May 196s II~HEVEKSIULE ENZYME INHIBITORS. CXVII

Irreversible Enzyme Inhibitors. CXT'II.')2 Hydrophobic Bonding to Dihydrofolic Reductase. XI.3 Comparison of the Enzyme from

4s3

Walker 256, Rat Liver, and L1210 Mouse Leukemia

B. R. BAKER

Thirty-one 4,6-dianiino-l,2-dihydro-s-t~riazines with varying substituents a t the 1 and 2 positioiis have been compared as inhibitors of the dihydrofolic reductase from Walker 256 rat tiinior, rat, liver, L1210 'FR8 mouse leitkemia, and pigeon liver. It was demonstrated t,hat a 1-phenyl groiip was complexed almost equally effectively to all four enzymes and that the interaction of the phenyl group with the enzyme was by hydrophobic bonding; similarly, all forir eiizyme,s showed ilboiit the same amoinit of hydrophobic bonding by a 1-(n-butyl) group. Dif- ferences in hydrocarbon interactioiis among the four enzymes were seen with larger hydrocarbon groups. The greatest differences in binding to the rat tumor and rat liver enzymes were seen wit.h a 1-phenylbutyl group and l-[p-(2,4-dichlorophenylb~ityl)pheiiyl] groiip, the tiimor enzyme being iiihibit,ed 100- and 40-fold better, reapec- tively, than the rat liver enzyme. In no case was a compound strikingly more effective on the L1210 eiizyme than the other three enzymes.

One of the key enzymes in cellular reproduction is dihydrofolic reductase which can reduce dihydrofolate and usually folic acid to the cofact'or form, tetrahydro- folate; t'he lat'ter is then involved in fifteen enzyme reactions catalyzing one-carbon transfer reactions including purine and pyrimidine bio~ynthesis.~ After the discovery of a hydrophobic bonding region on di- hydrofolic reductase,j an intense study of optimum conformation for binding to this region and the relative location on the enzyme surface xyas pursued.6 Strong evidence was found that this hydrophobic bonding region was just adjacent to the active site and is lo- cated near the 4 or S posit'ion of dihydrofolate when it resides on the enzyme.6 Since this hydrophobic region is not part of the active site, evolutionary changes of amino acids in this region would more easily occur without lethality than similar changes inside the active site.' Two studies have been previously made on species differences in reversible binding to t'he hydro- phobic bonding region, the first on Escherichia coli B LIS. pigeon live? and the second on T2 phage induced enzyme L I S . E . coli B vs. pigeon l i ~ e r . ~ , ~ I n this paper are described the studies on comparison of the hydro- phobic bonding region from four vertebrate sources, namely, Walker 256 tumor and liver from the rat, 1,1210 mouse leukemia, and pigeon liver.

(1) This work mas generously supported by Grant CX-08695 from the National Cancer Institute, U. S. Public Health Service.

(2) For the previous paper in this series see B. R. Baker, P. C. Huang, and R. B. hleyer. Sr., J . .Wed. Chem., 11, 475 (1968).

(3) For the previous paper in this series see B. R. Baker, ibid. , 10, 912 (1967).

(4) T. H. Jukes and H. P. Broquist in "Metabolic Inhibitors," R. 11. Hochster and J. H . Quastel, Ed., Academic Press Inc., New York, N. Y., 1963, pp 481-534.

(a) B. R. Baker, B.-T. Ho, and D. V. Santi, J . Pharm. Sci., 64, 1416 (lY65).

(6) For a review on the mode of binding to dihydrofolic reductase see B. R. Baker, "Design of Active-Site-Directed Irreversible Enzyme Inhibitors. The Organic Chemistry of the Enzymic Active-Site." John Wiley and Sons, Inc., New York, N. Y . , 1967, Chapter X.

( 7 ) For reviews on evolutionary changes in enzymes see (a) ref 6, Chapter IS, and (h ) V. Bryson and H. J. Vogel, Ed., "Evolving Genes and Proteins," Academic Press Inc., New York, N. Y . , 1965.

