1740
COMMUNICATIONS TO THE EDITOR
VOl. 54
TABLE I THEHETEROYOHIMBINE ALKALOIDS F -e-r--
Groups and stereochemistry
A ( d o ) C-19 CH3: a and e
B (epiallo) (2-19 CH3: a a n d e
Alkaloids
Tetrahydroalstonine ( I ) Aricine Reserpinine Isoreserpiline Akuammigine (11) Isoreserpinine Reserpiline Rauniticine (111) Raunitidine hlayumbine ( I V ) Isoraunitidine AjmaIicineb Tetraphylline Raumitorine"
basesC?! quinolizidine fusion C-19 methyl (from infrared chemical shift a t 3 . 4 ~ ) (and J in c.P.s.)
trans
czs
1.38 ( 6 . 1 ) 1.37 (6.3) 1.38(6.1) 1.39 ( 6 . 2 )
... 1.32 (6.5) 1.32 ( 6 . 3 ) 1.42 ( 6 . 7 ) 1.42 (7.1) 1.38 ( 6 . 3 ) 1.35(6.8) 1.16 (6.7) 1.16 ( 6 . 5 )
-Methiodide
'N-CHa chemical shift
3.42 3.43 3.39 3.41 ... 3.32 3.35 3.50 3.49
C!D fusion
saltsa---
c-1s
chemical shift (and J in c.p s . )
A p.p.m.
c-19 methyl base salt
trans
1.46 ( 5 . 8 ) 1.47 ( 5 . 4 ) 1.48(5.4)
... trans
...
...
1.39 ( 6 . 2 ) 1.39 ( 5 . 7 ) 1.40 ( 6 . 5 ) 1.41 (6.1)
-0.07 -0.07 $0.02 +O.Ol
-0.08 -0.10 -0.10
...
C ( a h ) (2-19 CH3: trans cis 0 and a D (epiallo) C-19 trans ... ... ... ... CH3: p and e 3.31 trans 1.43 (6.0) -0.08 E (normal) C-19 trans 3.35 -0.10 trans 1.26 ( 6 . 2 ) CH3: a a n d a 3.36 -0 10 1.26 (6.2) F (normal) C-19 truns ... ... ... ... CH1: p and e A11 a The n.tn.r. spectra of the methiodide salts were obtained in formamide solution with T M S as an internal solvent. spaces left blank signify that insufficient samples were available for the spectral measurements. * ,4n early suggestion t h a t the stereochemical proposals of Wenkert for the heteroyohimbine alkaloids needed to be modified was made by Professor E. E. van Tamelen a t the 139th meeting of the American Chemical Society held in St. Louis, Missouri. The C-3 hydrogeri in raurnitorine is alpha and axial as indicated originally by J. Poisson in his Docteur-Ps-Sciences thesis, "Kecherches sur les Alkaloides des Kacines du Rauwolfia vomitaria," Department of Pharmacy, University of Paris, p. 46 (1959).
To settle whether the above isomerization was of the allo to epiallo type or vice versa, we had resort to a new criterion based on the n.m.r. chemical shifts of the +N-CHs protons of the corresponding methiodide salts. W-e found that, in analogy with Katritzky's original findings in a series of N-methylquinolizidinium cations,s the +N-CH3 protons absorb a t lower fields in salts with cisfused C/D rings than in the tra.ns-fused analogs. Hence raunitidine must be as in 111, and must undergo a conformational change before methylation to yield ultimately I I I a which has a cisfused C,/D system. In other words, it is the conformer with the less hindered nitrogen atom that undergoes methylation. On the other hand, isoraunitidine must be represented by IV which methylates directly to the quaternary salt IVa. The conversion of raunitidine to isoraunitidine therefore involves an allo to epiallo isomerization. To determine the stereochemistry of the (2-19 methyl group, two lines of reasoning were used. First, it can be seen that the alkaloids of groups B and D although belonging to the epiallo series exist in different conformations, namely, I1 and IV, so that they can in each case accommodate the C19 methyl group in the preferred equatorial configuration. The second criterion is based on the chemical shifts of the C-19 methyl doublets in the n.m.r. spectra after treatment of the free bases with methyl iodide. -As can be seen from Table I, the C-19 methyl groups of the methiodide salts experienced a downfield shift in relation to the corresponding free bases, due to the deshielding of the C-methyl protons by the net positive charge on the nitrogen. However, in group C a small upfield shift (-0.015 p.p.m.) was recorded. This apparent anomaly (8) A . R . Katritzky, private communication. This work on t h e stereochemistry of the yuinolizidines will appear in the Joikmal of the Ciiaiiiicai S o i i r l y .
can be explained in terms of the immediate proximity of the C-19 methyl protons in I11 to Nb which leads to some hydrogen bonding and a consequent diamagnetic shift for the protons in question. \n:e have now defined the complete stereochemistry of two groups of a110 and two of epiallo heteroyohimbines. The remaining two stereochemical groups are E and F. The alkaloids of group E are now known to be of the normal configuration with the C-19 methyl group alpha and axial.' Hence by a simple process of elimination raumitorineg in group F must also be normul but with the methyl group beta and equatorial. Acknowledgments.-The authors wish t o thank S. B. Penick and Co. for a most generous gift of rauniticine, raunitidine, and isoraunitidine, and Dr. R. Goutarel for providing samples of the rare mayumbine and raumitorine. We are grateful to the National Science Foundation for financial support (Grant No. NSF-Gl9870). (9) There was insufficient sample of raumitorine for a n n . m . r . spectrum, b u t the infrared spectrum in carbon disulfide solution near 8 . 4 ~clearly indicated this alkaloid t o be different from any other heteroyohimbine here discussed. (10) Recipient of pre-doctoral fellowshig S o . GF-17,2'32 from the Division of General Medical Sciences of the National Institutes of Health.
MAURICE S H A M M A DEPARTMENT O F CHEMISTRY THEPENNSYLVAXIA STATE USIVERSITY JASE B. hloss1° UNIVERSITY PARK,PESNSVLVASIA RECEIVEDFEBRUARY 2, 1962 ALKYLATION OF PHENOL WITH A HOMOALLYLIC CHLORIDE
Sir: We wish to report on the uncatalyzed alkylation of phenol with 5-chloro-2-methyl-2-pentene(I). This reaction has uncommon features among which arc the facile formation of a seven-membered cyclic
ether and nucleophilic attack only a t the primary carbon of a tertiary homoallylic (or cyclopropylcarbinyl) cation.' When I was heated at 150' for several hours with excess (1 :4) phenol, hydrogen chloride was evolved.2 The three 1: 1 alkylation products3 were characterized by infrared, ultraviolet and n.m.r. spectra and by unequivocal independent syntheses.4 They are 5,5dimethylhomochroman (11),5 1,1-dimethyl-5-tetralol (111) and 1,l-dimethyl-7-tetralol (IV).'" Neither of the two remaining phenolic isomers of I11 and IV, nor the anticipated product, p-dimethylcyclopropylcarbinylphenol" was present.
6
+
1741
COMVIBIUNICATIONS TO THE EDITOR
May 5, 19ti2
(CH,),C=CH;H~&H~CI I
-+
OH arom
I
d'
I
-
Hod 1v
d
I1 The orientation in I11 and IV, with tertiary carbons meta rather than para to the hydroxyls was a t first surprising and suggested that I might ionize in phenol t o a homoallylic (V) or dimethylcyclopropylcarbinyl (VI) cation. Nucleophilic at-
Vb
vI
( 1 ) For leading references concerning charge distribution in homoallylic and cyclopropylcarbinyl cations, see S . Winstein and E. Kosower, J . Am. Chem. Sac., 81,4399 (1555); G . H. Whitham, Pioc. Chem. Soc., 422 (1961); R A. Sneen, K . M. Lewandowski, 1. A. I. T a h a and B. R. Smith, J . A m Cham. S O L .83, , 4843 (1961); A I . S. Silver, 11.C. Caserio, H. E. Rice and J. L). Roberts, J . A m . Chem. S o c . , 83, 3671 (1Y61). (2) No reaction occurred with a saturated analog under similar conditions (3) 4 t least 82% yield, formed in comparable quantities. Two very minor iroducts, with 2 X and 3 X t h e retention times of t h e phenols 111 and IV, probably represent 2: 1 alkylation; they were not investigated. (1) Reaction of I with phenol in acetone containing suspended b.p. 85-90° potassium carbonate gave ~-phenoxy-2-methyI-?-pentene, a t 1 m m . , n% 1.5112 which, with a drop of sulfuric acid gave 11, m.p, 47-48', in nearly quantitative yield. 111, m.p. 112.5-113.5°, was obtained from l,l-dimethyl-5-methoxy-2tetraloue~by WolffKishner reduction and hydriodic acid-acetic acid cleavage. I V , m p. 105-106O, was synthesized by dimethylation of 7-methoxy-2-tetraloneT with sodium isopropoxide and methyl iodide, Wolff-Kishner reduction and acid cleavage. All compounds gave satisfactory analyses. ( 5 ) T h e remarkable ultraviolet absorption spectrumg and basicity9 of I1 prepared in this way has been described. (6) J W. Cornforth, R. H. Cornforth and R . Robinson, J . C h o n . S o c . , 689 (1942). ( 7 ) B. W . Horrom and H. E. Zaugg, J . A m Chem. SOL.,72, 721 (1Y50). (8) H H a r t and C. R Wagner, Pvoc. Chem. Sac., 284 (1958). (9) E h1. Arnett and C . Y . W u , J . A m Chenz. S O L .82, , 6660 (1560). ( I O ) Identical products are obtained when hydrogen chloride is passed through a heated solution of 2,2-dimethyltetrahydrofuranin phenol. (11) Expected, because I is converted readily in high yield by aqueous potassium carbonate t o dimethylcyclopropylcarbinol;T . A . Favorskaya and Sh. A . Fridman, Z h w . Obshchei K h i m . , 15, 421 (1945).
Fig. 1.-3.m.r. spectra of unlabelled I1 (above) and of I1 obtained from I-5-da (below), in carbon tetrachloride, scale in r . The peak a t 8.66 r is at one-tenth the amplitude of the remainder of the spectrum.
tack by the oxygen or ortho or para carbon of phenol at the primary carbon of any of these intermediates would lead ultimately to the only observed products. To test whether an unsymmetrical (Va) or symmetrical (Vb or VI) intermediate was involved, 5-chloro-2-methyl-2-pentene-5-d2 was synthesized and subjected to the reaction. The n.m.r. spectra of unlabelled I1 and of I1 obtained from labelled I are shown in Fig. 1. Not only is there a change in the relative areas of H a : Hb :Hd:H : arom from 2:4:6:4 to 1:3:6:4, but Ha which is a triplet in the spectrum of unlabelled I1 is a singlet in the labelled product (implying that whenever H, is a CH2 in labelled 11, its neighbor must be CD?). There can be no doubt that the two methylenes (C, and C,) in I become equivalent during the reaction.12 The n.m.r. spectra of I11 and IV from labelled I1 show similar scrambling. Additional mechanistic details of this reaction and its general significance in the reaction of nucleophiles with homoallylic systems and in cyclialkylation^^^ will be investigated. Acknowledgments.-We are grateful to the Petroleum Research Fund of the =Imerican Chemi(12) Whether recovered I in an interrupted alkylation shows C r C a scrambling will be determined. (13) H. A. Bruson and J. W. Kroeger, J. A m . Chem. Soc., 62, 36 (1940); H. A. Bruson, F. W. Grant and E. Bobko, ibid., 80, 3633 (1958) ; L. R. C. Barclay, J. W. Hilchie, A . H Gray and N. D. Hall, C a x J . C h e m . , 38,94 (1560).
1742
COMMUNICATIONS TO THE
EDITOR
1701. 84
cal Society (Grant 48s-C) and to the National Science Foundation (G-14289) for financial support.
the systems approached epimeric equilibrium. This was measured by oxidizing each sample to KEDZIECHEMICAL LABORATORY JAMES L. CORBIU sulfone I, whose rotation and deuterium content MICHIGASSTATEUSIVERSITV HAROLD HART were measured by methods previously reported.?b EASTLANSING, MICHIGAS CHARLES K. LVAGSER LVith phosphine oxide, the polarimetric rates were RECEIVED MARCH15, 1962 measured in a conventional manner, and the deuterium content measured through use of the infraCONFIGURATIONAL STABILITY OF CARBANIONS red bands a t 10.98 and 14.29 p present in the deuterated and absent in the undeuterated compounds. STABILIZED BY &ORBITALS' Comparison of the steric courses of the exchange Sir: In an earlier investigation, the stereochemical reactions of sulfoxides I1 and phosphine oxide III course of base-catalyzed hydrogen-deuterium ex- with that of 2-phenylbutane indicates that the change of sulfone I was studied through deter- d-orbital containing functional groups exercise mination of k e l k , , in which k e was the rate constant even less stereochemical control over the course of for exchange and k , for racemization. Values of the exchange reaction than does a phenyl group this ratio varied between a high of about 2000 with Indeed, the data suggest t h a t what stereospecificity tert-butyl alcohol as solvent to a low of 10 with is observed with the sulfoxide and phosphine oxide systems is associated with asymmetric solvation dimethyl sulfoxide.2 rather than with carbanion asymmetry. This CH1 CHJ conclusion is in striking contrast to that drawn for * I *I sulfone system I, which provided high though %-C6Hia-C-H( U) n-C6Hia-C-H(I>) differing retention for exchange in all solvents. I SOC6Hj Two different hypotheses could account for the I1 difference in behavior of carbanions derived from CH3 sulfone I and oxides I1 and 111. (1) Non-stabilized *' carbanions probably have sp" hybridization similar n-CtiHia-C-H( D) to amines. Carbanions stabilized by d-orbitals 111 PO( CtiHj)? might have a tendency to rehybridize to sp'-p IVe have now prepared by conventional reactions since p-d overlap might be more favorable than both diastereomers of sulfoxide 113and phosphine sp3-d overlap. However, many investigators have oxide I I I ? in optically pure states, and have concluded that d-orbital stabilization of carbanions determined approximate values of the ratio of does not have rigid geometric requirements, although kelk, for the three substances in a number of dif- different geometries provide different stabi1ities.j ferent solvents (see Table I). Since the configura- Superimposed on overlap effects are electrostatic TABLEI EXCHAXGE STERICCOURSEOF HYDROGES-DEUTERIUX ---CompoundNat.
