J. Org. Chem. 1984,49,3652-3656
3652
other than 3a and 3b add to isodicyclopentadiene (1)with predominant or complete e n d o s t e r e o ~ e l e c t i v i t y . ~Its dehydro derivative 2 gives evidence of entirely similar
response, t h e high preference for below-plane c a p t u r e persisting with methyl p r o p i ~ l a t eb, e~n~~~y~n e , and ~ ~ t(p~ tolylsulfonyl)acetylene.g To what causative factor can the strikingly divergent behavior of 3a to 3b toward 1 be attributed? The point has been repeatedly made t h a t the ethano and m e t h a n o bridges in 1 appear too remote from the diene termini t o introduce meaningfully different steric perturbations.
However, both bridges are recognized to foster anti-Alder approach of the dienophile, at least in those cases where this feature can be recognized in the product(s). It is not unreasonable to presume that the high reactivity of the triazolinediones and the early timing of their cycloaddition states m a y cause these dienophiles to add preferentially in Alder fashion.24s25 A unique sensitivity to steric effects would now develop, which phenomenon is not particularly significant when less reactive reagents are involved. Although the directionality of alignment is n o t revealed in t h e urazole products because of facile pyramidalization at nitrogen, t h e above factors could be responsible for t h e observed crossover in a-facial stereoselectivity.
Experimental Section N-Methyltriazolinedione Addition to 2. To a solution of in dry ether (45 mL) N-methyltriazoliiedione (192 mg, 1.7 "01) cooled to -20 "C was added 2 (250 mg, 1.92 mmol) dissolved in the same solvent (5 mL). After 20 min, the reaction mixture was allowed to warm to room temperature where solvent and excess 2 were removed under vacuum (30 and 0.03 torr). There was obtained 347 mg (84%)of 7a as transparent prisms: mp 110 "C dec; 'H NMR (CDCl,, 300 MHz) S 6.50 (dd, J = 1.81 and 1.81 Hz, 2 H), 5.17 (dd, J = 1.8 and 1.4 Hz, 2 H), 3.59 (br s, 2 H), 2.77 (s,3 H), 2.32-2.04 (m, 4 H). This material was identical with that previously reported.s N-Phenyltriazolinedione Addition to 2. Treatment of N-phenyltriazolinedione (297.5 mg, 1.7 mmol) with 2 (250 mg, 1.92 mmol) in ether (50 mL) as described above afforded 440 mg (85%) of 7b as the only adduct. This substance, like 7a, decomposed relatively rapidly when allowed to stand in solution at room temperature: 'H NMR (CDC13,300 MHz) 6 7.5-7.28 (m, 5 H), 6.60 (dd, J = 1.8 and 1.8 Hz,2 H), 5.30 (dd, J = 1.8 and 1.4 Hz, 2 H), 3.66 (br s, 2 H), 2.42-2.15 (m, 4 H). N-Methyltriazolinedione Addition to 1. Reaction of 3a (50 mg, 0.44 mmol) with 1 (66 mg, 0.50 mmol) in dry ether (18 mL) at -20 OC for 30 min afforded 88.4 mg (82%) of 9a as transparent prisms: mp 131 "C dec (lit? mp 129 "C dec) (from ethyl acetate); 'H NMR (CDC13,300 MHz) 6 5.06 (t,J = 1.5 Hz, 2 H), 3.01 (m, 2 H), 2.91 (s, 3 H), 2.36 (dt, J = 8.5 and 1.5 Hz, 1 H), 1.81 (m, 3 H), 1.33 (d, J = 7 Hz,1 H),1.17-1.07 (m, 3 H). N-Phenyltriazolinedione Addition to 1. Reaction of 3b (50 mg, 0.29 mmol) with 1 (42 mg, 0.32 mmol) in ether (18 mL) at -20 "C furnished 71.2 mg (80%) of 9b as colorless crystals: mp 183 "C dec (lit.12mp 180 "C dec); 'H NMR (CDCl,, 300 MHz) 6 7.46-7.33 (m, 5 H), 5.18 (t, J = 1.43 Hz, 2 H), 3.10 (m, 2 H), 2.46 (dt, J = 8.0 and 1.4 Hz,1 H), 1.90-1.83 (m, 3 H), 1.36 (br s, 2 H), 1.18 (d, J = 6.9 Hz, 2 H). X-ray Analysis of 9a. A plate-like prismatic colorless crystal of approximate dimensions 0.125 X 0.30 X 0.375 mm was mounted on the tip of a thin glass fiber. Both X-ray examination of the crystal and data collection were a t room temperature by using an Enraf-Nonius CAD4 diffractometer with graphite-monochromated Mo Ka radiation. At room temperature the cell parameters and standard deviations were determined by leastsquares fitting from 24 reflections, well distributed in reciprocal (24)
(a)Sustmann,R.; Trill, H. Angew. Chem., Znt.Ed. Engl. 1972,II,
838. (b) Sustmann, R.; Schubert, R. Ibid. 1972,II,840. (e) Sustmann, R. Pure Appl. Chem. 1974,40,569. (25) (a) Huisgen, R.; Schug, R. J . Am. Chem. SOC.1976,98,7819.(b) Kiselev, V. D.; Miller, J. G.Ibid. 1975,97,4036.
space and lying in the 28 range between 25" and 30". Intensity data were collected by the w-28 mode with 28 range between 4" and 5 5 O . A total of 3036 reflections were measured with 1955 unique data having I > 3.0o(Z). Details of data collection are given in Table I. The data were corrected for Lorentz and polarization effects but not absorption. Solution and Refinement of the Structure. The analytical form of the scattering factors for neutral atoms were used throughout the analysis, and both w a n d i 4 t e r m s were included for all atoms. All the crystallographic computations were carried out on a PDP 11/44 computer using the SDP (Structure Determination Package). From systematic absences, the space group E 1 / n was determined unambiguously. Statistic distributions of reflection intensities also pointed to the likelihood that the space group was centrosymmetric. There are four molecules per unit cell and, therefore, one molecule per crystallographic asymmetric unit. The structure was solved via a combination of MULTAN, difference Fourier, and least-squares refinements on these heavy atoms. All of the hydrogens appeared on the difference electron density map. The function minimized during the least-squares refinement process was Cw(lFol - IFc1),2where the assigned weights are given as w = [ o ( n 2 (pr)2]-1/2; p = 0.02 was chosen to make CwAF uniformly distributed in IFol. The final full-matrix, least-squares refinement cycle with anisotropic thermal parameters for all non-hydrogen atoms (isotropicfor the hydrogens) gave R f = 0.037, / ~R& ~ F= o ~ and Rd = 0.045; where R f = xllFol - ~ F c ~ ~and ~ W ' / ~ ~-~IFcll/xw1/21FoI F,I for the 1955 reflections having I > 3.Ou(I) with 223 variables. The final difference Fourier map showed no significant features, with a maximum peak height of 0.177 e/A3.
+
Acknowledgment. W e t h a n k t h e National Cancer Institute for their financial support of this research (Grant CA-12115). Registry No. 1, 6675-72-5; 2, 6675-71-4; 3a, 13274-43-6;3b, 4233-33-4; 7a, 73321-36-5; 7b, 91391-11-6; 9a, 91463-92-2; 9b, 89689-35-0. Supplementary Material Available: Tables listing final positional and thermal parameters, bond distances, bond angles, dihedral angles between planes, and lFol and lFcl (23 pages). Ordering information is given on any current masthead page.
