952

Journal of Medicinal Chemistry, 1975, Vol. 18, No. 9

Scheme I

spectra and their thin-layer and gas chromatographic behavior. T h e infrared and N M R spectra were superimposable. T h e N M R assignments are as follows: (CC14) d 0.95 (d, J = 7 Hz, isopropylmethyl), 1.07 ( d , J = Hz, isopropylmethyl). 1.30 (peak of C-2 CH3, situated on top of various other protons), 2.35 (br, m, benzylic), 3.30 (br, C-6 proton), 5.95,6.12 (aromatic protons). T h e above correction applies also to t h e 1,2-dimethylheptyl homolog (at C-10) 2b obtained by t h e condensation of pinene (or limonene) with 5-(1,2-dimethylheptyl)resorcinol. T h e N M R assignments are as follows: (CC14) 6 0.75, 0.90, 1.05, 1.15, 1.20, 1.30 (methyl groups), 3.30 (br, C-6 proton), 5.90,6.10 (aromatic protons). Structure 2a is compatible with the structures assigned t o t h e condensation products of some alkylated hydroquinones with several monoterpenes (including l i m ~ n e n e ) as ,~ well as the products of the related condensation of orcinol with limonene.6s7 Acid condensations of phenols with limonene apparently proceed as s u g g e ~ t e d , not ~ . ~solely by attack of t h e aromatic compound a t C-2 of limonene but also a t C-3. This is probably due t o migration of t h e double bond in limonene to an endocyclic position. In t h e previous publication' we suggested t h a t a benzopyran moiety is not an absolute requirement for activity. This suggestion is not supported by t h e d a t a now reported, as all compounds tested by us actually contain such a moiety.

limonene

1

8

A R = C,H, b. R = CH(CH,)CH(CH,)CiH,,

2a.

H

3

'VOt P S

qoH References a n d Notes

5

4

pound 2a obtained from cannabidiol2 a n d from t h e condensation of olivetol with pinene' were shown to be identical by direct comparison of their infrared, N M R , and mass

(1) S. Houry, R. Mechoulam, P. J. Fowler, E. Macko, and B. Loev,

J . Med. Chem., 17,287 (19'74). (2) Y. Gaoni and R. Mechoulam, Isr. J . C h e m . , 6,679 (1968). (3) L. Crombie and R. Ponsford, J . Chem. SOC.C,796 (1971). (4) R. K. Razdan and B. A. Zitko, Tetrahedron L e t t . , 4947 (1969). ( 5 ) M. H. Stern, T. H. Regan, D. P. Maier, C. D. Robeson, and J. G. Thweatt, J . Org. Chem., 38, 1264 (1973). 16) K. L. Stevens, L. Jurd?and G. Manners, Tetrahedron, 30,2075 (1974). (7) The latter reaction has also been independently investigated in our laboratory. Although the reaction conditions employed by us were different than those of Stevens et al.,fi identical products in essentially identical yields were obtained. Our reactions conditions are those described in our previous publication' for the reaction between olivetol and limonene.

Synthesis and Hypoglycemic Activity of Phenacyltriphenylphosphoranes and Phosphonium Salts Benjamin Blank,* Nicholas W. DiTullio, Linda Deviney, J o h n T. Roberts, and Harry L. Saunders SrnithKlrne Corporation, Division of Research and Deuelopment, S m i t h Kline & French Laboratories, Phrladelphia. Pennsylvania 19101. Received March 6, 1975 Phenacyltriphenylphosphorane ( l a ) and several analogs substituted in the meta position of the phenacyl group lowered blood glucose levels in 48-hr fasted rats. The corresponding phosphonium salts had comparable hypoglycemic activity. Two compounds ( l a and lb) were also hypoglycemic in fed rats, but hypoglycemia could not be elicited in another species.

