J. Med. Chem. 1984,27, 1672-1676
1672
Trifluoroacetyl-c~-~-aspartyl-3,4-dinitroanilide (21). The 3,4-dinitroaniline has been obtained according to Nielsen et a1.l' The title compound was prepared according to method B yield D % ] . [ -58.5' (c 1,MeOH); 27%; TLC R (A) 0.47; mp 180-182 NMR (Mezdo-ds) 6 11.25 (s, 1H, NH Ar), 10.00 (m, 1H , NH,), 8.40-7.87 (m, 3 H, Ar), 4.83 (m, 1 H, CH), 2.90 (m, 2 H, CH3). Anal. (ClZH9N4O3F3) C, H, N. Trifluoroacetyl-c~-~-aspartyl-2,3-dinitroanilide (22). The 2,3-dinitroaniline was prepared according to Nielsen et al." The title compound was prepared according to method B: yield 15%; TLC Rf (A) 0.37; mp 182-185 "C; []."D -88.8 "C (C 1, MeOH); NMR (MezSO-d6) 6 10.03 (m, 2 H, NH,, NH Ar), 8.15-7.90 (m, 3 H, Ar), 4.93 (m, 1 H, CH), 2.90 (m, 2 H, CHZ). Anal. (Cl2H9N4O9F3)C, H, N. Trifluoroacetyl-cu-~-aspartyl-2-nitro-4-chloroanilide (23). This compound was prepared according to method B: yield 24%; "Lc R (A) 0.39; mp 15S161 'C; [a]%D -95.9' (c 1,MeOH); NMR (MezS6-d6)6 10.67 ( 8 , 1H, NH Ar), 9.97 (m, 1H, NH,), 8.04-7.77 (m, 3 H, Ar), 4.85 (m, 1 H, CH), 2.83 (m, 2 H, CHz). Anal. (CizHgN30&1F3) C, H, N. Trifluoroacetyl-cu-~-aspartyl-2-chloro-4-nitroanilide (24). This compound was prepared according to method B: yield 23%; "Lc R (A) 0.38; mp 162-164 "C; [.l2"D -69.5" (c 1,MeOH); NMR (Mezs6-d6)6 10.10 (m, 2 H, NH,, NH Ar), 8.33-8.20 (m, 3 H, Ar), 5.03 (m, 1H, CH), 2.85 (m, 2 H, CH,). Anal. (C12H9N30&1F3) C, H, N. Trifluoroacetyl-c~-~-aspartyl-3-methyl-4-nitroanilide (25). The 3-methyl-4-nitroaniline was prepared according to Wibaut.18
"c;
The title compound was prepared according to method B: yield -55.9 (c 1, MeOH); 24%; TLC Rf (A) 0.44;mp 167-169 "c; [.]%D NMR (MezSO-d,) 6 10.71 (s, 1 H, NH Ar), 9.88 (d, 1 H, NH,), 8.10-7.55 (m, 3 H, Ar), 4.83 (m, 1 H, CH), 2.85 (m, 2 H, CHz), 2.53 (s, 3 H, CH3). Anal. (C13H12N306F3) C, H, N.
