 1H), 3.44 (s, 2H), 4.78 (dd, 4.7 and 10.0 Hz, 1H),
7.05-7.39 (m, 8H). This acid was heated at 180 °C for 5 min
in a GKR 50 Bu¨chi apparatus to give the lactone 3a (66 mg,
85%) which crystallized from acetone: mp 246-248 °C. Rf 0.50
(ethyl acetate/methanol: 7/3). 1H NMR (CD3COCD3) δ 3.32 (dd,
12.8 and 15.6 Hz, 1H), 3.39 (dd, 7.5 and 15.6 Hz, 1H), 3.71 (s,
2H), 4.97 (td, 7.8 and 12.8 Hz, 1H), 7.49 (d, 7.8 Hz, 1H), 7.66
(d, 1.4 Hz, 1H), 7.84 (dd, 1H), 7.93 (d, 7.8 Hz, 1H). 13C NMR
(CD3COCD3) δ 32.2, 42.2, 48.1, 116.8-136.4, 152.0, 166.6,
167.6, 171.2. MS (ES, 30V) m/z 348 (MNa+); 326 (MH+), 208
(MH+ - PhCHdCdO). Anal. Calcd for C18H15NO5: C, 66.45;
H, 4.65; N, 4.30. Found: C, 66.36; H, 4.76; N, 4.14.
kcMat ) kcoat{1 + k4[MeOH]/k3(55.56 - 2.25[MeOH])} (1)
methanol, 55.56 M is taken as the concentration of water in
the absence of methanol, and 2.25 is the ratio of molar volumes
of methanol and water at 25 °C.
To obtain the structures of Figure 2, molecules were
constructed by means of the Builder Module of INSIGHT II,
version 95.0 (Biosym/MSI, San Diego, CA) run on an IBM 3CT
computer. Initial structural relaxation was performed by the
DISCOVER module. Final energy and structural minimization
was achieved by semiempirical AM1 calculations (MOPAC
6.0).
Syn th esis. Analytical thin-layer chromatography (TLC) and
preparative column chromatography were performed on Kie-
selgel F 254 and on Kieselgel 60 (0.063-0.200 mm), respec-
tively. Unless otherwise stated, the eluent used was the same
for TLC and for chromatography purification.
Meth yl 3-Acetoxy-4-(br om om eth yl)ben zoa te (7b). A
mixture of of methyl 3-acetoxy-4-methyl benzoate (1.57 g, 7.55
mmol), NBS (1.48 g, 8.30 mmol) and benzoic peroxide (50 mg),
in CCl4(30 mL) was refluxed for 3h under argon. The reaction
mixture was cooled to room temperature and filtered. The
filtrate was evaporated and the residue was purified by
chromatography (pentane/AcOEt: 9/1), affording product 7b
7-Ca r boxy-3-ben za m id o-3,4-d ih yd r o-2H-1-ben zop yr a n -
2-on e (3b). Compound 3b was prepared by the same experi-
mental procedure as employed for compound 3a . First, racemic
N-benzoyl-4-carboxy-2-hydroxyphenylalanine 10b (85%) was
1
1
obtained. H NMR (CD3COCD3) δ 3.21 (dd, 13.7 and 9.5 Hz,
(1.39 g, 64%): Rf 0.24. mp 111 °C. H NMR (CDCl3) δ 2.40 (s,
1H), 3.44 (dd, 13.7 and 4.9 Hz, 1H), 4.97 (m, 1H), 7.37-7.83
(m, 9H). This acid was heated at 200 °C for 40 min in a GKR
50 Bu¨chi apparatus to give the lactone 3b: mp 258 °C (dec).
Rf 0.62 (ethyl acetate/methanol: 7/3). 1H NMR (CD3COCD3) δ
3.46 (m, 2H), 5.19 (m, 1H), 7.42-7.57(m, 8H). 13C NMR (CD3-
SOCD3) δ 29.2, 47.0, 116.8-133.5, 150.8, 166.3, 166.4, 167.2.
