Crystal structures of the dihydro- and 2alpha;-hydroxydihydro-derivatives of phytuberin

摘要

1338 J.C.S. Perkin ICrystal Structures of the Dihydro- and Pa-Hydroxydihydro-derivatives ofPhytuberinBy David L. Hughes, Molecular Structures Department, Rothamsted Experimental Station, Harpenden, Hertford-The X-ray crystal structure analyses of two derivatives of the potato stress metabolite phytuberin, are reported.Dihydrophytuberi n (111) (8- (1 -acetoxy- 1 - methylethyl) - 3a,5a-dimethylperhydrofuro [3.2-c] i s 0 benzofuran) hasspace group P2,. a = 9.723(2), b = 10.722(5), c = 8.107(5) A, p = 101.29(3)", and its structure was derived bydirect methods. Full-matrix least-squares refinement, to R 0.060 for 1 650 reflections, showed a very compact,rigid molecule, with considerable conformational strain within the three-ring system.2a-Hydroxydihydrophytuberin (11) has two independent molecules in the triclinic P1 cell, a = 10.887(3),b = 10.057(8), c = 8.661 (5) A, a = 111.85(5), p = 89.63(4), y = 99.43(5)". This structure was determinedfrom the alignment of the rigid skeletal structure of (111) with small ring fragments in an E map.The final refinedstructure, R 0.1 34 for 2 845 measured reflections, shows a pseudo-two-fold screw axis closely relating the twomolecules. There are negligible differences in the conformations of the two molecules, and only in the ringsadjoining the hydroxy-groups do they differ from (111).It is suggested that, in (11) there is ahydrogen-bonding scheme spiralling parallel to thea axis; in (111). no such interactions are possible and all inter-molecular contacts are at van der Waals distances.shire AL5 2JQThe packing arrangements of the two derivatives are quite different.PITYTUBERIN is a liquid antifungal sesquiterpenoid whichaccumulates in potato tubers inoculated with the lateblight fungus Phyto9htkora infestans 1 or the bacteriumErwinia carotovora var.atrose$tica.2The structure of phytuberin (I) has been establishedfrom the crystal structure of its hydrogenated derivative,derivative, (11), 2~-hydroxydihydrophytuberin, the firstcrystalline derivative of phytuberin to be prepared andexamined ; the successful determination of its structurehad to wait, however, until that of the dihydro-deriva-tive (111) was complete.The crystallographic, chemical, and spectral,* databriefly reported re~ently.~ This analysis is now com- relate the structures of phytuberin (I) and its hydroxy-plete and is reported here in more detail, together with dihydro- (11) and dihydro- (111) derivatives.Thethe results of the crystal structure analysis of another absolute configurations shown here have been suggested1 J. Varns, ' Biochemical Response and its Control in the IrishPotato Tuber (Solanurn tuberosum L.): Phytophthora infestans D. L. Hughes and D. T. Coxon, J.C.S. Chem. Comm., 1974,Interaction,' Ph.D. Thesis, Purdue University, Lafayette, 822.Indiana, 1970. D. T. Coxon, R. F. Curtis, I. R. Price, and B. Howard,2 G. D. Lyon, Physiological Plant Pathology. 1972, 2, 411. Tetrahedron Letters, 1974, 23631976 1339from the postulated biogenetic relationships betweenphytuberin and other stress metabolites in pot at^.^EXPERIMENTALStructure of Dihydrophytuberin (111)Dihydrophytuberin was prepared by the catalytichydrogenation of phytuberin, and this sample was crystal-lised by slow evaporation of the solvent, n-pentane.Theproduct, m.p. 66 "C, comprised colourless masses from whichcrystals suitable for X-ray analysis were cut. Some ofthese fragments showed one good face but generally had nodistinct shape. The first selected crystal was mounted ona fibre but, after some days of exposure to air and X-rays,appeared to be covered by bubbles. A second crystal,0.12 x 0.57 x 0.19 mm3, was mounted in a fine glass tubefor the collection of intensity data; the longest dimension,the b axis, was arranged to be close t o the rotation axis.Monoclinic, a =U = 828.7 .k3, D , = 1.188 g 2 = 2, F(000) = 324.p(Cu-Kz) = 6.8 cm-l, h(Cu-Kz) = 1.5418 a (1 A = 10-lOm).Space group P2, (90. 