812 Journal of Medicinal Chemistry, 1971, Vol. 14, iVo. 8
BAKERA N D KELLEY
OD change of about 0.015 unit/min was followed o n a Gilford recording spectrophotometer with a 0-0.1 011 slidewire at 260 mp . In a 3-1111 quartz cuvette were placed 1.60 ml of 0.2 M arsenatesuccinate buffer of pH 5.9, 0.60 ml of 0.2 mill aq guanosine, 0.40 ml of HZO, and 0.30 ml of DMSO. The contents were thoroughly mixed, then a base line run; if the base line was not 0 due to inadequate mixing of the IIhfSO, further mixipg was made until the base line was 0. Then 100 p1 of enzyme soln was added, the contents were quickly mixed, and the initial 011 change was noted by drawing a tangent to 0 time. Since curvature is fairly rapid it is inadvisable to run more than 1 cuvette at a time. Inhibitor$ were dissolved in DSISO. The cuvette concn of guanosine was 0.04 mlW; its K , was found to be 57 p.2f by the reciprocal plot method.
amine followed by acid-catalyzed ring closure with ethyl orthoformate (method A).34 The intermediate 6-chloropurines and 6-substituted aminopyrimidines were not isolated. See Table 111.
Experimental Section Assay of Guanosine Ph9sphorylase.-Purine iiucleoside phosphorylase (1 mg/ml) from calf spleen was purchased from Calbiochem; for assay 75 pl of the soln was dild with 5 ml of HzO and kept a t 0". When assayed as described below the initial (34) C. Temple, Jr., C. L . Iiussner, and J. A Montgomery, J . .Wed. Pharm. Chem., 5,866 (1962).
Irreversible Enzyme Inhibitors. 188."' Inhibition of Mammalian Thymidine Phosphorylase B. R.
BAKER" AND
t
J 1,. I 3-fold increment in binding to the liver region on the E. coli B enzyme. The results are the enzyme compared t o 2 and a %fold increment in binding subject of this paper. to the E . coli B enzyme, the difference in binding beEnzyme Results.-Although Langen, et ~ l . could , ~ tween the 2 enzymes being only 20-fold-considerably detect no inhibition by 100 pi44 4 of a mammalian less difference than observed with 3, 4, or 9. Replacethymidine phosphosphorylase under their assay condiment of the 4-Et group of 16 by C1 (15) gives no change tions using 1 mi1.l F U D R as subst'rate, we could detect in binding to the E . coli B enzyme, but a &fold loss in inhibition of rabbit liver thymidine phosphorylase under binding to the liver enzyme. Replacement of the 4-Et our assay conditions of 0.4 mM F U D R and 10% DMgroup of 16 by 4-Et0 (17) enhances binding to both SO; 4 had 1 5 0 = 170 p M in our test system (Table I). enzymes about 2-fold. A striking difference between These 2 results are essentially in agreement, the importhe 2 enzymes occurs with the 4-C1Hs-n (18) and 4tant observation being the 490-fold less effective binding CsH5 (19) substituents; both changes lead to a loss in of 6-(2,3-dichloroanilino)uracil(4) to a mammalian enbinding (compared to 16) to the liver enzyme, but a gain zyme compared to the E . coli B enzyme.21 Similarly, in binding to the E . coli B enzyme. Similarly, the 36-(2,3-dimethylanilino)uracil (3) was complexed t'o t'he C6Hjsubstituent of 14 gives much better binding to the E. coli B enzyme21 270-fold better than the rabbit liver E . coli B enzyme than the liver enzyme. It would apenzyme. That this difference in binding to the enzyme pear that the inline, if not inplane, P h groups complex from the 2 species was due t o less effective binding of the well n-ith the large, flat hydrophobic regionz1on the E . anilino moiety was shown as follows. coli enzyme, but the liver enzyme is not flat in this area Uracil ( l ) , a substrate for the reversible reaction, was of the hydrophobic region. complexed equally effectively to both enzymes-a result' The effect of 2 substituents was then investigated and to be expected with a classical-type enzyme inhibithe 2-Me-anilino group (9) can be used as a