Structure-activity relations in antifungal agents. A survey - Journal of

Structure-activity relations in antifungal agents. A survey. Corwin Hansch, and Eric J. Lien. J. Med. Chem. , 1971, 14 (8), pp 653–670. DOI: 10.1021/ ...
0 downloads 0 Views 2MB Size
Journal of Medicinal Chemistry 0 Copyright 1871 b y the American Chemical Society

VOLUME14,NUMBER8

AUGUST1971

Structure-Activity Relationships in Antifungal Agents.

A Survey'

CORWIN HANSCH* AND ERICJ. LIEN Department of Chemistry, Pomona College, Claremont, California 91711 Received January 16, 1971 A survey of the literature has been made to find sets of congeneric antifungal agents whose biological activity has been expressed quantitatively. Linear free energy relations correlating 55 sets of data with hydrophobic and electronic parameters have been formulated. The intrinsic activity of various functional groups under isolipophilic conditions is given on a logarithmic scale. In a number of examples antifungal activity closely parallels antibacterial and hemolytic activity, suggesting that such fungicides hriiig about their action by membrane perturbation.

We have been int'erested in the recent efforts to place t'he discussion of biochemical structure-activity relationships in mathematical t'erms.2-7 A useful mat'hematical model can be constructed from the hyp ~ t h e s e sformulated ~~~ in eq 1 and 2 . It is assumed 1 log C

=

-k(log P)2

+

+ k" log kx + k"' = k1 log P + kz(elect) + ka(steric) + k, k' log P

log k x

(1)

(2) in eq 1 t,hat t'he first 2 terms on the right side account for hydrophobic interact'ions in the movement of drug from point of application t o the sites of action. C in eq 1 is the molar concentration causing a standard biological response (IiDjO, EDjo, etc.). Once the drug has reached the site of action, the biological response will be proportional to t,he rat'e or equilibrium constant (kx) of a crit'ical chemical or physical reaction. We have suggested t'hat a Hammett-like treatment can be applied to log kx as shown in eq 2. I n t'hese equations, P st'ands for the octanol-H20 partition coefficient of t'he un-ionized form of t'he drug unless otherwise noted, and electronic and steric effects of the different' members of a set of congeners can be approximat'ed by the use of suitable substituent constants. I n eq 2 , log P accounts for t'he last partitioning step of drug ont'o t'he active site or enzyme. Substitut'ion of eq 2 into eq 1 gives a model of some (1) This work was supported by National Institutes of Health Grant CA 11110. (2) C. Hansch, Accounts Chem. R e s . , 2, 232 (1969). (3) C. Hansch, J . M e d . Chem., 11, 920 (1968). (4) FY.Scholtan, Artneim.-Forsch., 18, 505 (1968). ( 5 ) W . P. Purcelland J. M . Clayton, J . M e d . Chem., 11, 199 (1968). (6) K . Boi.ek, J. Kopeck$, and M. Iirivucov6, Bzperientia, PS, 1038 (1967). ( 7 ) A. Cammarata, J . M e d . Chem., 11, 1111 (1968). ( 8 ) C. Hansch and T . Fujita, J . Amer. Chem. Soc., 86, 1616 (1964). (9) (a) C. Hansch, A . R . Steward, S. M. Anderson, a n d r). L. Ijentley, J . M e d . Chem.. 11, 1 (1968); (b) J. T. Penniston. L. Beckett, D . L. Rentley, and C. Hansch, M o l . Pharmacol., 6, 333 (1969).

general utility for which the disposable parameters (kl-k4) can be calculated by the method of least squares. I t may also be profitable to explore the use of higher order equations."'-l2 I t is possible that, for any given set of congeners, modifications have been made in such a way that not all substituent effects (electronic, hydrophobic, and steric) are evident or important. Thus one must employ regression analysis and analysis of variance to establish the validity of any given term. For example, although Taft's steric parameter E , may not generally be useful, in certain instances critical insight into biochemical reaction mechanisms can be gained through its use.13 In exploring the electronic effect of substituents on reactions, we have found the constants u , u+, u-, and h ' ~ t o be most useful.2 Since a very large amount of work has been done in studying fungicides and data are available on a variety of congeneric sets of fungicides, it seemed worthwhile t o attempt a survey and summary of the results in this field. While ive are interested in all aspects of the structure-activity problem, we are particularly interested at present in assessing the usefulness of the parameter log Po. This constant represents2 the ideal lipophilic character for a set of congeners acting via a common mechanism. I'or 16 different sets of hypnotics we foundg log P o = 2 . For nonspecific Gram-negative drugs we fourid14 log Po E 4 and, for Gram-positive organisms, log Po F. We are interested in comparing log Po for fungicides with the other established log Po values. As we have so often observed, while data are available from a wide variety of sets of fungicides, the derivatives chosen are usually not ideal for separating the various substituent effects. Nevertheless, some useful gen(10) C. Hanscli and s. hl. .Inderson, J . Med. Chem., 10, 745 (1Y67). (11) T. F u j i t a a n d C . Hansch, ibid., 10, 991 (1967). (12) J. A. Singer and W.P. Purcell, ibid., 10, 1000 (1967). (13) C. Hansch and E. J. Lien, Biochem. Pharmacol., I T , 70Y (1968). (14) E. J. Lien, C . Hansch, and S.M . Anderson, J . .Wed. Chem., 11, 430 (1968).

653

654 Journal of Medicinal Chemistry, 1971, Val. 14, Ma. 8

HANSCH A N D LIES

TABLE I ANTIFUNGAL DATAA N D PHYSICAL CONSTANTS USEDI N

THE

I~EGRESSION ANALYSES -Log 1 / C ohd"--

7--

R

Ia

H 2-Me 2,5-hIez 3,6-(OH )2-2,5-C12 2,5-C12 2,6-C12 2,3,5,6-C14 5,6-(C4H4) 2-Me-5,6-(C4H4) 2,3-C1&6-(CaH4)

0

Log P

Eq 11

Eq 13

0.2088 0.70 1.20 0.28 1.62 1.62 3.04 1.788s 2.28 3.20

4.19 4.00 5.00 2.96 3 . 52 5.00 5.40 5.10 5.10 7.00

4.43 3.19 5.00 3.96 5.10 5.00 5.40 5.22 3 05 6 70

Log P

Eq 37

1.05 -1.95 0,9588 0.03 1.05 2.03 3.03 4.03 IC

Id

RI

R2 +)=N R,--N

J n

I

HOEX HOEt

HOEt HOEt HOISt HOEt

HOEt HORt HOEt

H H H2NIGt Allyl Bu Hex le

H

BICH~CONHI:

Pr Allyl 2-Pr n-Bu i-Bu sec-Bu n-Am sec-Am Cyclohexyl n-Hex 2-(EtBt1) n-Hep n-0ct n-llec n-ClpH,,

0.84 0.54 0.64 1.34 1.14 1.14 1.84 1.64 1.85 2.34 2.14 2.84 3.34 4.34 6.34 Log

1.80

P

3.07 1.41 2.57 3.01 3.28 3.20 3.06

-Log l / r m cm-2 ob& Eq 38 Eq 39

3.09 0.91 1.62 2.66 2.94 3.51 3.32 2.93

Log I >

-----Log Eq 26

1.34 1.84 2.34 2.84 3.34 3.84

2.07 2.34 2.92 3.10 3.43 2.99

Loa P

Eq 31

0.07 3.07 4.07 5.07 6.07 5.77 6.27 8.07 10.07 3.55 6.55 6,07 7.75 8\55 9.55

1.76 3.65 4.43 4.88 3.04 4.76 4.59 4.98 3,l.i 3 12 4.31 4.84 4.83 4.87 ,5, 08

-0.12

0.13, -0.19 -0.13 -0.13 -0.21 -0.13, -0.21, -0.15 -0.13, -0.23, -0.13' -0.13 -0.13' -0.131 _-_____________Log Eq 6 Eq 7

4.4.i

3.65 1.18 2.38 3.15 3.39 3.73 3 , ,i1 3.08

0.33 1.91 2.93 3.06 3.17 2.97 3.00 l / C obsdC-7

Eq 28

2.29 2 76 3.15 3.55 3.7.;

Log 1 i C obsdd------Eq 32 Eq 3 3

2.09 4.13 4.64 4.91 4.99 4.7:) 4.82 4 , t57 2.97 3 , 33 4.41 4.74 4.54 4.60

Eq 40

2.19 4.43 4.89 4.90 4.99 4.77 4.64 4.29 3.31 3.60 4.37 4.96 4.77 4.67 4.34

7

t.;q 34

2.32 4.28 3.43 5.51 3.77

-Log Eq 66

1 /c obsde--\ Eq 6 i

-0.36 -0.22, -0.47 -0.39 -0.93 -1.13 -0.40 - 1.551 -0.79 -0.401 - 1.74, -0.40, -0.40 -0.40 -0.40

4.00 4.00 3.40 4.10 4.00 3.10 4.40 3.40 4.00 5.00 3.70 5.00 5.00 4.70 2.00

3.40 3.40

4.00 3.40 3.00

3.40 4.10 3.40 4.40 4.70

1,/C obsdt Eq 8

Eq 49

Eq 50

4,36

Journal of Medicinal Chemistry, 1971, Vol. 14, No. 8 655

ANTIFUNGAL AQENTS

TABLE I (Continued) Log P

If

1.50 1.30 0.70 0.80 0.50 0.30 0.00 -0.20 -0.70 -1.00 - 1.20 -1.70 - 2.20-2.50 -3.70 R R D

O

Ih

H

Eq 49

Eq 7

3.36 3.33 3.02 3.13 3.00 2.57 2.49 2.12

2.09 2.09 1.94 1.49 1.56 1.00 0.37 0.36

1.89 1.51 1.21 0.76 0.58 -0.53 Log P

2.62 2.66 2.46 2.35 2.42 1.98 1.64 1.37 1.66 0.39

Eq 50

3.45 3.86

4.36 4.16

3.76 3.82 3.57 3.66 3.36 3.20 3.35 2.98 2.64 1.89 1.88 1.17

4.00 4.12 3.79 3.96 3.54 3.35 3.49 3.20 2.82 2.11 2.15 1.39

a

Log 1/C obsd* Eq 54

2.38 3.35 2.70 3.70 2.68 3.70 3.35 4.40 3.26 4.22 3.35 4.30 3.70 4.30 4.00 4.40 4.10 4.52

H 4-C1 2-Me 2-Me-4-C1 3-Me 3-Me-4-C1 2,6-Me2 2,6-Me2-4-C1 3,5-Me~ 3,5-Me~-4-C1 2-i-Pr 2-i-Pr-4-Cl 3-Me-6-tert-Bu 3-Me-4-C1-6-tert-Bu 2-C yclohex 2-Cyclohex-4-C1 2-Ph 2-Ph-4-C1

1.4618 2.3986 1.96 2.89 2,02** 2.95 2.46 3.39 2.58 3.51 2.76 3.69 3.70 4.63 3.97 4.90 3.59 4.52

0.00 0.23 -0.14 0.09 -0.07 0.16 -0.28 -0.05 -0.14 0.09 -0.23 0.00 -0.59 -0.36 -0.23 0.00 0.00 0.23

R

Log P

a

Log 1/C obsdi Eq 29

0.00 0.21' -0.17 0.04 -0.14l 0.07 -0.15 0.06 -0.23'" -0.20 0.01 -0.16f 0.0.5

2.35 2.85 2.74 3.07 2.74 2.77 3.35 3.40 3.30 3.46 3 . .i2 4.14 3.82

0.01" 0.22 0.01 0.22

4.00 4.00 4.10 4.00 4.00 3.85 3.26 3.35 3.45 3.64 3.82 3.66 4.05

H 2-Cl 4-Me 4-Me-2-Cl 2-Me 6-Cl-2-Me 4-i-Pr 2-Cl-4-i-Pr 2-i-Pr 4-tert-Bu 2-C1-4-tert-Bu 4-CsH19 2-Cl-4-CsHls

1.4@a 2.1588 1.943a 2.63 1.96 2.65 2.86 3.55 2.76 3.14 3.83 6.94 6.65

4-Ph 2-C1-4-Ph 2-Ph 6-Cl-2-Ph 2-C yclohex 6-C1-2-Cy clohex 3,5-Mez 2-C1-3,5-hfe2 4-tert-Bu-2-Me 6-tert-Bu-3-Me 2-tert-Bu-4-Me 2,4-i-Pr2 2,4-Aml 6-Cl-2-i-Pr 6-C1-4-tert-Bu-2-Mek 2-C1-6-tert-Bu-3-Mek 6-C1-2-tert-Bu-4-Mek 6-C1-2-4-i-Przk