Anal. Calcd for C17H13NO5, 0.25 H2O: C, 61.20; H, 4.67; N,
4.20. Found: C, 61.27; H, 4.69; N, 4.29.
Met h yl 2,4-Dim et h yl-5-a cet oxyb en zoa t e (11). 2,4-Di-
methylbenzoic acid (485 mg, 3.23 mmol) was dissolved in
concentrated H2SO4 (2 mL). The solution was cooled to 0 °C
and stirred. Then, a mixture of concentrated sulfuric acid (1
mL) and concentrated nitric acid (1 mL) was added gradually.
Stirring was continued for 1 h, the mixture was poured on ice,
and the solid precipitate was filtered, washed with water, and
dried. The nitration reaction gave a mixture of 2,4-dimethyl-
5-nitro-, 2,4-dimethyl-3-nitro-, and 2,4-dimethyl-3,5-dinitroben-
zoic acid. A first purification was performed by column
chromatography (dichloromethane/methanol: 95/5). Then,
crystallization from ethanol gave the pure 2,4-dimethyl-5-
nitrobenzoic acid (236 mg, 35%): mp 196 °C (lit.47 mp 197.5-
198.5 °C). Rf 0.28. 1H NMR (CDCl3) δ 2.54 (s, 3H), 2.59 (s, 3H),
7.31 (s, 1H), 8.47 (s, 1H).
3H), 3.92 (s, 3H), 4.43 (s, 2H), 7.50 (d, 1H), 7.80 (d, 1H), 7.89
(dd, 1H). 13C NMR (CDCl3) δ 20.92, 26.70, 52.61, 124.59,
127.59, 131.05, 131.95, 134.71, 149.00, 165.90, 168.87. Anal.
Calcd for C11H11O4Br: C, 46.01; H, 3.86. Found: C, 46.31; H,
4.01.
Dieth yl (2-Acetoxy-4-(t-bu toxycar bon yl)ben zyl)ph en yl-
a ceta m id om a lon a te (8a ). To a solution of 205 mg (0.7 mmol)
of diethyl phenylacetamidomalonate 6a 19 in dry DMF (3 mL)
was added 28 mg (0.7 mmol) of NaH (60% in oil). The mixture
was allowed to react for 15 min. Then a solution of 223 mg
(0.66 mmol) of tert-butyl 3-acetoxy-4-(bromomethyl)benzoate
7a 12 in THF (2 mL) was added. After being stirred for 30 min,
the mixture was poured into a 10% HCl solution and extracted
with ethyl acetate. The ethyl acetate solution was dried over
Na2SO4 and the solvent evaporated. The residue was purified
by column chromatography (dichloromethane/methanol: 97.5/
1
2.5), affording 348 mg (95%) of a glassy product: Rf 0.59. H
NMR (CDCl3) δ 1.26 (t, 7.0 Hz, 6H), 1.59 (s, 9H), 2.23 (s, 3H),
3.54 (s, 2H), 3.58 (s, 2H), 4.23 (q, 7.0 Hz, 4H), 6.46 (s, 1H),
6.93 (d, 8.0 Hz, 1H), 7.58 (d, 1.6 Hz, 1H), 7.67 (dd, 1H). 13C
NMR (CDCl3) δ 20.4, 28.0, 32.3, 43.2, 66.4, 81.2, 116.8-136.4,
149.7, 166.9-170.0. Anal. Calcd for C29H35NO9: C, 64.31; H,
6.51; N, 2.59. Found: C, 64.22; H, 6.32; N, 2.62.
A solution of this acid (310 mg, 0.63 mmol) in methanol (10
mL) and 3 drops of concentrated sulfuric acid was refluxed
overnight. The solvent was evaporated and the product puri-
fied by chromatography (pentane/ether: 3/1), affording 315 mg
(95%) of methyl 2,4-dimethyl-5-nitrobenzoate: Rf 0.69. 1HNMR
(CDCl3) δ 2.64 (s, 3H), 2.66 (s, 3H), 3.93 (s, 3H), 7.24 (s, 1H),
8.61 (s, 1H).
