STEREOCHEMISTRY OF NETHADOLS
September 1969
in 25 ml of liquid NHI) suspended in 15 ml of PhMe, 1.96 g (0.0084 mole) of 2-chlorophenothiazine was added, the mixture was refluxed for 3 hr, then 1.7 g (0.0084 mole) of XI in 10 ml of PhMe was added, and the whole was refluxed with stirring for 6 hr. After cooling it was extracted with 30 ml of 10% HCl, the aqueous layer was made alkaline Tvith 50% NaOH, and the oil that separated was extracted (Et20). The extract was dried (Na2S04), the solvent was evaporated, and the oil residue was treated with dry HC1 in ether to give 2.5 g (63%) of II-l.2HC1, mp 189-192" (EtyO-EtOH). 10,ll-Dihydrodibenzocycloheptene Derivatives of 3,S-Diazabicyclo[3.2.1] octanes (111, IV). Synthesis of 8- [5-(10,11-Dihydrodibenzocycloheptenyl)propylidene] 3 -methyl 3,s diazabicyclo[3.2.l]octane (111). (a) Reaction of the Grignard of X with Dibenzosuberone. 8- [5-Hydroxy-5-(10,1 l-dihydrodibenzocycloheptenylpropyl)] 3 methyl 3,8 diazabicyclo[3.2.1] octane (XIV).-To a suspension of 0.9 g (0.037 g-atom) of hlg turnings in 10 ml of anhydrous THF, a crystal of I2 and a few drops of EtBr were added. As the reaction started, a solution of 7.5 g (0.037 mole) of X in 10 ml of T H F was added within 0.5 hr. The mixture was refluxed for 1 hr, then 3.9 g (6.085 mole) of dibenzosuberoiie n-as added portioiiwise, and the whole was refluxed for 16 hr. After cooling the resulting solution was dropped into 200 ml of a stirred 10% solution of x H 4 C l a t 0". The oil separated was extracted (CHC13), the extract was dried, and the solvent was evaporated to give 5.6 g (7970 of XIIT,mp 153-155' (EtyO-petroleum ether). Anal. (CZ,H~JY~O) C, H, N. (b) Dehydration of XIV to 8- [ (10,ll-Dihydrodibenzocycloheptenyl)propylidene]-3-methyl-3,s-diazabicyclo [3.2.1] octane(III).-
-
-
-
- -
- -
839
,4 mixture of 5.4 g (0.014 mole) of XIV, 3.3 g (0.017 mole) of p-toluenesulfonic acid monohydrate, and 200 ml of PhMe was refluxed 1 hr, then 100 ml of the solvent was slowly distilled. After cooling, the solution was concentrated to a small volume and washed nith two 20-ml portions of 10% NaOH. The organic layer was dried (Na2S04)and the solvent was evaporated. The residue was chromatographed on silica gel, eluting the impurities with EtOAc-cyclohexane ( 8 : 2), then eluting with MeOH to give 4 g (787,) of I11 as an undistillable oil which exhibited a single spot on tlc. The dihydrochloride melted a t 271-273". Anal. (C26X&12X2) N, C1. 3- [5 (10,ll Dihydrodibenzocyc1oheptenyl)propylidene methyl-3,8-diazabicyclo [3.2.1] octane (IV) was prepared from X I according to the procedure described for 111. The intermediate 3- [.!%hydroxy-5 - (10,ll- dihydrodibenzocycloheptenyl)propyl]- 8methyl-3,8-diazabicyclo[3.2.l]octane(XV), mp 167-169' (ether), was isolated in 42% yield. d n a l . (C2,H,NZO) C, H, N. Dehydration of 2.1 g of XV led to 1.8 g of an oil which was chromatographed on &03, eluting with EtOAc-cyclohexane (8:2) to give 1.2 g (60%) of pule IV as undistillable oil. The dihydrochloride melted at 239-242" (EtOH-EtlO). , I d . (C~SHuC1&;2)N, C1.