(8 ) B. R. Baker and B.-T. Ho. J . Pharm. Sci., 66, 470 (1966). (9) Further examples of species differences in the ability of dihydrofolic

reductase t o bind 2,4-diaminoheterocycles tha t have varying hydrophobic eroiipslo have been collated hy Hitchings and Burchall; see (a) G. H. Hitchings and J. J. Burchall, Aduan. Enzymol., 27, 417 (1965), and (b) J. J. Burchall a n d G. H. Hitchings. Mol. Pharmacol., 1, 126 (1965).

(10) I:or a similar study on the dihydrofolio reductase from Trypanosomu equiperdum see J. J . Jaffee and J. J . McCormack. Jr . , ib id . , 3, 359 (1967).

Enzyme Results.-The inhibition results with 31 1-substituted dihydro-s-triazines on the four sources of dihydrofolic reductase are collated in Tables 1-111. It is again noteworthy that the 1-methyl-s-triazine (1) (Table I) binds almost the same t o all the enzymes within a factor of 2.5 which also brackets the previous results with the enzyme induced by T, phage and the E . coli B e r i~yrne ;~ these results further support the concept3z8 that the 4,6-diamino-1,2-dihydro-s-triazine ring system complexes within the active site in the region that normally complexes the pteridine of dihydro- folate, a region unlikely to have undergone any muta- tion without lethality.

The increment in binding between the 1-methyl (1) and the 1-n-butyl (2) substituent due to hydrophobic interaction varies between 110- and 260-fold, depending upon the enzyme source. Similarly, the hydrophobic bonding increment between 1-methyl (1) and 1-phenyl (5 ) varies between 260- and 670-fold, indicating similar

TABLE I IKHIBITION OF DIHYDROFOLIC REDUCTASE BY

~-ALICTL-, l - h Y L - , . ISD 1-PHENYL.ILKYL-S-TRIIZINES

YHC' S-R

SH,L,+H, CH,

- -pM concn for 5Ov/ , inhib"---. Walker Rat Pigeon

N O . R 256 liver L1210/FR8 liverh 1 CHa 31 39 78 i 4 2 n-CdHs 0 . 1 2 0 . 3 7 0 . 3 9 0 . 3 6 3 i-CsH1i 0.0061 0.078 0.070 0.058 4 n-C6Hn 0.014 0.053 0,062 0 . 1 4 6 CEHS 0 . 1 2 0 . 1 5 0 . 2 1 0.11 6 (CHz)3CsHb 0.016 0 . 0 1 8 0.053 0.028 7 ( C H Z ~ C ~ H S 0,00030 0 , 0 3 3 0.038 0.041 8 CEHICOOH-P 2200 1100 9 C E H I C O O H - ~ 19 18 68 110

10 CoHaCOOC~Ha-p 27 42 11 CaHhCN-p 7 . 2 14 9 . 1 8 . 0 12 CsHaCH&Ha+-p 6 . 3 3 . 7 14 10 13 C8H40CH3-m 0 . 6 7 0 . 7 8 0 . 7 9 0 . 5 4 14 CsHaCFx-m 0.017 0 . 0 2 5 0 .041 0,080 15 C6H1No?-rn 0.084 0.16 0 . 2 2 0 . O i l

All assays were performed with 6 /~31 dihydrofolate niid 90 p X TPKH in pH 7.4 Tris bitffer as previously descrihecL'6 The techiiicnl assistance of Sharon Lafler, Jean lleeder, and 1)iaiie Shea is acknowledged. Data previoiisly siiinmariaed.3

T \ 1 3 LE I11

binding; these diff ereiices should be compared with the much smaller 23-fold i~icremerit in binding between 1 :it id 5 on the E . coli e~izyine .~

By iricreasing the ?[-butyl group (2) to ii-octyl (4). :I 3-9-fold increment in binding ih seen, the higher in- crement with Wallier 23G enzyme and the lower in- crement with the pigeon liver enzyme; this incremcnt is :il)out the same for Wallier 236 :iiid the rat livcr ('tizj mc. 111 contrtist, termilia1 substitution of the ) I -

1)utj.l group (2) with phenyl (7) gives :I 400-fold incre- metit in binding with the rat tumor enzyme, but only :1 tenfold increment with the enzyme from rat liver and the other two sources; about a tenfold increment was wx \vith the T r phage induced : i d I,.. coli B enzymes.?