(1 S )-la,4a,5/3,5a/3,6/3,8/3,9/3,9a/3-Octahydro-5,8-dih y d r o x y - 1,4-epoxy-6,9-met h a n o - 3 - b e n z o x e p i n - 7 ( 2 H ) - o n e ( a Levoglucosenone D e r i v a t i v e ) . The Product of a 1,4-Hydride Shift Maria G. Essig,* Thomas T. Stevenson, and Fred Shafizadeh'
Wood Chemistry Laboratory, University of Montana, Missoula, Montana 59812 Ronald E. Stenkamp and Lyle H. Jensen
Department of Biological Structure, University of Washington, Seattle, Washington 98195 Received March 23, 1984
The potential for obtaining useful chemicals from cellulosic waste materials has been investigated through t h e pyrolysis of acid-treated newsprint.l Levoglucosenone (1) is a major product of this pyrolysis1 and its synthetic utility h a s been widely studiede2 T h e [4 21 cycloaddition reaction between levoglucosenone a n d 1,3-cyclopentadiene yields 2 as t h e major product3 (Scheme I). T h e reaction of 2 with a n equimolar a m o u n t of osmium tetraoxide produces an osmate ester which can be reductively cleaved
+
Deceased October 1, 1983.
0022-326318411949-3652$01.50/0 0 1984 American Chemical Society
J. Org. Chem., Vol. 49, No. 19, 1984
Notes
3653
019)
Figure 1.
O R T E P ~stereoview ~
of 4. Scheme I
OH 1. -0,
1
z.Na,SO,
3 and 4
Table I. Hydrogen-Bonding Distances (Angstrom) and Angles (Degrees) for 4 atoma distanceb angleb
D O(2) O(9)
H H(2)
H(9)
A O(2) O(9)
D-H 0.95 (4) 0.90 (4)
H-A 1.75 (4) 1.75 (4)
D--A 2.679 (4) 2.650 (4)
D-H-A 166 (3) 175 (3)
acceptor symmetry Y , 1 - x , '14 + 2 Y , -x, I14 + 2
D = donor, A = acceptor. Standard deviation is shown in parentheses.
by two methods to yield different products. Cleavage of the ester with hydrogen sulfide at ambient temperature gives the cis diol 3 in 74% yield.2a However, cleavage of the eater in a refluxing (ethanol-water) solution of sodium sulfite gives 3 in only 20% yield with a second product isolated in 42% yield. This major product has not been found by single-crystal X-ray diffraction and 'H NMR spectroscopy to be the rearranged compound 4 [(lS)la,4a,58,5a8,68,88,98,9a8-octahydro-5,8-dihydroxy1,4epoxy-6,9-methano-3-benzoxepin-7(2H)-one] .4
Results and Discussion A stereochemical drawing of 4 (Figure 1)shows that it is composed of a hexopyranose residue [C(l)-C(6)] fused to a norbornyl moiety [C(7)-C(ll)] with hydroxyl groups at C(2) and C(9) (C-0 bond lengths of 1.436 and 1.429 A, (1)Shafiideh, F.; Furneaux, R. H.; Stevenson, T. T. Carbohydr. Res. 1979, 71, 169-191. (2) (a) Shafiideh, F.; Wig,M. G.; Ward, D. D. Carbohydr. Res. 1983, 114,71-82, and references contained within. (b) Stevenson, T. T.; Essig, M. G.; Shafizadeh, F.; Jensen, L. H.; Stenkamp, R. E. Carbohydr. Res. 1983,118, 261-268. (c) Mori, M.; Chuman, T.; Kato, K.; Mori, K. Tetrahedron Lett. 1982,23,4593-4596. (3) Ward, D. D.; Shafizadeh, F. Carbohydr. Res. 1981, 95, 155-176. (4) The numbering system used in this note is as shown by the hy-
drogen-atom labels for 4. This numbering is consistent with earlier publicatins (ref 2 and 3) and is different from the IUPAC numbering system used to name 4. Thus,C(1) of the IUPAC name for 4 is referred to as C(5) as in the text.
