Design, synthesis, and evaluation of the first mechanism-based inhibitor of glucosamine 6-phosphate synthase

Full text of DOI:10.1021/ja970254t

5748
J. Am. Chem. Soc. 1997, 119, 5748-5749
Scheme 1. Putative Three-Step Mechanism of GlmS
Inactivation by Compound 1
Design, Synthesis, and Evaluation of the First
Mechanism-Based Inhibitor of Glucosamine
6-Phosphate Synthase
Fre´de´ric Massie`re, Marie-Ange Badet-Denisot,
Lo¨ıc Rene´, and Bernard Badet*
Institut de Chimie des Substances Naturelles
Centre National de la Recherche Scientifique
91198 Gif-sur-YVette cedex, France
ReceiVed January 27, 1997
Scheme 2. Synthesis of Compound 1a
Glucosamine 6-phosphate synthase (GlmS, EC 2.6.1.16)
catalyzes the transfer of NH2 from the amide group of
L-glutamine (glutamine dependent amidotransferase) to D-
fructose 6-phosphate (6P).1 The participation of the N-terminal
cysteine residue in the catalysis and the structure of its glutamine
binding domain2 definitely established GlmS as a member of
the N-terminal nucleophile(Ntn)-hydrolase superfamily.3 The
formation of glucosamine-6P catalyzed by GlmS is a key step
in the biosynthesis of bacterial peptidoglycan and fungal chitin.
Therefore, GlmS has been considered as an interesting thera-
peutic target. Glutamine site-directed inhibitors of GlmS
actually display good in Vitro antibacterial and antifungal
activities when incorporated in oligopeptides.4 Most of these
inhibitors are affinity labels, i.e., glutamine analogs incorporating
one electrophilic function (halide,4fg,5 R-keto epoxide,6 R,â-
unsaturated carbonyl derivatives4) that interact irreversibly with
residue Cys1 of GlmS. The need to find more specific inhibitors
prompted us to design glutamine derivatives bearing a latent
electrophilic function. As a first result of this novel approach,
we present in this paper the first mechanism-based inhibitor,
L-γ-glutamyl-2-[((p-difluoromethyl)phenyl)thio]glycine, referred
to as compound 1.
a Conditions: (a) 2-methylpropan-2-ol, DCC, DMAP (cat), dioxane,
78%; (b) tert-butyl glyoxylate, DMF, cyclohexane, reflux, 89%; (c)
acetic anhydride, pyridine, 79%; (d) 4-mercaptobenzaldehyde, triethyl-
amine, DMF, 95%; (e) DAST, dichloromethane, 39%; (f) 20%
trifluoroacetic acid-dichloromethane, 61%.
step a) should lead to the formation of glutamate and N-
deacylated R-arylthioglycine which is known to be highly
unstable.9 Its decomposition (step b) would generate ammonia
and glyoxylate on the one hand as well as HF and 4-thioquinone
fluoromethide on the other hand. This latter powerful electro-
phile would then react with any active-site nucleophilic residue
(step c), resulting in enzyme inactivation. Although previous
approaches have been based on the generation of quinone
methide or quinonimine methide10,11 to achieve phosphatase10d-f
or elastase10c,11 inhibition, this is the first report of utilization
of 2-heteroatom-substituted glycines in the inactivation of a
glutamine-hydrolyzing enzyme.