4); since the compound is opticallyactive; [a], 41.1".Preliminary cell dimensions were determined fromoscillation, Weissenberg, and precession photographs, andlater refined from 8 measurements of 39 reflections on thediffrrtctonieter. Intensities of 1 657 independent reflectionsin the range 2.6 8 70' were measured on an Enraf-Nonius CAD 4 Kappa diffractometer by use of mono-chromated Cu-27% radiation.Measurement was by forwardand return scan in the 28--w scan mode a t varying scanningspeeds to obtain an approximately constant counting ratethrough the peak within a maximum scan time of 60 s.The 20 scan angle was set to 1.0 + 0.5 tan 8"; backgroundcounts were recorded from the first and last quarters ofeach scan.Seven reflections from a confined region in the reciprocallattice were later found t o have extraordinarily large andirregular background counts, suspected t o result frombeams scattered by the mounting table set on the gonio-mctcr arcs; these reflections were omitted from the refine-ment of the structure.Of the 1 650 remaining reflections,93 had a net intensity of zero. Lorentz and polarisationfactors were applied to the net intensity values, but noabsorption corrections were made.Stvzactuve A naZysis.-The structure of (111) was solved bydirect methods. E Values were calculated from the Icurve m e t h ~ d , ~ and their statistics were consistent withthe non-centrosymmetric space group P2, for the opticallyactive phytuberin derivatives.The 160 reflections with[El 2 1.50 were used in the automatic multisolutiontangent-formula program MULTAN,' which initially wasallowed to select a starting set of reflections. E Mapsbased on those sets of phases which showed the highestfigures of merit, ABSFOM, and lowest residual factors,RESID, all showed one outstanding peak, which was two tothree times as large as any other and one of a zig-zagfragment comprising eight of the strongest peaks. Varyingamounts of pseudosymmetry persisted around this fragment.Eventually, another set of starting reflections wasAfter J .and I. L. Karle, in ' Computing Methods in Crystal-lography,' ed. J. s. Rollett, Pergamon Press, Oxford, 2965,ch. 17.6 I . L. Karle, I. S. Dragonette, and S. A. Brenner, Acta Cryst.,1965, 19, 713.Cvysta2 Data.-C,,H,,O,, M = 296.41.9.723(2), b = 10.722(5), G = 8.107(5) A, p = 101.29(3)",selected and input to MULTAN; it was in an E map froma set of phases with only moderately high ABSFOM (1.18,vs. the highest value 1.25), but with the lowest RESID,that a fragment of the molecule, involving a five-memberedring, could be identified. Fourteen peaks were taken, ascarbon atoms, for a structure-factor calculation, and thewhole molecule showed clearly in a succeeding electron-density map.On the first refinement of the parameters of the twenty-one atoms (all designated carbon atoms), four had dis-tinctly lower temperature factors and were shown to be theoxygen atoms.The molecular skeleton found agreed withone suggested by Coxon,* except that the carbon andoxygen atoms indicated as C(l) and 0 ( 1 ) in Figure 1 wereinterchanged.With all the data (equally weighted) and scatteringfactors from ref. 8, refinement of the atomic parameterswas rapid. All hydrogen atoms were located in a FourierFIGURE 1 Projection of a molecule of dihydrophytuberin (111)showing the atom numbering schemedifference synthesis and were included, with scatteringfactors of ref. 9, in further calculations. In alternate cyclesof least-squares refinement, (i) the carbon and oxygen atomswere refined anisotropically, (ii) the hydrogen atom para-meters and the parameters of their bonded carbon atomswere refined.The y co-ordinate of O(2) was not refined:this defined the origin of the unit cell, as required by thespace group P2,. In the final cycles, a weighting schemewas introduced: w = l/a2 where o2 = 0.234 - 0.037 41F,I +0.002 041F,12 + 0.000 096jF,13. The final R was 0.060 andR' 0.066 for the 1 650 reflections. In a Fourier differencesynthesis, based on theo 1 557 observable reflections, thehighest peaks were 0.20 A-3.Final atomic parameters for (111) are in Tablc 1, and aprojection of the molecule (with the atom numberingscheme) is in Figure 1. In the numbering of the hydrogenatoms of the methylene groups the first digit, and of theG.Germain, P. Main, and M. M. Woolfson, ' MULTAN:A Computer Programme for the Automatic Solution of CrystalStructures,' University of York, 1971.International Tables for X-Ray Crystallography,' vol. 111,Kynoch Press, Birmingham, 1962, p. 202.B R. F. Stewart, E. R. Davidson, and W. T. Simpson. J .Chem. Phys., 1965, 42, 3175J.C.S. Perkin Imethyl groups the first two digits, indicate the carbonatoms to which the hydrogen atoms are bonded. Standarddeviations for bond lengths and angles (not involvinghydrogen atoms; Tables 2 and 3) were calculated from thevariance-covariance matrix, with inclusion of cell-para-meter errors. Where hydrogen atoms are involved, errorsTABLE 1Atomic parameters for dihydrophytuberin (111), withestimated standard deviations in parentheses.Theabsolute configuration, as