Table 11. Puckering Parameters for 4 atom number
ring pyranose cyclohexanone anhydro cyclopentane cyclopentanone
8 (A) 0.629 0.942 0.440 0.583 0.602
0 (deg) $J (deg) 135.3 195.3 88.8 115.5 58.1 6.2 1.1
1
3 C(2) C(4) C(7) C(1) O(5) C(5) C(11) C(l0) C(4) C(l1) C(7) C(8)
O(5)
2 C(1) C(3)
respectively) and a ketone group at C(8) (C-0 bond length of 1.206 A and bond angles of 106.2O, 125.7O, and 127.9O). The two cleavage products, 3 and 4, thus differ only by transposition of the hydroxyl and ketone groups at C(2) and C(8). In most l,Ganhydropyranoses, the outer two C-0 bonds of the C(5)43(5)-C(1)4(1)-C(6)bond sequence are longer than the mean C-0 bond length, while the inner two C-0 bonds are shorter than the mean.5 This is a consequence of the "anomeric effect"! In 4, the anomeric effect is also evident, with C-0 bond lengths of 1.439, 1.420, 1.422, and 1.446 A, corresponding to the above C-0 bond sequence. Both hydroxyl groups of 4 are involved in strong hydrogen bonds of the "donor-acceptor'" type where each (5) Noordik, J. H.; Jeffrey, G. A. Acta Crystallogr. 1977,87,403-408. (6) Jeffrey, G. A. ACS Symp. Ser. 1979,87, 50-62. (7) Jeffrey, G. A.; Takagi, S. Acc. Chem. Res. 1978, 11, 264-270.
3654 J. Org. Chem., Val. 49, No. 19, 1984
Notes
..........
Figure 2. Hydrogen-bonding network in crystals of 4
,
viewed ~ along the x axis.
hydroxyl oxygen acts as both a hydrogen-bonding donor (D) and an acceptor (A) (Table I, Figure 2). In crystals of 4, these hydrogen bonds form chains about the screw axes linking the molecules along the z axis. Since each molecule participates in two chains, the crystalline lattice is also linked along the x and y axes through hydrogen bonds. Calculation of ring-puckering parameters* (Table 11) from the X-ray data revealed that the pyranose ring has a conformation intermediate between a half-chair,'H0, and a sofa, Eo (five carbon atoms coplanar)? with significant distortion toward a chair 'C, conformation. As a result of the fused ring system in 4, this conformation differs somewhat from other 1,6-anhydropyranoses which usually exist between the 'C, and EO conformations: while the pyrannose ring of the norbmyl epoxide 5 [differing mainly by the hybridization state of C(2)l is in an Eo conformation." The cyclohexanone ring adopts a pronounced boat 'JOBconformation (with a large total puckering amplitude, Q = 0.94 A), held rigidly in place by the C(l1) methylene group. The anhydro ring adopts a twist C?7'oc5) conformation (4 = 54") with some distortion toward the envelope c(5)Eform (four atoms coplanar, 4 = 72'). This conformation is typical of 1,6-anhydropyranose~.'~The cyclopentane and cyclopentanone rings both assume an " E
conformation (where 4 = 0")with some distortion toward twist forms "Tl0and 'T,,respectively. The norlmmyl-ring conformations of 4 are very similar to those of 5.2b H(6no)
H(6-4)'P-O.
4
5
(8) Cremer, D.; Pople. J. A. J . Am. Chem. Soe. 1976,97, 1354-1358. (9) Jeffrey, G. A.; Yatea, J. H. Carbohydr. Res. 1979, 74, 319-322. (10) Jeffrey, G. & Park,Y. J. Corbohydr. Res. 1979, 74, 1-5.