base line for t ~ r . ~ Introduction ~ t ~ ~ , ~of ~the 6-anilino group (2) comparison. Insertion of a 3-Ale (3) gives a %fold gave less than a 2-fold increment in binding to the rabincrement in binding to the liver enzyme, but a 5-Me bit liver enzyme; unfortunately, an I b 0 could not be ob(21) gives no change. Insertion of a 4-Me group (20) tained for 2, the base-line compound, due to lack of gives an 8-fold increment in binding to the liver enzyme, solubility. I n contrast, 2 was 38-fold more effective but only a 2-fold increment t o the E . coli B enzyme; on the E. coli B enzyme than uracil (1); t'hat is, 2 was similar results were observed on insertion of a 6-Me (22). > 25-fold more effective on the E . coli B enzyme than on The fact that 6-(2,6-dimethylanilino)uracil complexes the rabbit liver enzyme. Since this anilino group is bebetter to both enzymes than 2-Xe (9) has been interpreted to mean that the P h and uracil are complexed (17) B. R. Baker and M. Kawazu, J. .Wed. Chem., 10, 313 (1967). (18) B. R. Baker and M. Kawazu, ibid.,10, 302 (1967). to the enzyme in perpendicular planes.z1 (19) B. R . Baker and M. Kawazu, ibid., 10, 311 (1967). Whereas a-naphthyl (23) and P-naphthyl (24) are (20) B. R . Baker and W. Rsesotarski, ibid.,10, 1109 (1967). (21) B . R . Baker and W.Rzesotarski, ibid., 11, 639 (1968). equally effective in complexing to the E. coli B enzyme, (22) B. R . Baker and S.E. Hopkins, ibid., 13, 87 (1970). p is at least 8-fold more effective than a on the liver en(23) B. R . Baker and M. Kamazu, ibid.,10, 316 (1967). (24) B. R. Baker, M. Kawazu, and J. D. McClure, J . Pharm. Sci., 66, zyme; since 2,3-h1e2 groups (3) add to binding, this re1081 (1967). sult hvith 23 indicates that there is not bulk tolerance (25) B. R. Baker and M. Kamazu, ibid.,66, 1086 (1967). within the enzyme-inhibitor complex for all 4 C's of the (26) B. R . Baker, J . Med. Chem., 11, 483 (1968).
814
Journal of Medzcznal Chonzstry, 1971, Vol. 14, LYo. 9
NO
1 2 3 4 1
I n h i h , p.11
Rabbit liver-------------1'0 '1.+ 150.' p.11
7*5~
0.90
1.40
s 500
1.32
,500 10@ 300
1.20 1.23 1,36
300 300
1.2 I .2
300 .?Oe
1 .:15
16
17 18 19
.!a
21 22 23 24 2.j) 261 27 2h 29 30 31 32
33, 34' 3 .i1 36, 371
E . rali Ij0,C
p.11
river I < . roli I3
1000
2.i0e
7 9 10 11 12 13 14 13
-
1.2
> 1000J 240 170 -200 .io0 2 100 ,560 130 -1 500
300 >2000 -.jOO -900 :300 170 >I200 > 1200 74 460 I .:o > 1200 1 1 .io
I200
100e 300 I 000e
1 . IO 1.0
I
..io
300
I .2:3
12.5~
1.2
> 1200 2000 600
-
1. N O .i20
> 600 I10 110
.id
306
1.24
220 2.30 39 \.3u 40 190 41 300 1.0 > 1200 42 1.01 300 > 1200 43 1 ,02 300 > 1200 44 1.01 300 > 1200 43 300 1.01 > 1200 " The technical ashistance of Maureen Baker, Julie Beardslee, Nancy Rliddleton, and Pauline hIinton with these assays is ac*knowledged. * V O / V I= velocity with no inhibitor/velocity with inhibitor; a t 1 5 0 , T'o,'TTi = 2. Rabbit liver acetone powder extd wit,h 10 ml/g of Tris buffer (pH 7.4) and stored in aliquots of 1.8 ml a t -1.5'. 15, = concn for .io%, inhibition \Then assayed with 400 p.4l FUDR in p H 5.9 arsenate-siiccinate buffer contg 107, DMSO as previously described.6 Data from ref 19. e >Inximum solitbiliiy. 1 Since 20'3, inhibition ( V o / V I= 1.23)is readily detected, the I;o is > 4 times concn meaiiired. Data from ref 20. j' 1)atn f r o m ref 21. ' Data from ref 22. See ref 27 for synth. 3x1
benzo moiety of 23-a. sharp contrast with the E . coli B enzyme. -4 series of 6-arylamino- and 6-alkylaminouracils available from another study in this laboratory27 was investigated for inhibition of rabbit liver thymidine phosphorylase. The n-C6Hg group of 25 gave a 2-fold enhancement in binding compared to the 6-Me" group of 8; any enhancement in binding compared t o 25 ( 2 7 ) T3. R . Baker and J. L.Iielley, J . X e d . Chem , 13, 456 (1970).
by increasing the chain length to ?2-C6Hll(26) could not be determined due to insolubility. KO gain in binding occurred with the C 6 H j C H J H group of 28 compared to M e S H (8), ns previously dlicussed. When the chain length was increased to C6H,(CH2)JH (33) or C6H5(CHs),?jH(34) binding was enhanced 19-fold compared to 6-SleSH (8). This enhanced to 40-fold with C&(CHI)~NH (35), but C,&(CH?)jSH (36) was not as good as 35. The CBH,O(CH2)JYH of 37 \vas 4-fold less effective than t h t >
Journal of Medicinal Chemistry, 1971, Vol. 14, No. 9 815
IRREVERSIBLE ENZYME INHIBITORS.188
INHIBITION^
OF
TABLEI1 FUDR PHOSPHORYLASE FROM RABBITLIVERBY
H RE
Rs
NO.
7 46 47 48 49d 50 51* 52f 53f 54f
Inhib, p M
VQ/VIb
H
H H
1
1900 500 43 180 200 25 110
2"
H H Br H CH3 NHn CHa NH(CHn)zCeHsh Br H 300 CsH5 H 300 CHzCsHe (CHz)zCe.He H 300 (CHz)3CsHe H 300 55f (CHz)aCsH:, H H 300 560 SCsHj H 300 570 OCeH:, H 300 580 NHCsH:, 590 NHCHzCsHa H 300 600 NH(CHn)nCsHj H 300 610 NH(CHz)aCsH:, H 300 620 NH(CHn)aCaHs H 100; 630 Son-NHCeHs H 300 640 CHzOCHnCsH H 300 f For synth see ref 29. e For synth see ref 28. d For synth see ref 4. a - - ~ See Table I. Max solubility. of 33 according to ref 27; yield, 54%, mp 199-200' from MeOH-H20.
NOn
C6H5(CH2)aNHof 35, even though the chain length was the same; this result can be attributed to the loss of binding of the CH2 group to the hydrophobic region when replaced by the more polar 0. Even though 37 does not bind as well as 35, 37 would be superior to 35 for st'udy of substituent effects on the P h binding due to easier synthesis. Although P h binding could not be detected with the 6-C6H61\:Hgroup of 2 due to insolubility, it could be detected with 6-C6H50 (39) or 6-CsH6S (40); the increments in binding compared to uracil were 2- and 10fold, resp, some contribution being made by the less polar S of 40. No iriteract'ion of 6-C6Hs (42) or 6-C6Hs(C"). (4345) could be detected with the rabbit liver enzyme, even though such an interaction was observed with the E. coli B enzyme. l9 A search was made to detect hydrophobic bonding to the rabbit liver enzyme with 5-substituted uracils (Table 11); all of these compds (51-64) were available from other studies in this l a b o r a t ~ r y . ~ Only ~~~~ the *~ 5-phenylbutyl substituent of 55 showed hydrophobic bonding, the increment being 13-fold over uracil (1). NO hydrophobic bonding could be detected with 17 1substituted uracils where the substituents were C6H5(CH,) n, 30 31 C& 5 0 (CHZ) 28 30.31 or alkyl. 3 2 The effect of small substituents at the 5 position of uracil was then made with the rabbit liver enzyme; the binding increments were similar to those previously ob1
%,
(28) B. R. Baker, M. Kawazu, D. V. Santi, and T. J. Sohwan, J . M e d . Chem., 10, 304 (1967). (29) B. R. Baker and J. L. Kelley, ibid., 11, 686 (1968). (30) B. R. Baker and J. L. Kelley, ibid., 11, 682 (1968). (31) B. R . Baker and J. L. Kelley, ibid., 18, 458 (1970). (32) B. R . Baker and T. J. Sohwan, ibid.,9, 73 (1966).
Iw,' p M
1.2 1.0 1.0 1.2 150
0
0.90 1.2 1.1 1.0 1.1 1.2 1.0 1.23 1.1 For synth see ref 8.
~1500
By bromination
served17with the E . coli B enzyme. The group of 46 gave a 44-fold increment in binding compared to uracil with the rabbit liver enzyme due to the increased acidity of the 1-NH.l' The 5-Me substituent of 48 gave a 9-fold increment in binding presumably due to a hydrophobic interaction;17 the 5-Br substituent of 47 gave a 10-fold increment in binding presumably due to hydrophobic bonding, the weakly increased acidity apparently having no effect compared to 5-11e.17 Introduction of a 6-KHz group (7) on uracil gave a 4-fold increment in binding, while a 6-NH2 group (49) introduced on thymine (48) gave an %fold increment. Similar results were previously reported by Langen, et al.,* with inhibition of thymidine phosphorylase from horse liver; they also noted that introduction of a 6-amino group on 5-bromouracil (47) gave a 6-fold increment in binding. However, introduction of 5-bromo group (50) on 6-(phenethy1amino)uracil gave no increment in binding, indicating an overlap of binding regions for Br and GHs(CH2)2". The following important conclusions are reached from the data in Tables I and 11. (1) Hydrophobic bonding adjacent to the active site of thymidine phosphorylase from rabbit liver and E. coli B can best be detected with 6-substituted uracils. (2) The hydrophobic bonding region differs considerably in the enzyme from the 2 species due to evolutionary changes. For example, (a) the C6H6 group of 6-C&&Tu", 6-C6H60,or 6-CeHsS binds a t least 40-fold better to the E. coli B enzyme than the rabbit liver enzyme, (b) there is hindrance to binding t o rabbit liver enzyme by large 2,3 substituents on the C6H6NHmoiety not seen with the E. coli B enzyme, and (c) the hydrophobic bonding region of the E. coli B enzyme is essen-
816 Journal of Medicinal Chemistry, 1971, Vol. 14, No. 9
tially planar whereas the hydrophobic region near where the para position of the 6-anilino group is complexed appears to be nonplanar. (3) Due to the differences in the hydrophobic bonding region of the enzyme in the 2 sources, several compds (4, 5, 14, 19, 23, 28) show 500-to 900-fold better binding to the bacterial enzyme than the mammalian enzyme. (4) The ability of the 4-Et0 group of 17 to bind to the liver enzyme better than 4-C4H9-n of 18 or the 4C6H5of 19 due to a bend in the hydrophobic bonding area in this region is due to the greater conformational flexibility of the Et0 group. If such is the case then
GINER-SOROLLA A N D BURCHENAL
higher alkyl or aralkyl groups may give highly potent inhibitors of the mammalian enzyme. Such studies are being pursued and already the most active inhibitor yet known has emerged; 6-(4-benzyloxy-2-methglanilino)uracil has 160 = 14 pLM and the C6H6CHzO substituent shows a 43-fold increment in binding over 9. 3 3 ( 5 ) Substituent effects on the 6-C6H6(CH,),NH (33), 6-CsHs(CH2)JXH (35), ~ - C ~ H ~ O ( C H Z )(37), ~XH and 5-C6Hs(CH2).r(55) uracils should be studied to determine how much binding can be enhanced. (33) B. R . Baker and S. E. Hopkins, t o be published.
