July 1968
S33
,~L~~YL~IETHOXA~~IS~S
distance separating the charges, ero is the effective dielectric between the charges, e is the electrostatic unit of charge, and RT has its usual significance. The quantity k in eq 23 is included as a parameter defining the relation between electronic polarizability CP and the r superdelocalizability X(E). That is, it is assumed that (y = kS‘E’ (24) The ratio of the coefficients of eq 20, nccording to equations 21-23, is given by
a/b
2971/253
=
=
DrO3
(25)
in which it iS assumed that k is equal to 1. The bond distance is therefore Dro = 2.05 A (26)
which is of the same order of magnitude as could be assigned to a hydrogen bond. Thus, despite the crude nature of the 110 calculations on which eq 20 is based, the relatice magnitudes of the coefficients in this equation are not unreasonable from a physical stand1)oint. The absolute magnitudes of the coefficients, hon ever, may be greatly in error. It should be noted that the correlation provided bv eq 20 gives no indication of whether the 3-hydroxy group is functioning as a proton donor or a proton acceutor. Either mode of bondine is consistent with the correlation obtained. This work, however, provides substantiating evidence for the dominant mode of irlteraction of 3-HPTA derivatives Lvith AChE aIld indicates a method whereby hydrogen-bonding interactions may be investigated in biological systems.
-
N-sec- and N-t-Alkyl Derivatives of Methoxamine and Related Compounds’ RICHARD BALTZLY ASD SARIMAS B. XEHTA The Wellcome Research Laboratories, Burroughs W’ellwme & Go. ( U.S.L4.) Inc., Tuckahoe, .Yew I’ork
10707
Received October 6 , 1967 Revised Manuscript Received Febmary 19, 1968 ,4 number of N-sec- and l\’-t-alkyl derivatives derived from or related t,o methoxamine [ei.yfhro-a-(2,.i-dimethoxypheny1)-0-aminopropanol] have been prepared. Some of these compounds exhibited the physiological properties of “p-blockers” and antiarrhyt’hmic agents. Several had a marked tendency to lower the blood levels of glucose and free fatty acids. In vivo N-sec-alkyl compounds were found t o be degraded metabolically to the parent methoxamine (among other products) but the N-t-alkyl system was stable as regards this degradation. The sec-alkyl derivatives were prepared mainly by reductive alkylation of methoxamine, t-alkyl compounds from t,he appropriat,e amine and bromo ketone followed by reduction. When reductive alkylation created a third point of asymmetry, the physiologically inferior enantiomeric pair D-alkylamino-( - )-methoxamine-Lalkylamino-( )-methoxamine was formed preferentially. Some conclusions are possible as t,o t,he spatial requirements of “receptor” sites.
+
AI et hoxamine [erythro-a-(2,Sdimethoxyphenyl) -paminopropanol] has been regarded pharmacologically as a pure a-adrenergic stimulant. Interest having been expressed as to the fashion in which this property would be altered by, e.g., S-isopropyl substitution. a considerable number of such derivatives were prepared by reductive alkylation of methoxamine base in the presence of available aliphatic ketones, cycloalkanones, and aromatic aldehydes. 111these reactions, the aromatic aldehydes probably form Schiff’s bases but the ketones presumably give only alkylolamines 11‘
I RNHC-Ifed., I n t e r n . Congr. Ser., 48, 1205 (1962); (c) C. H. Ellis a n d S.Gross, F e d . Pror., 2 2 , 2 4 i (1963); (d) R . -4, Salvador, K . I. Colville, L. -4. Lindsay, a n d J . J. Burns, i b i d . , 2 2 , 508 (1963); (e) K . I. Colville, L. A . Lindsay, R. A . Salvador, a n d J. J. Burns, i b i d . , 23, 542 (1964); (f) R. A . Salvador, K. I. Colville, S.A. April, a n d J. J. Burns, J. Pharmncol. E x p . Ther., 144, 172 (1964); (9)J. J. Burns, K . I. Colville, L. A . Lindsay. a n d R. A . Salvador, i b i d . , 144, 163 (1964); ih) C. H. Ellis, Arch. I n t . Pharmacodun. Ther., 160, 144 (1964); (i) J. J. Burns, K. I. Colville, L. A. Lindsay, S. A. April, a n d R. A. Salvador, Pharmacologist, 6 , 186 (1964); ( j j J. J . Burns a n d L. Lemberger, F e d . Proc., 24, 298 (1965); R. A . Salvador a n d S . h. .ipril, i b i d . , 24, 298 (1965); (k) K. I. Colville, L. .\. Lindsay, a n d ,J, J. Burns, I-’/iarmncoloyist, 7, 178 (1965); (1) K. A. Salvador, S.A. April, a n d L. Lemberrer, ibid.. 8, 181 (1966); (in) K. .i. Salvador, 8. .4. .lpril, a n d L. Lemlierger, Fed. Proc., 25, 500 (1966).
\.()I.
534
OK'
I
11
July 196s
&~LKYLMETHOXAMISES
but rather irregularly. For the 2,;-dimethoxy series maximum activity was found with the 2’-nonyl homolog XIII, arid this compound was seriously considered for clinical trial although intravenous (though not oral) administration produced appreciable hemolysis. The 2,S-diethoxy-S-isopropyl derivative XVIII was then found to be a t least equally active while lacking hemolytic tendencies. Some other homologs were investigated. Antiarrhythmic activity perbisted in the 2,sdiethoxy series through the S-?’-pentyl derivative XX without appearance of hemolysis, in the 2,5-dipropoxy series through the S-?’-butyl compound XXV, and was present also in XXVII. Hemolysis, however, was produced by XXT’ and XXT’II, marking the limits of useful activity. The second physiological property of interest was a blocking of the hyperlipoidemia and hyperglycemia evoked by catecholamines (Id-lg, li-lm). This behavior was shown by 1-111, XI, and XVIII but not by IV, IX, and XXVIII. The higher homologs showed no advantage over I in this effect and were, in fact, generally less potent. Since therapeutic employment of any compound for this purpose would involve extended treatment, the metabolic fate of isopropylmethoxamine was investigated rather intensively. Initially, search was made for dehydrogenation and deamination products. Several possible neutral substances and also the aminoketone XI,VI12 were prepared for comparison. I t soon became apparent, however, that most of the compound was being converted, at least in the rat, to more hydrophilic bases. The obvious mechanism for this (aside from ring hydroxylation) is alkyl group hydroxylation, a process known to be catalyzed by certain microsomal enzymes. Hydroxylation on a carbon bound to oxygen or nitrogen results in a hemiacetal or an alkylolamine either of which would be expected to ArOCH2R +.-lrOCHOHIl IINHCH?I:’ +ItXHCHOHR’
disijociate readily under physiological conditions. Identification of these metabolic products accordingly required the 2- and 5-demet hyl-S-isopropylmethoxamine:, (XLVIII and IJV). The first of these 15 as prepared quite readily by reductive alkylation of 2-deniethylmetho~amine.~The preparation of 5-demethyl-S-isopropylmethoxamine was more complicated ; the scheme of synthesis arid identification is shown in Chart I. The necessary starting material, L, was 1)repared by a modification of the method used here ~)reviously for the corresponding acetophenone der i v a t i v e ~ . ~This mas converted to the isonitroso ketone, I,I, which was subjected to catalytic hydrogenation. This last step was rather unsatisfactory since it mas difficult either to complete the hydrogenation (absorption of 4 molar equiv of H2)or to interrupt the procesi with isolation in satisfactory yield of 2~ part idly h) drogenated intermediate. Eventual Iy, :I two-stage reduction with addition of acetoiie irl the second step permitted isolation of IAVidentical 15ith the ( 2 ) Prepared essentially h y t h e method of J. F. Hyde, E. Browning, a n d S a c . , 50, 2287 f1928), I‘rom isopropylamine a n d 2,A-dimetho?t~-ol-iirolnoprogiopl1en~ne [.I E...\rdis. R . Baltzly. a n d I\-, dciioen, i b i d . . 68, 591 (1946)). ( 3 ) I T - . S. I d e a n d R . Baltrly. iiiid., 70, 1084 (1948). ( 4 ) R. Baltsls, J. S. Buck, a n d I\-. S . I d e , iiiid., 72, 382 (1990).