Acknowledgment. We thank Joseph Tsang and Judd Berman for helpful discussions. We thank the National Institute of Denatl Research for their support of this research through Grant DE 05476. Registry No. 1, 41696-59-7; 2, 41566-94-3; 3, 92398-66-8; 4, 48193-99-3; 5, 39219-30-2; 6, 92398-67-9; 7, 92398-68-0; 8, 61980-46-9; 9, 92398-69-1; 10, 92398-70-4; 11, 41567-05-9; 12, 92398-71-5; 13, 92398-72-6; 14, 41567-06-0; 15, 41567-00-4; 16, 92398-73-7; 18, 92398-74-8; 19, 92398-75-9; 20, 92398-76-0; 21, 92398-77-1; 22, 92398-78-2; 23, 92398-79-3; 24, 92398-80-6; 25, 92398-81-7;N-(tert-butyloxycarbony1)-L-aspartic acid P-tert-butyl ester, 1676-90-0;trifluoroacetic anhydride, 407-25-0; 4-nitro-Nmethylaniline, 100-15-2;3,4-dinitroaniline, 610-41-3; 2,3-dinitroaniline, 602-03-9; 3-methyl-4-nitroaniline, 611-05-2; m-chloroaniline, 108-42-9; o-chloroaniline, 95-51-2; m-cyanoaniline, 2237-30-1; o-cyanoaniline, 1885-29-6; m-nitroaniline, 99-09-2; o-nitroaniline,88-74-4 p-toluidine, 106-49-0; m-toluidine, 108-44-1; o-toluidine, 95-53-4; 2-methyl-4-nitroaniline, 99-52-5; 2-methyl3-nitroaniline, 603-83-8; 2-rnethyl-B-nitroaniline, 99-55-8; 4chloro-2-nitroaniline, 89-63-4; 2-chloro-4-nitroaniline, 121-87-9. (18) J. P. Wibaut, Reel. Trav. Chim. Pays-Bas, 32, 287 (1913). (19) We repeated the synthesis of the o-chloroanilide4, which La-
(17) A. T. Nielsen, R. T. Atkins, W. P. Norris, C.L. Coon, and M. E. Sitzmann, J. Org. Chem., 45, 2341 (1980).
pidus and Sweeney described as "not sweet", and found this compound to be slightly bitter.
A 'H NMR Study of the Interactions and Conformations of Rationally Designed Brodimoprim Analogues in Complexes with Lactobacillus casei Dihydrofolate Reductase B. Birdsall,? J. Feeney,*vt C. Pascual,? G. C. K. Roberts,+I. Kompis,* R. L. Then,* K. Muller,* and A. Kroehns Physical Biochemistry Division, National Institute for Medical Research, Mill Hill, London NW7 IAA, England, and F. Hoffman-La Roche & Co., CH 4002, Bade, Switzerland, and Roche Products Ltd., Welwyn, England. Received March 20, 1984
A consideration of the detailed structural information available from X-ray crystallographic and NMR studies on complexes of dihydrofolate reductase with inhibitors has led to the design of trimethoprim analogues with improved binding properties. Computer graphic techniques have been used to predict which substituent groups were required a t the 3'-0 position of brodimoprim (2,4-diamino-5-(3,5-dimethoxy-4-bromobenzyl)pyrimidine) to make additional interactions with the enzyme. NMR spectroscopy provided a convenient method of assessing if the analogues were binding in the predicted manner. On the basis of this approach, the C4,CG-dicarboxylicacid analogue IX was designed to interact with Arg-57 and His-28 in the enzyme, and this analogue was found to bind 3 orders of magnitude more tightly than the parent brodimoprim.
The antibacterial drug trimethoprim (2,4-diamino-5(3,4,5-trimethoxybenzyl)pyrimidine)acts by selectively inhibiting dihydrofolate reductase in bacterial cells. In the past, many trimethoprim analogues have been investigated in attempts to find inhibitors that are either more selective or more active against resistant strains. With the recent availability of detailed structural information on complexes of dihydrofolate reductase with inhibitors from both X-ray crystallography1-" and NMR spe~troscopy,~~~ we now have a framework for designing trimethoprim analogues with modified binding characteristics. These techniques also provide methods of monitoring the complexes formed with new inhibitors to assess whether or not they are binding in the predicted manner. Kuyper and co-workers' have recently used this structural approach to design a series f
National Institute for Medical Research.
* F. Hofffman-La Roche & Co.