predicted in ref. 3, is definedwith respect to a right-handed system of axes(a) Non-hydrogen atoms : co-ordinates (fractional x 104)X598(3)820(2)-4 327(2)-4 536(3)887(4)1825(4)1 003(4)- 59(4)- 324(3)-1 652(3)-3 013(3)-2 778(4)-1 567(4)- 169(3)- 4 287(3)-5 676(4)-4 198(5)-1 302(4)- 4 465(3)-4 499(6)389(4)Y-1 087(4)-3 719(-)-4 780(3)-3 602(4)-3 600(5)- 3 173(5)-2 277(4)-2 858(4)-3 634(4)-2 897(4)-1 903(4)-1 047(5)-1 752(4)-3 736(5)- 3 028(5)-4 412(5)-2 111(5) - 2 041(5)-5 822(6)-947(4)-4 597(5)(b) Hydrogen atoms : co-ordinatesisotropic thermal parameters *X189(4)226(5)263(5)152(7)82(4)55(6) - 173(4)- 144(4)- 326(4)- 363(5)- 252(4)- 181(4)- 147(4)- 581 (6)- 648(6)- 419(5)- 495(6)- 336(5)- 19(5)-577(7)52(6)129(7)- 195(5)- 179(5)- 98(5)- 361 (12) - 509(8) - 462( 7)Y-118(4)4(5)- 443(5)- 297(5)- 264(7)-376(6)- 405(4) - 407(5)- 247(4)- 142(5)- 230(4)-63(4)-31(4)-263(7) - 237(7)- 369(6)- 373(6)-491(6)- 481 (5)- 279(5) - 133(6)- 256(7)- 147(5)- 282(6)- 157(6)- 598( 13) - 647(9)- 575(6)z4 063(3)2 428(3)1791(3)4 034(4)2 421(5)5 232(5)4 207(4)2 408(4)1 840(4)1276(4)751(5)1252(4)614(4)401(6)3 979(5)- 16(5)-1 013(5)- 310(4)5 042(5)3 389(5)4 265(6)(fractional xz238(5)211(5)417(6)381(6)613(8)561 (6)279(6)225(5)84(6)- 24(5)- 107(5)- l l ( 5 )- 66( 8)- 196(6) - 124(7)- 97(6)- 103(5)- 103(7)- 13(8)181(4)136(9)W7)431(6)514(6)612(7)433(13)363(9)532(9)* From the expression : exp[ --B (sin'%) /A2].lo3), andB/A23.8(8)4.5(9)5.6( 11)5.4( 10)7.7(15)5.9(12)3.9(9)4 4 9 )3.3(7)3.9(8)5.1 (10)2.9(7)4.0(8)7.5(17)8.5(15)6.6(12)5.9(11)6.5( 13)4.4(10)4.6(9)6.4(13)8.3 (1 5)5.7(12)5.5( 11)6.9( 1 2)17.4(35)6.9 (1 5)10.2 (20)in bonds were calculated from the atomic variances in thebond direction, but spherical atomic variances wereassumed in the calculation of bond angle m values.The absolute configuration of (111) has been predicted afrom the suggested biogenetic relationship of the parentcompound phytuberin with other compounds isolated frommicrobially stressed potatoes.By use of the anomalousdispersion factors of ref. 10 for the oxygen atoms, structurefactors of the 1475 acentric reflections were calculated forthe predicted configuration, using the final atomic co-ordinates (x, y , z ) , and for the structure inverted throughthe origin, with co-ordinates (Z, 9, 2 ) . On comparison, nosignificant difference between the IFo/ values were found,and the difference between the R values was negligible,0.063 46 vs. 0.063 55. Distinction between the twoTABLE 2Bond lengths (A) in the phytuberin derivatives, withstandard deviations in parentheses0(1)-C(1)0 ( 1)-C(4)0 (2)-C(2)0(2)-C(5)0(3)--c(1 6)0(5)-C(2)C(l)-C(10) c (2) -c (3)C(3)-C(4)C(4)-C(5)C(5)-C(6)C(6)--C(7)C(7)-C(8)c (f9-c (9)C(9)-C(10)C(1 1)-C( 12)0(3)-C(11)0 (4)-C( 1 6)C (4)-C (1 5)C (5)-C (1 0)C(7)-C( 11)C( 1 0)-C (1 4)C( 1 1)-C( 13)C( 16)-C (1 7)C( 1)-H( 1 1)C( 1)-H( 12)C( 2)-H (2 1)C( 2)-H (22)C( 3)-H (3 1)C (3)-H ( 3 2)C( 6)-H( 6 1)C(6)-H( 62)C( 8)-H (8 1)C (8)-H (8 2)C (9)-H ( 9 1)C( 9)-H (92)C( 12)-H( 121)C( 1 2)-H ( 1 22)C( 12)-H (1 23)C( 13)-H (1 31)C( 13)-H( 132)C(13)-H(133)C( 1 4)-H (1 4 1)C( 14)-H (142)C( 14)-H( 143)C( 15)-H( 151)C( 1 5)-H( 152)C( 15)-H (1 63)C(17)-H( 17 1)C( 17)-H (1 72)C( 17)-H ( 1 73)C(2)-H(2)C(7)-H(7)(111)1.42 1 (5)1.441(5)1.439(5)1.4 76 (4)1.343(5)1.196 (5)1.5 22 (5)1.483(6)1.631 (6)1.660(5)1.51 8 (5)1.52 9 (5)1.536(5)1.532 (4)1.542(5)1.540(5)1.526(5)1 .539 ( 5)1.52 1 (6)1.629 (5)1.523 (6)1.496( 7)1.02(4)1.09(5)0.99 (6)1.06(5)0.99 (7)0.86(6)0.91(4)0.99( 5)0.98(4)0.96(5)1.03(4)1.04(4)1.08(4)0.91(7)1.10(7)1.05(6)1.06(6)0.90( 6)0.91(6)1.03 (5)1.05(6)0.99(7)0.98(5)0.97 (6)1.00 (6)0.88( 12)0.98 (9)0.89( 7)1.444(4)1.425 (1 2)1.439( 12)1.409( 12)1.448 (1 1)1 .509( 1 1)1.275( 12)1.206( 12)1.397 (1 3)1.5 15( 13)1.525(16)1.57 1 ( 1 5)1.548( 12)1.490( 13)1.539 (1 2)1.557 (1 2)1.553( 12)1.51 1 (12)1.533 ( 12)1.508( 13)1.520(13)1.519(13)1.627 (1 4)1.467( 13)1.529(16)0.93 ( 7)1.02(7)0.94 (7)1.06(10)0.87(8)0.98(6)0.97 (5)0.98(6)0.89(8)1.08(7)0.93(6)0.9 7 (6)1.07(9)0.95 (1 3)1.15 (8)1.02(8)1.08( 10)0.89( 11)0.99(8)1.01(8)1.03(6)1.23( 10)1.04(13)0.87 (1 0)0.98( 12)0.84 ( 10)0.91(10)1.388 (1 3)1.444( 12)1.37 8( 12)1.402(11)1.503( 11)1.314( 11)1.396 (1 3)1.5 17 ( 14)1.508(17)1.518(15)1.591 (1 1)1.492(13)1.540( 13)1.570( 13)1.540(12)1.501 (13)1.547( 12)1.5 13 (1 3)1.5 1 O( 13)1.532 (1 3)1.499 ( 13)1.485( 13)1.530(16)1.02(9)l.OO(8)1.05( 1 1)0.96 (9)1.05(6)1.05(5)0.9 5 (7)1.02(7)0.99(4)1.04(4)0.87(6)1.02(8)0.95(10)1.06( 11)0.96( 10)1.07(9)1.03( 11)0.95(9)0.92 ( 9)0.88( 12)1.2 1 O( 1 2)1.09(9)0.99(11)1.0 1 ( 10)1.12(11)1.01( 