It. .\dams, .I. .im. Chrm.
S33
material obtained by the alternate route through L I I I and LIV. When the isonitroso ketone LI was reduced with LiA1H4, a high yield was obtained of a mixture of bases, presumably the thieo and eiylhro forms. These were separated by distillation after which the fractions solidified and could be recrystallized. The eiythio configuration was established for the /i? isomer (LIII) by conversion t o methoxamine by acetylation, debenzylation, methylation, arid hydrolysis. Reductive alkylation of 1,111 over platinum in the presence of acetone and 2-nonanone, respectively, afforded the S-isopropyl and S-2’-nonyl compounds LIV and LVI. These in turn were deberizylated over palladized charcoal t o form the desired 5-demethyl-N-isopropylmethoxamine (LV) and also the S-2’-nonyl homolog LT‘II. Compound LV was positively identified among the metabolic products of N-isopropylmethoxamine, arid XLVIII with considerable probability. ~lethoxamine itself was also present and some of its 0-demethylated d e r i v a t i ~ e s . ~The formation of methoxamine from S-isnpropylmethoxamirie in rats was paralleled in preliminary clinical trials by a delayed rise in blood pressure. The above phenomenon is riot acceptable in a drug for chronic administration, and therefore an attempt was made to devise a drug that would not be susceptible to S-dealkylation. Obviously a tertiary alkyl group could not be hydroxylated to an alliylolamine by a process involving only replacement of hydrogen by hydroxyl. (Hydroxylation elsewhere in the alkyl group might, of course, take place but would not result in a compound readily cleaved by simple chemical operations.) ilccordingly it was decided t o prepare the N-2-alkyl derivatives of methoxamine. Since the route employed hitherto in this project was not available for this purpose, t-alkylamino ketones mere prepared by reaction of the t-alkylamines with the appropriate a-bromo ketones in acetonitrile. The amino ketone hydrochlorides were isolated without difficulty. Their catalytic hydrogenation appeared t o give exclusively erythw amino alcohols, as is usual in such reactions. While such catalytic reductions are feasible on a small scale, they are extremely slow. It was found that reduction with SaBH4 was more expeditious and satisfactory although not completely stereaspecific. The t-alliylamino alcohols were all active in suppressing hyperglycemia and hyperlipoidemia. Since the higher homologs showed no advantages most of the study was on S-l-butylmethoxamine (1,XI) butoxamine). I n rats there was no indication of S-dealkylation although some 0-dealkylation did take place. The 2-demethyl analog LIX was prepared for comparison but was not clearly identified as a metabolite. Hydroxylation on the t-butyl group would, of course, form a more hydrophilic substance. TIC reve:tlrd :L rather complicated tem of such m:tteriwls lrhose ConipoiieIits \%‘erenot rt: Compound L I X was prepared from cu-bromo-2benzyloxy-5-methoxypropiophenorie by reaction with t-butylamine in acetonitrile, and borohydride reduction to I,VIII which w i 5 dcbenzyhted over I’d-C. D:it:i ( 5 ) .\. Klutcli a n d A I . Bordun, J. .Wed. Chem., 10, 860 ( 1 9 6 i ) . (6) 1.Iilutch, private communicatlon; see also ref 9.
a isomer
(threo)
+ O W NOH II H j
+ H NHOCH-Ph LII
hCH,Ph
1,111
/
CHJOC-H,,
OMe I
Pt
Q Q
+
H,
1
1 Ac.0 2 Pd-C + H, 3 Me1 5aOH 4 alkaline hvdrolysis
+
t
OH NHCHCH OCH-Ph
I
I
B B OCH,Ph
LII'
LYI
OH NHL
p
OMe H
OMe
OMe m etho xa m ine
H
OH NHCHCH
I
OH
C-H,
LYII O I I coinpoundh syrithesizcd a q metabolites :uid Intermediates imlated along the way :ire shown in Table 111. Data o n the S-f-all;~lniethox:nii~ies arid their h o m olog.: are presented in Table I t . Thc f-all,~lamino 1;etolies I\ ere alw found to ha-\.(. :Iiitih~~~ergl3-ceiiiic and hyperlipoidemic propcrtlcxs though in less degree than the aniiiio alcohol.. Stereochemical Aspects. --lIcthox:iniiiie 1 1 : ~ ~~ ~ rciii-idcrcd to be thc ri ytlri o pair of c > i i : i i i t i o n i c r + ( C O I rcspotiditig to ephedrine) OII the ba of the method ot preparation? Over a period of rq :I riuniber oi attempt.; at resolution had provcd fruitle propertieh of S - i s o p r o p ~ l m e t h o x u " whicl -:irne c o i l figuration sugge-ted c~ means of settling this questiou An approximatc sepwration of the isomers of roved quite easy although I through the acid tartrat complete optical purity lebs easily secured. Thc optical propertie:, of these and of the other stereoiaoniers isolated in this project are shown in Table T.'.
I
I
1 Ft + H..HAc 2 Me.CO Pt + Hz
ALKYLMETHOXAIIINES
July 196s
8337
TABLE111 D.i'P.4 Ozi
COMPOUhUS OF CHAIlT
1AhD
OTHER SUBSl'AhCES PItEPMtEJ) FUR IIhl'W O L I C
Sr UlIIES
Sol\ e n t
Structure
NU.
SLVII XLVIII SLIX L LI LII
RIp, " C
2,3-(iLIeO),C6H3COCH?rIeSHCHMe2~ IICl 2-OH-5-?\leOCsH~CHOHCH~~eVIICII?\Ie~ . HCI.O.JIILO
for crjstnn
164-165 A-BC 213-214 dec: A-E 2,5-(PhCH,O),C6H3COEt 82-83 E-11 41-42 c-P 2-lIe0-5-PhCH20C6H3COE t 2-lIe0-5-PhCH20C6H~COCRIe=NOH 124-125 E 2-lIeO-5-PhCH2OC6H3CHOHCHhIeNH~thi LO-base 1JO-lX3 E-I' threo HC1 2 18-2 10 -4-AC 2-lIe0-5-PhCH20CsH3CHOH.CHMeNH2 er& o-base 108-1 09 C-P LIII erylhro HC1 195-196' A-AC 2-MeO-5-PhCH20C6H&HOHCHiLleX€€CHh1e~.HC1~0.5H20 223-224 HZO LI v 2-MeO-5-HOC6H3CHOHCH1\IeNHCHhIe~~ HC1 236-238 11-E LV X-?vIeO-5-PhCH20C6H~CHOHCHiCIeXHCHhTeC~€I~, .HCl 127-12!) .&-E LVI 2-lIeO-5-HOC6H~CHOHCH?\IeNHCH~leC~H~~~ €IC1 164-163 JI-AC LVII 'J2-'34" L I I LVIII 2-PhCH20-5-lIe0C6H~CHOHCHhIeNHC11e~ 2-OH-5-lIeOC6H~CHOHCHlIeXHClIe~~ HC1 262-26.j ri-Ac LIS A = Absolute ethanol, .4c = acetone, E = ether, H = hexaiie, AI = rnethaiiol, P = peiitaiie. 243' dec. A hydrate melts a t 153". 1'
It
KO.