3.59 4.28 3.59 4.28 3.97 4.67 2.46 3.15 3.64 3.64 3.64 4.06 6.46

-0.15O

0.06 -0.14 0.07 -0.34 -0.59p -0.699 -0.38m -0.30'

656 Journal of Medicinal Chemistry, 1971, Vol. 14, No. 8

II.\NSCH

L1b.N

TABLE I (Coqtinued) Log l / C

x H 4-C1 3-C1 4-Br 3-Br 2-Br 4-1

3-1 4-Me 3-Me 4-Me0 2-Me0 4-NOs

3-NO2 4-CN

Oll.~iI~

Log I'

c

E < [10

2.8315 3.33 3.59 3.8.5 3.77 3. Ti8 4.09 3.9s 3.3.5 3.34 2.79 2.30 3.07 2.94 2 31

0.00 0.23 0.37 0.23 0.39 0.20 0.28 0.33 -0.17 -0.07 -0.27 -0.27 0.7s 0.71 0.63

4.80 ..i 09 * 5 ,33 5.12 .j,38 5.17 5.62 5.59 4.37' .i, 30 4,392 4.60 4 . so 4.80 4.83

---Log

-

1 / c oI1sci~--

lt

Eq 42

Eq 4:1

Eq 44

1.:q 45

1Sq 46

Eil 47

.0.38 .0.08= 0.42 0.92 1.42 1.92 2.42 2.92 3.42 3.92 4.42

2.74 2.46 3.09 4.11 4.12 4.14 3.16 3.18 2.89 3.21

2.54 2.74 3 , ,59 4.09 4.11 4.12 4.14 4.16 4.18 3.19 3.21

2.72 3.04 3 . 39 4.09 4.11 4.82 4.s4 ..i16 4.8s 4 . 89 4.61

2,54 3.04

2 ,54 3.04 4 . .;!I 4.09 4.63 4 8% .i14 4.3G 4 16 3. 3.61

2.72 3.04 3.76 4 , O!) 4 , .io 4.82 .1,x4 .i 16 .

J

x

Log I'

2.38 3.26 3.36 2.7% -0.74 1.16 2.38 1.18 -1.72

p-NOs p-c1 m-C1

0-c1 p-HO p-Me0 p-Me p-CH3CO p-HzNSO;!

3.59

4.09 4.63 .i 12 , .i. 14 4.86 4.88 4.19 4.21

d

Log I'

H H

2.7s I .88 2.38 1.8.5 2.59 1.21 1.92 1.71 1.71 1.71 1.71 1.71 2.21 1.19 3.01 2.96 2.96 4.20 4.46 3.96 4.38 5.38 6.38

H

H H H I3

H H H H H H H H H H

H H H H

4.91

0,78 0.23 0.37 0.21 -0.36 -0.27 -0.17 0 , *;2 0.62

It?

H

4.88 4 , 89

XO*(R?,R3)L

1 .oo 0.9% 0.98 0.98 0.98 0.9%

0.98 0.98 0.98 0.98 0.98 0.98 0.98 0.98 0.98 2.69 2.69 5.14 2.69 2.69 2.69 2.69 2.69

Log RBIl olr* follows

=

1.09 - 2.13

+ 0.50 - 0.20 = -0.74

Table 1x.-The appiopriate value of K for R was added to log I> of 0.13 i 0.03 foiuid for pyrazole t’o give log P for the derivs ill t.his table. Table 1y.-Log P was fouild as iii Table If. Table Iz.-Log I’ was found as in Table Ig. The calcd log P vdiies for the diortho-s~ibstit,ut,edphenol are probably a little low since shielding of the OH has not been considered. Table Ia’.-Log P valiies for t.he set are calcd from log P of 0.70 i 0.01 of the 1-Ph ether of glycerol and log P of 1.16 k 0.01 from 2-phenoxyethanol. K values from the phenoxyacetic acids = log were employed. Log P(2-CII&,H,OCH&HOIICH,) 1.16 P(C,N~OCI-IZCHZOH) T ? - C H 3 KCHI Tbranchinn 0.68 0..iO - 0.20 = 2.14. Table Ib’.-Where possible, log P for the benzyl alcohol deriv15 wai. used. When this value was lacking, TX from the benzene or the phenoxyacetic acid system was added to log P of benzyl alcohol. Table IC’.-Log P(C,HjOCOC(CH3)=CH2 = log P(CsH,OCOCIl3) - T C H z i T ( ‘ € i t = C I i ? TCH3 T b r a n c h i n g = 1.49 - 0.50 0.70 0.50 - 0.20 = 1.99. The log P values for the derivs were calcd iising T X from the phenoxyacetic acid system. P values are based on dodecylguanidinium Table Id’.-Log acetate. Table Ie’.-Log P values were calcd by adding the K value for t’he 2-alkyl moiety t’o 1.03 for the parent compd. The value of 1.0.5for the .Y-dodecyl derivs comes from Table Ib. Table If‘.-Trti2 was added to the corresponding compd from Table Ii. Table Ig’.-For these molecules, t’he apparent partition coeffiP = 1.85 f 0.11) was used as cient of ~ - C I ? H & I T ~ + C I (log the reference. This was obtd by partitioning between 0.01 N HC1 and octanol. For the details of the testing procedures used in obtaining the relative biol activities of the different compds, the original work24-52should be consulted.

+

+

+

+

+

+

+

+

+

(24) R . G . Oxvens, Contrib. Boyce T h o m p s o n I n s t . , 17,273 (1953). (25) J . C . LoCicero, D. E. H . Frear, and H . J. Miller, J . B i d . C h e m . , 172, 689 (1918). (26) R . G . Rossand R . A . Luduig, C a n . J . B o t . , 86, 6 5 (1967). (27) R . 1 1 . TVellman and S. E . A . RicCallan, Contrib. Boyce T h o m p s o n Innt., 14, 151 (1946). (28) J . 11. Leonard and V. L. Ijlackford, J . Bacteriol., 67,339 (1947). (29) 0. TVyss, I3. J . Ludwig, and R . R . Joiner, A r c h . Biochem., 7, 415 (1945). (30) H . G . Shirk, R . R . Cores, and P. L . Poelma, A r c h . Biochem. B i o p h y s . , 82, 392 (1951). (31) ?,I. G . Shirk and R . R. Cores, itid.,38, 417 (1952). (32) L. Ilrohnica, ti. Zemanovi, P . S e m e c , K . Antbs, P. K r i s t i h , A . Stullerorir, V. KnoppovB, and P. Nemec, J r . , A p p l . Microbiol., 16, 701 (1967). (33) R . .I. Cutler, E . I3. Cimijotti, T. J. Okolowick, and F. Wetterau, S o a p Chem. S p e c . , 43, 102 1967. Presented a t t h e 53rd Annual Meeting of t h e Chemical Specialties Rlanulacturers Association, Hollywood Beach, I’lorida, 1966. (34) ji’. 13. Geiger. A r c h . Biochem., 16,423 (1948). (35) G . A . Carter, J. L. Garrawas, D. M . Spencer, and R . L. Wain, Ann. A p p r . B ~ ~ si, z . ,135 (1963). ( 3 6 ) M .Huppert, Antihiot. Chemother., 7,29 (1957). (37) D . l’laehova and L. Drohnica, Collect. Czech. C h e m . C o m m u n . , 81, 997 (1966).

Results and Discussion The object of this survey was to uncover as many self-consistent sets of congeners acting on different fungi as possible. At this stage of the development of quantitative structure-activity correlations, we sorely need more equations correlating relatively simple systems so that some of the A B C’s of quantitative structure-activity relationship work can be established before going on to more difficult problems. The large amount of work done with microorganisms in vitro appears to be a good area in which t o gain experience before, say, attacking difficult stereochemical problems in whole animals. The equations obtained in Table I1 can be compared with those obtained for hemolysis of red cells53 and those for antibacterial action. In Table I I a are assembled those equations in which antifungal activity is linearly dependent on the single variable, log P , and in Table IIc are the structure-activity relationship equations parabolically dependent upon only this variable. The intercepts of such equations are useful parameters of reference for comparing the activity of different sets of congeners acting on totally different systems. For example, we have recently shown that 15 different sets of congeners causing hemolysis of red cells yield linear correlations of log 1/C us. log P having a mean slope of 0.93 f 0.17. The small standard deviation (0.17) was unexpected for work from inany different laboratories employing different kinds of red cells. For 7 sets of neutral compounds the mean value and standard deviation of the intercept was -0.09 f 0.23. Using this information, we can construct eq 57 for our expectation of membrane perlog

1

=

0.93 (zk0.17) log P - 0.09 (k0.23)

(57)

turbation by more or less neutral molecules such as alcohols, esters, ketones, phenols, etc. Of course the intercept of eq 57 can be made to vary somewhat by the kind of hemolysis one elicits (e.g., loo%, 50%, etc.) as well as the time of the experiment, temp, etc. The value of the intercept is determined by the sensitivity of the test and the intrinsic pharmacophoric (38) E. Klarmann, L. W.Gates, V. A . Sliternov, and P. H . Cox, J . A m e r . C h e m . S o c . , 66, 4657 (1933). (39) D. B. Reisner and P. R I . Borick, J . A m e r . P h a r m . A s s . , 44, 149 (1955). (40) N. E. Rigler and G . A . Greathouse, A m e r . J . Bot., 27,701 (1940). (41) J. H a t a , hf. Tsurukawa, and h‘I. Kakuma, T a n a b e S e i y a k u K e n k y u N e m p o , 1, 32 (1956); C h e m . Abstr., 61, 5191c (1957). (42) G . Weitzeland E. Schraufstatter, Z . P h y s i o l . C h e m . , 286, 172 (1950). (43) T . Kosuge, H . Okeda, Y.Teraishi, H . I t o , and S. Kosaka, Yakugaku Z a s s h i , 74,819 (1954). (44) E. Klarmann, V. A. Shternov, and L. W . Gates, J . A m e r . C h e m . Soc., 66, 2576 (1933). (45) F. M . Berger, C . V. Hubbard, and B. J. Ludwig, A p p l . Microbiol., 1,146 (1953). (46) D. V. Carter, P. T. Charlton, A . H . Fenton, J . R . Housley, and B. Lessel, J . P h a r m . Pharmacol., 10, S u p p l . , 149T (1958). (47) E. Woodside, hf. Zief, and G. Sumrell, Antibiot. Chemother., 9, 470 (1959). (48) I . F . Brown and H . D. Sisler, Phytopathologv, 6 0 , 830 (1960). (49) R . W.Finholt. M . Weeks, a n d C. Hathaway, I n d . Eng. C h e m . , 44, 101 (1952). (50) D. Ghosh, J. Med. C h e m . , 9 , 4 2 3 (1966). (51) R . Crosse, R . McWilliam, and A . Rhodes, J . Gen. Microbiol., 84, 51 (1964). (52) F. A. Barkley. G . W. M a s t , G. F. Grail, L. E. Tenenbaum, F. E. Anderson, F. Leonard, D. M . Green, J. J. D. Hart, D. P. Kronish, S. Yohimura, and 0. L. Ittensohn, Antibiot. Chemother., 6 , 554 (1956). (53) C. Hansch a n d W.R . Glave. Mol. Pharmacol., in press.

664

Journal of Medicinal Chemistry, 1971, Vol. 14, N o . 8

&?

+

ea103mQ,

w t - m m m

t - o. a .w t. -

o ea ea ea ea.

0 0 0 0 0

+i+isi++

u v u v -

Journal of Medicinal Chemistry, 1971, Vol. 14, No. 8 665

ANTIFUNGAL AGENTS

;I!

2

m M

m

a m

i

0

0

m

?

2

w 3 m o a

O b M m i N m

?????

???????

C

0 0 0 0 0

?

- ~ m c a c o

2

*

N * m h w a *

a

0 0 0 0 0 0 0

0

0 3 3 3 3 3 m

w

3 3 3 3 3 3 3

?

n

0 h

h

09

?

E9

m

19

v

3 v

3

?

N

W

W

N

h

h

5

0

? 0

0 0 0 0 0 0 0

$1

v v v v v v v

siSii-Hiisisi

v

g

C

$i

v

3 00

*

N

0 3 a c w t-*3m

t-

???+

m

4 3 0 0 0

0

3

I m

cl

3

0

I

a B C

H

assess . . . I

I

I

I

I

I

. .

E8

+

2

% .C

-ii

h

2

Journal of .lfedzcmal Chemistry. 1971, Vol. l i , AYo.8

666

character of t'he set of congeners under consideration. T h e only equations in Table IIa lvhich are comparable to eq 57 are eq 3 and 9-13. Equations 4-8 are foi, molecules which ivill be largely ionized under the esperimerital coriditioris. I:or the neutral molecules onl\. oiie set is close in form to eq 3 i ; that is ecl for the phenols. The confidence intervals 011 thc slope, arid wpecinlly thr. intercept of eq 3 , are rather largc. 111 fact, they essentially overlap with those of cci 5 7 . iiidicutirig the considerable similarity in t hc> ti\-()proce, This c:~iibe takeii w one sniall piece of evideiicc that the fungicidal action of phenols is through niemt)r:inc. pert urh $1t.1011. Ai inore direct comparison can be made ~ ' i ueq 38.

interval on the intercept of eq 30 i. too large t o allon its ube 111fruitful compariqoiii. Uiifortunatel? . most of the eiiormoui amount of J\ orlk done n i t h phenolb 011 bacteria has been reported in terms of phenol coefficient. arid hence is not direct]? comparable u i t h log 1 'C data. Exception5 t o thi5 arr t h e re4ult. mibodied in eq 59 and BO The intc>rcepts :mcl coefficieiiti n i t h log I' iri eq 59 arid GO :ire quite c l o ~ eto thoit of eq 3 arid 38, again uiidclw ~ r i i i gthe c l o v relat ronihip bet\\ een hernol 1):ictcrial. ,uid ailtifurigal action of phenol. aiid dcoholy T h e v r c d t . c a n be comuared 1 ?a our extrathermokinds of p r o c e w s

s

0 037 I'heiiol. A c tiiig

1 log C ,

=

0 GS (*O

24) log I'

-

0 92 (*O

i2)g

011

1'scttrlo)itonas

amrglnosa'

+ 0 27 ( i o 49)

21

0 S47

0 222

9

0.979

0 087

5

0.995

0 082

14

0.979

0 087

0.975

0 182

Xlcoholi Acting on Staphylococcus a u r e m I 4

log

C'

=

+ 0.06 ( 1 0 . 0 s )

0 . 6 5 ( 1 0 . 1 2 ) log P

1lc)II Caiwiiig

log

=

0.S6 (zkO.16) l o g P

-

-,!

mV Change in Hest Potential of 1,ohster Ason

0 . 1 0 (zkO.13)

('

HOH Toxicity t o Red Spidei

log

c

=

0 69

( i o 09) log P

+ 0 . 1 6 (+O.OS)

IOO", Iiihitiitioii of Frog Heart by JIisc Seiitral ( " n p o u n d ~

log

I'

=

0 . 9 3 (2cO.09)

+ 0 . 1 1 (zkO.12)

2s

0 2%

0 190

0 20s

0 193 rl'h(, yiiiiilarit\. 1)c.t \\xy,r i cq 58 and :< is strikirig, XIt lioiigti tlicl corrc1:ition of cci :ii h not :is sh:irp as that of ('(1

.iX,it

of

1 ) l i c ~ r l o lor

: i ~ ) p ~ : ithat rs

:I some\\-liat higher concentration :~lcoliolis requircd t o c:iuse hemol? than is ~ i c w l ( dt 0 inhiliit C, a,lDicuti.y. Compnrisoii of the iiitc,rcc,pts of :i: ~ n d.i8 \\-it11the more complex eq 29 The confidencc : l i l t 1 .i-t ;.how r(>:i,soii:il)lcs:igiyenieiit.

:is ,sho\vii in e q 61--63, The m e common mechanism ivhich might be used to explain the great similarity iii nctioii of phenols arid alcohols in hemolysis, nnrcotic action ( a s in eq 61-63), antibacterial actiori, arid :intifurignl actiori is that of membrane perturbation. This need riot necessarily mean rupture as in hemolysis. Sirice a11 exterisive oxidative enzyme system is part of

ANTIFUNGAL AGENTS

the membrane structure, disturbing the membrane structure could easily turn off or diminish this vital system. Turning now to the other neutral molecules of Table IIa, higher intercepts are found for the quinones and arylalkyl isothiocyanates. The average intercept for the quinones is 3.6 and that for the isothiocyanates is 3 . 2 . This indicates their much greater intrinsic toxicity. Since these equations are linear in log P , more toxic members of each of these series could be prepared by making more lipophilic derivatives. The aliphatic isothiocyanate function is much more toxic than the aromatic analog. The intercept of eq 27 is what one expects t o find for the nonspecific membrane perturbation discussed above. Before considering the acids of Table I I a in which either the ionic or neutral form of the molecule may be the active species, it is important to consider the completely ionized benzyl ammonium derivatives of eq 42-47. Unfortunately, the pyridinium compounds of eq 37-41 cannot be compared with respect to intercepts since these data are not on the log l / C scale. For the three sets of ammonium compounds killing fungi, the mean intercept is 3.03. For the 3 sets inhibiting fungi, the mean intercept is 3.21. For our purpose, the difference between the killing and inhibiting concentration is small and we shall ignore it by taking the mean and standard deviation for the 6 sets as 3.12* 0.26. The mean values for these 2 parameters can be compared with results obtained by quaternary ammonium compounds of 3 types causing hemolysis of red

cells,:3 lcor 6 such equations we find a mean intercept of 2.91h0.21. This is very close indeed to the value for fungicidal activity. The antibacterial action of a large set of quaternary ammonium compound5 of varying alkyl chain length and different ring substituents is summarized in eq 64. The intercept of eq 64 agrees very well with the average found for hemolysis as well as that found for antifungal action. Since the action of the benzylammonium compounds against fungi appears to parallel their hemolytic action, one might expect the same to be true for fatty acids. For this reason we have used log P values for the ion pair, RCOONa, in correlating these compounds instead of log I-’ for the neutral RCOOH. Equation 35 will not be considered with the others since this work was done at pH 5.6 rather than the p H 6.3 employed in the other work. Two equations (6 and 7) are for killing action and have, as expected, a lower mean intercept of 2.3 while five equations (8, 48-51) are for inhibitory action and have a mean intercept of 3 . 5 ~ t 0 . 7 . The intercepts for 2 sets of acids, RCOOH and RCHBrCOOH, causing hemolysis53are 2.60 and 2.58, respectively. These figures are closer to the intercepts for the killing log 1/C equations, indicating that killing action more closely resembles hemolysis. There are a number of equations in Table I1 (15, 23-25, 37-40, 56) which cannot be compared with the others either because activity could not be placed on the log 1/C scale or because ?r values had to be used instead of log P values. In addition t o these, eq 4 and 5 for the

Journal of Medicinal Chemistry, 1971, Vol. 14, N o . 8 667

diamines are not comparable to the others because under test conditions these molecules are protonated and we have had to employ log P values for the neutral compounds, It seems most likely that it is the protonated amine which is the pharmacophore. From a practical point of view, one of the most interesting sets of data in Table 11is that correlated by eq 15 for the griseofulvin analogs. These derivatives cannot be compared directly with the other sets of Table I1 in terms of intercepts since the activity of these compounds was expressed on a relative basis rather than as log 1/C. Although the correlation with this group of very complex molecules is not as sharp as one would like, it is reasonable considering the fact that we do not have ideal substituent constants, Suitable steric parameters are not available, and for u we have had to assume that up for aromatic functions is suitable for the groups (X, Table Im) directly conjugated with the carbonyl group as well as the ether function. A finding of importance is the large coefficient with u, indicating that activity is highly dependent on the electron-attracting groups cy to the carbonyl function. a,@-Unsaturated ketones react via addition n-it h mercaptans and inactivate enzymes such as succinic, alcohol, and triosephosphate dehydrogenases and urease which have essential S H I t is known that griseofulvin kills young and actively metabolizing cells but not the older, more dormant elements.54 Also, it has been hypothesized that the antifungal activity of griseofulvin is due at least in part to its inhibition of nucleic acid synthesis at steps either prior to or at the polymerization stage. The partial reversal of gtowth inhibition by purines and purine derivatives has led to the suggestion that griseofulvin may be a structural analog of a purine n u ~ l e o t i d e . ~ ~ Interferences with the replication mechanism of fungal cells have also been suggested, although at present no clear answer to the mechanism of action is generally accepted.56 The spiro linkage has also attracted attention in structure-activity ~ ~ r Since k . the ~ role ~ of u in our analysis supports the idea of reaction of the a,@-unsaturated ketone linkage with an S H group, it would be interesting to investigate the interaction between griseofulvins and SH-possessing enzymes involved in the synthesis of intermediates for nucleic acids; e . g . , inosinate dehydrogenase.58 The fact that activity for the set of griseofulvins at hand depends linearly on electron withdrawal and lipophilic character indicates that replacing X (Table Im) by functions such as CF,, SICj, or C6H4CNshould give derivatives with higher in vitro activity. I n trying to get higher in vivo activity one should first establish log POfrom in vivo studies. l:or a variety of drugs acting in vivo, log POhas been found to be about 2 . Griseofulvin itself has log P of 2.18. The benzyl alcohols of eq 16 also show a modest degree of specificity. Recent work with benzyl derivat i v e ~indicates ~~ that for this function one often finds better correlations in biological work using the radical parameter E R instead of u . Making this change in ey 16, we obtain eq 63. The positive coefficient with the (54) H.Blank, D.Taplin, and F. J. Roth, Arch. Dermatol., 81, 667 (1960). (55) E.G.McNall, Anfibiot. A n n u . , 1969-1960, 674 (1960). (56) R. 13. Angier, Annu. R e p . M e d . Chem., 1986, 157 (1967). (57) H.Newman and R.B . Angier, J . O m .Chem.. 31, 1462 (1966). (58) A. IIampton, J . B i d . Chem.,338,3068(1963). (59) C.Hansch 8ndR. Kerley, J . M e d . Chem.. 13,957(1970).

668 Journal of Medicinal Chemistry, 1071, Vol. 14,,Yo. 6

term of eq 65 indicates that' free radical stabilizing subst>it'uentsyield more active derivatives. The N-phenylbenzylamines of eq 18-20 show a rather high degree of specificity, in so far as the intercept' is a measure of this property. The coefficients with the log P terms in these equations are quite low, indicating that variation in log P has less than half the usual effect on activity. Hence, although the equatioris are linear in log P , not much increase in activity is to be expected by further increases in lipophilicity. The highest log P in this set is 5.41 and, from general experience n-ith neutral molecules, it is rarely found that log Po is much above 6. In this set there are two pharniacophoric functions to consider. Is the phenol or the benzylamine function the active one? The intercepts of equations correlating the toxicity of phenols are usuallj- in the range of 0.5-1.0. Therefore, the high activity appears t o reside in the benzylamine moiety. Again our interest is dran-n t o the highly active benzylic hj-drogens as a source of toxicity. Unfortunately, lack of ER constants and the small variation in the substituents studied prevent our study of this interesting poiiit. The benzyl function is also present in the isocyanates of eq 10. The isocyanate function has a much greater intercept than the benzyl alcohol, indicating t'he much greater toxicity of this function. The replacement of u by E l t or simply the addition of a11 E n term to eq 10 does not result in an improved correlation. This nould seem to be the result of t'he fact that the toxic character of henzylic hydrogens is 2 orders of magnitude lower than the isocyanate function and that their activation might, if anything, lon-er activity via metabolic loss. 1,ulceris and Horsfall, in :irecent study of ant'isporulaiits, made the interesting observation bhat phenoxyacetic acids inhibit glycolate oxidase of A . ~ o l u n i . ~ " Jloreover, the inhibition closely paralleled the antisporulation activity. Both kinds of inhibition paralleled thc ZTT for the substituents. Unfortunately, only G derivatives Ivere tested and so little variation \vas made iri the attached groups (all but one were polychloro compounds) that \ve cannot subject the set to regression to see if the radical stabilizing ability of the sub,qtituent#plays a discernible role. Since it h:ts been possible to shoiv"' t,hrough substituent constant analysis that radical stabilizing ~ubst~ituents have pronounced effects 011 a number of oxidase reactions;, it would be wort,hnhile to make such a study of the glycolate oxidasc-phenoxyacetic acid int eract'ion. The most specific antifungal agents in Table H a are the hisaiiilinopyrimidiiles of eq 14. The reason for the high specificit,?. is not otivious. l l o r c insight' could tie gained by the study of a bett'er selectioii of subst'itut>iits. Adding a term in ER does result in considerable improvement (Fl,b= 3.9; F1,5a.l = 4.1) in correlation but does not quite reach our arbitrary cutoff level of sigiiificaiice at 01 5 0.1. In Table IIc, the confidence interval 011 the alkylpyrazoles of eq 21 is ext,remely wide. Even so, it seems safe tmosa). that, lit,tle specificit'y resides in the pyrazole furictioii. Equat'ions 31-34 with the imidazolines are interesting from the point of view of the intercepts. The mean for the 4 equations is 1.93h0.3. For these combinations \\-e have employed t,he log P for the hydrochloride E'R

(60) Ii. .J. Liikens a n d J . G . Ilorsfal!, I'hytopnthology, 68, 1671 (1968)

IT INSC'H

.\ND LII.N

since, because of their basicit)., they would be esse11tially completely ionized under test conditions. The intercept for imidazolines is close to that of 2.4 found for simple aliphatic amines (eq 36). This ~vouldindicate no special toxicity for the heterocycle function. These intercepts can be compared with the value of 1.6 found for hemol by HSH,.HCI. The agreement is close enough to suggest membrane perturbation as the cause of toxicit).. This is supported by Rich and Horsfall's suggestion t'hat alkylimidazolines disrupt thc membrane permeability of Coi~idia. lIiller, et al. have shoivn that spores of Neuimpom sitoph ila accumulated a lO*-fold concentration of 2-heptadecyliniidazoliiie from :HI :iqueous solutioii of 2 &ml. Our rcsults would indicate that the high lipophilic character of this molecule is the primarj. driving force for its accumulation rather than anj. special liind of active transport . f i 2 In Tthle I I d are listed equations in which activity depends linearly on the electroriic effect of the molecular modification arid parabolicall?. 011 log P. The dithiocarbamates cannot be compared Jvith t'he other sets because for these salts 11-ehave had to use instead of log P. Since their activity depends so little on lipophilic ch:iracter ( r 0= -0.4), it \\.auld appear that they must bring ahout their effect iii an aqueous phase. AlbertG3has discussed the import'ance of the connectior~ of the chelating poiver of diniet hj.ldit hiocartiamic acid ivith Cu2+ for fungicid:il action. The relative uriimport aiice of hydrophobic bonding apparent from eq 36 is also clear from the a o r k of It'euffe~i~~ on antifung:tls of the type C6HjCH,CH,SHC:(S)SII (11 = CH, to H - C ~ H ~ J.ictivit>. . in this series is practicallj. bidependent of the nature of 11. Of interest is the negative sign of the coefficient associated ivith u * in eq 56. This indicates t'hat electron-releasing groups it-hich raise the electron densitp 011 the S atom increase activity. This is t o b e expected if chelating ib the primary cause of toxicity. I t must lie kept in mind for this set of compounds that K for the alkyl groups and the steric effects of these groups ( E s ) tend to vary i n :i similar manner. l'roni the limited set of derivatives I\T have not bceii able to dissect out the ttvo independent roles for 11. Wherever appropriate, attempts \vert made t o evaluate steric effects of substituents using Taft's parametcr. Statistically valid results \\-ere not, obtained rx-ith the set of congeners iri Table 11. I'or the bromoacetanilides of Table Ie, the steric nature of the It attached to the :imide S appears significant. Equation G(i correlates the rciults n i t h A . ~ ~ i q eand r ey 67 those with 7'. vii-irle. I'or the ivork summarized in cq 67 \ v i t h 2'. vii,ide, fc\\-er derivatives lvcrc studied and it is not possible to estimate log Po or to estimat>ethe role of u*. The steric effects of H , as revealed by the E , term, are essentially the same in each equatiori. The positive coefficient wit,h this term indicates that bulky groups hinder activitl-. Log Pofor eq GG is considerably lon-er than that fourid for most of the other sets of congeners arid indicates the possibilit'y for a diff erelit mode of activity for the bromoamides. The large inter(61) L. 1'. Miller, S. K , .\, IIcCallan, ancl K . M , \Veed, Conlrih. B o g r e ThomphonInnt.,17, 173 (1958). (62) R . J. M-,I3yrde in "The Vungi," Tal. I , G . G . Aninnortll a n d . \ , S. sussman, E d . , Academic Press, Ne\\. l - o r k . K . Y,,1Y0.5, p 5 2 6 . (63) A . .\lbert, "Selective Toxicity," 3rd erl. \Vile?, N e \ \ . York. 5 . Y., lY6.5, p 259. (64) \I-, TVeuffen, I'iiurmarie, 21, 686 (1966).

Journal of Medicinal Chemistry, 1971, Vol. 14, No. 8 669

ANTIFUNGAL AGENTS

ions of Table IIc. While the intercepts of the imidazcepts indicate the high intrinsic activity of the bromoolines are close to those found for the amines of eq 36 amide function. which are assumed t o be acting in their protonated I n addition t o the intercepts, another generally useful forms, the log Po values are very much different. The parameter in the correlation equations is log PO. This log Povalues for the imidazolines are like those found is the optimum lipophilic character for a given set of for neutral molecules and it may be that this is their congener^.^ For the 6 sets of eq 21, 23, 27, 29, 54, and active form. If one uses log P for the neutral form of 55 where we have reasonably sharp 95% confidence inthe imidazolines in, say, eq 34, an intercept of -0.81 f tervals on this parameter, a mean value of 5.6*1.0 is 1.0 is found. This would indicate no specificity for this found. This compares with a mean value for 8 sets of function. neutral drugs acting on Gram-negative bacteria of When dealing with very long aliphatic chains at4.4k0.4. For 6 sets of neutral drugs acting on Gramtached to a polar function, one cannot be sure that the positive cells, a mean value of 5.7 f0.5 was found. By additivity principle of calculating log P by the addition this crude measure the fungi resemble Gram-positive of 0.5 for each CH2 unit holds. We have made some cells more closely than Gram-negative cells. Of course attempt to investigatefi5this problem by studying the it is well known that fungi like C. albicans give the apparent partition coefficients of N-alkylpyridinium broGram-positive test. mides. The difficulty of getting accurate log P values of The mean log Pofor 6 sets (eq 42-47) of quaternary the higher members of the series precludes any statement ammonium compounds having antifungal activity is a t present about additivity for members of the series 2.6k0.5. This figure agrees n-ell with the value of 2.6 beyond CI4. However, it is clear when one works a t found in eq 64 for antibacterial action. However, these values are lower than the mean figure of 3.7 *0.4 found very low concentrations to avoid micelle formation or for quaternary ammonium compounds causing hemolypremicelle dimerization that additivity is almost constant in the Cl4-CI8 range. Whether this small desis, The higher log Po for hemolysis indicates that parture from additivity is the result of molecular oil inore lipophilic derivatives can be made before reaching droplet formation as Kauzmann6fi has suggested or maximum activity for a given series. Log Pois highly time dependent ; that is, more lipophilic molecules rewhether it is due to some premicellar dimerization6' quire a longer period of time to reach their sites of is not clear. I t does not seem to be a serious problem action. Log Pois also dependent on the nature of the for our present level of comparison of log Po values. material in the system. The red cell is a much simpler I n summary, the intrinsic antifungal activity of isosystem in which the partitioning of the drug directly lipophilic functions can be tentatively ordered on a onto the surface of the cell is essentially the same as logarithmic scale according to intercepts of Table I1 reaching the site of action. The process is much more as in Table 111. The ordering is of course crude and i t complex with the fungi and bacteria, in part because of involvement with the growth media and in part because T.iljLIc; I11 of the more complex nature of the organisms. The LOGARITHMIC sC.\LI,: O F ISOLIPOPHILIC ANTIFUNG.~L ACTIVITY hydrophobic surface of an ammonium salt of log P I~NHC(=NI1z)NHz+ > = 2.6 is large (e.g., log P ( C M H ~ ~ S + ( C H ~ ) ~ C H =& ~ H ~ ) 2.92). The fact that such molecules form micelles 4 PhHN NHPh easily means that they tend to bind tightly to any 1,4-Cyclohexadieiioiie 3 ,5 hydrophobic area with which they come in contact. ltaN 3.2 This greatly hinders their random movement to the It CHzN CS 3 ,2 critical sites of action in the membranes. Br CHICONH I< 3 Keutral molecules acting on fungi and Gram-positive PhNHCHzPh 3 cells have log Povalues of about 5.6 which is about 3 log ItNHs 2 units higher than charged ammonium ions. The CHzCHz + greater number of C atoms for lipophilicity using the log 1 )NHR 2 P scale for charged compounds plus, very likely, the X=CH interaction of the charge itself with the proteinaceous PhCHzOH 1 material of the cell must somehow combine to set lower PhOCHzCHOHCHzOII 1 log Po values on the charged molecules. Phenols 0 . .i PliNCY O..i The same appears to be true for the anions. HowCH2=C(CH,)COzPh 0.5 ever, with the fatty acids of eq 48-51 there is such wide RNHC(=S)NHIt 0.2 variation that the mean value of 1.7 has little meaning except that the values are much lower than for neutral will vary by at least 0.5 log unit, depending on what compounds. kind of response one uses in measuring activity (ie., From the point of view of log Po, the iniidazolines are inhibition or killing action). I t will of course vary most interesting. Since the pKa of this compound is somewhat from one type of fungus to another. Neverrather high (9.6 for E), log P for the protonated form of theless, it does enable one t o compare different sets of the amine has been employed. The log Po found using congeners, setting aside the sometimes confusing factor these values is not at all close to that found for the other of nonspecific toxicity due to simple lipophilic character. This factor alone can account for a large amount of +

+

'

R.iY.Smith, U . Soderberg, a n d C . llanscll, unpublished results. (66) W.Kauzmann, Aduan. Protein C k e m . , 14,37 (1959). (67) P. hlukerjee, J . Phus. Chem., 69, 2821 (1965). (65)

E

JAMESAND WILLIAMS

670 Journal of Medicinal Chemislry, 1971, Val. 14,No. 8

variation in the activity of a set of congeners. For example, in a set of neutral congeners having log Po of 5.5 and a dependence of activity on log P of 0.6 (slope) in the linear relation between log 1/C and log P , the difference in activity of derivatives of log P = 0 and log P = 5 will be 3 log units. Unless this large variation in activity can be separated in structureactivity relationship discussions, it is quite difficult to begin t o mechanistically classify different functional groups, especially when one gets beyond simple homologous series. How valuable such scales as that in

Table I11 will ultimately be will not be known until more extensive studies have been made. Acknowledgment.-We \vis11 to thank Miss Catherine Church (Smith Kline and E'rench research associate), Dr. William Glave, Dr. William D u m , and Mr. David Soderberg for determining a number of the partition coefficients employed in this uork. We thank Dr. Paul Craig of Smith Kline and 1:rench for the pK, of N-2-hydroxj.et hylimidazoline.

Crystal Structure of dl-Brompheniramine Maleate

[l-(p-Bromophenyl)-l-(2-p~ridyl)-3-N,N-dimeth~lpropylamine maleate] 11. S.G. JAMES A N D G . J. €3.

U-ILLIAMS

Dcpnrtnienl of Bzochemislry, T h e I.niversity of Alberta, Edmonton, Alberta, Canada Kecewed January 6 , 1 9 ? 1

B,

l t a c e ~ i ot)roniphetiiraniitie maleate crystallizes iti space group 1'21,'~with a = 9.863 0 = 10.836 A, c = 13 = 115.83". The crystal structure was solved by convetitiorial Patterson and Fourier techniques mid refined by least square5 to weighted and unweighted IZ factors of 6.3i atid 4.55c&, respectively. The propylamine chain is fkilly extended atid adopts an asymmetrical dispoiition with respect to the 2 aryl moieties. The p-bromobei~ylgroup ia partially occluded by the asymmetry and the 2-pyridyl ring is exposed, thus giving the molecrile ail open side. The maleic acid is i n the nioiioaiiion form atid is H bonded to the YlIe2 group. Molec.ular pwanieters are close to expected values with the exception of the location of the second ba2e dissociable proton of the maleate, which is engaged in a very short asymmetric iiitraiori H bond of length 2.415 -4. 21.494 A , and

Tlic aiiti1iist:imiiiic drugs as a cl wert their :ictioii bj. successful competition with histaniiiic for the :illergic ( H l ) receptor site on the ~valls of smooth muscle tissue.' The title compound is :1 potciit histamine :iritagoiiist and, because receptor sites arc difficult to study directly, it \\-as thought that uscful iriformatioii regarding molecular conformations of :intihist:imiriic drugs could be obtained by defining t h e structurc of this effector molecule. Thc structure of histamhe has recentlj, been complet cd by t w o independent groups? and, more recently, one of these groups has published their preliminary results of the first X-ray study of :in ,, I h p present ivorl; i v ~ begun s in an attempt to delineate sonic of the seemingly relevant structural parameters for :intihist:iminic action. It is reasonable to suppose that if the, :ititihistamiiie acts :is a competitive inhibitor of 1iist:imine then there should be some points of sigiiific:int structur:il similarity between them. In particulnr it secmcd jmportunt t o know if the S-S dist:iip in brompherii mine vas comparable to the 3.%-A (list :iI ice I