BaseConcn. .\-
IIa-11 Ila-d IIb-11 111-h 111-d IIa-d 111-d 111-h 111-d 111-d 995; deuterated
0.17 0.17 0.17 0.20 0.20 0.17 0.20 0.20 0.21 0.21 (ref. l b ) .
Solvent
( CH3)3CODU
*
(CH3)jCOH (CHa)3COD" (CH3)aCOD" (CHa)aCOH ( CH3)2S06 (CH3)2SOh DOCH2CH?ODa HOCH2CHzOH CH30H 1.2 -11 in methanol.
tion a t sulfur of I I a and I I b was unaltered during the exchange at carbon, k , for these compounds actually is associated with the rate a t which 11) W e are pleased t o acknowledge financial support of t h e research reported here by t h e National Science Foundation. ( 2 ) (a) r). J. Cram, W. D. Nielsen and B. Rickborn, J . A n t . Chem. S u r . . 82, 6415 (1960); (b) 13. J. Cram, D. A . Scott a n d W. D. Sielsen, i b i d . , 83, 3696 (1961). (3) Carbon, hydrogen, sulfur and phosphorus analyses deviated from theory by a maximum of 0.2870. Deuterated I I a , I I b and I11 contained 9 7 4 8 % deuterium in t h e position indicated (combustion analysis and falling drop method by J . S e m e t h ) . Isomer I I a had m . p 20.7-21.54, [a]% -192 i 3' ( c 2 , 9.5% ethanol), and isomer 1 I b mas an oil which was separated from I I a b y chromatography o n silica gel, [ e ] % b + I 6 6 i :j0 i t 2 , 95;; ethanol). These isomers possessed the Dame configuration a t carbon, hut dibered a t sulfur. Compound I11 possessed m.p. 94-9S3, [e]%t. - 14.7' i c 0 . carbon tetrachloride). All three compounds uere prepared frc,m optically illire h i c t y l to%yl:it?
Nature
(CH,),COK (CHa)aCOK (CHahCOK (CH3hCOK (CH3)sCOK CHaOK CH30K DOCHzCHzOK HOCHzCHzOK CHaOIi
~-
Concn \
0 78 I1 79
0 78 0 34 0 34 0 23 0 21 0 28 0 29 0 30
I ,
oc.
60 00 00 100 100 60 75
176 175 175
k,lka
2 3 3 3 3 1 1 1 1 1
effects. In the most favorable conformations, repulsion between the negative charge on carbon and the partial charges on oxygen attached to sulfur or phosphorus would be less for sp8 than for sp2-p hybridization a t carbon. Thus the overlap and electrostatic effects might oppose one another and a continuous spectrum of configurations a t carbon ranging from trigonal t o tetrahedral can be envisioned. Possibly no configurational stability is associated with trigonal carbon, whereas (1) I n tevl-butyl alcohol, k , ' k , > 10 lor t h e exchange of 2-phenylbutane-2-d [ref. l a and 11. J. Cram, C. A . Kingsbury and B. Rickborn, J . A m r h e i n . S o c . , 83, 3688 (196111. ( S ) ( a ) W. E. I h e r i n p and I,. K . l.evy, ibid., 77, 511 (19,X)); (1)) H E. Zimmerman and B. S. Thyagarajan, ibid., 82, 2505 (1960); ( c ) K. Breslon and Is. hfohacii, i b i d , 8 3 , 4100 (1961): (d) S . One, 8'.Tapaki :ind .4. Ohno, i h i d , 83,30.:T (1981).