LX LXI LXII LXIII LXIV
LXV LXVI LXVII LXVIII LXIX
LXX LXXI LXXII
x
R'
;\le hIe
hle Me
0
&Ie AI e
Et
0 HOH 0
Me Me Ale Me Et
Et n-C4TTg CII2ClIe3 CHsClle3
Et Pr Pr
Bri Bit hIe hIe lle hIe ?\le
HOH
HOH
0 IIOH 0 IIOH HOH 0 HOH
isfactory with the K-acetyl derivatives. However, although as a preparative operation, DL-Igave about a GOYo yield of threo isomer, and the (+)-threo isomer of I was obtained under Welsh conditions from (-)-I, the Welsh inversion mechanism is clearly not the only process involved. Under the conditions of the rearrangdment, racemic acetylated n'-isopropyl-+-methoxamine also gave around 30% of I (whereas N-acyl$-ephedrines are not altered sterically). When refluxed 1-1.5 hr in 1 N HCI, both DL-Ihydrochloride and the nL-threo hydrochloride corresponding were isomerized to the extent of 5-10%. This phenomenon could be due to solvolytic exchange in the following sense.
Such a process was rejected by Welshiob as significant with the K-acylephedrine isomerization, in part,
Analy1urmula
CiaH2iS03. HCl
Cl3H211103.I€CI C23H2203 C'7H18OJ
ses
C, I1 C, H c, H C, I1 c, Ir c, H c,I I C, H C, H c, H
Ci71&7NO4 Cl,IILlllOJ C I ~ H ~ I N IICI OJ cijII21so3 CnH,iSOa.IICl C,oH,,N03 IICl 0.5H20 C1jH,1SOJ.IICl CJ H c?GII3Si\;03 'IIc1 C, H c1qII3jL-O~.I€CI c, €I C', €i CIIIIILINO, ('lalI,JNOJ'I1c1 c, Ii The hydrochloride i i i d t , at 244-
lip, " C
184-186 259-260 194-196 205-206 eff 203-204 246-247 175-1 76 220-230 eff 157-159 eff 229-230 215.5-216.5 175-176 235-236
since in his system the process was sterecspecific. I n the methoxamine series, the o-methoxgl group mould be expected to favor a process iri which positive charge is developed on the benzylic carbon atom. The degree of isomerization found with the amine hydrochlorides is not alone capable of accounting for our observations, but the same process should be more facile in the 5acylated form which would not need to form a bivalent cat ion. Although thesc operations do not pre.;ent the elegant picture of a stereospecific conversion of the (-)erythro to the ( + ) - t h o isomer, they do show conversion of (-) isomer having a molecular rotation comparable to that of (-)-ephedrine into (+) isomer whose AIDis also comparable to that of (+)-+-ephedrine. Resolution of methoxamine itself was eventually accomplished using dibenzoyltartaric acid. This gave only a partial separation after which complete resolution was possible with n- and L-tartaric acids. Inspection of the data of Table V shows a reasonable coiisistency in the molecular rotations of the optical isomers with each other arid with ephedrine and
+-
s3s
\'()I, 1 I
S 1 1 1 u i i 1 i imade i~ u p fiir rotatioiis wei'e i i o t the~niohlatetlhut were i i i the raiige 25--27'. 111dl c:i>e., ( 1 Where possible, c~oiii.eiiti.:iiiciii.: wei'e 4 15 m i i i ) after s o l u t i o i i a had been made up to vcilume. methcisaniiiie arid of SVIII were not dfirieiit1)- so1uk)le to give solutiolis more c*oiic.eiiir:ited than 2
ervat i o i i h were mntle pr,oiiil)i1.1.
. The iwmer- of S-isopropyl-
(wit.hiii (
+ ) - ? j - i ~ o ~ ~ i ~ ~ ~ ~ ~ ~ l - $ - ~ naiid e t h oseveral x a r n i ~ofi ethe isomers of
a litlle exresi HC1 (0.03 rnl of coiiren I 1)oteiit in :holi,hiiig arrhl thmiai but i t i artiori \ t i i s ~)rololiged:it least three tinlei I ' There :are ~onic. itidic:itioris that t h r ( T ) iioiiier ha\ a n effect antagoiiistic :is regard> its aritil)ode n'heii the compounds of Table I n e r e screened for their efi'ect or1 ctirdiac arrhythmias. activity showed U I J in it roniiderablr number of them but no structurecyhcdriiir.
. The nvailable quantities of butylmethosamiiie required operati iii more dilute solutioiis. A11 ' HCI i i i 2.5 ml). Theye dat:i :ire o i i the anhydrous , d t a . io i eft'),aiid 111)ic: c~:~ic~iii:ite~l o i i the basi. of the molecxlai,
:ictivi t) wl:ttioii~hil)s\wre dcducihlc. Sotnbl!.. scvcrnl ronil)ouiid+i w r e IC- active t hail would be exllected a i d thi- j)ecnliaritv u :L' tlcquciitl) a-sociated with thct i)rescnce oi :I*) mriietrJ- i n the S-nl :I definite ilobsihility that the ruriniiig through the I)rcsunil)tive intermediate A gavc exclusively or ~)redorniii:intl\-the less active of the t \ \ o theoretically poshible 1)airs of enantiomers derivable from r)r,-niethosaniiiic. Thi- could be either becwsr one of thew t u o Linirs of coiifiguratioiii formed morcx
11
:I I ) 'l'lir nature (11 tlir j~li.vsiolugiwlt r b t r ~ r i ~ k limit8 i ~ ~ d t \ r r l i y t l i ~ n i a spru\-oked I I V digoxin in t i l e ( l o p last a l l a r p not rrprntalhlP n n tlir same animal. Raremir. .\\'I11 ahulislieii arrhyt 11inias irir atioiit 5 min, ( - ) - X V I I I , a t tlie same dose. was effecti\-e for 15 inin ( t l i e diiration of t l i test). ~ Tlie time factor of :3 is tliereiore a minimum. I n connection iritii tliis, it seems \\ orthn-liile t o record a n observation that mal- liar? some general significance. Tlie hydrochlurideP oi ( + ) - a n d ( - ) S V I I l a r e very sparingly soluhle a n d lor injection purposes more soluble salts. prpferably crystalline, w o ~ i l dbe desiraijle. .\ number of salts were iirepared. Tlie lactate a n d citrate a r e relatively soluble h u t do not crystallize reailily. T I I P l i i d r o l m m i i l e (mp 2 M 0 ) is not more soluble t h a n t h e hydroi , l i l i J r i d e o n ii molar liasiz. The neutral siilfate ( m u 243-247') is soluble to in n a t r r a n d tlie a c i i i phosyhate (B.lIaPO4, mp 2 2 2 O ) is soluble t o . 'The pliospliate, sulfate, a n d n-tartrate Trere found t o 'UP fiignificantli. less artivp physiologically rhan the hydrwhlorirlr ani1 li~-iir,~lir,,iiii~le on a molar basis. I n 57c aqueous solution t h e phosphate gave [ a l l ) - 2 2 . 4 1- 0.2' corresponding t o 1 I n -85.0 & 0.!q0. T h e hydroillloride has 111, -81.3 f 1 . 6 O wliicli is significantly Ies. One ir led t o tlie Iiyimtlieiis t h a t assoriation 1,etv e r n t l i i s amino alcohol cation a n d imlyf,in r(~:iw~i:il)lo t o extend thi. findirig to the rest of the zerie. further Imsibilit!? n a 5 that iiolatioii iiiid purrhcation had prefcrentially coiiceritrated one of the ])air+ of isomer.. This nould be more probable wheli lhv concentrations were comparable if the physic:tl pro1)erties differed appreciablj . Examination of :t coilsiderable quantitjr of 11 a. the hydrochloride failed t o establish homogeneit) or the lack of it. (3lelting points were fairly high and probably attended hy decomposition .) Thc only imambiguoub approac7h appeared to be to zj tithesize the isomers individually. Preliminary cupei.iment.i with racemic niaterial. wtablished sati5factorj conditions for separating 1111rwcted methoxumiiic from its S-sec-butyl derivative.