-
-
-
Acknowledgment.-The authors wish to thank Dr. A. Vigevani for spectroscopic determinations and ah. A. Campi for chemical analyses.
Stereochemical Studies on Medicinal Agents. V1I.l Absolute Stereochemistry of Methadol Isomers and the Role of the 6-Methyl Group i n Analgetic Activity2l3
P.S. POIZTOGHESE A N D D. A. ~ V I L L I A A I S ~ Uepui h e n t of V e d z c c v ~ a Chenustry, l College of Pharmacy, Universzty of Minnesota, Xinneapolis, Mznnesota
56456
Receibed Narch 28, 1969
The stereochemistry of optically act'ive a- and p-methadol has been deduced by the asymmetric synthetic procediire of Prelog. The apparent, dissociation constants of the title compounds, when compared with those of methadone and 3-deoxymethadone, suggest the absence of substantial intramolecular H bonding in aqueous medium. Nmr and ir studies point t'o the presence of strong intramolecular H bonds in nonpolar media, with the p isomer being more strongly internally associated. Possible preferred conformations which are consist,ent with the spectral dat'a are depicted. The ( 6 R ) , (6S), and (6R) receptor stereoselect'ivities for methadone, amethadol, and acetyl-a-methadol, respect'ively, have been rat'ionalized in terms of differing modes of interaction.
I t is generally believed that strong analgetics exert their effect by interacting with specific sites in the CSS, and that these sites possess asymmetric topographies which enable them to distinguish between e1iantiomorphs.j The more active enantiomers of methadone (1) and certain related analgetics possessing a common asymmetric center have been determined to have the ( R ) configuration.h It subsequently was pointed out that there itre also :I tiumber of strong :Inalgctics whoic configurations are in the (8)series. and that the reversal (1) P a r t VI of this series: P. S. Portoghese, A. A. Afikhail, and H. J. Kupferberg, J . Med. Chem., 11, 219 (1968). (2) We gratefully acknowledge support of this work b y Piational Institutes of Health Grant N B 05192. ( 3 ) (a) Presented in part a t t h e 152nd National Meeting of the American Chemical Society, New York, X. Y . , Sept 1966, Abstract P-4. (b) For a preliminary report on this work, see P. S.Portoghese and D. A. Williams. J . Pharm. Sci., 65, 990 (1966). (4) NIM Predoctoral Fellow 8-E'l-G&I-20515, 1963-1966. (5) P. S. Portoghese. J . Pharm. Sci., 61, 865 (1966), and references cited therein. (6) A . H. Beckett and A. F. Casy, J . Chem. Soc.. 900 (1955).
of stereoselectivity may be due to differing modes of drug-receptor interaction.' PhzCCOEt
PhzCCH(0R)Et
I
CH2CH(NMe2)CH3 1
CH&H(NMe2)CH? 2,R=H 3, R = Ac
One of the most dramatic and interesting examples of this phenomenon has been in the literature8s9for some time arid ib illustrated in Table I. The more potent a-methadol enantiomer [( -)-a-21 is derived from (6s)methadone which has a low order of activity. ?\loreover, conversion of a-2 to CY-3again reverses the stereoselectivity so that the more potent enantiomer, (+)a-3, now has the (6R) configuration. With the optically active 0 isomers there is no inversion of stereoselectivity and, consequently, the activity is found in the (6R) series [( -)-@-2, (-)-p-3].9 ( 7 ) P. S.Portoghese, J. M e d . Chem., 8 , 609 (1965). ( 8 ) A. Pohland, F. J. Marshall, and T. P. Carney, J .4m. Chem. Soc , 71, 460 (1949). (9) N. B. E d d y and E. L. M a y , J . 078. Chem., 17, 321 (1952).