Compound 1 was prepared according to Scheme 2. Con-
densation of suitably protected glutamine and tert-butyl gly-
oxylate12 afforded the 2-hydroxyglycine 2 in 90% yield, as
expected from earlier studies performed with simple amides.13
After conversion of 2 into the acetoxy derivative, nucleophilic
substitution with freshly prepared 4-mercaptobenzaldehyde14
afforded N-(γ-glutamyl)-2-(arylthio)glycine 3 (76%). This
aldehyde was converted with DAST15 into the difluoro deriva-
tive and deprotected by trifluoroacetic acid to give compound
1 as an undefined mixture of epimers16 in 12.4% overall yield
from Boc-L-glutamine. Attempts to prepare the 2-(phenoxy)-
glycine and the 2-(anilino)glycine analogs failed because of the
instability of these products. Compound 1 which turned out to
be perfectly stable under the conditions of enzyme assay even
The good inhibitory properties of N3-fumaroyl-2,3-diamino-
propionate derivatives7 had previously shown that the glutamine
site of GlmS is large enough to accommodate a bulky group
that would be linked to the N5-amide nitrogen of substrate
glutamine. During the course of our investigations in the
chemistry of R-heteroatom-substituted glycines,8 compound 1
was designed as a possible mechanism-based GlmS inhibitor.
Enzyme-catalyzed hydrolysis of its peptide bond (Scheme 1,
* Author to whom correspondence should be addressed at the follow-
ing: phone: 33 01 69 82 31 06; fax 33 01 69 07 72 47; e-mail:
badet@icsn.cnrs-gif.fr.
(1) Badet-Denisot, M.-A.; Rene´, L.; Badet, B. Bull. Soc. Chim. Fr. 1993,
130, 249-255.
(2) Isupov, M. N.; Obmolova, G.; Butterworth, S.; Badet-Denisot, M.-
A.; Badet, B.; Polikarpov, I.; Littlechild, J. A.; Teplyakov, A. Structure
1996, 4, 801-810.
(3) Brannigan, J. A.; Dodson, G.; Duggleby, H. J.; Moody, P. C. E.;
Smith, J. L.; Tomchick, D. R.; Murzin, A. G. Nature 1995, 278, 416-419.
(4) (a) Kenig, M.; Abraham, E. P. J. Gen. Microbiol. 1976, 94, 37-45.
(b) Chmara, H. J. Gen. Microbiol. 1985, 131, 265-271. (c) Milewski, S.;
Chmara, H.; Andruszkiewicz, R.; Borowski, E.; Zaremba, M.; Borowski,
J. Drugs Exp. Clin. Res. 1988, 14, 461-465. (d) Shoji, J.; Hindo, H.;
Sakazaki, R.; Kato, T.; Hattori, T.; Matsumoto, K.; Tawara, K.; Kikuchi,
J.; Terui, Y. J. Antibiot. 1989, 42, 869-874. (e) Andruszkiewicz, R.;
Milewski, S.; Zeniawa, T.; Borowski, E. J. Med. Chem. 1990, 33, 132-
135. (f) Andruszkiewicz, R.; Chmara, H.; Milewski, S.; Zeniawa T.;
Borowski, E. J. Med. Chem. 1990, 33, 2755-2759. (g) Andruszkiewicz,
R.; Chmara, H.; Zeniawa, T.; Borowski, E. Eur. J. Med. Chem. 1994, 29,
61-67.
(5) Auvin, S.; Cochet, O.; Kucharczyk, N.; Le Goffic, F.; Badet, B.
Bioorg. Chem. 1991, 19, 143-151.
(6) (a) Chmara, H.; Za¨hner, H. Biochim. Biophys. Acta 1984, 787, 45-
52. (b) Chmara, H.; Za¨hner, H.; Borowski, E. J. Antibiot. 1984, 37, 1038-
1043. (c) Baldwin, J. E.; Adlington, R. M.; Mitchell, M. B. J. Chem. Soc.,
Chem. Commun. 1993, 1332-1335. (d) Crossley, M. J.; Stamford, A. W.
Aust. J. Chem. 1993, 46, 1443-1446.
(9) (a) Kingsbury, W. D.; Boehm, J. C.; Mehta, R. J.; Grappel, S. F.;
Gilvarg, C. J. Med. Chem. 1984, 27, 1447-1451. (b) Kingsbury, W. D.;
Boehm, J. C.; Perry, D.; Gilvarg, C. Proc. Natl. Acad. Sci. U.S.A. 1984,
81, 4573-4576. (c) Kraus, J.-L.; Attardo, G. Eur. J. Med. Chem. 1992, 27,
19-26.
