Tetracyclic triterpene synthesis. Part 2. Deduction of ring fusion configurations incis- andtrans-7-methoxy-3a,9b-dimethyl-1,3,3a,4,5,9b-hexahydrobenzeinden-2-ones by nuclear magnetic resonance spectroscopy

摘要

1978 117 Tetracyclic Triterpene Synthesis. Part 2.l Deduction of Ring Fusion Configurations in cis-and trans-7-Met hoxy-3a,9 b-di met hyl-I,3,3a,4,5,9b- hexahydrobenzeinden-2-ones by Nuclear Magnetic Resonance Spectro- scopy By Keith G. Orrell, Richard A. Packer, Vladimir Sik, and J. Stanley Whitehurst," Department of Chemistry, The University, Exeter EX4 4QD Detailed analyses of the lH and 13C n.m.r. spectra of the cis-and trans-3a,9b-dimethyl derivatives of 7-methoxy-1,3,3a,4,5,9b-hexahydrobenze inden-2-one enable a clear distinction to be made between the two ketones. The lH spectra show that the cis-ketone exhibits rapid inversion of the cyclohexene ring whereas the trans-ketone is conformationally rigid. The spectral parameters for these two compounds, and their 3a-CD3 and 3a-CD2H ana- logues, are discussed.THE preceding paper describes the stereospecific HX 90E spectrometers, respectively. Ileuteriocliloroforni conversion of 6-methoxy-1-tetralone (1) by way of ( compound (9) into 7-methoxy-trans-3a,9b-dimethyl-1,3, 3a,4,5,9b-hexahydrobenzeinden-2-one(10). The estab- lishment of the trans-configuration in this compound was not immediately apparent from an n.m.r. spectro- scopic study, and in order to aid the assignment of configuration the corresponding ketone of opposite configuration (12) was prepared. This paper describes how the configurations of the two ketones may be established with considerable certainty from the lH and 13C n.m.r. spectral features of the ketones and their 3a- deuteriomethyl analogues.The conclusions based on the spectral evidence were subsequently confirmed by chemical evidence as described in the preceding paper. EXPERIMENTAL ,Waterids .-The cis- and trans-dimethylhexahydrobenz-indenones were prepared as described in the preceding paper.l The dideuteriomethyl compound (11) was readily prepared by substituting dideuteriomethylene iodide for methylene iodide in the cyclopropanation of the un-saturated alcohol (9). The deuteriomethyl ketone ( 13) was prepared from the tetralone (1) by way of the com-pounds (2), (4),and (6). The action of trideuteriomethylmagnesium iodide on the ketone (3) gave compound (7). If the acid-catalysed cyclisation step (5) (12) which yields the cis-ketone is a concerted step then the corresponding reaction with (7) ought to lead, by way of the ion (8),to the deuteriated compound (14).In the event, when compound (7) was treated with formic acid it gave the undeuteriated com-pound (12). It follows that the cyclisation is not a con- certed process Spectra.-100 MHz N.1n.r. spectra were recorded with a JEOL MH-100 spectrometer operating under internal field- frequency lock conditions. When accurate comparisons were made with computed spectra, experimental line posi- tions were measured with a frequency counter. Variable temperature measurements were carried out with the standard J ES-VT-3 temperature controller. 220 MHz lH spectra and 22.63 MHz 13C spectra were recorded at the P.C.M.1'.Harwell with Varian HI-200 and Briiker Part 1, I. 1.Packer and J. S. Whitehurst, preceding paper. RR.8y0) was used as solvent for all the samples. Me0@:: Me Me0 1(1)R=O,R 2= H2 (9) (21 R1=0,R2= H ,CD3 (31 R1=0,R2= Me,CH?CCL:CH, (4) R1=0, R2= CD3 I CH?CCL:CH, 1(5)R =CH, R2=Me ,CHiCCI:CH2 ~ (6)R'=CH,, R2=CD, ,CH,CCL:CH, (7) R~=cD,, ~2Me ,CH~XCH~ 1(8)R =CHD, ,+ charge ;RZMe ,CHiCCL:CH, A-H are protons (10)R1= R2= Me (11) R'=Me , R?CHD2 Me0 A-ti are protons 12(12) R=R=Me (13)R =Me, R2=CD3 1 (1L)R1=CHD2 ,R2=Me RESULTS AND DISCUSSION The lH 100 MHz spectrum of the cis-ketone (12), prepared from (1) by way of (3) and (5),was examined first. The spectrum was essentially first order in appearance with the exception of the 4-and 5-H, region, which exhibited lines typical of an ABCD system simplified by virtue of certain chemical shifts and coup- ling constants being