111 :in effort t o explain thebe remarkaI)lv cl~inges111 receptor itereoselectivity. n n investig:ition of the complete stereochemistry of the methndol iwmers was uridertakcti.J S m r . ir. :md pKa studies were also cxrricd out i i i order to itivc>.:tig:\tethc conformation:il ltreferericcs of thcw 1nolcc1111~i. Chemistry.-Application of I ' r e l o g ' ~ ~procedure ~ for determining the absolute configuration of asymmetrir carbinols has been emploj edll nith great success, arid I\ therefore used thi.: method for our htereochemical studies of the methadol isomers. ith benzoylformyl Esterification of ( -)-0-mcthadol rhloride :ifforded henzoylfoi~mateebter hydrochloride 14.HC1) i n yicltl. Ttic. f i w 1 x 1 (4). ~ upon reaction (3
I
I
OH
I
H
I
with J I e ~ I g I ,gave ribe to the atrolactate ester (6, 13 = S.\Ie2) which when wponificd 2 n S-(+)-ntrolactic acid (7) (T.(i' optic:d purity) in :in over-all yield of 20c/, (based 011 4). The luw yield of 7 was likely due to tlie formation of pyrrolidiriium quaternary salt 8 (X = PhCOCOO-) since this compound was isolated when 4 wa5 allowed to stand a t room temperature for 3 days. Presumably, this wa< formed by displaccnient of the benzoyli'ormyloxy group in 4 tiu iritraniolecular Sx2 attack by the baqic nitrogen. Treatment of 8 (X = PhCOCOO-) with HC1 gave the quaternary chloride (8, X = C1-) which pobsessed hpectral proprrties identical with thv compound isolated by Perrine arid 1Iay,l2 who had oht:iincd this material in racemic form from reaction of
The configurations of the methadols have beeii corroborated by pyrolytic cyclization to isomeric 2ethyl-3.3-diphenyl-5-niethylfurans"" whose stereochemistries were elucidated ti)- iimr m u l y studies. a pKa Studies.-- Dissociat ioii cotistaxits have b ~ i i employed as criteria for assessing the abilit,y of :in c+ctroneg:ttive group to stabilize the conjugate acid o/' mi amiiie, as thij often gives rise to enhanced basicity due to intraiiiol(.ciil:ir hydrogeli horidirig.l" Pinr(,
( I O ) V Prelog, Bull. Soc Cham. France, 987 (1956). (11) (a) V Prelog and H L hleier Hrli Chzm. Acta, 36, 320 (1953), ib) W. Q Dauben, TI F Ilickel 0 Jeger, and T' Prelog zhzd., 86, 329 (1954). ( c l 1 J, Hirctl, .I \\ C'laik-Imi\ uiirl i i'lioliertion I C ' h F n i \or 3586 11957) (12) I 1) Perrinn and t L ;Lla). J O r y L i t e m 19, i i 3 (1954)
(13) P. S. Portogliese and I). .\. \Villiams, lS4tli National A,Ieet,inp uf 1 1 1 ~ . American Chemical Society. Chicago, Ill., Sept 1967, Abstracts S-YO; 1'. S . Portoghese antl D. A . Williams. J. Hqterocycl. Chem.. 6 , 307 (1969). (14) A . F. Caay and 11. AI. .I. klaasan, Z'elrukedvoii. 23,4075 (1967). (15) J . F. king in "Tecl~nic~ue of OrFanic ('hemistry," Part I . Val. XI, lnterscience Publishers. Inc., New Tork, N . 1..1963, Chapter V I , i i 318.