The branching of the /!-butyl group (2) to isoamyl (3) gives :L 20-fold increme~it in binding to the rat tumor enzyme, but only about a sixfold increment with the other three enzymes.

Thc compound in the series 1-7 showing the greatest specificity toward inhibition of the Walker 256 enzyme over the rat liver enzyme is the I)hen3.lbutyltrinzirie (7) where thc difference is 100-fold; the next most specific c~m~1)ourid is the isotimyltriazine with a 20-fold dif- ~ C ~ O I I C V t)etween rat tumor and liver. S o compound

showi t o caiibe :t 1.500-fold decreaqe in binding to t Ii(1

Iligeoii liver cmz>-nie;" qiniilarlj-, when the phcl i~ I group of 5 was replaced by the large flat 9-fluorcnori-L'- yl group (18), :LI~ 800-fold loss in binding occurrcd.'2 This hindrAiice to binding to the other thrce enzyme\ was le rs severe with 16, being only 150-240-fold; sinii- larlj-, with 18, the loss was only 4.5-fold with the 1,1210 crizynie :md 120-fold with the two rat tissue enzyme\. 1 [ore hindrance to binding to the pigeon liver cnxymc wa$ also seen with the m-phenyl substituent (17) oti 5 th:m with the other three enzymes; the loss i n h i d - ing to the pigeon liver enzyme was l2-fold, to tlic. J,1210 enzyme about twofold, but no cliaiige in hindiiig or- curred with the t\io rat enzymes. It w t ~ s I)rcviouslJ noted3 that the in-phenyl substituerit of 17 gnvc : L I I

increase in bindirig to the Ts phage induced and k'. tol i B enzymes.

substituc'iit\ (Table 11) on the 1-phenyl group of 5 were previously observed to give six and twofold increments, respec- tively, in hiriding t o t lie pigeon liver ciizymc. Thc, binding increment to the other three enzymes n'~.

The m-benzyl (19) l 2 arid p-benzyl (20)

(11) H I{. Baker and B -T Ho, .I. l i e t r i o c y c i . Chem , 2, 3 J 5 11062) (12) R It Baher and €3 -1'. I {<) , bid., 2 , 7 2 (lYb5).

larger, a 10-17-fold increment being observed with 19 and a 7-14-fold increment with 20. Longer phenyl- alkyl substituents such as phenylbutyl were observed13 to give even better binding than benzyl to the pigeon liver enzyme; the 2,4-dichlorophenylbutyl substituent on the ineta position (23) gave a 14-fold increment and 011 the para position (22) a 21-fold increment. There was considerable difference in binding to the four en- zymes with these two compounds (22, 23). The para compound (22) gave a 190-fold increment in binding to the rat tumor enzyme, but only a tenfold increment in binding to the rat liver enzyme, and no increment in binding to the L1210 enzyme. I n contrast, this sub- stituent at the meta position (23) gave only a three fold binding increment to that rat tumor enzyme, an eightfold increment with rat liver enzyme, a 14-fold increment with pigeon liver enzyme, but a 43-fold in- crement in binding to the 1,1210 enzyme. Little dif- ference in binding among the four enzymes was seen with the 3-chloro and 4-phenylbutyl substituents (24) on the 1-phenyl of 5. 41~0, little difference in binding among the four enzymes was observed with the 111-

chloro (25) and 3,4-dichloro (26) substituents, although the chlorines did increase binding.

The greatest difference in binding to the two rat enzymes with the compounds in Table I1 was observed with the p-(2,4-dichlorophenylbutyl) substituent of 22 where the tumor enzyme was inhibited 40-fold better. No compound in Table I1 was more than twice as ef- fective on rat liver enzyme than rat tumor enzyme. The pattern observed with the 1,1210 enzyme was quite similar to the pigeon and rat liver enzymes except for 22, which was tenfold more effective on the rat liver enzyme and 40-fold more effective on the pigeon liver enzyme; no compound in Table I1 was more than three- fold more effective on L1210 enzyme than on the two liver enzymes.