fortuitously equal (see later).The three aromatic proton signals were assigned on the basis of the relative magnitudes of ortho-, meta-, and para-coupling constants. The two signals in the methyl region could not be unequivocally assigned at this stage. The 4-and 5-H, region showed an unexpected spectral pattern. The low-field portion was assigned to the protons (c and D) on C-5 on the basis of their proximity to the aromatic ring and their smaller lanthanoid-induced shifts (LIS) as compared with the C-4 protons (A and B). It was not possible to distinguish between J.C.S.Perkin I however, for this parameter to be exactly zero. In the present case it appears that A~(AB)0.01. The 4,5-H2 region spectrum was accurately analysed, using a modified version of the LAOCNR program,2 on the basis of an ABCD system with the nuclei A and B both isochronous and magnetically equivalent and with all TABLE 1 220 MHz lH Spectral parameters for ketones (10) and (12) Position trans-Ketone (10) cis-Ketone (12) Ketone (10)-r-Ap ketone (12) ASi(p.p.m.)si(P.P.m.) i,j JijlHi si(p .p .m.) J~~/H~9 6.87 8,9 9.2 7.20 8.9 0.33 8 6.69 6,8 3.0 6.76 2.9 0.07 6 6.70 6,9 0.4 6.60 0.4 -0.10 7-OCH, 3.74 3.80 0.06 4 1.81 (A) A,B -13. 1.77 ? -0.04 2.20 (B) A,C 10.5 5 2.97 (c) A,D 0.8 3.01 (D) B,C 9.0 3 2.15 (E) B,D 9.0 2.43 (F) C,D -18.2 1 2.39 (G) E,F -18.0 2.71 (H) G,H -18.0 3a-CH, 0.81 ~~-CH,,F 0.5 9b-CH, 1.18 I)b-CH3,H 0.5 A and B nor between c and I) in view of the small internal shift in the former case and the essentially zero shift difference in the latter.This shift difference, A~(AB), 1.77 6.7 -0.43 2.79 6.7 -0.18 2.90 6.7 -0.11 2.21 6.7 0.06 2.29 -17.8 -0.14 2.49 -18.0 0.10 2.68 -18.0 -0.03 1.14 0.33 1.35 0.17 four vicinal coupling constants equal (Figure 1). No other basis of analysis would explain the observed spec- trum. Such an analysis is incompatible with a rigid ring structure and a rapid ring inversion process appeared to be operating.A Dreiding model of the cis-ketone (12) revealed that a rapid ring inversion process was indeed likely. In contrast, the model of the trans-ketone (10) suggests considerable strain in the cyclohexene ring which cannot be relieved by any conformational inter- conversion. This observation was of decisive import- ance in distinguishing between the ketones (10)and (12). Experimental evidence for the presence of a rapid ring inversion process in the compound presumed to be the cis-isomer (12) was sought from the low temperature (ca. -80 "C)spectrum of the compound in CS, as solvent. At this temperature appreciable broadening of all the methylene proton lines was observed (Av~ca. 6 Hz) but no coalescence phenomena were seen.Cooling to ca. -100 "C would presumably have revealed a further slowing of the inversion process but experimental difficulties were encountered on cooling the n.m.r. probe below -80 "C. Evidence supporting the view that the inversion barrier is too low to be easily detected by lH t1.ni.r. methods was found in some studies on 9,lO-dihydropl~enanthrene,~where inversion was rapid even at -90 "C. Rapid ring inversion would explain why the coupling = Jr3c3and JA,(l Jltn, since the observed couplings would be averaged over the dihedral angles, 1'1 B-III~IIlI~IIII~IlII~l111~111 t----150Hz ____( FIGURE 220 MHz lH Spectrum of the 4-and 5-H2 of the1 cis-ketone (12); (a) experimental, (b) simulated 1~ 0 = 60"-0"-60" and 0 = 18O0-12Oo-6O0 respect-ively; but this does not explain why JAkO = JA1) and = J1,,.However, the experimental equality of the chemical shifts 6, and SI, (even at 220 MHz) remains surprising.The methylene protons at C-1 and C-3 3 RI. van Gorkom and G. E. Hall, Quart. Rev., 1968, 22, 14. 4 P. bsol;V. Kabideau, R. G. Harvey, and J. B. Stothers, Chem. Comni., 1969, 1005. was not distinguished from zero even in the 220 MHz spectrum (Figure 1, Table 1). There is no reason, S. Castellano and A. A. Bothner-By, J. Chem. Ph?ls., 1964, 41, 3863. 