July 1968
839
L\LHYLlIETHOX.4MINES
and for utilizing alkylating agent efficiently. l2 For the purpose of specific syntheses all operations were performed with the sec-butyl mesylates thereby avoiding the partial racemization that appears unavoidable in conversion of alcohol to bromide. JIesylate from pure D-sec-butyl alcoh01'~was first treated with (+)and (-)-methoxamine bases giving the L-(+) and L-(-) isomers. After the determination of the physical properties, crystal form, etc., of these two diastereomers, similar reactions were run with L-enriched mesylate (ca. SOYG)and the D-(+) and D-(-) isomers were obtained. A priori it was anticipated that activity would be a function of the (-)-methoxamine portion; however, the I)-(+), D - ( - ) , and L - ( + ) isomers were all of about the same low activity. The L-(-) isomer was active and long acting, comparable to the (-) isomer of XVIII. JIeanrvhile, i~ careful fractional crystallization of I1 as the base resulted in approximate separation into a major component crystallizing in plates and melting at S3-86" and a minor component crystallizing in needles arid melting at 96-98', in a ratio of tibout 3 : l . Resolution of both was attempted with g-tartaric acid and the less soluble D acid tartrates were separated in both cases in sufficient quantity for identification of the corresponding hydrochlorides. The major component afforded the L - ( + ) isomer and the minor, the D - ( + ) isomer. I t was thus apparent that the reductive alkylation method had indeed produced preferentially the less active of the possible pairs of enantiomers. Knowing the rotations of the individual isomers it was possible to gain a somewhat more precise figure for this preference. Tn o parallel experiments were run. I n the first, (+)methoxamine was allowed to react incompletely with a large excebs of DL-sec-butyl bromide. In the second, (-)-methoxamine was hydrogenated in the presence of but:tnone. The total N-sec-butylmethoxamine fractions in each experiment were isolated and rotations were determined. The rotation of the material from the :dkylation experiment indicated that 70-7.5YG of the L - ( + ) isomer had been formed. In the reduction experiment the product was S0-90% the D-(-) isomer. In each case about 40yG of the reacted material was isolated in pure form as the hydrochloride of the predominant isomer. On the assumption that the reductive alkylation proceeds through an intermediate alkylolamiiie (as A ) , (-)-methoxamine would give rise to the two forms I3 :tnd C. Examination of Fischer-Hirshfelder models of these reveals considerably greater crowding in B than i n c'. Replacement of OH by H with retention of configuration \rould lead to the I)-(-) isomer from C
and to the L-(-) isomer from B. To our knowledge there is no evidence as to the mechanistic peculiarities CH3
CH3
I I
C2H5
CH3
I I H--C-NH--C-OH
I
H-C--OH Ar
B
C
of such catalytic hydrogenations but there would seem to be only two obvious paths for a selective process: frontside attack with retention of configuration and backside attack with inversion. In molecules as crowded as these the latter process (as in an S N 2 substitution) would seem most improbab1e.l4 For the displacement reactions, the transition states of bimolecular displacements would be D and E (written for (-)-methoxamine with D- and L-sec-butylX, respectively). D has an obvious correlation with B, and E with C. E would be less crowded than D though the disparity probably would be less in these transition states than in the intermediates (B aiid C) whose bonds are fully covalent. This also agrees with the fact that the displacement was less selective than the reductive alkylation.
Ar
Ar
D
E
Some further correlations of the stereochemical peculiarities with the antiarrhythmic activities seem reasonable. The absolute configuration of L-sec-butyl(-)-methoxamine is shown in F (Z = Ollle, R = Et). The first requirement for high activity is, of course, the correct ( - ) configuration for the methoxamine moiety. However, the essential physiological equivalence of this isomer with the ( - ) isomer of XVIII (F, Z = OEt; R = Me) indicates that the magnitude of R cannot be sharply critical. Rather it is the methyl group (relatively unhindered in models) which cannot be replaced by something bulkier. This suggests that whatever the drug acts upon (receptor, enzyme surface?) has a slot which must be occupied for firm attachment but which cannot accommodate a group larger than methyl. I t is also tempting to suggest, though positive evidence is lacking, that the isolated specimen of the highly active XI11 fortuitously contained a relatively high proportion of the I)-(+)-L-( -) pair of isomers. CH3
I H+-NH--C-H I
(12) Reactions were run with racemic methoxamine a n d DL-sec-butyl
liromide a n d methanesulfonate in MeCN. Contrary t o expectations, t h e mesylate mas not much more t h a n twice a s reactive as t h e bromide. Very approximate calculations give khr (bimolecular) a t reflux as 0.05 for t h e bromide a n d 0.12 for t h e mesylate. These figures are approximate not only because no a t t e m p t a t precision was made b u t also because i t is uncertain w h a t form of t h e bimolecular equation should b e employed when t h e basicity of t h e primary a n d secondary amine m a y not differ greatly. (13) D a n d L a r e here applied as systematic notations t o t h e isomers of .su-butyl alcohol. As applied t o t h e stereo isomers of 11. D - ( - ) , D - ( + ) , etc., are used, t h e D referring t o t h e configuration of t h e hutyl a n d (C) or ( - ) t v tile configuration of t h e methoxamine moiety.
CzH,
I I H+-NH--C-OH I I H-C--OH CH3 I Ar
H-C--OH
&
CH3
I I
R
2
F
(14) Especially since t h e widence on t h e hydrogenations leading to I a n d S X I S suggests t h a t once alkylolamine is formed its liydropenation is rapid.