11)1.02(10)1.19( 11)enantiomers is thus not possible, and, by this method,are not able to confirm the absolute configurationphytu berin.lo D.T. Cromer and D. Liberman, J . Chew. Phys., 1970,1891.weof581976 1341TABLE 3Valence angles (") in the phytuberin derivativesC ( 1 )-0 (1)-C( 4)C (2)-0 (2)-C (5)C( 1 1 )-0 (3)-C( 1 6)O( 2)-C( 2)-O( 5)0 (5)-C( 2)-C( 3)O( l)-C(4)-C(3)O( 1)-C( 4)-C( 5)0 ( 1 )-C (4)-C( 15)O( 1)-C( 1)-C( 10)0(2)-C(2)-C(3)C( 2)-C( 3)-C(4)c (3)-c (4)-C( 5) c ( 3)-c ( 4)-C( 1 5)C( 5)-C( 4)-C( 15)0 (2)-C( 5)-C (4)0 (2) -C ( 5) -C ( 6)0 (2)-C (5)-C ( 1 0)C(4)-C( 5)-C( 6)C(6)-C(5)-C(lO)C( 5)-C(6)-C (7)C( 6)-C (7)-C (8)C (6)-C ( 7)-C ( 1 1 )C(4)-C (5)-C( 10)C(S)-C(ir)-C( 11) c (7)-c (8)-C (9) c (q-c (g)-c ( 10)C(1)-C( 10)-c(5)C(l)-C(lO)-C(9)c~l)-c(lo)-c(l4)C(5)-C(lO)-C(9)C(5)-C(lO)-C(14)C( 9)-C( 10)-C( 14)0 (3)-C ( 1 1 )-C( 7)0(3)-C(ll)-C(12)O(3)-C( 11)-C(13)C(7)-C(ll)-C(l2)C(7)-C( 11)-C( 13)C(12)-C(ll)-C(l3)O( 3)-C ( 16)-0 (4)0(3)-C(16)-C(17)0 (4)-C( 16)-C( 17)Range of 0Mean cs(111)109.9110.1122.2106.9105.1102.6109.7105.8107.0103.2112.0118.9105.0105.0109.4118.3104.1114.5116.0109.0113.1112.5110.1112.5101.1109.8110.8110.7114.0110.2109.9108.3102.3112.4113.7109.7125.2110.2124.60.25-0.3 80.29Molecule (A)109.0109.3120.2104.8112.6105.0112.5104.5108.8104.2107.3103.3112.1120.4106.2105.8107.3118.8105.5112.6115.1108.2111.3111.5110.1113.798.9110.1112.2110.3113.7111.1108.6107.8101.0112.9115.6110.1131.1109.2119.70.790.69-1 .OOhiIolecule (B j108.7111.5119.3107.1112.6104.6109.7106.8110.6104.3107.8102.0113.0118.8104.9107.5108.1118.2103.9113.6114.6110.9110.9112.6110.2114.298.4109.7113.4110.7112.4111.5108.8110.6101.5111.4114.7109.4131.1107.7121.10.72-1 .OO0.81The values of angles involving hydrogen atoms all lie in thenormal ranges.Stvuctuve of 2a-Hydroxydihydrophytzcberin (11)2a-Hydroxydihydrophytuberin was prepared by acid-catalysed hydration of phytuberin.The major a-epimerwas separated by fractional crystallisation. Recrystallis-ation from n-hexane yielded needles, some almost fibrous,m.p. 119 "C.Several crystals were mounted before asuitable single specimen was found ; this had dimensions0.35 x 0.10 x 0.12 mm3 and was enclosed in a fine capillarytube.Crystal Data.-C,,H,,O,, M = 312.4. Triclinic, a =10.887(3), b = 10.057(8), G = 8.661(5) A, a = 111.85(5),g cm-3 (by flotation), 2 = 2, D, = 1.20 g ~ m - ~ , F(000) =340, ~ ( M o - K E ) = 0.9 cm-l, A(Mo-KE) = 0.710 69 A. Spacegroup P1 (No. 1).Preliminary oscillation, Weissenberg, and precessionphotographs indicated the triclinic crystal system and thecell dimensions. The dimensions were determined moreaccurately by refinement from Guinier powder photographmeasurements. The non-centrosymmetric space groupwas considered more probable since the suggested molecularformula of phytuberinp = 89.63(4), y = 99.43(5)", U = 866.8(8) Pi3, D, = 1.16had four asymmetric centres.Intensity data were collected on a Stoe STADI 2 diffracto-meter which has Weissenberg equi-inclination geometry.The crystal was mounted with the u axis coincident with therotation (w) axis.Monochromated Mo-KE radiation (withattenuators) and an w-scan were used. Of the 3 021reflections with 28 50°, 176 were considered weak fromprescan searches, and intensity measurements were notmade; these reflections are included in the data-set withF, 0. Of the remaining 2 845 reflections, 1647 wereconsidered observed, having I 3 201, where I is the netintensity count, and 01, = S2(T + (b, + b2)12/1 600) whereS = attenuation factor, T = scan count, b, and b, arebackground counts, each ~ O S , n = number of scan steps,each step being 0.01" in w and 1.0s.Reflections havingI G 0 were given the value I = 1. Corrections forLorentz and polarisation effects, but not for absorption,were made, and structure amplitudes calculated. Thedesign of the diffractometer provides a constant crystal-to-detector distance, and all data are therefore scaled by thesame factor.Some reflections with low Y (i.e. 20') values could not bemeasured accurately in this setting on the diffractometer.For purposes of structure determination only, these re-flections were added to the data-set with intensities esti-mated visually from photographs.Structure A naZysis.