(10) (a) Walsh, C. Tetrahedron 1982, 38, 871-909. (b) Abeles, R. H.
Chem. Eng. News 1983, Sept. 19, 48-56. (c) Wakselman, M. NouV. J.
Chim. 1983, 7, 439-447. (d) Myers, J. K.; Widlanski, T. S. Science 1993,
262, 1451-1453. (e) Born, T. L.; Myers, J. K.; Widlanski, T. S.; Rusnak,
F. J. Biol. Chem. 1995, 270, 25651-25655. (f) Stowell, J. K.; Widlanski,
T. S.; Kutateladze, T. G.; Raines, R. T. J. Org. Chem. 1995, 60, 6930-
6936.
(11) Vergely, I.; Boggetto, N.; Okochi, V.; Golpayedani, S.; Reboud-
Ravaux, M.; Kobaiter, R.; Joyeau, R.; Wakselman, M. Eur. J. Med. Chem.
1995, 30, 199-208.
(12) Carpino, L. A. J. Org. Chem. 1964, 29, 2820-2824.
(13) (a) Zoller, U.; Ben-Ishai, D. Tetrahedron 1975, 31, 863-866. (b)
Schouteeten, A.; Christidis, Y.; Mattioda, G. Bull. Soc. Chim. Fr. 1978,
248-254.
(7) (a) Chmara, H.; Andruszkiewicz, R.; Borowski, E. Biochim. Biophys.
Acta 1986, 870, 357-366. (b) Badet, B.; Vermoote, P.; Le Goffic, F.
Biochemistry 1988, 27, 2282-2287. (c) Kucharczyk, N.; Denisot, M.-A.;
Le Goffic, F.; Badet, B. Biochemistry 1990, 29, 3668-3676.
(8) Qasmi, D.; Rene, L.; Badet, B. Tetrahedron Lett. 1993, 34, 3861-
3862.
(14) Young, R. N.; Gauthier, J.-Y.; Coombs, W. Tetrahedron Lett. 1984,
25, 1753-1756.
(15) Diethylaminosulfur trifluoride. For a review, see: Hudlicky, M. Org.
React. 1988, 35, 513-637.
S0002-7863(97)00254-0 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 24, 1997 5749
production was also detected by lactate dehydrogenase and
NADH,21 but the exact amount could not be quantified because
of excessive background noise.22 The 4 equiv (5.25-1) of
4-thioquinone methide released in the medium could add a water
molecule (1,6-addition) to generate an unstable R-fluoro-R-
(hydroxymethyl)phenyl thiol precursor of 4-mercaptobenzalde-
hyde. A new 214 nm absorbing compound was indeed detected
in the HPLC profile of the inactivation mixture, but the minute
quantity formed precluded its identification.
Although the four observations above meet the criteria
required for suicide inhibition,19 GlmS inactivation was pre-
vented by nucleophiles such as dithiothreitol (5 mM), cysteine
(10 mM) or tetra-n-butylammonium phenylthiolate (1.25 mM)
which are not competitive inhibitors of the glutamine site. The
possibility of enzyme reactivation by breaking an enzyme/
inhibitor species adduct was discarded since extensive dialysis
of inactivated enzyme in the presence of 5 mM dithiothreitol
(5 mM) did not result in regain of activity. This protective effect
may then suggest that the species responsible for inactivation
was released into the solution (where it can be quenched by
nucleophiles) before alkylating the enzyme.19
Figure 1. Time-dependent inactivation of GlmS. GlmS (13 µg) was
incubated at 20 °C and pH 7.2 in 100 µL of 100 mM KPO4 containing
10 mM fructose-6P and 1 at varying concentrations. At different time
points, aliquots (5 µL) were withdrawn and the reaction was stopped
by 200-fold dilution at 0 °C in 100 mM KPO4 at pH 7.2 containing
both saturating substrates (10 mM fructose-6P; 6 mM glutamine). Then
residual activity is assayed at 37 °C by the described procedure.7c The
enzyme remained 100% active during 120 min of incubation in the
presence of fructose-6P and in the absence of 1. Curves show time-
dependent inhibition (semilog plot) with ( ) 11.6, ( ) 14.5, ( ) 17.4,
and ( ) 22.6 mM. Inset: double reciprocal plot of similar inactivation
experiments performed with 1 at 14.5, 16.0, 17.4, 20.3, and 22.6 mM.