Evidence of flattening of the pyranoid ring in solution was also observed hy 'H N M R speetmscopy (seeTable ID),
Notes Table 111. Proton Dihedral Angles (Degrees) and 'H NMR Coupling Constants (Hertz) protons dihedral coupling protons dihedral coupling involved angle" constant involved angle" constant 0 4.0 5,6endo 93 192 54 293 139 4.0 7,llsyn 58 smb 8 12.2 7,llanti 68 sm 394 397 50 4.5 9,lO 79 0 495 85 0 l0,llsyn 65 sm 4,lO 50 4.3 10,llanti 62 sm 5,6exo 37 4.3 Calculated from the X-ray data. Observed as a sharpening of one signal upon spin-spin decoupling of the other proton signal involved.
where the coupling between H(3) and H(4) is large (12.2 Hz), indicating a small dihedral angle between these cis hydrogen atoms (8' in the crystalline state). In other 1,6-anhydropyranoses in a chair conformation, J3,4(axialequitorial) is usually within the range of 4.2-5.8 Hz." In addition, J4,5 of 4 is unusually small (-0 Hz), indicating a dihedral angle of approximately 90' (found to be 85' in the crystalline state), whereas 1.5-2.5 Hz (equitorialequitorial) would be expected for a chair conformation.'l The 'H NMR spectrum of 4 was interpreted with the aid of spin-spin decoupling experiments as well as by comparison with the spectra of similar compounds.2a The relative chemical shifts of H(3) and H(7) of compounds 4 and 3 are consistent with the location of the carbonyl group. The resonance for H(3) is upfield by -0.5 ppm in 4 (2.28 ppm) relative to that of 3 (2.73 ppm), indicating the absence of a C(2) carbonyl group in 4. Also, the resonance for H-9 is downfield by -0.5 ppm in 4 (4.03 ppm) relative to that of 3 (3.54 ppm), indicating the presence of a C(8) carbonyl group in 4. The S configuration of the C(2) hydroxyl group of 4 is consistent with the coupling constants Jl,z= - 4 Hz and J 2 , 3 = -4 Hz (corresponding to dihedral angles of 54' and 139' in the solid state, Table 111). The R isomer would have coupling constants Jl,z small and J 2 , 3 large. This is seen experimentally for derivatives of 2 in which the carbonyl group at C(2) was reduced, yielding the R configuration. For these compounds, the resonances for H(1) appear as singlets in the 'H NMR spectra.12 Long-range coupling occurs between H(9) and H(l1anti) (Jl,llanti = 3.1 Hz) which can be attributed to the "W" ~onfiguration'~ between the two protons. The difference in the geminal coupling constants ( J 6 e ~ ~ , b= ~ d6.9 ~ Hz, Jllsyn,llanti = 10.7 Hz) is consistent with the geminal Karplus orr relation.'^ Formation of 4 also occurs when cis diol 3 is refluxed with sodium sulfite in ethanol-water under the same conditions used for osmate-ester cleavage (TLC and 'H NMR evidence). When 3 is refluxed under these conditions, production of 4 is incomplete after 45 min, but after 2 h, an equilibrium mixture of -75% of 4 is established and remains unchanged further refluxing. The same ratio is established in less than 3 h when 4 is used as the starting material. This ratio was determined by integration of the signals resulting from H(1) of 3 and 4 in the 'H NMR spectra. Transformation of 3 to 4 is a base-catalyzed 1,Chydride shift (similar to an intramolecular Meerwein-PonndorfVerley/Oppenauer reaction) and has been found to occur (11)Czerny, M.; Stanek, J., Jr.; Adu. Carbohydr. Chem. Biochem. 1977,34,23-177. (12) Essig, M. G.; Shafizadeh, F., unpublished results. (13) Sable, H. Z.; Ritchey, W. M.; Nordlander, J. E. J. Org. Chem. 1966, 31, 3771-3775, and references contained within. (14) Gutowsky, H. S.; Karplus, M.; Grant, D. M. J. Chem. Phys. 1959, 31, 1278-1289.