-After normalisation of structureamplitudes, the E statistics confirmed the non-centro-symmetric space group P1.The high IEl value reflections,plus 50 low IEI value, $o, reflections (taken from the weakunmeasured reflections) were entered into the programMULTAN.' In several runs, normally with 32 sets ofphases produced, there was very little variation in theconsistency indicators of the different sets. E Mapscalculated from sample sets showed little sign of any likelymolecules. Eventually, in a run of MULTAN using 210reflections with /El 2 1.70 and in which the starting set ofreflections was selected manually, more variation in theconsistency indicators was achieved ; the set of phaseshaving the lowest absolute figure of merit (1.21 comparedwith all the rest in the range 1.31-1.38), the lowest t,b0-value and the highest RESID value, yielded the mostpromising E map in which substituted six-membered ringswere recognisable.A pseudo-centre of symmetry confusedthe map; a sufficient proportion of the atoms of the twomolecules was not found, and electron-density maps basedon these fragments did not yield a complete solution.The solution of the crystal structure of dihydrophytuberinhad by this time shown that the molecules of phytuberinand its derivatives should a11 be rigid and of similar con-formations, except perhaps in the acetate group. Conse-quently the co-ordinates of the atoms of the six-memberedring (Figure 1) were aligned with the peaks of a ring in themost promising E map (see earlier) and structure factorswere calculated for 18 atoms, O(1)-(3) and C(1)-(15), ineach of the six possible conformations.One arrangementlooked more feasible than the others and, after a cycle ofrigid-body refinement of this whole fragment, structurefactors were again calculated and an electron-density mapdrawn. Two possible substituted five-membered ringswere found in this map and from electron-density mapsbased on one of these fragments (plus the first moleculefragment), the second molecule of the cell eventuallyemerged, and the remaining atoms of the first molecule werelocated.In the course of a lengthy refinement process, 54 oJ.C.S. Perkin Ithe hydrogen atoins of the two molecules were located indifference maps, or their co-ordinates calculated in idealisedtetrahedral positions about the carbon atoms.The twomissing hydrogen atoms are of the hydroxy-groups in eachmolecule; they could not be positively identified in anyelectron-density or difference map, and their co-ordinatescould not be calcuIated. From examination of inter-molecular contacts, i t is thought that these hydrogenatoms should be involved in hydrogen bonds between thehydroxy-groups and the carbonyl groups of neighbouringacetate groups, but this has not been proved from the re-finement process. The numbering of the hydrogen atomsfollows the same pattern as that in dihydrophytuberin.In the least-squares refinement, all the carbon andoxygen atoms were refined with anisotropic vibrationparameters ancl the hydrogen atoms with isotropic para-meters.Scattering factors were as for (111). The refine-ment of 6 13 Parameters was performed in stages, refiningI C(17)o0 1 $iFIGURE 2 Projection of one molecule of Za-hydroxydihydro-phytuberin (11). This is molecule (A), with atomic labels asshown; molecule (B), has an almost identical view, in projec-tion, and its atonis are asterisked C* (1) , O* ( l ) , etc.three groups of atoms in successive cycles: in the firstgroup atoms O(1)-(2) and C(1)-(10) of both molecules, inthe second group atoms O(3)-(5) and C(l1)-(17) of bothmolecules, and in the third group all the hydrogen atoms(refer to Figure 2 for the atom numbering scheme). Itwas not necessary to maintain one atom’s co-ordinatesconstant (in order to define the cell origin) throughout therefinement process, since a t no time were all the heavieratoms’ parameters being refined simultaneously.In the final data set of 3 021 reflections, the weak re-flections were given zero-weighting and effectively excludedfrom the refinement process.The 2 845 diffractometermeasurements were initially given weights derived fromintensity counts, we = l/ac2 where oc2 = 01~/(41.Lp) andLp is the Lorentz-polarisation factor: in the final stages ofrefinement, the weights were modified so that the mean* See Notice to Authors No. 7, in J.C.S. Perkin I, 1975, Indexissue.l1 IBM 360 and ICL 4/70 programs: NUCLS, R. J. Doedensand J. A. Ibers; ORFFE, W. R. Busing, K.0. Martin, andH. A. Levy.values of w(lFol - I F , ( ) 2 in several ranges of lFol wereapproximately constant. The scheme used was: w =wc/(-0.145 + 0.2211F01 + 0.0511F012), but for lFol 1.0for the calculation of w, lFol was adjusted to 1.0.The refinement was concluded with R 0.134 and R’ 0.087for the 2 845 measurements, and I? 0.072 for the 1 6 4 7observed measurements. Final atomic parameters are inTable 4. Errors in bond lengths (Table 2) were calculatedfrom the atomic variances in the direction of the bonds; inthe bond-angle cr calculations (for Table 3) symmetricallyspherical atomic variances were assumed.Measured and calculated structure amplitudes for theanalyses of both derivatives, and the anisotropic thermalparameters, Uzj values, for the carbon and oxygen atomsin both structures are listed in Supplementary PublicationsNo.SUP 21697 (25 pp., 1 microfiche).