Dieth yl (2-Acetoxy-4-(m eth oxyca r bon yl)ben zyl)ben -
za m id om a lon a te (8b). Compound 8b was prepared in 79%
yield in the same manner as compound 8a , using, however,
diethyl benzamidomalonate instead of diethyl phenylacetami-
domalonate as starting material. Rf 0.32 (dichloromethane/
1
methanol: 99/1). H NMR (CDCl3) δ 1.27 (t, 7.0 Hz, 6H), 2.05
(s, 3H), 3.70 (s, 2H), 3.86 (s, 3H), 3.87-4.41 (m, 4H), 7.14-
7.53 (m, 9H). 13C NMR (CDCl3) δ 14.2, 20.6, 32.7, 52.3, 63.1,
The nitrobenzoate (265 mg, 1.27 mmol) was dissolved in
absolute ethanol (10 mL). Cyclohexene (1 mL) and 10% Pd-C
(45) McCracken, P. G.; Bolton, J . L.; Thatcher, G. R. J . J . Org. Chem.
1997, 62, 1820-1825.
(46) Zimmerle, C. T.; Frieden, C. Biochem. J . 1989, 258, 381-387.
(47) Fisher, C. H.; Walling, C. T. J . Am. Chem. Soc. 1935, 57, 1700-
1702.

720 J . Org. Chem., Vol. 64, No. 3, 1999
Cabaret et al.
(20 mg) were added, and the mixture was refluxed vigorously
for 16 h. After filtration and washing of the catalyst with
ethanol, the solution was evaporated and the product purified
by column chromatography (pentane/ether: 1/1), affording 193
mg (85%) of methyl 2,4-dimethyl-5-aminobenzoate:48 Rf 0.55.
mL) was added a mixture prepared by adding 17 mg (0.42
mmol) of NaH (60% in oil) to a solution of 124 mg (0.42 mmol)
of diethyl phenylacetamidomalonate 6a in dry DMF (3 mL)
and stirring the mixture for 45 min. The whole mixture was
then poured into a solution of 10% HCl and extracted with
ethyl acetate. The organic layer was washed with water, dried
over Na2SO4, and then evaporated. The residue was purified
by chromatography (pentane/ethyl acetate: 1/1), affording 151
mg (72%) of the 5-(2,2-bis(ethoxycarbonyl)-2-phenylacetamido)-
ethyl-6-acetoxyphthalide 14: Rf 0.52. mp 73 °C. 1H NMR
(CDCl3) δ 1.24 (t, 7.0 Hz, 6H), 2.27 (s, 3H), 3.54 (s, 2H), 3.66
(s, 2H), 4.18 (q, 7.0 Hz, 4H), 5.13 (s, 1H), 6.53 (s, 1H), 7.20 (s,
1H), 7.54 (s, 1H). 13C NMR (CDCl3) δ 20.7, 33.0, 43.5; 66.4,
1
mp 68 °C. H NMR (CDCl3) δ 2.17 (s, 3H), 2.46 (s, 3H), 3.85
(s, 3H), 6.92 (s, 1H), 7.26 (s, 1H).
A mixture of 112 mg (0.63 mmol) of the aminobenzoate and
20% sulfuric acid (4 mL) was treated at 0 °C with a solution
of 53 mg of sodium nitrite in water (2 mL). To the resulting
solution of the diazonium salt, concentrated sulfuric acid (4
mL) was added slowly with an efficient cooling. The final
solution was added to boiling 50% sulfuric acid (20 mL). The
temperature was maintained at 100 °C for 5 min then the
solution was poured on ice and extracted with ethyl acetate.
After being dried over sodium sulfate, the solvent was evapo-
rated and the product purified by chromatography (dichlo-
romethane/methanol: 9/1) to give the 2,4-dimethyl-5-hydrox-
ybenzoic acid (89 mg, 86%): mp 185 °C (lit.49 mp 187-188 °C).
69.3; 119.7-143.4, 150.6, 167.1-170.3. Anal. Calcd for C26H27
-
NO9: C, 62.76; H, 5.47; N, 2.87. Found: C, 62.66; H, 5.51; N,
2.76.