Above 25 mM, precipitation of 1 occurred. The data points represent
the average of three determinations.
However, this assumption would be in contradiction with the
absence of a lag time before the onset of inhibition and with the
low partition ratio of about 7 ((5.25/0.66) - 1, see above) which
was deduced from the quantitation of glutamate formed during
inactivation. It is then possible that in the protection experi-
ments the thiols enter the active site and compete with the active-
site nucleophiles for the activated thioquinone methide. The
design of a bulky thiol unable to enter the glutamine site of the
native enzyme would certainly be of interest to elucidate this
phenomenon. A similar behavior was actually observed in the
inhibition of â-glucosidases by difluoromethyl arylglucosides.23
Although the structure of native GlmS has not been determined
yet, the active site has been consistently located at the bottom
of a cleft widely exposed to solvent on the recently reported
three-dimensional structure of the glutamine binding domain.2
The nature of the adduct between 1 and GlmS remains to be
elucidated. The participation of cysteine1 in this adduct is
highly unlikely since it is assumed to be still engaged in the
γ-glutamyl thiol ester resulting from inhibitor processing by the
catalytic thiol (Figure 1, step a).1 The observed absorption at
308 nm, a wavelength higher than the typical value (∼260 nm)
of phenylthiols or -thiolates24 but lower than that reported for
a doubly conjugated thioketone25 (330 nm), might reveal further
evolution of the methyl-substituted 4-fluoromethylphenyl thiol
adduct resulting from nucleophilic 1,6-addition26 on the thio-
quinone methide formed according to Scheme 1. Since elec-
trospray mass spectroscopy failed to give reproducible results
in the analysis of the inactive protein, further investigations are
required to understand the molecular details of this inhibition.
Despite moderate efficiency, further use of glutamine deriva-
tives releasing reactive leaving groups may constitute a general
approach to the design of irreversible inhibitors of glutamine-
hydrolyzing enzymes. The application of this concept to
protease inhibition is currently under investigation.
in the presence of 20 mM dithiothreitol was tested on Escheri-
chia coli GlmS purified as described.17
The conclusion that compound 1 behaved as a mechanism-
based inhibitor was drawn from the following results. (i) When
GlmS was incubated with 1 at concentrations 11-23 mM and
saturating D-fructose-6P, a first-order loss of enzyme activity
was observed (Figure 1) over a period of about 1.5 h.18 This
inactivation process displayed saturation kinetics, as shown from
the positive y-intercept in the double reciprocal plot of the
apparent inactivation rate Versus inhibitor concentration (Figure
1, inset). Therefore, reversible enzyme/inhibitor complex
formation is much faster than the inactivation step itself.19
Values found for inactivation rate constant (kinact) and irreversible
inhibition constant (Kirr) are 0.054 ( 0.002 min-1 and 35.8 (