8
September 1969
STEREOCHEMISTRY OF NETHADOLS
841
following possibilities: (1) tlicrr i5 110 ~uh~t:tritial intramolecular H bonding of the type, -(H)O * * H-;R; +, or (2) internal 0-He . .N bonding competes with -(H)O. .H-S+ and thereby decreases the stability of the conjugate acid. Factors contributing to 1 might be related to the lower proton acceptor capacity of the carbinol oxygen (when compared to the carbonyl 0 of methadone) and to greater steric hindrance to intramolecular association when the H-bonding proton acceptor group is attached to a tetrahedral center. In case 2 , OH. . . S bonding might offset stabilization gained from -(H)O. . .H-N+ association. Of the two possibilities, we presently favor l.19 I n support of this, existing evidencez0 suggests that (H)O. . . H-S+ bonds are stronger than 0-H. . ~3and hence should substantially increase basicity of tertiary amines by stabilizing the conjugate acid. Since the Ila methadols are less basic than 3-deoxymethadone, it is conceivable that both diastereomers are not internally H bonded in aqueous solution to any great extent. If this is the case, the different pKa values for a- and pmethadol may be due to differences in accessibility of solvent to stabilize the conjugate acid. Infrared Studies.-Although evidence suggests that Me the methadol isomers are not substantially internally Ilb H bonded in polar solvent\, ir studies indicate the acidic proton are coplanar so as to allow maximum presence of strong iritr:r"ecul:ir associ:ttion in 11011stabilization. If this is the case, the phenyl groups then polar media. The high-rcsolution ir spectra of 0.4 JI would be oriented in "quasi-axial" and "quasi-equasolutions (CC1,) of a- atid p-methadol bases revealed torial" conformations. A4ccordingly,the C-5 and C-6 a broad bonded OH absorption overlapping the C H substituents would be staggered in "axial" and "equastretching vibrations a t approximately 3000 cm-1.31 torial" positions to minimize vicinal nonbonded Examination of a 0.005 A l l solution of the bases showed interactions. no change of the absorption characteristics in the 3000-cm-' region. However, some concentration deTABLE I1 pendence mas observed for the a isomer in that a very weak (e 2.5) free O H absorption appeared at 3580 APPARENTDISSOCIATIOX CONSTANTS FOR METHADONE A N D RELATED COMPOUNDS cm-I. No free OH band could be seen with /3-methadol. Compd PKB The ir data suggest that both a- and p-methadol are llethadone 8.62 strongly internally H bonded in nonpolar solvent and 3-Ileoxymethadone 7.93 that the former diastereomer forms weaker H bonds. a-JIethadol 7.86 It has been reportedz2that a chain length of four carbons p-LIethadol 7.59 between electronegative groups gives maximal stability to intramolecular H bonds. The strong internal H The stereospecificity of t'he catalytic r e d u ~ t i o n ~ ~ ~bonding ~ ' ~ ~ ' observed ~ in the methadol isomers might be due of methadone t,o a-methadol may be explained in terms in part to the fact that they possess the optimal number of the cyclic conformat~ions. StuarkBriegleb models of carbon atoms between the OH and amine functions. indicate the upper face of t'he carbonyl carbon to be Nmr Studies.-Additional evidence regarding the less hindered and hence more accessible t'o the hydrorelative intramolecular H-bonding strengths of a- and genation catalyst. A similar explanation can be inp-methadol in nonpolar solvent was obtained by studyvoked t'o rationalize the stereospecificity of LAH ing the temperature dependence of the hydroxylic reduction.18 I n t'his case a cyclic conformation could be proton chemical shift. According to H ~ n e , the * ~ diamaintained by coordination of a n alurhinum hydride stereomer which is intramolecularly H-bonded more species between N and 0 functions. strongly should show the OH proton resonance as It can be noted (Table 11) that the methadol diabeing less temperature dependent. The results of our stereomers are less basic than methadone and its 3variable temperature study (Figure 1) show that the deoxy analog. The decreased dissociation of the regression corresponding t o p-methadol has a lower methadol diastereomers can be ascribed to one of the slope than that of the a isomer. The lower temperature
niethadont: m d the m e t h d o l s poisess a n electronegative group a t C-3 which conceivably might stabilize the protonated dimethylamino group, i t was of interest to compare the pKa values of these compounds with :i-deoxymethadonc (Table 11). The substantially greater basicity of methadone when compared to the deoxy analog suggests that the conjugate acid of methadone assumes a cyclic conformation which is stabilized by i ntern:d H bonding. Ifi Two possible i iitertinlly bonded rotamers :we depicted by conformation$ l l a and l l b . It is awtmed that C=O arid the
(16) I t has been suggested [A. H. Beckett, J. Pharm. Pharmacol., 8 , 848 (1956)l that intramolecular association of this type could occur with methadone, although it was found that 3,3-diphenyl-N,N-dimethylpropylamineis a stronger base than 6-desmethylmethadone. A possible explanation for these results might be related to the fact that the former is not a good model compound, since it does not yomess a substituent coniparahle t o the size of the propionyl group. (17) E. L. May and E. hlosettig. J . O w Chem.. 13, 459 (1848). (18) M. E. Speeter, W . M. Byrd, L. C. Cheney. and S. B. Binkley. J . Am. Chem. Soc., 71, 57 (1948).