J. Org. Chem., Vol. 49, No. 19, 1984 3655 Scheme I1
across cyclohexanone rings, where a boat conformation can be assumed16 and certain fused ring systems16 including a norbornyl system similar to that of 4.'6b 1,4-Hydride shifts can also occur in straight-chain compounds." These reactions are generally catalyzed by strong base (concentrated hydroxide, alkoxide, or dimsyl ions). The isomerization of 3 to 4, however, is accomplished under mildly basic conditions using a sulfite catalyst. The same equilibrium can also be established by using dilute sodium hydroxide as a catalyst. A solution of 3 in ethanol-water was adjusted to pH 10.5, the same pH as was used in the previous experiments,ls and refluxed for 18 h. Again, the 'H NMR spectrum of the reaction mixture showed a 3:4 ratio of 1:3. This indicates that the reaction is occurring through a base-catalyzed mechanism (Scheme 11),as has been hypothesized for other 1,6hydride shifts. A specific interaction between 3 (or 4) and sodium hydrogen sulfite (e.g., a hydrogen sulfite addition intermediate) is not required.
Experimental Section General Methods. Melting points were determined with a Fisher-Johns melting point apparatus and are uncorrected. Infrared (IR) spectra were obtained with a Nicolet MX-1 fourier transform infrared spectrometer. 'H NMR spectra (360 mHz) were recorded at the Regional NMR Center, Colorado State University; 'H NMR (90MHz) were recorded on a JEOL FX-9OQ instrument. Thin-layer chromatography (TLC) was performed on Baker-flex silica gel IB2-F plates using ethyl acetate as the eluting solvent. Visualization was achieved by spraying with 1:227 anisaldehydmulfuric acid-ethanol solution and charring. Column chromatography was performed on silica gel 60 (70-230 mesh). la,4a,5t9,5at9,6B,8B,9B,9aB-Octahydro-5,8-dihydroxy1,4-epoxy-6,9-methano-3-benzoxepin-7(2H)-one (4). A solution of osmium tetraoxide (0.983 g, 3.9 mmol) in p-dioxane (15 mL) was added to 3 (0.739 g, 3.8 mmol) dissolved in p-dioxane (10 mL) a t ambient temperature and stirred. After 1 h, TLC showed the absence of starting material and the p-dioxane was removed by rotary evaporation. The osmate ester was then dissolved in an ethanol-water solution (l:l, 30 mL) saturated with sodium sulfite, and heated under reflux for 2 h. After this time, TLC showed the presence of the major product 4 (Rf0.25), the minor product 3 (R, 0.42), and a very minor component (R,0.50), which was not isolated. The solution was filtered to remove the black osmium residue and the solvent was removed by rotary evaporation to yield an oil (1.099g, 128%). The oil was applied to a silica column (30 x 240 mm) and eluted with ethyl acetate, giving 3 (0.172 g, 20%) and 4 (0.365 g, 42%). Recrystallization of 4 from hot methanol gave white needles: mp 182-183 "C; IR (KBr) 3234, 2884, 1751, 1155, 1092, 1053 cm-'; 'H NMR (acetone-d,) 6 5.14 (d, 1 H, H-1), 4.68 (d, 1 H, H-5), 4.03 (d, l H , H-9, J g , 1lanti 3.1 Hz), 3.84 (d, lH, H-Gendo, Jeendo,aero 6.9 Hz), 3.72 (dd, 1H, H-Gexo), (15) (a) Warnhoff, E. W. Can. J. Chem. 1977,55,1635-1643. (b)Watt, I. Tetrahedron Lett. 1978, 4175-4178. (c) Warnhoff, E. W.; ReynoldsWarnhoff, P.; Wong, M. Y. H. J.Am. Chem. SOC. 1980,102,5956-5957. (16) (a) Lansbury, P. T.; Saeva, F. D. J. Am. Chem. SOC.1967, 89, 1890-1895. (b) Shepherd, J. M.; Singh, D.; Wilder, P., Jr. Tetrahedron Lett. 1974, 2743-2746. ( c ) Craze, G.-A.; Watt, I. J. Chem. SOC.,Chem. Commun. 1980, 147-148. (d) Craze, G.-A,; Watt, I. Tetrahedron Lett.