0.1 mM, respectively. (ii) Extensive dialysis of inactivated
enzyme did not result in any recovery of activity, demonstrating
that the inhibition is an irreversible process. Furthermore, the
ultraviolet spectrum (not shown) of this dialyzed dead enzyme
was different from that of the native one as shown by appearence
of a broad peak at λmax ) 308 nm. (iii) The presence of
glutamate semialdehyde (0.27 mM), a potent glutamine site-
directed reversible GlmS inhibitor (Ki ) 55 µM),20 totally
protected the enzyme from inactivation. This demonstrates that
inactivator 1 is directed against GlmS glutamine site. (iv) The
inactivation resulted from inhibitor processing by the enzyme:
glutamate dehydrogenase analysis of a GlmS (4.0 µM) solution
inactivated to 66% by compound 1 (28 mM) revealed the
formation of 21 ( 2 µM glutamate that is a 5.25-fold excess
over the protein. Consistently, a simultaneous glyoxylic acid
(16) It has never been possible to separate at any stage diastereomers
on reverse-phase HPLC. Nonetheless, all compounds gave satisfactory
analytical data. For compound 1: 1H NMR (300 MHz, DMSO-d6) δ 2.02
(m, 2H, â-CH2), 2.39 (m, 2H, γ-CH2), 3.70 (t, J ) 6.4 Hz, 1H, R-CH),
5.74 (d, J ) 8.8 Hz, 1H, 2-CH), 7.09 (t, J ) 55 Hz, 1H, CHF2), 7.63 (m,
4H, CHarom), 8.92 (d, J ) 8.8 Hz, 1H, 2-NH); 13C NMR (75 MHz, DMSO-
d6) δ 26.4 (â-CH2), 30.9 (γ-CH2), 52.3 (R-CH), 57.4 (2-CH), 114.6 (CHF2
(t, JCF ) 234 Hz)), 126.0 (3′-CH and 5′-CH), 131.2 (2′-CH and 6′-CH),
132.4 (4′-C), 137.1 (1′-C), 168.9 (CONH), 170.5 and 170.9 (2 COOH);
UV-vis (c 10-4 M, H2O) λmax 262 (3400) nm; MS (FAB+, glycerol +
HCl) m/e calcd for C14H16F2N2S + H+ 363.0826, found 363.0808; 363
Acknowledgment. Financial support from Roussel Uclaf to one
of us (F.M.) is gratefully acknowledged.
JA970254T
(21) Pineda, M.; Piedras, P.; Ca´rdenas, J. Anal. Biochem. 1994, 222,
450-455.
(22) The enzyme activity remained unaffected by glyoxylate concentra-
tions within that range (20 µM). Under these conditions it exhibits a IC50
of 3 mM.
(23) Halazy, S.; Berges, V.; Ehrhard, A.; Danzin, C. Bioorg. Chem. 1990,
18, 330-344.
(24) (a) Jones, R. A.; Katritzky, A. R. J. Chem. Soc. 1958, 3610-3613.
(b) Albert, A.; Barlin, G. B. J. Chem. Soc. 1959, 2384-2396.
(25) (a) Metzner, P.; Vialle, J. Bull. Soc. Chim. Fr. 1972, 3138-3145.
(b) Barton, D. H. R.; Choi, L. S. L.; Hesse, R. H.; Pechet, M. M.; Wilshire,
C. J. Chem. Soc., Perkin Trans. 1 1979, 1166-1175.
(26) Wagner, H.-U.; Gompper, R. Quinone methides. In The chemistry
of the quinonoid compounds; Patai, S., Ed.; J. Wiley and Sons: New York,
1974; Chapter 18.
(100, MH+); [R]D + 3.4° (c 1.0, MeOH).
20
(17) Dutka-Malen, S.; Mazodier, P.; Badet, B. Biochimie 1988, 70, 287-
290.
(18) In the absence of fructose-6P, inactivation occurred at a slower rate.
GlmS incubation with 20.3 mM 1 under the conditions reported in Figure
1 resulted in a 28% loss in activity compared to 70% in the presence of
fructose-6P.
(19) Silverman, R. B. Mechanism-based enzyme inactiVation; CRC
Press: Boca Raton, FL, 1988; Vol. 1.
(20) Bearne, S. L.; Wolfenden, R. Biochemistry 1995, 34, 11515-11520.