ALKYLJIETHOXAMISES
July lOGS
and, after removal of the catalyst, crystallization from 1leOII-
Et20afforded 1.5 g of pure LV. Preparations of N-&Alkyl Derivatives.-Previous operations in these laboratories in the preparation of N-methylamino alcohols have usually involved preparation of benzylmethylamino ketones and subsequent debenzylation in order to avoid handling the characteristically unstable secondary amino ketones as bases. For the present application, however, it seemed likely that use of a benzyl t-alkylamino intermediate would involve serious hindrance problems and 2,5-dimethoxy-cu-bromopropiophenone was allowed to react' directly with t-butylamine. Even in this instance hindrance effects were apparent and deserve some attention. I n our previous experience (e.g., ref 2-4) reactions of such bromo ketones wit,h benzylmethylamine had been very rapid. Ethereal solutions about 0.1 J l in bromo ketone and 0.2 M in amine began depositing crystals of benzylmethylamine bydrobromide almost a t once and precipitation was effectively quantitative within 30 min, although reaction mixtures were often allowed to stand longer than that. With t-butylamine under similar conditions the reaction rat'e was negligible. Satisfactory condit.ions were found to exist in l\IeC?i, a very "fast" solvent for this type of reaction, in rather concentrated solution (about 1 M in bromo ketone). Once obtained, 2,5-dimethoxy-cu-t-butylaminopropiophenone ( L X ) showed unusual stability for this type of compound. The base even survives distillation around 0.01 mm without much decompositiori. 2,5-Dimethoxy-~-t-butylaminopropiophenone(LX).-2,5-Dimethoxy-a-bromopropiophenone (9 g, 0.033 mole) was dissolved in 25 g of MeCN. (-Butylamine (10 g) was added and the solution was allowed to stand for 21 hr. At that time some solid (t-butylamine hydrobromide) had separated. Absolute Et20 (200 ml) was added and after a few minutes the solution was filtered from the separated salt which was washed with more EtzO. The filtrate was washed with H 2 0 ( p H ca. 7 . 5 ) , dried (K2C03), and evaporated in vacuo. The residue weighed 8.5 g. It was redissolved in Et20 and added to 15 g of 107, HC1 in MeOH. More EtvO was added to incipient turbidity and crystallization was induced. The first crop weighed 5 g and melted Two grams more was obtained from the mother at 181.5-183.5'. liquors. For analysis, the solid was dissolved in 20 ml of hot absolute EtOH and an equal volume of Et20 was added. The crystalline precipitate now melted a t 184-186". ~-(2,5-Dimethoxyphenyl)-~-t-butylaminopropanol (LXI). ( a ) By Catalytic Hydrogenation.-LX'HCl ( 4 g) in 40 ml of bIeOH was hydrogenated over Adams' catalyst. Absorption of Hz was complete in 4 hr. The solution was filtered, evaporated t o about 10 ml, and diluted with Et2O. The solid so obtained melted a t 249-251" dec. I t was further purified by recrystallization from HzO containing a trace of excess HCI. ( b ) By Borohydride Reduction.-To 3 g of LX.HC1 in 15 ml of 50% EtOH, solid XaBH4 was added in small portions with stirring until excess reducing agent was still demonstrable after 30 min. The solution was then acidified with HCl and boiled down to half-volume in the hood. On cooling and basification, the base precipitated as a solid that melted at 111-112' after recrystallization from EtzO-hexane. I t was then dissolved in hot dilute HCl from which the hydrochloride separated on cooling. The other compounds of Table I V were prepared by the same general method as LX and LXI. Borohydride reduction consistently gave some isomeric amino alcohols but in small-scale preparations these were not isolated. I n large-scale reductions the amount of threo isomer was of the order of The ketone corresponding to the amino alcohol LXX was not obtained as a crystalline hydrochloride. After reduction of a solution of the basic product, LXX . HC1 crystallized readily. Resolutions.-Compounds I, XVIII, LXI, threo-LXI, and methoxamine itself were resolved. The stereoisomeric secbutylmethoxamines were prepared specifically, but o-sec-butyl(+)-methoxamine and L-sec-butyl-( )-methoxamine were also isolated by resolution. Except for the resolution of methoxamine itself all of these were done with D-tartaric acid (either in the first step or entirely). Usually the (+)-base D-acid tartrate was the salt that separated first. Compound XVIII came down as the (-)-base D-acid tartrate with high efficiency (90% of the calculated amount of salt as a trihydrate in the first crop), With threo-LXI also the D-acid tartrate of the ( )-base separated. Methoxamine cannot be resolved directly with tartaric acid owing to the very low roliibility of a neutral tartrate, containing
loco.
+
-
541
both isomeric forms ot base. The acid D-dihenzoyltartrate of (+)-methoxamine separates readily from ble2C0 in a rather impure form. Recrystallization is feasible by dissolving in warm S O % ,\Ie2C0 and boiling off hleYCOuntil solid begins to separate. However, we have not been able to obtain optically homogeneous material by this method. Conversion of such dibenzoyltartrates to the hydrochloride followed by careful crystallization from ethanolic solution by gradual addition of Et20 always resulted in appearance of the characteristic elongated prisms of DL-methoxamine hydrochloride. After the bulk of this had been separated, the glossy leaflets of (+)-methoxamine hydrochloride would appear. Similar operations with ( - )-enriched hydrochloride from the more soluble dibenzoyltartrate fractions gave ( - )methoxamine hydrochloride. Later developments indicated that these salts were not better than about 98% of the predominant isomer. The following process proved eventually to be most advantageous. Thirty mmoles of DL-methoxamine hydrochloride was dissolved in 50 ml of HzO in a 500-ml wide-mouth erlenmeyer flask equipped with a stirrer. To this was added 16 mmoles of D-dibenzoyltartaric acid in 50 ml of J\IelCO and 7 ml of 2.4 S YaOH solution. The solution was stirred under a gentle air stream, and a t intervals over 3 hr a few milliliters of 11e2C0was added to wash down the sides of the flask. Two 20-ml portions of HzO were also added. The weight was now 95 g. The precipitated solid was filtered and dried, 5.8 g. The aqueous layer was washed with Et20 and bahified, and the base was taken into ether and converted to hydrochloride, 4.7 g. The rotation of this hydrochloride indicated that it contained 62% of the ( - ) and 387, of the ( ) isomer. The 5.8 g of n-enriched dibenzoyltartrate was dissolved in 100 ml of 50yo lZIeaCO and to it was added 8.7 mmoles of hydrochloride estimated to be two-thirds ( + ) komer and 4 mmoles of D-dibenzoyltartaric acid. The evaporation and equilibration were repeated as before. The precipitate weighed 5.6 g, mp 175.5-176.5'. The hydrochloride obtained from the aqueous ayer was estimated to be 55yo ( + ) and 45% the ( - ) isomer. With (+)-enriched dibenzoyltartrate of this quality it was advantageous to proceed to resolution with D-tartaric acid. Thirteen mmoles of hydrochloride estimated t o contain 11 mmoles of D and 2 mmoles of L isomer was dissolved in HzO. To this was added 1.8 g (12 mmoles) of D-tartaric acid and the solution was warmed while 5 ml of 2.4 iY NaOH was added. The volume was adjusted to 25 ml and the solution was allowed to cool. After 2 hr the crystalline precipitate was collected and dried, 2.5 g ( 7 mmoles) as the D-acid tartrate. The above solid was redissolved in H 2 0 together with another batch of 6 mmoles of hydrochloride, 1 g of D-tartaric acid (6.7 mmoles), and 4.3 mmoles of NaOH solution. The volume was adjusted to 25 ml and the solution was allowed to cool. The precipitated solid weighed 4.5 g. It was recrystallized from the minimum of HzO together with 0.7 g of n-tartaric acid, 4 g of solid melting a t 162-164" being obtained. D-Tartrate (from the above and several similar batches) (15 g) was dissolved in H,O. The solution was basified and the base was taken into Et20 and the solution was dried (K2CO3). The ethereal solution was poured slowly into a flask containing 8 g of 45% (w/w) HCl in absolute EtOH and 100 ml of Me2C0. The hydrochloride separated as glossy leaflets which were filtered o f f and washed with Me2C0, then with EtzO; 8.7 g, mp 180.5181.5". Of this, 1.0742 g was dissolved in Hz0 plus a drop of concentrated HCl and made up to 25 ml; OD +2.45', [.ID t-28.5 f 0.3". A similar sequence of crystallizations of heavily L-enriched material with L-tartaric acid et seq. afforded ( - ) methoxamine hydrochloride. Application of the same techniques to the resolution of #methoxamine20 was unsuccessful. The dibenzoyl tartrate that precipitated was n-enriched but in our hands did not yield hydrochloride much better than one-third D (two-thirds DL). Further operations through the very soluble n-tartrate or through the hydrochlorides afforded no material worthy of confidence. N-Isopropylmethoxamine (I).-The (+)-base D-tartrate and ( - )-base L-tartrate crystallize readily from HpO in leaflets but in hydrated forms that melt in the vicinity of 90". The melting points are of little value in determining steric homogeneity. The following procedure for the ( + ) isomer was the most successful.
+
(20) I. Satoda. F. Kusuda, T. Omoto, M. K a w a m a t a , a n d Y. Tamamoto, Yakugaku Zasshi, 79, 989 (1969); C h e n . Abstr., 64, 4475 (1960).
The solid base liberated from 130 ninioles of I)L hydrorhloride was dissolved together with 170 nimoles of o-tartaric arid i r i 200 ml of H20. On cooling, the precipitate was collec-led, washed with ice-water, and dried; 28 g. Of this 1 g w : i h holiihle i n 18 nil of 17; D-tartaric acid solution. The total solid was I from 90 g of 1120containing 2 g o f 1)-tartaric, acid ai g of tartrate with a solubility trf 1 g in 21 in1 of I f , t)-tart:irii, acid solution. The second re( allization w a , ~f r o m 65 g o f ,solution containing 2 g of o-tartaric acid aiid gave 9.5 g of ciiliti with a solubility (as before) of 1 : 2 3 . The third and fourth rerrystallizat,iotis from 54 and 41 g of bolut,iori, each with 1 g of excess acid, gave 6.7 and 5.2 g, reupectively, earh with rolubilitieb of 1: 26. The last portion of o-tartrate was cwtiverted to the base which was crystallized from 30 ml of hexane giving hairy crystals that This base war then conre-formed as large prisms, mp 85-86'. verted to t,he hydrochloride arid c~ystallizedtwice from Hd)! mp 242-244' dec. Interconversions between the w y t h r o and thi.eo Forms of I.The acylations of these compounds enroiintered difficulties probably from hindrance. The same factor may have been 0 shift in aqiieoiis involved in the relative slowness of the N acid, but low solubility roiild a1.w have been responhible. X(i attempt was made to isolate the hydrochlorides of the 0-acyl derivatives. Rather, after all the material had tlibsolved i i i aqueous phase, solvent way removed and the residue was- ,subjected to acid-catalyzed ester exchange t c i remove 0-acyl group. :tnd convert all the material to smiiio alcohol hydrochloride. The separat,ion of threo from erythro isomer is easier with racetnict material since the I)L-threo hydrochloride is readily mluble in hot l\Ie2C0and crystallizes from the cooled solution slowly iri elongated prisms. l l o s t of the erythro hydrochloride can be removed first because of its low solubility i n H 4 . The separation of ( + ) A r e 0 from ( - )-erythro hydrochloride follow .similar lines b u t is less facile as the solubilitier are c l o ~ etogether. r DL-N-Isopropyl-1C.-methoxamine.---Ten grama of I base wadissolved in 30 ml of JIeCN together with 25 nil of Et& and 10 ml of Ac20. The sollition was allowed t o stand overnight after which volatile material was removed I'n metro i n a watei, bath at 40-45". The residiie was taken into Et20 and washed (H20, 5% NaOH, € 1 2 0 , .ire HCI, H&). The Et20 was evaptr rated, 200 ml of ,5!;. FICl was added, and the flask was heated under reflux. All material went into iolutiori in about, 15 niiii and refluxing was continued for 1 hr thereafter. The wlutiori was taken down to dryness in vacuo and the residue was dissolved in 200 ml of RIeOH to which 10 ml of concentrated HC1 had bee11 added. The resultant solution was alloTved to si arid overnight and again evaporated in v a c ~ o . The residue wa.: dissolved i r i 30 ml of boiling H 2 0 from whirh 3 . 5 g of 1. HCl cry.stallized O I I c~roling. The supernatant (and hIe,CO washes from the solid) were evaporated in vacuo and the remaining material wa< dihsolved in 32 g of hot Me2C0. This solution was refrigerated yielding 5.7 g of crystalline material melting at 129-132". The solid was redissolved in 50 ml of hot, lIe2C0, a little undihsolvrd material (I'HCI) was filtered off, and Et& was added to cry.;tallixation. Solid ( 3 g ) melting at, 130-132.5" wai' obtained. The best sample, obtained from the Dbo-chloroberizoyl derivative ( i f I, melted a t 129-129.5'. Anal. (ClaH,,?J03~HC1) C, €1. The above procedure was certainly the most convenient and probably the most efficient of those iiivestigated. The product of acetylat,ion is probably a mixture of moiio- atid diacetj-1 derivatives. Experiments (base-catalyzed ester exchange) interided t o remove the 0-acetyl usually removed both acetyl groups. 011 consideration, i t seemed best, to conduct the rearrangement with the crude material. To some extent it would be expected that the ester acetyl should be removed more rapidly than the amide acet.yl. Further, i t is arguable that an S,O-diacetyl derivative hhould undergo the Welsh rearrangement ai' readily as the Nnionoacetyl compound. The original Welsh modification of the Schotten-Baumaun react ion using benzoyl chloride or o-chlorobenzoyl chloride gave hardly any amide from I. Using considerable excess of reagents in repeat'ed additions, mainly diacyl derivatives were formed: oL-diacetyl-N-isopropylmethoxamine, inp 121.,5--122" from EtOAc-hexane [Anal. (ClsHz;NO,) C, 111, oL-dibenzoyl-Nisopropylmethoxamine, nip 119-120" from EtYO [ . I d . (CW &N%) C, HI, DL-bis(o-chlorobenzoy1)-N-isopropylmethoxamine, nip 165-166' from Ale,CO [Anal.(C?a€I?c2ClnN05) C, HI. When allowed to st,and3 days in MeOH with a trace of h-aOlIe: the l)i.~(o-rhlorobenzoyl)derivative was converted t i i the mono +
derivative, nip 170.5 -17%', which i ~ ~ r n laibo d b e ohtainetl tlirect I? t i c r i i :is described above for the ac~etylwtiori. Froin o - ~ ~ h l ~ r r r i b e n z o y l - ~ - i s o p r i i p y l ~ i ~450 ~ t hnig ~~s:~~ni~~~~ of ~ i r , - ~ - i s ~ ~ ~ ~ r ~ i ~ ~ ~ ~ I - ~h~~rli~iichlorid~~ - n i ~ l h o s a n IV:I> i ~ r i ti i~I ~ i : ~ i i ~ r d ill 1.5- 20'; j.ield, f -i- I-N-Isopropyl-+methoxamine.. - J - ~ - h p I ' i i pI >r i i r i liox:iiiiine \v:is :ii~yl:ite11 a n d rearranged liy the procediuc~ givrii iii tlri ail ahtivc.. Thr solid resicliie frorii the final evaporntioii in i'nc')to W:I> PSIi~:ic.tetI t \vice wiih h r i l i n g lIezC0 :rnd i i r i re:irraiiged ~i~.droc~liIiii.i(~f~ \vas ciillec*trtl OII cvoling. Thr iniit 1it.1. Iiqiiirr+, on trdditiori of Et:(), deposit ed miall prisms and rowi t t" of fine iieedles. The prism. were wnioved tnoc~hanirallyn r i t l t lit) were dih,solved i i i H 2 0 . The i ~ ~ i i : i t i ~indi(xteti rii thnt ihe.v were the tlesird t h w o isiinirr. The : i q i i ~ ~ ~ -iiiiit i u ~i o i i piirated i'n u " o arid the residue was taken up i i i i u IlclC'( ), a little iindiwilved ni:rterial (pri>ni IYpeated twice i i i .i 111. a r i d 1.: mi mor(' ojl I l e C S was :uideii I o )Iiiiiirri, After standing 24 h i . 1 he r n i x t i i i ~ e~ t i ai:rkrri c i r o i i r i ii ~ i e a t i ihath and the reaidlie was p a r i i t i ~ r n r ~ l tyO t111t1 I€,O ( p H (;a. 6). Thr etheiwl layrr \\:I,. washed with 1 . \ T:tOll until the w:i.liiiig-. were clr:idy iilkiiiiiit~. then with € J 2 0 u r i t l finally with 1 . \ TIC1 ( X nil ). 'I'hi- :ic:itl tract yielded 0.4 g of base involatile i i i wntei'-piiinp v:ii'ii~iiiio i i 1 l i t , steam bath ( w .2 nimole-). The ethereal la>.er w i s evaporated ( T I ~ ( i r u o( S K J . To i t w:il added 90 nil of 5' IICl, arid the flask was heated uiidcr reflux. All material had tfiasolvetl after 25 miri of boiling. The .oliitioii was refluxed 0.5 Iir loiiger and iheii takeri domn in 'owrlo. T h e residue was diwilved i l l I00 ml ( i f lIeOII, and I O ml of I > ( ; TI was added. Sot all the material disbolved. 11owed to -land ovrrnight, warmed t o .i5" i o :ind t t l l o ~ ~tod h t : i i i d 24 h i . longer. .I i e a large prism.