3,4-Dih ydr o-3-(ph en ylacetam ido)2,5-dih ydr o-2-oxofu r o-
[3,4-d ]-2H-1-ben zop yr a n -2-on e (5). To a solution of mal-
onate 14 (114 mg, 0.23 mmol) in ethanol (10 mL) was added
20% sodium hydroxide (10 mL), and the mixture was stirred
at room temperature for 16 h. The ethanol was evaporated,
and the aqueous solution was washed with diethyl ether and
then acidified with 50% hydrochloric acid. Extraction of the
aqueous solution with ethyl acetate and then evaporation of
the solvent gave a product which was heated for a few minutes
at 180 °C for decarboxylation. The product was again dissolved
in a mixture of ethanol (5 mL) and 20% sodium hydroxide (5
mL) and stirred at room temperature for 2 h. Evaporation of
the ethanol, acidification with hydrochloric acid, extraction
with ethyl acetate, and evaporation of the solvent gave a
mixture of acid 10c and phthalide 15. Compound 10c: 1H
NMR (CD3COCD3) δ 3.02 (dd, 7.0 and 12.6 Hz, 1H), 3.29 (dd,
4.6 and 13.6 Hz, 1H), 3.51 (s, 2H), 4.74 (s, 2H), 4.80 (m, 1H),
7.17-7.38 (m,7H), 7.42 (d, 7.7 Hz, 1H). Compound 15: 1H
NMR (CD3COCD3) δ 3.05 (dd, 7.0 and 13.8 Hz, 1H), 3.30 (dd,
5.2 and 13.8 Hz, 1H), 3.48 (s, 2H), 4.85 (m, 1H), 5.14 (s, 2H),
7.17 to 7.29 (m, 7H), 7.80 (d, 7.8 Hz, 1H). This mixture was
heated at 180 °C for 15 min to give the dilactone 5 (64 mg,
83%). 1H NMR (CD3COCD3) δ 3.34 (dd, 12.4 and 15.8 Hz, 1H),
3.41 (dd, 7.6 and 15.8 Hz, 1H), 3.66 (s, 2H), 4.94 (dt, 7.7 and
12.5 Hz, 1H), 5.37 (s, 2H), 7.44 (s, 1H), 7.62 (s, 1H), 7.82 (d,
7.7 Hz, 1H).13C NMR (CD3COCD3) δ 31.6, 43.6, 48.2, 70.5,
123.7-143.7, 167.4-171.6. Anal. Calcd for C19H15NO5, 0.25
H2O: C, 66.75; H, 4.57; N, 4.10. Found: C, 66.69; H, 4.63; N,
4.11. MS (ES, 70 V) m/z 360 (MNa+), 338 (MH+), 242, (MHNa
- PhCH2CO+), 142 (PhCH2CONa+), 120 (PhCH2COH+).
6-(Br om om eth yl)-7-ca r boxy-3-(p h en yla ceta m id o)-3,4-
d ih yd r o-2H-1-ben zop yr a n -2-on e (4). To a solution of 40 mg
(0.12 mmol) of dilactone 5 in acetonitrile (2 mL) were added
40 mg (0.37 mmol) of sodium bromide and 3 mg of tetrabutyl-
ammonium bromide. The mixture was cooled to 0 °C in an ice
bath under a dry argon atmosphere before the addition of a
solution of boron tribromide in dichloromethane (1 mL, 1
mmol). The mixture was stirred for 24 h at room-temperature
The reaction mixture was quenched with ice/water, and
dichloromethane was added. The dichloromethane layer was
washed with water and dried over sodium sulfate. Evaporation
of the solvent left a pale yellow solid (35 mg). The NMR spectra
showed the presence of a mixture of the bromo derivative 4
(85% yield) and the starting dilactone 5 (15% yield). Compound
4: 1H NMR (CD3COCD3) δ 3.37-3.51(m, 2H), 3.66 (s, 2H),
4.90-4.99 (m, 1H), 5.06 (s, 2H), 7.51 (s, 1H), 7.60 (s, 1H), 7.87
(d, 7.4 Hz, 1H). 13C NMR (CD3COCD3) δ 30.4, 31.5, 43.2, 47.9,
119.8-136.7, 167.5-172.9. MS (ES, 70 V) m/z 417 and 419;
416 to 420 (MH+).
1
Rf 0.32. H NMR (CD3OD) δ 2.25 (s, 3H), 2.50 (s, 3H), 7.01 (s,
1H), 7.38 (s, 1H).
This acid (564 mg, 3.40 mmol) was esterified by refluxing
in methanol (20 mL) with sulfuric acid as catalyst to give
methyl 2,4-dimethyl-5-hydroxybenzoate (522 mg, 86%): mp
89-90 °C (lit.49 mp 88-89 °C). Rf 0.50 (pentane/ether: 1/1).