*Computing.-The full-matrix least-squares refinementprogram NUCLS,ll the program ORFFE for determiningstatistical errors from the full correlation matrix, and theniultisolution program MULTAN for phase determinationby the tangent formula, have been adapted for use on theIBW 360/65X system of University College, London, andTABLE 4Atomic parameters for (11), with estimated standarddeviations in parentheses. The absolute configuration,as predicted in ref. 3, is defined with respect to a right-handed system of axes(u) Non-hydrogen atoms: co-ordinates (fractional x lo3)X- 28(5)2 861(6)2 635(6)550(6)2 164(7)968(9)1895(10)525(10)479(7)1871(8)2 476(8)2 144(6)2 430(9)1685(8)1998(8)2 856(7)2 432(10)3 263(9)4 220(10)1533(9)-342(7)1566(13)-4 052(6)-1 551(6)-3 203(5)-5 299(6)- 2 994( 10)-3 309(12)- 2 500(7)-3 104(7)-2 784(9)-3 068(8)-2 345(8)-2 878(8)-3 629(9)-1 550(9)-4 873(9)-4 366(9)-4 424(10)-1 999(7)-1 991(11)-3 749(7)-2 394(7)- 973(9)Y-1 428(7)-1 192(7)4 215(7)4 104(8)-904(8)- 1 732(10)-1 781(12)-1 971(11)-754(11)- 204(9)1332(10)2 307(10)871(10)4 070(10)5 353(11)4 235(11)2 86 ( 1 0)4 188(10)4 231(14)4 745(8)2 595(9)-429110)-755(12)4 399(7)-1 025(7)-1 002(8)4 095(8)4 989(11)5 149(12)4 035(10)3 450(10)1890(10)688(9)2 464(10)3 722(10)5 029(11)99 1 ( 1 0)-830(10) - 2 015( 10)-1 012(11)4 043(11)2 969(12)-1 323(15)-1 074(11)z1 866(8)146(8)1464(8)1586(10)2 669(12)-2 346(8)-1 417(13)-1 024(13)743(13)1 345(11)1477(11)4 561(11)4 459( 11)3 017(11)3 108(11)4 476( 14)3 213(14)3 347(13)713( 13)950(12)3 021(10)-796(16)-2 884(9)-1 314(8)-2 753(8)-3 024(9)-3 691(13)-73(14)-1 798(12)-2 473(12)-2 659(11)-4 216(11)-5 756(12)-5 575(11)-4 140(11)-4 354(12)-6 804(12)-4 443(13)- 4 508(12)-1 786(15)-2 327(13) - 687(15)1 153(9)249(131343 1976TABLE 4 (Contimed)(b! Hydrogen atoms : co-ordinates (fractional xtropic thermal parameterstH(11)H(12)H(2)H(31)H(32)H(61)H(62)H(7)H(81)H(82)H(91)H(92)H(121)H(122)H(123)H(131)H(132)H(133)H(141)H(142)H(143)H(151)H(152)H(153)H(171)H(172)H(173)H * ( l l )H*(12)H*(31)H* (32)H*(61)H*(62)H* (2)E:g\) H*(82)H*(91)H*(92)H*(121)H*( 122)H"(123)H*(131)H* (132)H* (1 33)H*(141)H* (142)H*(143)H* (1 51)H*(152)H*(153)H*( 171)H*( 172)H*( 173)X121(6)227(6)58(5)30(8)7(7)229(5)336(4)126(5)223(7)342(7)184(5)144(9)280(11)289(6)456(6)475(8)403(6)341(6)350( 7)83(6)449(9)- 17(8)- 13(12) - 113(9)11 1 (10)9 3 w231(9)- 242(8) - 343(8) - 138(7)-318(9) - 396(8) - 258(5) - 143(5)- 329(6)-lS9(4) - 402(4)-295(5)-448(7)- 366(9)-116(9) - 141(8)-92(7)-55(8)-36(9) - 504(7)- 465 (10)-55218)-465(8)- 380( 11)- 52 1 ( 10)- 399(7)-333(9)- 94(9)Y- 260(7) - 197(7)- 260(8)-299(10)- 195(8)141(7)136(6)256(7)300(9)231(8)82(7)77(6)503(10)541(14)641 (8)408(9)529( 11)359(11)1(7)- 80(8)- 166(9)133( 10)91(14)327(13)443(10)464(12)595( 10)51 6( 10)583(9)625(11)488(10)181(6)194(6)-4(11)53(8)22(8)96(5)246(5)251(7)- 183(10)- 244( 11)-278(11)-64(12)- 197(10)-34(9)316(11)458(11)472(12)231(10)269( 12)346(10)- 51( 11)-170(13)- 233( 12)t See Table 1.z200(9)362(8)- 197(9)-89(11)- 184(10)41 (8)163(6)275(8)548( 10)469( 10)428(9)551(7)436 ( 1 2)549(16)434(10)422( 10)337( 12)2 3 2 (1 4)445( 10)248( 11)206(13)40(13)339(9)-1(16)-150(14)-103(12)- 99( 13)-294(10)- 474( 12)lOl(11)5(W57(11)- 150(7)-277(7)- 432( 10)-678(9)- 596(6)-550(6)-655(8)- 609( 11)- 694(14)-540(13)- 537 (14)-460(11)- 340( 11)- 489( 12)-543(13)-352(14)- 291( 11)- 84 ( 1 4)- 140(12)-49(15)-80(13)32(12)lo3) and iso-B/A23.5 ( 16)2.8( 14)4.1 ( 18)5.8(25)4.7( 19)2.1 (14)0.5(10)1.5( 13)5.4 ( 1 9)5.3 ( 19)3.1(16)2.4 (1 2)6.0 (25)9.6(35)3.3( 19)5.0( 20)7.5 (25)10.2 (26)3.4( 15)4.7( 17)4.5 (1 9)6.3 (2 6)12.3 (36)9.3 (28)1 0.4 (3 3)7.8(25)5.5( 2 7)8.5( 23)6.9 (24)4.8(21)8.2(27)4.9(23)1.5(11)2.0(13)4.0( 18)4.7(17)0.6 ( 10)3.1 (1 4)4.6(20)8.3(28)5.3 (2 8)10.3 (32)7.8 (24)5.3(21)7.8( 26)8.8 (28)7.9( 21)6.7 (21)9.6(32)7.7(25)7.0 (25)13.5( 35)10.4(32)0.8(9)on the ICL 4/70 computer a t this Station.An IBM 1130computer and our X-RAY ARC l2 library of programs wereused for the remainder of the computing for these analyses.DISCUSSIONThe two independent molecules in the crystal form of(11) are remarkably similar in conformation, and apseudo-twofold-screw axis, parallel to the a axis, closelyrelates the two molecules. Figure 2 is a projection ofmolecule (A) ; the corresponding projection of molecule(B) differs only slightly from that shown, in the acetategroup. Comparison with the projection of (111) (Figure1) shows that differences in conformation are small, andonly about C(2) is there any significant difference.These two Figures show the atomic numbering schemeswhich are based on that of eudesmane to which, it hasbeen suggested? phytuberin is related.Torsion angles (") inTABLE 5the rings of the phytuberin derivatives(11)C(9)-C( 1O)-C(5)-C( 6)C( 1 0) -C (5)-C( 6)-C ( 7)C (5)-C (6)-C( 7)-C (8)C (6)-C (7)-C (8)-C( 9)C( 7)-C( 8)-C( 9)-C( 10)C( 8)-c (9)-C( 1 0)-c (5)C( 1) -C( 1 0)-c (5)-C( 4)c.(4) -0 ( 1 ) -c ( 1 ) -c (1 0)O( 1)-C( 1)-C( 10)-C( 5)C(4)-C( 5)-0 (2)-C(2)C(5)-0(2)-C(2)-C(3)0 (2)-c (2)-C( 3)-C( 4)C( 2)-c (3) -c (4) -c (5)C ( 1 0)-C( 5)-C (4)-0 (1)C( 5)-C(4)-0 (1)-C( 1)C(3)-C(4)-C( 5)-O( 2)Range of CJMean (T(1111 'Molecule (A) Jiolecule (Bj44.4 45.3 43.6-45.0 -47.1 - 43.950.3 53.4 50.7-57.6 - 59.1 -57.161.7 62.9 60.7-53.7 - 54.2 -52.730.0 29.6 28.8- 16.1 - 8.5 - 8.6- 6.0 - 18.4 - 18.226.3 39.1 39.7- 34.9 -41.2 -42.1- 7.0 28.2 28.228.5 - 36.0 - 33.9-37.7 28.7 24.932.6 - 11.8 -8.5- 16.3 - 8.5 - 10.40.35-0.43 0.87-1.05 0.88-1.120.39 0.97 1.01TABLE 6Mean planes.Deviations (A) of atoms from mean planesin the phytuberin derivatives. Values in italicsindicate atoms used to define the plane(11)(111) kolecule (A) Molecule (Bj- 0.54 - 0.57 - 0.54- 0.007 - 0.002 - 0.0100.011 0.002 0.0070.71 0.71 0.68-0.012 - 0.003 - 0.0090.010 0.002 0.009- 0.002 - 0.001 -0.001 - 0.003 - 0.001 - 0.0020.004 0.010 0.013-0.002 -0.005 - 0.0050.026 - 0.095 -0.0270.080 - 0.004 - 0.2010.025 -0.78 0 -0.52 O -0.510.52 0 -0.22 0 -0.23 0-0.037 -0.76 0 0 0 0-0.028 0 0.42 0 0.41 00.51 0.019 0 0 0.50 0-0.014 0.85 -0.21 0 -0.24 00.046 0 0.30 0 0.21 0-0.54 0 0 -0.46 0 -0.39-0.031 0 0 0 0 00.41 0.045 0.82 0 0.41 0The molecular dimensions of all three molecules arein Tables 2, 3, and 5.Mean planes, describing theconformations of the six- and five-membered rings andthe acetate groups are in Table 6.l2 ' X-Ray ARC,' Library of Programs for the IBM 11 30 Com-puter, J. Afipl. Cryst., 1973, 6, 3091344 J.C.S. Perkin IIn the molecules of both derivatives, the skeletalstructures are rigid, consisting of three rings which havecommon edges or vertices. There is considerable con-formational strain in each system; all the bond lengthsthe same molecule, ca.3.00 A in every case. Similarlythe O(4) - - H(7) and O(4) - . H(121) distances lie inthe range 2.37-2.66 A. The dimensions of the acetategroups are normal,13 and in each molecule, the fourare as expected but the valence angles show the dis-tortions of strain.The cyclohexyl ring in each molecule has a chairconformation; the angles within the rings range from108.2 to 116.0", and the torsion angles (60" in an idealsystem) vary from 44 to 63". The variation in anglesfollows very similar patterns in the three molecules.From the best-mean-plane of those of four ideally-planarcarbon atoms in the six-membered rings, C(5) and C(8)are displaced ca. 0.55 and 0.70 A in each molecule.From the cyclohexyl ring, the substituent bondsC(5) -0( 2), C( 7)-C( 1 1) , and C (10)-C( 1) are equatoriallyarranged; C(5)-C(4) and C(lO)-C(14) are axial sub-stituents.In (111) the two five-membered rings are each bestdescribed as having envelope shapes.One atom in eachring [viz. C(10) and C(3)] is ca. 0.5 A displaced from themean plane of the other four, but in each case, the fouratoms defining the plane cannot be considered strictlyplanar, some deviations from the plane being ca. 10 C.The torsion angles around each ring also illustrate thelack of planarity. The angles between the normals tothe four-membered planes and their envelope-flapcomplement planes are 32.9 and 36.0" in the two rings.The hydroxy-group substitution in the molecules of(11) affects the shapes of the five-membered rings; allnow have conformations closer to ' half-chairs.' Inthe 0(1 rings of both molecules, O(1) and C(5) areca.0.5 B above and below the plane of the other threeatoms, whilst in the O(2) rings, C(3) and C(5) are ca.0.4 A, either side of the corresponding plane.The acetate group in each of the three molecules foldsaround, so that O(4) lies close to both C(7) and C(12) ofl3 J. F. McConnell and J. D. Stevens, J.C.S. Perkin 11, 1974,345.atoms 0(3), 0(4), C(16), and C(17) form a good plane;C(11) and C(13) are not far displaced from this plane.TABLE 7Some shorter intramolecular