Substituted Hydroxylaminopurines and Related Derivatives. Synthesis and Screening Tests' ALFREDOGIKER-SOROLLA* AND JOSEPH H. BURCHENAL Divisions oj Cell Biochemistry and Applied Therapy, Sloan-Kettering Institute for Cancer Research, A-ew Ymk, S e w York
100.31
Received March 16, 1971 Treatment of 6-chloro-2-fluoropurine (I) and its 9-P-D-ribosyl derivative (11)with NH20H at low temp yielded 2-fluoro-6-hydroxylaminopurine (111) and its B-P-D-ribosyl derivative IV, resp. Compds I11 and I V were reduced to 2-fluoroadenine (VIII) and 2-fluoroadenosine ( I X ) with Raney Ni. The known 2-fluoro-6-mercaptopurine ( X ) was conveniently prepared from I and thiourea and transformed into 2-hydroxylamino-6-mercaptopurine ( X I ) with NH20H. I and methanolic NH3 at 100' gave VI11 or 2,6-diaminopurine (XVIII), depending on the duration of the treatment. I and X yielded, by reaction with the appropriate amino derivatives, several A'-methylhydroxylamino-, -methoxyamino-, and -methylaminopurines. Reaction of I with hydrazine gave 2-fluoro-6-hydraxinopurine (XXI). Equimolar amounts of I and XXI afforded 6,6'-bis(2-fluoroadenine) ( X X I I ) which on Raney Ni treatment was reduced to VIII. Interaction of VI11 with NH2OH resulted in the synthesis of 2-hydroxylaminoadenine (XXIII). Reaction of 2-fluoropurine (XXIV) with NHLOH yielded 2-hydroxylaminopurine (XXV). 111 and X I possessed inhibitory activity against P815 mouse leukemia, and X caused marked reduction of Ridgeway osteogenic sarcoma; I V was inactive.
S e w substituted purines have been prepared in conSynthetic Studies.-Treatment of 6-chloro-2-fluoropurine (I)6 (Scheme I) with an excess of ethanolic tinuation of studies of biologically active purine derivaSCHEME I tives. NHOH "OH The synthesis of 2-fluoro-6-hydroxylaminopurine and its 9-p-D-ribofuranosyl derivative was accomplished in 9 H order to study the effect of the F atom at C, on the V V I growth-inhibitory effect and toxicity of 6-hydroxylaminopurine2 and of its 9-p-D-ribofuranosyl d e r i ~ a t i v e . ~ Similarly, 2-hydroxylamino-6-mercaptopurinewas prepared in order to investigate the influence of the HONH t'his compd may group at' the Cz of 6-mer~aptopurine;~ also be considered as the 2-N-OH derivative of thioguanine.j Several substituted N-methylhydroxylr' amino, methoxyamino, N-methylamino, hydrazino, and 6,6-bis(2-fluoroadenine) derivatives were synt'hesized as well as the 2-N-hydroxylamino derivative of adenine, NI [HI 2-hydroxylamino-6-aminopurine,t o study their potential grow t h-in hibit'ory activity . r.
H
i
~
L
(1) This investigation waa supported by National Cancer Institute Grant CA 08748; Atomic Energy Commission Contract AT (30-l),910; the Harder Foundation G r a n t for Cancer Research from t h e American Cancer Society, Grant T-128K; t h e American Cancer Society Grant T45; a n d t h e Hearst Foundation. (2) (a) 8. Giner-Sorolla and A. Bendich, J . A m e r . Chem. Soc., 80, 3932 (1958); (b) A. Giner-Sorolla, Ph.D. Thesis, Cornel1 University, Ithaca, N. Y., 1958; Diss. Abstr., SO, 1148 (1959); (c) A. C. Sartorelli, A. L. Bieber, P. K. Chang, and G . A. Fischer, Bzochem. Pharmaeol., 18, 507 (1964); (d) P. F. Pecora and M . E. Balis, zbid., 18, 1071 (1964); (e) A. Giner-Sorolla, Galenica Actu, 19, 97 (1966). (3) (a) 4 . Giner-Sorolla, L. Medrek, and A. Bendich, J . M e d . Chem., 9, 143 (1966); (b) J. H. Burchenal, M . Dollinger, J. Butterbaugh, D . Stoll, a n d A. Giner-Sorolla, Biochem. Pharmacol., 16, 423 (1967). (4) G. B. Elion, E. Burgi, and G. H . Hitchings, J . A m e r . Chem. Soe., 74, 411 (1952). (5) G. B. Elion and G . H. Hitchings, ibid., 77, 1676 (1955).
XVII
x ill1
1
xv
XIX,n-2 xx.nr3
HONH, a t 5" afforded 2-fluoro-6-hydroxylaminopurine (111)in 83y0yield. A minor by-product, 2,6-dihydroxylaminopurine (Ir) ,' was also obtained. When (6) J. A. Montgomery and K. Hewson. i b i d . , 83, 463 (1960). (7) A. Giner-Sorolla, C. Nanos, J. H. Burchenal, M. Dollinger, and A . Bendich, J. Med. Chem., 11, 52 (1968).