1H NMR (CDCl3) δ 2.26 (s, 3H), 2.50 (s, 3H), 3.88 (s, 3H), 7.27
(s, 1H), 7.40 (s, 1H).
A solution of 198 mg (1.1 mmol) of the hydroxybenzoate in
a mixture of acetic anhydride (0.12 mL, 1.2 mmol), triethyl-
amine (0.17 mL, 1.2 mmol), DMAP (20 mg), and dichlo-
romethane (10 mL) was stirred at room temperature for 2 h.
The reaction was quenched with methanol (0.5 mL), and the
mixture was washed twice with water, dried over sodium
sulfate, and evaporated. The product was purified by chroma-
tography (dichloromethane), affording 225 mg (92%) of methyl
2,4-dimethyl-5-acetoxybenzoate 11: Rf 0.33. mp 54 °C. 1H
NMR (CDCl3) δ 2.18 (s, 3H), 2.56 (s, 3H), 2.32 (s, 3H), 3.86 (s,
3H), 7.11 (s, 1H), 7.61 (s, 1H). Anal. Calcd for C12H14O4: C,
64.85; H, 6.35. Found: C, 64.81; H, 6.34.
Meth yl 2,4-Bis-(br om om eth yl)-5-a cetoxyben zoa te (12).
A solution of 605 mg (2.72 mmol) of acetoxybenzoate 11 in
carbon tetrachloride (15 mL) was refluxed for 1 h in the
presence of N-bromosuccinimide (1.07 g, 6.0 mmol) and benzoyl
peroxide (20 mg). After removal of the succinimide by filtration,
the solvent was evaporated and the product purified by column
chromatography (dichloromethane), giving 385 mg (38%) of
methyl 2,4-bis-(bromomethyl)-5-acetoxybenzoate 12 which was
1
used directly in the following step. Rf 0.68. H NMR (CDCl3)
δ 2.40 (s, 3H), 3.94 (s, 3H), 4.39 (s, 2H), 4.93 (s, 2H), 7.53 (s,
1H), 7.77 (s, 1H).
6-Acet oxy-5-(b r om om et h yl)-3H -isob en zofu r a n -1-on e
(13). A flask containing the methyl 2,4-bis-(bromomethyl)-5-
acetoxybenzoate 12 (85 mg, 0.22 mmol) was immersed in an
oil bath at 160-180 °C for 15 min. The cyclized product was
purified by chromatography (dichloromethane/methanol: 99/
1). A yield of 61 mg (95%) of 6-acetoxy-5-(bromomethyl)-
phthalide 13 was obtained as a powder: Rf 0.21 (dichlo-
1
romethane). mp 152 °C. H NMR (CDCl3) δ 2.43 (s, 3H), 4.48
(s, 2H), 5.27 (s, 2H), 7.57 (s, 1H), 7.70 (s, 1H). 13C NMR (CDCl3)
δ 21.0, 26.6, 69.4, 120.3-149.6, 168.5, 169.7. Anal. Calcd for
C
11H9O4Br: C, 46.34; H, 3.18; O, 22.45. Found: C, 46.35; H,
3.36; O, 22.28.
5-(2,2-(Bis(eth oxyca r bon yl)-2-p h en yla ceta m id o)eth yl-
6-a cet oxy-3H -isob en zofu r a n -1-on e (14). To a solution of
120 mg (0.42 mmol) of (bromomethyl)phthalide 13 in THF (5
(48) Roth, B. D.; Blankley, C. J .; Hoefle, M. L.; Holmes, A.; Roark,
W. H.; Triverdi, B. K.; Essenburg, A. D.; Kieft, K. A.; Krause, B. R.;
Stanfield, R. L. J . Med. Chem. 1992, 35, 1609-1617.
(49) Schultz, A. G.; Hardinger, S. A. J . Org. Chem. 1991, 56, 1105-
1111.
(50) Fisher, J .; Belasco, J . G.; Khosla, S.; Knowles, J . R. Biochemistry
1980, 19, 2895-2901.
Ack n ow led gm en t. This research was supported by
the National Institutes of Health (R.F.P.). We acknowl-
edge the guidance provided by preliminary kinetics
experiments with 3a carried out by Dr. Yang Xu.
J O980564+