= tvere i i o w pre-ent , The wliitioii TVRS rd'riger:+t(~iI .i hr arid then decanted frnm the i.ryst:il> which were idriitifieti iia i)1,-1. The IIt~OH-olutiori wis ev;ipornted in m c u o a n d thi, residue wab dissolved i n warm lIeyCO. Several sinall r r o p of I werr obtained totalliiig 1.7 g 16 m m o l e ~ ) . .I total of :3.:3 g i l 1.:; ninicile,+)of uL-thrr.o h~.drcichloi~ide wa.: ~ r t ~ t : i i iI)y i ~ ~:i(lditioii l o). I < t ? oto the S ~ J ~ I I ~ ~ O I I ~ . I,* o(. i > i , - f , r ~ / / h r o:iild nI>-//zrco Ii~drociil1~rit1~~111, wit t i 20 nil of 1 .\' II('l! a i i c l rrlfluseti 1 ..i -IIII \\a+ iicrt quite houiid i)ecarise niiich (it' itic, erythro hydroctiiiJritfr iwnniried undissolved t hrciughout 1 litL heating perilid. Tht~i,tvidioii mixtures were worked up aloiig i h t ~ lines desrrihed a h o w . The origiiinlly eiythro ninirrial yirlded .i1 riig o f nc~etciirr-.iiluhle hydrochloride ( f h r c d . The oi,iginally ihreo niaieiinl g:ivc' 77 nig of :~c.etoiiP-iiisi,liiblehytirochlr~ridr icr!lthro i. N-t-Butylmethoxamine (LXI).~- - T h c A rr+oluiic (~onsiderableexrew of t:rrtaric arid ( 50-100r~) wliible salt is the iir.-base neiitrd u-t;irii,ate, rnp I X I babe 141 ninioles, 11.2 9 ) arid 65 inniole tartaric acid in about 50 in1 o f hot HnO gave, on ( d i n g , wellformed iieedles nieltirig about 100°. a f t e r recrystallizatioii frciiii 20 nil of hoiling f f , o i~ontair~ing 2 g C J f excess tartaric. acid, 1 hrrtX was obtained 6 g 114 mmoles) of halt, nip 196---197..i". 'rlii, tart,rate was dissolved i n II,O and the base w a s liberated. Thi. was crystallized from lSt~lO-per~tarleat 4" giving w e l l - f ~ r ~ ~ n i ~ ~ l needles melting at 7 2 -2".The cry to the hydrochloride which caii he cr or from H20. The melting point, va (2.53..i--2.53 and 261L262' have been obtained) arid iu not reliabltx. The I + ) hydrochloride and Ur, hydrochloride do mil difyer 211)preciahly i i i melting point but the DI, hydrochloride is much lc+hq soluble ( ? a . I r i and at least 4";, wspectively) in HrO and milst h a racemic conipoii~~d. The rricithei, liquor f r o m the separatioii of the ( t )-t)a>f, utartrate wa: basified arid the ba.se was taken into Et20 a n d dried IK2COaj,and the solntioii was boiled down with hexane t o abou I 20 nil. On cooling, fine crystals (i)r,-base) separated first, followed hy long needles / ( - )-baxej. Solvent war added t c r di..oIvr~ thiL rreetllr- : i n r i t tie ~oliitioiiwaq decariied from thcl .I( lit1
+
+
~~
is\-
(
l h I p ( ' ( )
July 1968
ALKYLMETHOXAMINES
soluble crystals, leaving about 2 g of DL-base. The mother liquor, estimated to contain about 1.5 g of Dkbase and 3.5-4 g of (-)base, was evaporated to small volume and dissolved in 13 ml of H 2 0 t,o which 4 g of ktartaric acid had been added. The solution was heated to boiling, and H 2 0 was added to a weight of 24 g and allowed to cool. The ( - )-base t t , a r t r a t e that separated weighed L.5 g (13 mmoles) and melted a t 196.5-197.5'. The btartrate was dissolved in H2O and the base was liberated and crystallized from hexane. I t weighed 3.8 g and melted at) 74". The base in turn was converted to the hydrochloride, 3.5 g, glossy leaflets from EtOH-Et20; mp 259-261" dec. threo (Pseudo) Form of LX1.-The starting material was the basic oil remaining after crystallization of LXI base from hexane when LXI was being prepared on a relatively large scale (ca. 1 kg) by borohydride rediiction. This base formed a hydrochloride crystallizing from dilute HC1 and melting at 234-23.;" dec. I t had the same composition as LXI.HC1 hiit the honiogeneity (as the oL-threo isomer) was uncertain. The oily base when dissolved in H20 with a 507, excess of n-tartaric acid gave, from cold solutions, needles melting a t 56.59", probably hydrated and rather soluble in cold H20 and in Me2C0. As a consequence, the loss in washing was severe and it was expedient to keep the first filtrates separate from the washings. After several recrystallizations, material melting at 71-72.5" was obtained. Six grams of this quality was obtained from 44 g of base in three experiments but much partially beparated material remained in the washings. The base in the first mother liquors was liberated and converted to hydrochloride. The solution was dextrorotatory showing that the u-tartrate was that of the (-)-base. When the dextrorotatory solution of hydrochloride was evaporated to small volume, crystals separated and were observed to be of two types: one of irregular shape and holding onto a brownish contaminant present originally in the solution melted a t 234-236" dec; the second type, nearly colorless lath-like crystals, melted a t 222224" and showed [ a ]+47.4' ~ (4% in H2O). This being in the range expected for a threo ibomer, lath-shaped crystals (6.5 g) were separated where possible from several such preparations. This material was recrystallized twice and then melted at. 219223' dec and had the composition of a monohydrate. The anhydrous form, obtained by drying in high vacuum, melted a t 22.7-226". The hydrate was used for rotations. The (-)-base was recovered from the D-tartrate salt and converted to the hydrochloride which was also obtained as the monohydrate. Small portions (0.5 g each) of the (+)- and (-)-threo hydrochlorides were combined in aqueous solution and allowed to tallize. As first obtained, the salt melted a t 220-223" (eff) and had the composition of a hemihydrate. On stringent drying, the melting point rose to 232-233'. This material was not clearly distinguishable from the original hydrochloride (mp 234-235 O dec) but the lat,ter may have contained appreciable amounts of erythro material. Compound XVIII was resolved with minimal difficulty. The use of ktartaric acid to purify the (+) isomer was unnecessary since the (+) and ( - ) hydrochlorides are less soluble (and higher melting) than the DL hydrochloride. Stereospecific Synthesis of the Isomers of N-sec-Butylmethoxamine (II).-Model experiments were carried out with racemic material to determine the proper reaction conditions. In these, an excess (1.3-1.4 molar equiv) of methoxamine base was treated in 1IeCN with sec-BuBr or mesylate (e.y., 10 mmoles of base plus 7 mmoles of bromide in 10 ml of LIeCN). After 3 days at. room temperature about 77, of the bromide had reacted ( b y titration). I n a similar experiment at reflux for 21.3 hr, 57% had reacted. With sec-butyl mesylate, 8-97, of the ester had reacted after 1 week. At reflux for 22 hr, 5.7 out of 7.1 mmoles had reacted. Since a considerable port,ion of this reaction could have been elimination, unreacted methoxamine was separated from sec-butylmethoxamine. The titration of the latter gave a figure of 3.68 mmoles. It later became apparent that in these conditions the principal reaction was between (+)-base and D reagent or (-)-base and L reagent. I n the specific reactions of the (+)-base with L reagent, et,c., considerable elimination may have interfered. The separation of methoxamine from sec-butylmethoxamine depends upon the fact that methoxamine base is fairly soluble in HzO and very little soluble in petroleum &her fractions. (The distribution ratio of met.hoxamine for hexane-Hz0 is 0.004. For the system 1 : 1 Et20-hexane/H20, the ratio is 0.08.)