In a continuing search for hypoglycemic agents which a c t by novel mechanisms, a variety of chemical structures was screened for hypoglycemic activity in t h e 48-hr fasted rat. A result of this practice was t h e finding t h a t 2-triphenylphosphoranylideneacetophenone (phenacyltriphenylphosphorane, la) and the phosphonium salt precursor 2a in-

duced significant hypoglycemia in t h e 48-hr fasted rat. Several additional analogs were prepared and studied in t h e 48-hr fasted r a t t o determine what structural features were needed t o evoke this hypoglycemic response. T h e compounds were prepared according t o t h e method of Aksnes' (Tables I and 11). T h e phenacyl bromides used

Journal of Medicinal Chemistry, 1975, Vol. 18, No. 9 953

Notes

T a b l e I. Phosphoranes

R Hypoglycemic a c t , in 48-hr fasted r a t b

bLH=p(c6H5),3 No.

R

Mp, "C

la

H III-CF3 o-CF~ P-CF3 ///-CH3 P-CH3 /I/-OCH~ P-OCH, //t-Cl p-c1 p-C02Me p-CO2H 3,4-C12

175 140-141 170 195 152 179-181 164 140 151 195-197 180-182 164-166 178-180 165' 15 7-1 59' 22 7-2 2 gh

lb IC

Id

le If 1g lh

li 1j lk 11 lm

3 4

5

Recrystn solventa % y i e l d A-B A-C A-C A-C A-C A-C A-C A-C A-C A-C D E-F A-C

92 60 77 80 72 91 92 50 57 84 20 38 76

Formula C26H210P

c2 7HZOF30P C27H2 OF,oP C'27H20F30P C27H?30P

c 2 7H230P

c?i H 2 3 0 2

1hr

2 hr

4 hr

-16" -1 5' -1 5 -24" 1 -36" -9

-24' -2 0' -2 1 -28' -1 -38" -1 -39' -13' -1 -7 0 -4 0 9

-24e -22d -1 2 -4 -3 0' -2 -42" -4 42" -21"

c Z ?HZ 3O2 C26H2 ,clop -4 c ~ ~ H ~ o C ~ O P-2 C28H?303P -3 c 2 7H2103P -1 C26Hi8C120P 4 -3 1 3

-ac -8 2 4 2

4

"The abbreviations have the following meanings: A, CeHe; B, EtzO; C, ligroine (bp 40-60"); D, MeCN; E, MeOH; F. H20. bResults are expressed as the percent difference between the mean change in control and treated groups after a drug dose of 100 mg/kg PO.' p 5 0.001. dp 5 0.01. ep 5 0.05. fLit.11 mp 171O. gLit.lOb mp 156-158". hLit.ll mp 205". T a b l e 11. Phosphonium Salts

R

0

(Ct,H;),

x-

brCH:;@Rl No.

R

2a 2b 2c 2d 2e 2f 2g 2h

H //j-CFj o-CF, D-CF, //7-CH3 P-CH, I / / OC H, p-OC H, ///-Cl p-c1 p-C02Me 3,4-C1, P-F p-OH 3,4-(OH), H H H H

2i 2j 2k 21 2m 2n 20 2P 2q

2r 2s

-

R1

X

Mp, "C

H H H II/-OCHJ D-OCH, II/-OH p-OH

Br Br Br Br Br Br Br Br Br Br Br Br c1 c1 c1 Br Br Br Br

290 115e 188 136 2 78 2 7 7-2 78 178 240 222 264-266 198-19gC 283-284 295' 303-305" 2 7 7-2 79 127-1 29 130 275-277 260-262

H H H H H H H H H

H H H

Recrystn % solventa yield A-B C-D A-B C-D A-B A-B C-D A-B A-B A-B A-B E

63 76 33 80 50 63 99 48 50 87 20 88

A-D F A-D F-D A-D A-D

55 50 92 80 74 70

Hypoglycemic act. in 48-hr fasted r a t b Formula

1hr

2 hr

4 hr

C2,H2,BrOP C2,H2,BrF,0Pf C2,H2,BrF,0P C2;H? , B r F,OPf C?7H24BrOP C2,H2,BrOP C2,H2,Br02P C ziH, , B r O ? P C,,H,,BrClOP C ,!H, B r C1OP* C,,H2,Br0,P C2,iH,,BrC1,0P C!,H,,ClFOP C:,H,:ClO?P C?,H2,C103P' C, ;.H2jBr0,Pj C21H2,Br0,Ph C2,;H,,BrO2P' C2,;H,,Br02Pk