- -
a
(19) I n our preliminary reportab w e suggested the presence of ( H ) O . H-N + bonding. Subsequently, however, the apparent pKa value determined for 3-deoxymethadone indicated this t o he no longer tenable. ( 2 0 ) H. Rapaport and 9. h,lasamune. .I. A m . Chem. Soc.. 77. 4330 (1955). (21) A similar observation was reported [.I.BI. DeRoos and 0. .I.Bakker. Kec. ?'Tau. Chim.. 81, 219 (1962)] with several hydroxyyropylaminee. (22) A. B. Foster, A. H. Haines, and M. Stacep. I'rtrnhedrorr, 16, 177 (1961). (23) J. B . Hyne, Can. J . Chem., 38, 125 (1960).
420
K I
90
80
100
110
120
I
OC
Figrii e
1 .--The tenipeiatiire dependelice of the OH iewiiaii((~ foI
0-methadol (0) and d-methadol ( 0 )111 C2C1,.
dependelice ioi the OH rcw>ii:iiice of P-meth:idoI -.uggryts t h a t it i-. moie trongly H bonded 'I'hc cliiditative di m " c ~111H-bond strength ciiii tic. 1 atiorialized 111 t c ~ t n +of "chair-lihe" cwiforniiitions 12 :tnd 13 for a- :*rid 3-metliadol. respectively. If " d t :txial" I'h .\re interaction- are important i r i de-tnbiliziiig 12b :mcl 13a. then 12a :irid 13b \hoiild he the primary coiitrihut 011 to thc rot:tmcric populations. Stu:ti-t I3riegleh model^ r w w l t h a t 13b i-. less hindered thLm 12a. sirice' thc !jauc.hc iiiteructiori between the two phenyl rings and thc ethyl group 111 t h e lnttcr i s coii4dcwbly greater than the iirigle gaicc4u iriteractiori betweeii t h e v groups iti the former It might he eupected th:it :i H
Ph
13a
13b
tlouhle i/auOre iiiterxction \vould not simply be double t h t of :I hingle intertiction. hiit greater. -.irice otic phenyl group should influelice t h e oricntLitiori of t h e second. . \ I o I e c ~ ~ l amodel< i~ ~ i iiidic:itc' 1 ~ that :I "di:ixi:il" inter:ictioit hct\veeii the E t :ind the C-5 proton 111 13b pi~olxthlyi < much le- than that found iri cyclohesniic I)cv:iiis- of the flexibility of thc, molecule. Hcticcl, the more stnble "ch:iir-IiL.e" coriforni:itioii of d-methiidol (13b) should be less hiridered than that of cy-rneth:itlol (12a), and this is consi-.tent with the rey u l t i oht4itcd froin the proton-exchange studies which indicnterl thc 9 isomer t o be more strongly intratiiolecularly H bonded. I n CDC13. the OH protons of a- and B-methadol arc seen a t 8.3 (TYH = 30 cps) and 7.9 ( W , = 23 cp5) p p m , rmpectively. Thr appearance of OH ahsorptioii at iiiil l ~ l l lo\$ ~ l ~ field ~ ~ hUggt'St5 thP I)IYSf'Ilc(' O f ~ t l ' O l l S11
September 1969
STEREOCHEUTSTRY OF ~\IETH;\DOT,~
S43
minor role. This is consistent with the stereochemistry of the active isomethadol antipode (15)27which also possesses thc (:
