used it1 this study: lh. liolaiid Iebrateenzymes; this observation along with comparative "inhibitor profiley" with the other compounds gives support to the possibility t'hat the TP phage DX.4 particle divergedfrom a higher form of life than bacteria.
Ilihydrofolic reductase is one of the key erizynies i n cdlular reproduction; the enzyme reduces dihydrofolic :wid (I) and usually folic acid to tetrahydrofolir acid, the cwfwtor form of the vitamin whirh is then in-
plexed to the enzyme;'" since t,he hydrophobic* hoiicliiig region is outside of the active-site, evolutionary changesI4 of amino acids in this region might hc (3spect'ed to occur more easily, without destroying t,lic function of the enzyme, than in the active site. ' I h t binding t'o the hydrophobic bonding region could tliffcr among species was demonstrated with dihydrofolic reductases isolated from Escherichia coli B arid pigcori liver furthermore, these diff ererices in hydrophohic. bonding could be corisiderably amplifiedR by use of active-site-directed irreversible inhibitors. 'j.'' When E.coli is infected by T,,.,, bacteriophages, 1110 Teation or stimulation of at least 12 enzymatic : es ocvxm,i8 including dihydrofolic reductase;'" t,lic dihydrofolic reduchse induced by the phage has Iwcn established to be a new enzyme genetically c.oritrollcd by the viral DNA arid not just, higher producotiori of enzyme coded by the bacterial DP\'A.18119 Furthermoro, when mammalian cells in tissue culture are infected hy such tumorigenic viruses as polyoma virus or SY40, the levels of some of the enzymes involvecl i i i IlS.1 synthesis :ire iiicre:rscd. including dihydrofoli(. rv;i5s*6
I L: sp:Lte of erizyniatitr reactions iiicluding The discovery purine Ltnd pyrimidine biosynthe dihydrofolic reof a hydrophobic bonding region ductasefi led to an intense study of the conformational arid its relative locaaspects of binding t'o this tion on the enzyme. Evidence has been that this hydrophobic bonding region is located near the 4 or S position of dihydrofolate (I) when it is comVIJlVCd i i i
(I)This work was generou3ly supported by Grant C.\-08696 from tlie National Cancer Institute, U. 9. Public Health Service. ( 2 ) Paper S('VI1 in tlie series on Irreversil,le Enzyme 1nliil)itorsJ anrl l'aiwr S o n tiydroplioiiic Ihntlinlr t o 1)illydrofolic IiedirctaPe.4 (:{I F o r the previous paper of this series see 13. R. Haker. d . C i i r m . K , l i i c . . in
lIrf~P.3,
( 4 ) I.'cir tlie urevii~iisuaper of this series see 13. R . Raker and G . .J. Louren*. . I . l'ii,lrm. S r i , , in press. i.51 T. 11. .Jukes and 11. P. I3roquist in "Metabolic Inhibitors," H. A l . lloclister and .J. €1. Quastel, Ed., .Icademic Press Inc., New York, N . Y.,
1963, pp 431-534. 16) 13. R. Baker, E.-T. 110, and D. \'. Santi, J . Pharm. Sei., 64, 1415 (l'Jfi5). ( 7 ) E. R.Baker and J. H. Jordaan, ibiri.. 64, 1740 (1965). (8) 13. R. Baker and 13.-T. Ho, J . Heterocyclic Chem., 2 , 335 (1965). ( Y i B. R. Baker and G. J. Lourenu, i b i d , , 2 , 3 4 4 (1965). 10) 13. R . Raker, T. .J. Schwan, .J. Novotny, a n d l3.-T. 110. J . P h u r m . ., 66, 29.5 (1966). ( 1 I ) Ti. I t . ]laker. .I. K . Coward, P,.-'r.110, and 1). V. Santi, ibid., 66, 302 (I!Jfiti). ( I l ' j I:. I:. 1:nkt.r
Wellcome & Co. The remainder of the compounds have been previously prepared in this laboratory and references are cited in Tahles 1-1..
Results The inhibition results of the dihydrofolic reductases from T, phage, E. coli B, arid pigeon liver with 49 inhibitors are presented in Tables I-V; these inhibitors can be divided into five discreet classes. Series A. 1-Alkyl and 1 -Phenylalkyl Variants.-It is noteworthy that the 1-methyl-s-triazine (11) binds
\'Ill. i o
A' A N-K N H -N~+CH
C'H
Pigeon T2 T1 liver1 E . coli pliagelh'. coli pliage,'pigeon 13 liver a
- -
('
p.lf concn for .50Yc inhib---(1"
It
[Id
111
XL\ I SL\.II S L \ I11 SLIS
IT TI CHJ CH,
Rfi
T?phage
E . coli I3
CHJCeFl, n-C4H1 n-C4H 1 s-CJH, OCHj
0.32 0 . 1% 0,047 0 . 00'30
0 . 000.5 0 ,032 0,032 0 . OOO!)
D
Pigeon liver
0.43 0 . 16 0 . 07.;
0,027
/:z
1/>,6 1 I .). .) 1.8 1/5.2 1/9.2 1.0 1/2.8 1p.o 1.5 11 7.1 48 4.2 1/12 l h t a previously reported,16tuiless
Pigeon l i i e r ' E , coli 13
4;. *;. 0 2.3 30
'r g11age;Id. coli I3
I
34 3 .7 1 . .i 10
I
OCH, trimethoprim
ivith all three enzymes. Thus, since the T, phage and E . coli B enzymes can give a near equal increment in binding with a terminal phenyl (as in XIII) or a terminal butyl (as in VIII) on the n-butyl group, it appears that this part of the hydrophobic bonding region can complex the staggered butyl group or a flat phenyl group with equal efficiency; in contrast, the pigeon liver enzyme has a greater affinity for the flat phenyl group of XI11 by a factor of 3 compared to t h e terminal butyl group of VIII. The branching of butyl (V) to isoamyl (VI) gives a 3-6-fold increment in binding to all three enzymes. The two most potent rompounds in this series for iiihihition of the T2 phage enzyme are the n-octyl(T'III) aiid pheriylbutyl-s-trinziiie (XIII). The noctyl cwmpourid (YIII) i b about fourfold more effective oti the T, phage enzyme than the E . coIi B enzyme and about t1vic.e as effecative a s on the pigeon liver enzyme. The phenylbutyl-s-triazine (XIII) is about three times more effevtive on the T, phage enzyme than the E . coli enzyme, hut :tbout the same n i t h the pigeoti liver enzyme. The two compounds in this series shoii ing greatest qpecificity toward iiihihition of T, phage enzyme over E . coli enzyme are the 1-phenyl (IX) and 1-benzyl (X) derivatives where 23-Wfold diff ereiicer in bindirig are Iiotcd (Tahle I).
Series B. Polar-Substituted 1-Phenyl Variants.Strong evidence was previously presented from this laboratory8 that the 1-phenyl group of the s-triazine IX (Table 11) was complexed to pigeon liver dihydrofolic reductase by hydrophobic bonding. The main line of evidence for hydrophobic bonding was the large decrease in binding observed when the phenyl group was substituted by either a rationic group (XX) or an anioriir group (XIV, XV); that this decrease was not due to steric reasons was shown by acetylation of XX to XXI and esterification of XIV to XVI which allowed increased binding by blocking the polar group. Other polar groups such as C S (XI'II)>C=O (XVIII, XIX), C H 2 0 H (XXII), and OCH, (XXIII) also led to loss in binding; no correlation with the Haniniet u constant was observed. As can be rioted in Table 11, similar results n-ere obtained from the dihydrofolicreductases from E . coli B and induced by T, phage; thus the 1-phenyl group of IX is most probably coniplexed by hydrophobic bonding to dihydrofolir reductase from :dl three soiir('es. In no case was uiie uf the coiiipouiids iii Table I1 more effective on the E . coli B enzyme than the T2 ph:igr ititlucctl ciizymc~. Thc grcntwt y)rwcl i i i biiiditig with T, enzyme aiid the E . coli B enzyme was 77-fold obswved with the trifluoroniethyl group (XXIV); iiiost of thi> diff'erence is due to the 21-fold stronger
BIXDING TO DIHYDROFOLIC REDUCTASES
September 198i
2-benzyl substituent (XLII) gave a pattern similar to 2-phenyl (XLI). When the 2-(p-acetamidophenyl) group (XLIV) was introduced on the 1-phenylbutyl-s-triazine (XIII, Table I), little change in binding was observed with the E . coli B enzyme, but a 15- and %fold loss, respectively, occurred with the pigeon liver enzyme and the T2 enzyme; thus, the T2 phage induced enzyme was affected the most by this structural change. The considerable variation in patterns with 2-substituents suggests that synthesis of further analogs in this area followed by enzymic evaluation might be rewarding. Series E. Trimethoprim and Deazapteridines.-In an elegant study, Burchall and Hitchings16 compared trimethoprim (L) (Table V) and some alkyl-substituted 2,4-diamino-5-deazapteridinesas inhibitors of dihydrofolic reductases from four species of mammalian liver and three species of bacteria; clear “inhibition profiles” were obtained that distinguished the mammalian enzymes from the bacterial enzymes. Through the generosity of Dr. Hitchings in supplying the compounds to us, we were able to obtain comparative “inhibition profiles” for the T, phage induced, E. coli B, and pigeon liver enzymes; the results are presented in Table V. One of the most striking contrasts observed by Burchall and HitchingslGbwas the wide spread of affinity of trimethoprim (L) for bacterial enzymes us. mammalian enzymes; L was found to be 50,00070,000 times more effective on E . coli dihydrofolic reductase than the dihydrofolic reductase from rabbit, rat, or human liver. Note that in our test system, trimethoprim (L) was 53,000 times more effective on E . coli B enzyme than the pigeon liver enzyme. Of high significance is the closer similarity in the binding of trimethoprim (L) to TZphage induced enzyme and pigeon liver enzyme; only a 23-fold difference in binding between these two enzymes as ohserved. I n contrast, E. coli B enzyme was affected 2300 times more strongly than the T, phage enzyme. This, toward trimethoprim (L), the T, phage induced enzyme is much more like a vertebrate enzyme than a bacterial enzyme. With the other four compounds in Table V, the biggest difference in binding to the E . coli B and T2 enzymes is with the 6-benzyl-5-deazapteridirie (XLVI) ; the latter was 34 times more effective on the E. coli B enzyme. n’one of the compounds in Table V was more effective on the T, phage induced enzyme than the E . coli B enzyme; similarly, none of the compounds in Table V was more effective on the pigeon liver enzyme than the E’. coli B enzyme. Thus, with the compounds in Table V, TZphage induced arid pigeon liver dihydrofolic reductases have closely similar “inhibition profiles,” but the T, phage induced and E. coli B enzymes have differerit “profiles,” particularly with trimet hoprim. Discussion From the above results. it is clear that certain parts of the hydrophobic bonding region uf T, phage induced dihydrofolic reductase is different from the hydrophobic bonding region of the dihydrofolic reductase from E. coli B. There are some close similarities of the
917
T2phage enzyme to the E . coli B in some areas, but also close similarities to the pigeon liver in other areas of the hydrophobic bonding region. The greatest difference in selectivity of inhibition of the T, phage enzyme over the E. coli B enzyme was 70-90-fold with the 1-(m-trifluoromethylpheny1)-s-triazine(XXIV, Table I I) and the 1- (nz-benzylphenyl) -s-t riaz ine (XXVII, Table 111). Now consider the T, phage infection of E . coli B as a model system for infection of a mammalian cell by a tumorigenic virus. The 70-90-fold difference in binding to the virus-induced enzyme with XXIV and XXVII would most probably not be a big enough spread for effective chemotherapy. S o t e that the effective antibacterial, trimethoprim (L) (Table V), has a 50,000-70,000-fold difference in binding between bacterial and mammalian enzymes;16 also note that the analogous 1-benzyl-s-triazine (X, Table I) without the three methoxyl groups is actually complexed tenfold better to the liver enzyme than the E. coli B enzyme. Thus, it is in the realm of possibility that continued study of 1-substituted s-triazines or 5-substituted pyrimidines could lead to reversible inhibitors several thousand fold more effective on a virus-induced enzyme than the host cell enzyme. Furthermore, even small differences in binding to the hydrophobic bonding region of virus-induced and host cell enzymes should be greatly magnified with active-sitedirected irreversible inhibitors. 3,13 The answer to the first question posed at the beginning of this paper is therefore a promising there are sufficient differences i n the hydrophobic bonding region of the tu-o enzymes in the model T, phage-E. coli B system to warrant extensive study with tumor enzymes, particularly if induced by tumorigenic viruses. The possible origin of viral genetic particles has been a frequent topic for speculation at coffee breaks and cocktail hours. On the paleontological time scale it is estimated that vertebrates arid yeast evolved from a common ancestoral form about 1200 million years ago ;28 since bacteria were initially anaerobic they probably evolved even earlier than yeast. The poor inhibition of T2 phage induced enzyme, mammalian, and pigeon liver enzymes by trimethoprim (L) compared to bacterial enzymes gives strong evidence that the viral particle arose in a period of time vhen higher forms of life such as vertebrates had already arisen; the “inhibition profile” of the T, phage enzyme with the other compounds in Tables I-V is intermediate between E . coli B and pigeon liver “irihibition profiles,” although there are distinctive differences it1 the T2enzyme from the other two species. Of an even more speculative nature at this stage of knowledge in molecular paleontol~gy’~ is that the T, phage may have been produced by a higher form of life as a protective mechanism against E . coli infection. Studies o n the linear sequence of dihydrofolic reductases from bacteria, vertebrates, arid virus induction will someday be done as is being done 011 cytochrome c:28 then (me should be able to make better judgnieiit as to when on the time scale a virus particle diverged from some particular form of life. (28) E. llarguliasli and .\. Schejter, A d a m . Protein Chem., 21, 191 (1966).