-9' -18' 3 3 -25' -5 -30' -s.5 -37' -2 4 11 -4

-22' -30g 2 -5 -26g -7 -456 0 -36' -7 -1 -14 -17d 2 -11' -lld -17d 0 0

-24' -30g 1 -9 -2gg -4 -4F +12 -42' -16" -14d -22' -11 -2 -8 -12 -33" -8 -2

-4

-10 -6 -4

6 4

"The abbreviations have the following meanings: A, MeOH; B, HzO; C, CHC13; D, EtzO: E, absolute EtOH; F, MeCN "Results art' e l pressed as the percent difference between the mean change in control and treated groups after a drug dose ot 100 mg/kg po except tor en-s which were administered at a dose of 150 mg/kg PO. C p 5 0.05. dp 5 0.01. ?With decomposition. 'Hydrate. g p 5 0.001 hHemihydrate 'Purchased from NEW-CHEM Co., ref 8. IHzO,0.75 mol. &H20,0.25 mol. i n these syntheses were prepared as described by Cowper and Davidsonz and were used directly without extensive purification.

An a-methyl (3) a n d a-chloro ( 4 ) analog of la were prepared together with a vinylog of 3 (5). Two possible metabolites of 2 a (21-a n d 2s) were synthesized from diphenyl-m-

.vii l I ’

954 Journal of Medicinal C h r m z i t r t , 1975, Vol 18, No 9

2

1

a n d -p-methoxyphenylphosphine I a n d phenacyl bromide. T h e intermediate methyl ether4 2 p a n d 2q were cleaved with hydrobromic acid in w e t i c acid t o give 2r a n d 2s.

Q

Melting points were determined in a Thomas-Hoover apparatus and are uncorrected. Elemental analyses were performed by the Analytical and Physical Chemistry Department of Smith Kline & French Laboratories. The structures of the compounds listed in Tables I and I1 are supported by elemental analyses and ir and NMR spectral data. Analyses (C and H) for compounds reported in this paper were within f0.4% of the theoretical values. The procedures presented below are representative of all the syntheses. p-Carbomethoxyphenacyltriphenylphosphonium Bromide. To a stirred solution of triphenylphosphine in dry CCHs (100 ml, 0.1 mol) was added an equimolar solution of the phenacyl halide in CsHs. The mixture was stirred 45 min under reflux and cooled to l o o , and the solid was filtered, washed with CeHs, dried, and recrystallized. 2-Triphenylphosphoranylidene-p-carboxyacetophenone.

C

H COC(X)=P((‘ I1 3X=CH 4.

(‘

H < Il=r“COC(CH

)=P(C,H,)

5

x = c1

Discussion From t h e results shown in Tables I and I1 i t can be seen t h a t la, its meta-substituted analogs, and t h e corresponding phosphonium salts have hypoglycemic activity in t h e 48-hr fasted rat. Whjlr only a limited number of compounds have been synthesized, this has produced a series of compounds in which substituents (bot,h electron releasing a n d electron withdrawing) have been placed in ortho, meta, a n d para positions of t h e phenacyl portion of 1. Simple changes in t h e nature of t h e substituent had n o effect o n hypoglycemic activity. Compounds wih meta substituents (of any nature) were active while ortho- a n d para-substituted compounds were inactive. Replacement of t h e hydrogen on t h e carbon atom adjacent t o phosphorus in la with a methyl or chloro group ( 3 and 4 ) gave inactive compounds. T h e disubstituted analogs lm, 21, a n d 20 were also inactive as was t h e vinylog of 3 ( 5 ) . Two possible metabolites of 2a, 2r and 2s, were inactive b u t the precursor methoxy compounds 2p a n d 2q had hypoglycemic activity comparable t o 2a. la was hypoglycemic when tested in fed a n d alloxan-diabetic rats. However, it did not lower glucose levels when tested in fed guinea pigs. 1b was also hypoglycemic in fed rats but was inactive when given t o fed rabbits or fasted dogs.;’ I t is difficult t o explain t h e differences noted with comparable meta- a n d para-substituted analogs in these series. Simple steric effects d o not seem t o account for t h e differences. If t h e size of t h e para substituent is a prime determin a n t of hypoglycemic activity. t h e n t,he p-fluoro analog 2m and t h e unsubstituted compound 2a would be expected t o have comparable activity since t h e atomic sizes of these atoms are similar (bioisosteric).6 However, this is not t h e case. From t h e available d a t a we are forced t o assume that t h e “active site” in t h e affected biological system m u s t have very specific steric requirements. T h e observations reported here a r e comparable to previous findings in which 3-mercaptopicolinic acid, alone of many closely related compounds, displayed significant hypoglycemic activity in t h e 48-hr fasted r a t after oral administration.’; Experimental Section (:omrnercially available phenacyi halides were examined spectrally and were used withiiiit further purification. Methyl p-brorncmcetylbenzoate was prepared by hromination of methyl p-acetylbenzoate in CHCl.3-E 0 in the usual way.’ 2m was purchased from NEW-CHEM c‘o.; 3-5 were prepared on a custom basis by Orgrnet. Inc.,9using published procedures.’”.”