1982,975-978. (17) Warnhoff, E. W.; Wong, M. Y. H.; Raman, P. S. Can. J. Chem. 1981,59,688-696. (18) While pH meaurements in ethanol-water cannot be assumed to be accurate, the relative hydroxide activity should be approximately equal in both cases.
3656
J. Org. Chem., Vol. 49, No. 19, 1984 atom
c-1 c-2
X
Notes
Table IV. Fractional Atomic Coordinates"and Thermal ParametersbvC for 4 Y 2 U J U atom X Y -7.
4376 (2) 3746 (2) 3577 (2) 3737 (2) 4221 (2) 5366 (2) 2474 (2) 1738 (2) 1882 (2) 2669 (2) 2362 (3) 5381 (1) 4217 (1) 3855 (1) 1106 (1)
2859 (2) 3242 (2) 2397 (2) 1256 (2) 1175 (2) 1411 (3) 2362 (2) 2311 (2) 1242 (2) 769 (2) 1241 (2) 2527 (1) 4114 (1) 1983 (1) 2944 (2)
1740
55 47 47 48 55 73 58 56 53 51 62 64 53 57 84
solution
reflux time ( h )
%3
%4
1 0.75 51 49 2 2 29 71 3 3 32 68 4 4.5 20 80 5 18 26 74 This procedure was repeated with 4 (20 mg) dissolved in the sodium sulfite solution (25 mL) and heated under reflux for 3 h. 'H NMR integration gave equilibrium values of 21% for 3 and 79% for 4. This procedure was repeated with 3 (15 mg) dissolved in a 1:l solution of absolute ethanol-water (10 mL) with the pH adjusted to 10.5 with sodium hydroxide and heated under reflux for 18 h. After reflux was ended, the solution was treated as before and the equilibrium percentages, determined by integration of the same 'H NMR signals, were 23% of 3 and 77% of 4. X-ray Structure Determination. Weissenberg photographs of crystals of 4 showed them to be tetragonal, space group P4, or P4, (001 reflections present for 1 = 4 n); z = 4. X-ray intensity data were collected from a single crystal (0.22 X 0.62 X 0.18 mm) with a computer-controlled, four-circle diffractometer (Table IV). Cell constants obtained from 20 centered reflections were a = b = 12.9505 (7) A, c = 6.3054 (3) A. Data were collected in the w-28 mode, scan width = LOo, scan rate = 2O/min, X = 1.5418 A (Cu K a radiation), maximum 28 = l l O o with 10-s backgrounds measured on both sides of each reflection. Six standard reflections measured periodically throughout the data collection indicated no deterioration of the crystal in the X-ray beam. Six reflections (111,101,200,210,211, and 301) overflowed the counter and had significant coincidence losses even at reduced beam intensity and were deleted from the data set.
(19) North, A. C. T.; Phillips, D. C.; Mathews, F. S. Acta Crystallogr., Sect./A 1968, A24, 351-359.