S43
The procedure used for isolation of secondary bases was consequently to remove the solvent in vacuo and then to partition t.he residue between Et20 and dilute alkali (two funnels in series). After the total bases were taken into the ethereal layers, these were diluted with equal volumes of hexane and washed serially with small amounts of HzO until the washings were a t about, p H 8. As a guide, these washings could be titrated with standard acid. The secondary bases could next be extracted from the EkOhexane layers with standard acid and the quantity of secondary base was determined approximately. The predominant isomeric hydrochlorides were isolated by fractional crystallization from aqueous solution. The procedure was complicated by t,he fact that earlier specimens of (+)- and (-)-methoxamine used were riot in fact sterically pure and by the relatively poor yields in the reactions of D-mesylate with ( - )-methoxamine and L-mesylat,e with (+)-methoxamine. The isolation was a ences i n crystalline form, the D-( + ) and I.-( - ) salt,s forming rather heavy rectangular prisms while their diastereomers tended to crystallize in needles. Separation of I1 into Component Racemates.-Various attempts to fractionate I1 .HCl had been unsuccessful. Eventually it was found that when the base was refrigerated in EttOpentane solution two types of crystals appeared. A partial separation was accomplished: from 9.5 g of mixed bases 7 g of the preponderant base was obtained. This crystallized in heavy plates melting a t 82-84'. The minor component cryst,allized in needles melting a t 96-98". This amounted to 2 g and seemed fairly pure. The major component, which was much more likely to be contaminated, was recrystallized repeatedly and two frartions, I (2.6 g), mp 86-87', and I1 ( 2 . 3 g), mp 84-85.5') were finally accept,ed as fairly reliable. The two specimens were resolved with D-tartaric acid. From the 96-98" base there wa.* obtained through the o-tartrate, 120 mg of D-( +) hydrochloride, [ a ] D +19.3 i 1". The lower melting base was similarly resolved yielding 435 mg of L-( + ) HCl, [ a ]+~l 5 . 2 + 0.6". (+)-Methoxamine and DL-sec-Buty1bromide.-Pure ( )methoxamine (20 mmoles) (from 5.0 g of pure hydrochloride) and 13 g (0.1 mole) of DL-sec-BuBr were dissolved in 100 ml of MeCIC'. The solution was refluxed 16 hr after which most of the solvent was boiled off. The residue was dissolved in 50 ml of .iOYc MeOH and titrated to pH 5 (outside indicator) with standard HC1; 7.42 mmoles of acid was required. As there was a tendency for solid to separate, 5 ml of concentrated HCl was added and more Hz0 and the solut,ion was evaporated to 47 g. On cooling, 1.8 g of crystals separated ( 6 mmoles if hydrochloride). The filtrate was subjected to the separation described previously. The washings appeared to contain 5.1 mmoles of methoxamine. The subsequent acid extracts (containing secondary amine) were evaporated to dryness in vucuo, the earlier crop of crystals was added and the total solids were dissolved in 100 ml of RIeOH to which 20 g of 40% (w/w) HCl in EtOH had been added. This solution was boiled for 2 hr allowing solvent to evaporate slowly (to remove any bromide present,)Z' and then evaporated in vucuo. The residue was dissolved in H 2 0 with a trace of added HC1, charcoaled, filtered, transferred to a tared beaker, and evaporated to dryness on the steam bath. The weight of the residue was 2.83 g. This solid was redissolved in HZ0 with 1 ml of 1 S HCl and made up to 50 ml a~ +1.82 =k 0.02", [ a ]+16.1 ~ =k 0.2". On the basis of specific rotation of $14.7 and +19.6" for the L-( + ) and D-( +) isomers, respectively, this corresponds to 70-75y0 of the former. The solution was evaporated to 15 g giving a first crop of flattish needles and then to 10 g giving a second crop of the same type. The two crops were combined and recrystallized from 14 g of aqueous solution. There was obtained 1 g of solid identical with the L-( + ) isomer previously prepared. Hydrogenation of ( - )-Methoxamine and Butanone.-The base was liberated from 4 g (16 mmoles) of (-)-methoxamine hydrochloride and taken into Et20 and t'he Et20 was evaporated. The residue was dissolved in NeOH in a hydrogenation bottle containing 1 g of freshly prepared lOY0 Pt-C and 7 g of freshly distilled butanone and shaken under Hz. The absorption of Hz was 13.7 mmoles in 140 min. The solution was allowed to stand overnight with a Hz uptake corresponding to 4 mmoles. Shaking for 1 hr further gave an uptake of only 0.2 mmole. The solution was acidified with HCl, removed from the catalyst, and evaporated in oucuo. The residue was subjected to the separation as
+
(21) A . P. Phillips and R. Baltzly. J . A m . Chem. Sac., 74, 5231 (1952)
Adrenergic Neurone Blocking Agents.
1II.l
Heterocyclic Analogs of Guanoxaii
Analogs of guaiioxari niodified in t'lle h e t e r o q d i c ring have been cotnparetf with respect to their adrenergic iieurone blocking poteiick, their effects7's. epinephrine atrd riorepirtephrinr iir the (,:it, s11d their :~ntihypeiteiisive effects in the dog.
I t has been reported? that the niitihyl)erterisivc agent guanoxan (1) disl)layed, like guanethidine, adrenergic ricurorie blocking properties in the cat, but in roritraqt t o guanethidine, guanoxan also displayed a classical blockade of CY receptors. The effects on these properties upon introduction of wbstituent5 t o the aromatic ringla or t o the dioxane riiig,lh and of' modific:ttion of the side chain, l a have already been discu\sed. This paper is concerned n ith the adrenergic neuroii(1 Mocking activity of 2-10 in which the heterocyclic ring o f guanoxan ha5 been modified. Syntheses of most of these cornpounds have already been described in the literature. -4dditional features of interest, and details of those syntheses not already reported, are described in the Experimental Section. Biological Results and Discussion.-Xdrenergic neurorie blocking activity of the compound- was measured (1) (a) Paper I in this series: J . Airgrtein, S. A I . Green, .\, AI. 3Ionro. G . W.H. Potter, C . R . Worthing, and T. I. Wriglep, .I. .\fed. Chem.. 8, 440 f,lRXj; (1,) paper 11: .IRI. . U o n r o , (;. IV. I t . Potter. and \ I . ,J. S e \ \ - ~ l l
zbid., 10, 880 (1967). f 2 ) .\I. , J . I>avey a n d f t . Iieinert, Brit. ./. / ' / ~ o ~ m f l c d24, . , 29
(l!ltj.i).
1
5
4
7
10
3
2
6
8
9
11
12
/NH
G = -NHC/ "13,