To a solution of 4 mmol of the above salt in 25 ml of MeOH was added 10 ml of cold 40% NaOH. The cloudy solution was stirred 1 hr at room temperature and the MeOH was removed in vacuo. The aqueous residue was acidified with HC1 and the mixture was extracted with CHC1:3. The CHC13 phases were washed with H20. dried, and concentrated. The residue was triturated and/or recrystallized. 2-Triphenylphosphoranylidene-p-carbomethoxyacetophe-

none. The above phosphonium salt was dissolved in MeOH and made basic by the portionwise addition of NaOMe (NaBr precipitated). The mixture was stirred for llj min at room temperature and the solvent was evaporated. The residue was dissolved in a mixture of CHC13 and HzO and the layers were separated. The aqueous phase was extracted with CHC13 and the combined organic phases were washed with H20 until neutral, dried, and evaporated. The residue was triturated with Et?O, filtered, washed with EtsO, and recrystallized. Hydroxyphenyldiphenylphenacylphosphonium Bromides (2r a n d 2s). 2p (and 2q) was stirred under reflux with a 1:1 mixture of 4896 HBr and HOAc (10 ml/g of methoxy compound) for 20 hr. The solution was cooled and concentrated and the residue was filtered and recryst,allized. Biochemistry. Hypoglycemic activity was measured in 48-hr fasted male rats weighing -200 g. On the morning of the test day, a zero-time tail-vein sample was obtained, followed by the oral administration of the test compound suspended in 0.5% tragacanth at a dose of 100 mg/kg. A similar group of animals receiving only the vehicle served as controls. Tail-vein samples were obtained at 1, 2, and 4 hr after drug administration. Glucose determinations and the significance of the values reported in Tables I and I1 have been described previously.’,” Tolbutamide, after an oral dose of 200 mg/kg, lowered blood glucose levels in this test system 28% at 1 hr. 47% at 2 hr, and 48% at 4 hr after treatment. These values were significant at the p 5 0.001 level. References a n d Notes G. Aksnes, Acta Chc.m. Scand., 15,692 (1961). R. M. Cowper and I,. H. Davidson in “Organic Syntheses”. Collect. Vol. 11. A. H. Blatt, Ed., Wiley, New York, N.Y.. 1943, p 480. A. E. Senear, W ,Valent, and J. Wirth, J . Org Chem.. 25, 2001 (1960). E. N. Tsvetkov, M. M. Makhamatkhanov, D. I. Lobanov, and M. I. Kabachnik, Zh Obshch. Khim., 40, 2387 (1970): Chem. Ahstr., 74, 1 4 7 1 0 2 (1971). ~ N.W. DiTullio, B. Blank, V. Kostos, and H. I,. Saunders, un. published observations. A. Burger in “Medicinal Chemistry”, Part I, 3rd ed, A. Burger, Ed.. Wiley-Interscience, New York-London-Sydney-Toronto, 1970, p 74. B. Blank. Pi. W. DiTullio, C. K. Miao, F. F. Owings,