Introduction of larger groups at the 2 position of the dihydro-s-triazine system was observed to be more detri- mental in binding to the enzyme from pigeon liver than from E. coli B.378 Similar losses in binding to the three mammalian enzymes have now been observed (Table 111). Introduction of a p-acetamidophenyl substituent (27) at the 2 position of 25 gave a 22,000- fold loss in binding to the pigeon liver and L1210 enzymes; the loss in binding to the two rat enzymes was also large, being 3700-7700-fold. The loss in binding by the p-acetamidophenyl substituent (29) was less when the s-triazine was substituted with a 1- phenylbutyl group (7); the least effect was on the L1210 enzyme where only a threefold loss in binding occurred, but was most dramatic on the rat tumor enzyme where a 1700-fold loss in binding occurred. The 2-benzyl substituent (28) showed much less steric hindrance to binding than the 2-(p-acetamidophenylj substituent (27), 28 causing only a 7-27-fold loss in binding to the four enzymes, in contrast to the 3700- 22,000-fold loss with 27. Of the compounds 27-31 in Table I11 none showed more than a fourfold difference in binding to the four enzymes.

Conclusions.-It has been shown that the l-phenyl- s-triazine (5) has essentially the same ability to bind to the four vertebrate enzymes and that the 1-phenyl

(13) E. R. Baker, B.-T. Ho, and G . J. Lourens, J . Pharm. Sei., 66, 737 (1967).

group is complexed by hydrophobic interaction with these enzymes. However, substitution of additional nonpolar groups on the 1-phenyl moiety of 5 showed differences in binding to the four enzymes. 1Iany of these differences were due to a stronger hydrocarbon interaction to the rat tumor enzyme than the other enzymes.

The biggest differences in reversible inhibition of the enzymes from Walker 256 rat tumor and rat liver are observed with the 1-phenylbutyl (7) and the p - ( 2 , 4 - dichlorophenylbuty1)phenyl (22) substituents, being 100- and 40-fold, respectively. These differences are not sufficient to expect chemotherapeutic effectiveness on the tumor; note that the antibacterial agent, tri- methoprim (32), shows 50,000-70,000-fold more ef-

OCH, FH2 I I 2

"AN N@cH2qocHJ OCH,

32

fectiveness on bacterial dihydrofolic reductases than 011 mammalian dihydrofolic reductase^.^ Of the 31 com- pounds listed in Tables 1-111, 22 shows the greatest spread in effectiveness between the pigeon liver and E. coli B enzymes, being only 230-fold. Since tri- methoprim resulted from years of study in the pyrim- idine area by Hitchings' l a b ~ r a t o r y , ~ i t is clear that our limited studies on seeking a reversible inhibitor of dihydrofolic reductase that is 50,000-fold more effective on Walker 256 than rat liver enzyme has been mini- mal in comparison. From the evolutionary standpoint, it is unlikely that such large differences between the two rat enzymes will be found, though some additional attempts have been made.14 However, i t is clear that small differences exist in binding to the hydrophobic bonding region of dihydrofolic reductase from Walker 256 rat tumor, rat liver, and pigeon liver. It nas predicted3 that i t should be possible to greatly amplify these relatively small differences in hydrophobic bond- ing by use of the bridges principle of specificity15 with active-site-directed irreversible inhibitors.6v1b This pre- diction was first borne out with the dihydro-s-triazine (33) bearing a terminal sulfonyl fluoride group; even though differences in reversible binding were small, 33 was a rapid irreversible inhibitor of the pigeon liver di- hydrofolic reductase, but showed 110 inactivation of the

33, meta isomer 34, p a r a isomer

enzyme from Walker 256 or L1210.16 The para isomer showed rapid inactivation of the enzyme from Walker 256 and liver from the rat, L1210 mouse leukemia, and

(14) B. R. Baker and hl. .L Johnson, J . .Wed. Chem.. 11, 486 (1968);

(15) See ref 6, Chapter IX. (16) B. R. Baker and G . J. Lourens, J . .lied. Chem., 10, 1113 (19671, paper

paper CXVIII of this series.

CV of this series.

of nil-rninialian dihydrofolic reductases, but shomd selective irreversible inhibition; a 1 &lI concentr, d t ' io11

of 35 could rapidly inactivate the Walker 2-56 rat turnor ciizyme :ind tlic 1.1210/FKS mouse leukemia ctizynie biit did not illactivate the rat or mouse liver enLynie.:

N & u u - 0 0 mco 9 "2(N NH-

SO,F

35 at thiq conceutr:itioii.lj

Irreversible Enzyme Inhibitors. CXl*III.l ' Hydrophobic Bonding to Dihydrofolic Reductase. XII.' Further Comparisons with

the Enzynie from Walker 256 Rat Tumor and Rat Liver

I~ti\irtceii oi?ho-diit)stitiit et1 1-plien~-l-4,8-dinniii io-l~~-dih~~~o-s-t riiieiiiej h:tve kieeii nmisiired as iiihibiturs of the diliylrofolic reductase from Walker 256 rat tumor aiid rat liver: i io appreciable tiifferelice iii t~iiidiiig to the enzyme from the two soi i rces was foriiid. The 600-fold lo>, i i i biiidiiig nheii 4,6-diamiiio-2,2-dimethq-l-l,2- tiih?.dro-l-pheii)l-s-triazirie (4) is srihtitriteti with ail o-chloro groiip ( 5 ) raii he recoiiped ti>- frirther substittitioil of a 3-chloro groiip (16) or 4-pheiiylbLityl group (20).

In the previous paper of this series a comparison of the binding to the hydrophobic bonding region of the dihydrofolic reductase from Walker 256 rat tumor a d rat liver were made with 31 substituted 4,6-diarnino-1.2- dihydro-s-triazines. The largest difference in binding between the two eiizynies was observed3 with 1 and 2, the tuinor enzyme being inhibited better by 100- and 40-fold, respectively. Even though o d i o substituent.: 011 the 1-phenyl group of dihydro-s-triazines give a great loss in biological a ~ t i v i t y , ~ ~ ~ this area has no\v been further studied particularly since it was possible that

IVY x2 N-R

N H ~ ~ , , - ~ c H ~

CH,

the lowered activity could be recouped by additional substituents on the benzene ring5 arid it was possible that t'issue specificity might be uncovered. The re- sult,s of this study on binding of 3 and related coin- pounds to Walker 2 3 and liver dihydrofolic reduct'ase of the rat is the subject of this paper.

The inhibition of dihydrofolic reductase froin two sources with 14 ortlio-substituted 1-phenyl-s-triazines aiid ihree related compounds are collated in Table I. S o specificity in binding to the two enzymes n-as ob- served; nevertheless, some intoresting correlat'ions on

(1) This work ,vas penerorialy sripporteti 1)s Grant C.1-0869d from the

(2) For t he previous paper of this series, PPC B. R. Baker, .l. Iferl. Ctiern., Sational Cancer Institute, U. S . Public Iiealtli Service.

11, 48:+ (19681. . I:. I 'olpy, l ' r o c . Ani. . lisnc. C

O r g . C/ie?n., 21, 1 (1956). m. Sci . , 53, 1137 (1864).

(r)) P,. R. Baker, B.-'l. 1-10, and (;. J. Loiirens, zbid. . 56, 737 (19671, l ~ a l i f ~ i 1.XS.Y V I of tliis serira.

the ortho effects caii be made. For discussion purposes only the Walker 236 data will be used.

The o-chloro sub*tituciit of 5 causes 3 GOO-fold loss in hindiiig t o t he c~izynic; I his loss corresponds to 1 he t otiil increment iii 1-phenyl binding compared to 1- iiietli) I , ? indicating that the 1)henyl ring of 5 is now out-