8 Roche Products Ltd.
0022-2623/84/1827-1672$01.50/0
of trimethoprim analogues with aliphatic a-carboxylic acid substituents arranged to interact favorably with the con-
s. T.; Hamlin, R.; Xuong, N.; Kraut, J.; Poe, M.; Williams, M.; Hoogsteen, K. Science 1977,197,452. (2) Matthews, D. A,; Alden, R. A.; Bolin, J. T.; Filman, D. J.; Freer, S. T.; Hamlin, R.; Hol, W. G. J.; Kisliuk, R. L.; Pastore, E. J.; Plante, L. T.; Xuong, N.; Kraut, J. J. Bid. Chem. 1978, (1) Matthews, D. A,; Alden, R. A.; Bolin, J. T.; Freer,
253,6946. (3) Bolin, J. T.; Filman, D. J.; Matthews, D. A.; Hamlin, R. C.; Kraut, J. J.Biol. Chem. 1982, 257, 13650. (4) Matthews, D. A.; Alden, R. A.; Freer, S. T.; Xuong, N.; Kraut, J. J.B i d . Chem. 1979, 254, 4144. (5) Cayley, P. J.; Albrand, J. P.; Feeney, J.; Roberts, G. C. K.; Piper, E. A.; Burgen, A. S. V. Biochemistry 1979, 18, 3887. (6) Birdsall, B.; Roberts, G. C. K.; Feeney, J.; Dann, J. G.;Burgen, A. S. V. Biochemistry 1983, 22, 5597. (7) Kuyper, L. F.; Roth, B.; Baccanari, D. P.; Ferone, R.; Beddell, C. R.; Champness, J. N.; Stammers, D. K.; Dann, J. G.; Norrington, F. E. A,; Baker, D. J.; Goodford, P. J. J.Med. Chem. 1982,25, 1120.
0 1984 American Chemical Society
Brodimoprim Analogues
Journal of Medicinal Chemistry, 1984, Vol. 27, No. 12 1673
Table I, Inhibition Constants (Ki) and His-28 pK Values for Complexes of L. casei Dihydrofolate Reductase with Brodimoprim and Its Derivatives (I-IX)g
glass electrode: the notation pH* denotes a pH meter reading uncorrected for deuterium isotope effects. Transfer of saturation experiments, used to detect the resonances of protons in bound ligands, were carried out by using selective irradiations a t 20-Hz intervals over the frequency range of interest as described previou~ly.~ The irradiation was applied for 1.0 s and then gated off before applying a 90' observing pulse. Saturation of a nucleus of the bound ligand on irradiation a t its resonance frequency results in a decrease in intensity of the signal from the corresponding nucleus of the free ligand if the exchange rate of the latter is faster than its relaxation rate. Dihydrofolate reductase activity was determined by using a spectrophotometric assay by measuring the decrease in absorbance a t 340 nm a t 37 "C.A microcomputerized Kontron Uvicon 810 spectrophotometer was employed, and reaction rates were either traced on the recorder (usually for 5-10 min) or directly computed over the time intervals in which the reaction proceeded in a linear fashion, as judged by correlation coefficients close to 1.0. Ki values were determined according to Henderson13 and as described in detail by Baccanari and Joyner14 with the following modifications of the assay: the final volume of 2 mL contained 30 WMdihydrofolate, 55 WMNADPH, approximately 0.6 nM enzyme and inhibitor, if required. Enzyme was preincubated with inhibitor and NADPH and the reaction started by diluting aliquots of this mixture into the cuvette, containing buffer, NADPH, and dihydrofolate. Imidazole hydrochloride (100 mM), pH 7.0, containing 2 mM ascorbate and 2 mM EDTA was used as the buffer, and all reagents were dissolved in this buffer. The Km of dihydrofolate used to calculate Ki was determined to be 2.2 fiM under these conditions. The data were evaluated by using the equation It/(l- ui/uo) = K i ( l S/Km)uo/vi+ E,, where It is the total inhibitor concentration, vo is velocity of the uninhibited reaction, vi is the velocity of inhibited reaction, S is the concentration of competing substrate, and Et is the total enzyme con~entrati0n.l~ The early molecular modeling was carried out on a PDP11/40 host computer interfaced to a Megatek 7000 display processor and later molecular modelling on a VAX-11/780 host computer with an Evans Sutherland Color M P S and a Tektronix-4113/4662 color terminal system using software developed in the Roche research laboratories.
R
comDd
K:.nM DK His-28
R
IV o-w-co2
V VI VI1 VI11
O-COzH
0 -COzH
0-Co2H
11.3 4.1
6.80 6.71
1.7
7.12
5.3
6.71
H
0.9 0.2 0.4 0.6
6.80