1978 exhibited AB quartets and the signals were assigned on the assumption that the C-1 methylene absorbed at lower applied field as a result of deshielding by the aromatic ring. Assignment of the individual proton signals in the C-1 and C-3 methylene groups was made by analogy with the other isomer, where long-range couplings between the methyl protons and one of each pair of C-1 and C-3 methylene protons enabled an unambiguous assignment to be made (see later).The 3a-CD3 deriv- ative (13)was subsequently prepared to confirm further the above assignment and (particularly) to enable a distinction to be made between the two methyl signals at 61.14 and 1.35. The deuteriated derivative exhibited a single signal at 6 1.35and thus it was possible to conclude that the 3a-Me group absorbs at higher field than the Sb-Me group. In an attempt to confirm the cis-configuration of the CH, groups a nuclear Overhauser effect (n.0.e.) experi- ment was performed.Despite careful control of the experimental decoupling conditions no conclusive results were obtained on irradiating at the frequency of either CH, group and observing the other CH, signal. It thus appears that the protons are affected by a variety of relaxation mechanisms (i.e. the dipolar mechanism cannot be the dominant one) which are tending to minimise the n.0.e. factor between the two CH, groups. The I3C chemical shift data of the cis-compound (12) are given in Table 2. The assignments of the low-field bands were fairly straightforward. The small chemical shift difference between the quaternary C-5a and C-9a makes definitive assignment difficult but, on the basis of the high-field substituent shifts of a Me0 group on ortho-and para-carbon atoms, the higher field signal is TABLE2 I3C Spectral data for ketones (10) and (12) trans-Ketone (10) cis-Ketone (12) Ketone(l2)-ketone(l0)rph--__ -7si(p.p.m.) i 8dP.P .m.) i Aamp;(p.p.ni.) 23.5 3a-CH, 23.1 3a-CH3 -0.4 25.3 9b-CH, 25.3 9b-CH, 0.0 ;;:;)26.5 44,5 30.7 5z::;} 3a,9b 3a,9b 1,3::::I 1,3 53.7 55.2 7-OCH3 55.1 7-OCH3 -0.1 111.1 8 113.2 8 2.1 114.6 6 113.2 6 -1.4 125.4 9 129.3 9 3.9 136.6 134.9 136.71 5arsquo;9a 135.6) 5a,9a 157.9 7 157.6 7 -0.3 218.0 2 217.5 2 -0.5 ascribed to C-9a.An unusual feature of the aromatic carbon absorption is the exact equality of the C-6 and C-8 shifts. No explanation is apparent and the effect is not present in the other isomer (10).In the saturated J. H. Noggle and R. E. Schirmer, lsquo; The Nuclear Overhauser Effect.rsquo; Academic Press, New Vork, 1971. carbon region distinction between the pairs C-1 and -3 and C-3 and -9b is not conclusive, but the proximity of C-l and C-9b to the aromatic ring suggests that these nuclei provide the low field members of the two pairs of signals. A similar argument leads to the assignment of C-4 and C-5 signals. Assignment of the methyl signals at S 23.1 and 25.3 was made by analogy with the trans- isomer, where the higher field signal (623.5) was the one modified on deuteriation of the 3a-CH3 group. The 100 MHz lH spectrum of the trans-ketone (10) proved rather intractable, and detailed analysis was based on the 220 MHz spectrum. The aromatic region of the 220 MHz spectrum is depicted in Figure 2(a).L +lOO Hzl FIGURE 220 MHz IH Spectruni of the aromatic protons 2 of the tvans-ketone (10); (a) experimental, (b)simulated Unlike the cis-ketone the aromatic region of this isomer constitutes a strongly coupled ABC system. Iterative analysis using a modified LAOCNR program yielded the spectrum Figure 2(b) corresponding to the para- meters listed in Table 1. The chemical shifts v~~ and vc, assigned to H-8 and H-6, respectively, are very similar, and are the main cause of the second-order nature of the spectrum. These shift values reflect the usual upfield shifts of protons ortho to a methoxy-substituent. The methylene region of the lH spectrum Figure 3(a) posed considerable analytical problems.It basically comprises two overlapping AB quartet patterns and a four-spin ABCD system, one portion of which overlaps with one of the AB quartets. Assign-ments of the AB systems to the pairs of protons on C-1 and C-3 and the ABCD system to the protons on C-4 and C-5 are straightforward, but individual proton assignments are less obvious. However, the presence of long-range couplings involving the CH, groups and one of the C-1 and C-3 proton pairs enabled a complete assignment to be made with some confidence. In order to utilize this coupling evidence it was necessary to assign unambiguously the two methyl signals (8 0.81 and 1.18). This required the preparation and n.m.r. study of a deuteriomethyl analogue of the ketone (10).G. C. Levy and G. L. Nelson, lsquo; Carbon-13 Nuclear Magnetic Resonance for Organic Chemists,rsquo; Wiley-Intersciencc, New York, 1972, p. 81. H. Spiesecke and W. G. Schneider, ,I. Chern Phjis , 1961, 35, From the lH spectrum of the 3a-CD2H analogue (11) the bands at S 0.81 and 1.18 in the spectrum of (10) were assigned unambiguously to 3a-CH3 and 9b-CH,, respect- ively. Both these signals exhibited long-range couplings (4Jca. 0.5 Hz) to the nuclei which provide the lower field halves of the AB quartets. This was confirmed both by B+ (bl I 350 0 Ht I I 350 f3-b 0 Ht FIGURE 220 MHz lH Spectra of the methylene protons of the3 trans-ketone (10); (a) cxperimental without Eu(fod),, (b) simu- lated, (c) experimental with 10 mg Eu(fod), showing the mul- tiplet nature of the R methylene absorption decoupling experiments and by study of the 3a-CD2H derivative (11).Previous data indicate that four-bond trans-H,CH, coupling constants averaged over the equilibrium conformations of the methyl group are appreciably larger (4Jca. 0.6 Hz) than gauche-H,CH, coupling constants (*J ca. -0.2 Hz). The low-field members of the AB quartets are accordingly assigned to the protons (F and H) trans to the CH, groups. Both AB systems experienced large and equal lanthanoid-induced shifts, the system centred at lower applied being assigned to the protons on C-1 by virtue of the proximity of the aromatic ring. The ABCD system arising from the protons on C-4 and C-5 was analysed in the following way.The low field (CD) region was M. Barfield and B Chakrabarti, Chern. Rev., 1969, 69, 757. M. Karplus, J. Chew. Phiis , 1959, 30, 11; I. r4mnP?/.Chew. SOC.,1963, 85, 2870. 1111''7''''''''1'''''''''''''''''''I between equal energy conformations. J.C.S. Perkin I assigned with certainty to the hydrogens on C-5 as they will be deshielded by the aromatic ring and also since they exhibit a smaller lanthanoid-induced shift as compared with the AB pair. The addition of 10 mg of Eu(fod),, revealed the multiplet pattern of nucleus B as that of a perturbed doublet of triplets Figure 3(c). This enabled an accurate iterative analysis of this four-spin system to be performed from a trial set of parameters based mainly on a first-order treatment of the AB regions.After three iterations an acceptable fit (r.m.s. error 0.323 Hz) of computed and experimental spectra was obtained Figure 3(a) and (b). The strongly coupled nature of the CD region AS(CD) 0.041 caused a number of the final iterated parameters (Table 1) to deviate appreciably from the first-order values. The values of the four vicinal coupling constants are of interest in that they provide insight into the conform- ation of the cyclohexene ring. First, the set of values is compatible only with a fixed ring conformation in distinct contrast to the cis-ketone (12). This is also borne out by the absence of any line broadening in the spectrum at low temperatures which would have indi- cated the slowing of a rapid ring inversion process. Furthermore, a molecular model of this trans-isomer suggests the presence of considerable strain in this six- membered ring with no possibility of interconversion The values of the four vicinal coupling constants suggest, by reference to the Karplus eq~ation,~that the saturated CH,CH, fragment has adopted an approximately eclipsed con-formation with dihedral angles of 20, 20, 100, and 140" for the two pairs of protons.Considering the upfield pair (AB) HI{exhibits long-range coupling (*J ca. 0.5 Hz) with the 3a-CH3 group. On the basis of sizeable long- range trans-H,CH, couplings, HI{ was assigned as the proton trans to 3a-CH3, and HA as cis to 3a-CH3. Since Jau is small (0.8 Hz) HAL and HI, must be related by a dihedral angle of ca.100". It follows that the dihedral angle relating H-bsol; and H,: must be ca. 20". The hydrogen pair c and D may now be assigned such that c is cis to 3a-CH3 and D trans to this group. An additional point in this portion of the spectrum is the appreciably smaller geminal coupling constant for the AB methylene pair as compared to the other geminal couplings in both this compound and the cis-isomer (12). This is a further indication of distortion in this six-membered ring particularly in the vicinity of C-4. A decrease in 2Jg,,, indicates a decrease in the C-CH,-C valence angle,1deg; assuming that other factors contributing to the measured JqenL values are either weak or relatively constant for the two isomers (10) and (12).A decrease in the 3a,4,5- angle is most likely to occur by slight extensions of the 3a,4- and 4,5-bonds, particularly the former. Such an effect would necessarily be accompanied by a slight increase in the 4,5,5a-angle, with a consequent increase in magnitude of JcD(gem) as compared with geminal couplings of methylene protons in unstrained rings lo K. C. Cookson, J. J. Fraenkel, J. Hudec, and T. A. Crabb, Tetvahcdron, Suppl. 1966, 7, 355. (e.g. the cis-isomer). The data, vix. JcD(cis) = -17.8 Hz and JcD(trans) = -18.2 Hz, support such an inter- pretation. Details of the 13C spectrum of the trans- ketone (10) are given in Table 2. The majority of the assignments are based on the arguments used for the cis-ketone.In the aromatic region this isomer, in contrast to the cis-isomer, exhibited the expected three carbon signals. Distinction between C-6 and C-8 was made on the basis of n.0.e. factors, the more intense band being assigned to C-8, which will be appreciably relaxed by the adjacent C-9 proton. As with the cis-ketone, distinction between the pairs C-1 and -3 and C-3a and -9b was not conclusive. The assignment of CH3 carbon signals was based on the observation that the 3a-CD2H analogue (11) displayed a 1: 2 : 3 : 2 : 1 quintet (Jcn 19.1 Hz) at 6 22.9 implying that the highest field signal (6 23.5) in the trans-isomer spectrum be assigned to C-3a. Tables 1 and 2 also show the changes in lH and l3C shifts on going from the trans- to the cis-ketone.The aromatic lH and 13C shifts show analogous trends. The very appreciable downfield shift of the 9-position l3C and lH nuclei on going from the trans- to the cis-isomer is very indicative of strain in the aromatic ring of the trans-isomer. It is noteworthy that the lH methyl shifts are apparently more sensitive than the 13C methyl shifts to the stereochemical relationship of the methyl 121 groups. The lower field shifts for the cis-isomer are surprising, as mutual shielding of the methyl groups in this isomer would have been expected to cause higher field shifts than in the trans-isomer. It is also surprising that no nuclear Overhauser enhancements were detected for the CH, lH signals in the cis-ketone where the configuration of the groups would seem highly f avourable to positive n.0.e.s. A further point of note is that the CH, 13Csignals do not appear to be particularly sensitive to the configuration of the ring fusion. In contrast to this, Gough, Guthrie, and Stothers I1 found that the 13C shielding of the C-19 protons in 5a-and 5p-steroids (i.e. single bridgehead methyl groups) varied by 11-12 p.p.m. For example, the shift difference for 5a-and 5p-androstanes is 12.0 p.p.m., with the latter steroid appearing at lower applied field. The authors concluded that the C-19 methyl carbon shift can therefore be a reliable indicator of the stereochemistry of the AB ring junction. Our present findings suggest that this observation cannot be extended to fused alicyclic structures containing two bridgehead methyl groups. We thank the S.R.C. for use of the 220 MHz n.m.r. facilities at the P.C.M.U., Harwell. 7/978 Received, 8th June, 19771 l1 J. L. Gough, J. P. Guthrie, and J. B. Stothers, J.C.S. Chem. Comm., 1972, 979.