non-bonded interactions (A)in the phytuberin derivatives(11)(111) kolecule (A) Molecule (BjO(4) * * * C(7) 3.009 3.026 3.100O(4) * * * C( 12) 3.000 2.987 2.977C(1) * - * C(2) 3.173 3.659 * 3.606 *C(6) * * * C(16) 3.386 3.386 3.480C(7) + * * C(15) 3.308 3.356 3.3460(1) * * * H(91) 2.72 * 2.48 2.56O(2) - * * H( 143) 2.53 2.50 2.39O(4) - - - H(7) 2.41 2.37 2.49O(4) - - - H(121) 2.49 2.37 2.66 *C(6) * * - H(133)C(7) * * H(151)C(8) * - H(131)C(12) * - - H(81)C(13) * - - H(62)C(13) - - - H(82)C(14) * - - H(82)C(15) - * * H(7)2.842.912.712.762.832.792.792.702.77 2.842.62 2.852.77 2.702.79 2.652.70 2.762.73 2.832.83 2.882.67 2.92H(11) * * * H(22)ZH(2) 2.29 3.61 * 3.66 *H(32) * * * H(153) 2.85 * 2.68 * 2.30H(62) * * H(133) 2.28 2.24 2.31H(7) - - - H(121) 2.41 * 2.33 2.26H(7) * * - H(151) 2.18 1.77 2.25H(82) * * - H(131) 2.24 2.17 2.13H(91) - * - H(151) 2.25 2.47 * 2.54 *H(32) - * * H(152) 2.44 * 2.76 * 2.10* These distances are not considered short but are includedfor comparison.Space-filling models show each molecule to be verycompact with only the C(17) methyl groups and, in(11), the hydroxy-groups capable of any uninhibite1976 1345vibration.Some short intramolecular distances are inTable 7.The two compounds differ most in their packingarrangements. Figure 3 shows the packing of theTABLE 8Intermolecular distances (A) in (111)O(4) * - * C(17I) 3.489O(4) * - C(29 3.530C(14) * * * C ( 1 P ) 3.806C(2) * * - C(15IV) 3.828C(2) * * - C(16v) 3.880C(12) * * H(14311) 2.94C(14) - H(l63111) 3.00C(15) * - * H(21VI) 3.05H(123) * - H(14311) 2.45H(11) .H(133VII) 2.48H(142) * * - H(163III) 2.60Roman numeral superscripts denote the following equivalentpositions relative to the reference molecule a t x, y, 2:I - x - l l , y + ; ) , - - z + l I V - x , y - ~ , - z + lI11 x , y, z - 1 VI - x . y + *, -2 + 1I1 x - 1 , y , 2 v x + L Y , ZVII -x, y + &, -zTABLE 9Intermolecular contacts in (11)(a) Dimensions of suggested hydrogen-bonding scheme (dis-tances A, angles "1Atoms Dimensionsa-b * - c-d b . . . c a-b...c b.*.c-dC(2)-0(6) * * 0*(4)-C*(16) 2.830 103.9 137.4C*(2)-0*(5) - * * 0(4)-C(16) 2.800 101.9 143.3(b) Other short intermolecular distances (A)Cell translationsAtoms Distance of atom b froma... bO(2) - * C*(17)0*(4) - * - C(2)0*(2) * * - C(17)C(l) * * * C*(13)O(4) * * C*(2)C(2) - * * C(12)C*(3) * * C*(12)C(13) - - * C*(l)C(13) * * * C*(12)C*(12) * - H(132)C*(3) - * - H*(122)C(13) * * - H*(12)C*(13) .* - H(12)C(14) - * * H*(121)C*(14) * H(121)C(l) * - - H*(131)H(132) * * - H*(123)H(131) - * H*(12)H*(l2) - - - H*(123)a ' . . b3.3703.3763.4433.4663.6763.7683.7933.8393.8942.822.892.932.932.952.952.982.292.302.39x10-1000011-1010100110Y00000-1101-1100000101Z00001-1111-111-11-11110dihydro-derivative (111) in which the binding of mole-cules is solely by van der Waals interactions; shorterintermolecular contacts are in Table 8. In the hydroxy-dihydro-derivative (11) , however, there is a possibilityof hydrogen bonding, and, although the hydroxy-grouphydrogen atoms have not been located, intermoleculardimensions (Table 9) indicate that there should behydrogen bonding between each O(5) atom and the O(4)atom of the other molecule; the hydrogen bonding isshown in Figure 4, spiralling parallel to the a axis (theneedle axis of the crystal). Other short intermoleculardistances are also in Table 9.Although the molecular structures are not readilyapparent in either packing diagram, each arrangement1 :I , I\ Y \ 1'FIGURE 4 The packing arrangement of (11), showing the spiral-ling hydrogen-bonding system and the pseudo-two-fold screwsymmetry parallel to the a axis. The molecule labelled I,which is a type (A) molecule, is arranged in a similar orientationto that marked I in Figure 3. The projection is down theb axishas the acetate group lying roughly parallel to theplane of the drawing, and the molecules labelled I ineach diagram are in similar orientations; the inter-molecular contacts and packing arrangements are thusshown to be quite different.I thank Dr. D. T. Coxon (A.R.C. Food Research In-stitute, Nonvich) who initiated this work and supplied thecrystals, Professor A. C. T. North and Dr. B. Sheldrick(Astbury Department of Biophysics, University of Leeds) ,and Dr. G. H. W. Milburn (formerly at Sheffield Poly-technic) for use of their diffractometers and assistance inintensity-data collection. I also thank Professor M. R.Truter for valued discussion, the Director of the UniversityCollege Computing Centre for facilities, and the RoyalSociety for some equipment.[6/1893 Received, September 30th, 1976