903 (1) 443 (2) 306 i2j 403 (2) 422 (2) 402 (2) 575 (3) 566 (3) 231 (2) 214 (2) 271 (2) 165 (2) 286 (2) 407 (2)
2
u.. I u ~~
99 (7) 77 -111 (7) 283 (5) 55 (8) -1769 (7) c-3 45 i4j c-4 -982 (7) -285 (5) c-5 1195 (7) -197 (6) 1189 (11) C-6 186 (6) c-7 -2694 (7) 262 (8) c-8 -847 (8) -15 (8) c-9 149 (8) -358 (6) C-10 -1328 (8) 158 (6) C-11 -3464 (7) -126 (5) -1 (2) 0-1 1138 (7) 107 (2) -394 (6) 0-2 -1132 (7) 107 (2) -456 (6) 0-5 2569 (6) 467 (3) -19 (8) 0-8 -318 (7) 19 (3) 88 (2) 99 (8) 4 ~ 1 0 for 4 C and 0; x103 for H atoms (standard deviation in parentheses). *Values in A* X lo3. cEquivalent isotropic temperature factors Ue,, = l/Q(Ull+ U,, + U33)given for non-hydrogen atoms; isotropic temperature factors U for hydrogen atoms. 3.18 (dd, 1 H, H-2), 2.82 (m, 2 H, OH), 2.64 (br d, 1 H, H-7), 2.57 (m, 1H, H-lo), 2.39 (dd, 1H, H-4), 2.28 (ddd, 1H, H-3), 2.13 (dm, 1H, H-llsyn,Jll, ,Ilanti 10.7 Hz), 1.61 (dm, 1H, H-llanti). Anal. Calcd for CllH14& C, 58.40; H, 6.24. Found: C, 58.25; H, 6.28. Equilibrium Measurements. A 1:l solution of absolute ethanol-water was saturated with sodium sulfite a t room temperature, giving a pH of 10.5. Five samples of 3 (- 15 mg each) dissolved in the sodium sulfite solution (10 mL) were heated under reflux for varying amounts of time, the solvent was removed by rotary evaporation, acetone-& was added, and the flask was set in an ultrasonic bath for 5 min. Equilibrium values were determined by integration of the lH NMR signal for the doublet at 6 5.14 [H(1)]of 4 and the multiplet a t 6 4.90 [H(l) and H(5)] of 3.
0-9 H-1 H-2 H-3 H-4 H-5 H-6exo H-6endo H-7 H-9 H-10 H- 1lsyn H-llanti H-2' H-9'
733 (2) 337 (2) 349 i2j 251 (2) 89 (2) 50 (2) 115 (3) 117 (3) 293 (2) 128 (2)
The structure solution was carried out in space group P4, and resulted in a model with the absolute configuration opposite to that dictated by the synthetic work. The correct space group for the molecule is therefore P4,. The direct methods programs, M U L T A N ~and O ~ RANT AN^^ did not yield structure solutions when run with default input parameters. However, use of the weighted tangent formula option in RANTAN~Oproduced a structure solution containing all 16 non-hydrogen atoms in the asymmetric unit. The structure was refined against F with 1/u weights by using the least-squares program of the XRAY76 system.21 All the hydrogen atoms were located in subsequent difference electron density maps. In the final least-squares cycles, thermal parameters for carbon and oxygen atoms were refined anisotropically and hydrogen atoms were refined isotropically. The least-squares refinement converged to an R ( = CllFol- IFcII/CIFol) of 0.023 for 732 reflections with F, > 40(F0). Four reflections had F, < 4u(F0). The weighted R was 0.029 and the maximum shift/error in the last refinement cycle was 0.13.
Acknowledgment. We acknowledge the support of the National Science Foundation for equipment g r a n t PCM 76-20557 to L.H.J. and research g r a n t PFR 80-23854 to F.S. We also thank Dr. J. P. Wehrenberg for use of the Weissenberg camera, Dr. T. G. Cochran for reviewing this manuscript, and the Rocky Mountain Laboratory (NIAID) for use of the 90-MHz 'H NMR spectrometer. The highfield 'H NMR spectra were obtained from the Colorado State University Regional NMR Center, funded b y National Science Foundation Grant No. CHE-8208821. Registry No. 2, 79849-64-2; 3, 85427-14-1; 4, 91238-75-4. Supplementary Material Available: Figures 3 and 4 showing bond lengths and bond angles and Tables V and VI listing thermal parameters and torsion angles (7 pages). Ordering information is given on any current masthead page.
(20) Main, P.; Fiske, S. J.; Hull, S. E.; Lessinger, L.; Germain, G.; Declercq, J. P.; Woolfson, M. M. "MULTAN80, A System of Computer Programs for the Automatic Solution of Crystal Structures From X-Ray Diffraction Data"; University of New York, 1980. (21) Jia-Xing, Y. Acta Crystallogr.,Sect. A 1981, A37, 642-644. (22) Stewart, J. M., Ed. "The X-ray System-Version of 1976"; Technical Report TR-446 of the Computer Science Center, University of Maryland, College Park, Md.
