2 Physical Methods and Techniques Part (ii) X-Ray Crystallography By S. NEIDLE Department of Bioph ysics Kings College (University of London 1 26-29 Drury Lane London WC2B 5RL The literature of organic crystal structures continues to grow at a steady pace of about 2000 newly reported structures per annurn. As in previous years the reader is referred to the latest extensive bibliographies from the Cambridge Crystallographic Data Centre.’ The detailed files of the Data Centre are now available on computer installations world-wide and are extensively used for speedy access to details of all published organic molecular structures.2 As detailed below this powerful facility is being utilised as a structural tool in its own right capable of extracting new and meaningful data by means of structural correlations.Several trends in organic crystallography have become apparent during the past several years. The power of spectroscopic methods is now such that X-ray crystal- lography only rarely needs to be called upon as a determinant of molecular structure its traditional role in organic chemistry. Thus crystallographers are increasingly defining new areas for their attention. Studies relating structure to theoretical calculation and prediction either for bonding geometry or molecular conformation are of particular interest at present as are analyses involving solid-state reactivity uis h uis structure. It is also increasingly apparent that small-molecule as well as protein crystallography can provide important data on biological molecules and such molecules serve as models for biological systems.Thus organic crystallography is now moving rapidly into the phase of its development where the crystallographer himself defines his problems and performs his own chemistry and biochemistry as well as solving and analysing his structures. 1 Molecular Structure and Bonding A number of studies have been concerned with the influence of crystal structure (i.e. non-bonded forces including inter-molecular hydrogen bonds) on molecular con- formation. In one approach the conformational polymorphism of N-(p-chloro-benzy1idene)-p- chloroaniline has been e~amined,~ by combining molecular orbital calculations with an analysis of the molecular packing in the two space groups found ‘Molecular Structures and Dimensions’ Crystallographic Data Centre Cambridge University 1978 Vol.9; 1979 Vol. 10. * F. H. Allen S. Bellard M. D. Brice B. A. Cartwright A. Doubleday H. Him,T. Hummelin B. G. Hummelink-Peters,0.Kennard W. D. S. Motherwell,J. R. Rodgers and D. G. Watson Acra Cryst. 1979 B35.2331. J. Bernstein and A. T. Hagler J. Amer. Chem. Soc. 1978,100,673. 14 S. Neidle for this molecule. Extensive conformational polymorphism has been noted in the three crystal forms of iminodiacetic acid with each showing a distinct molecular conf~rmation.~ Intermolecular forces in hydrogen-bonded crystals of carboxylic acids and amides have been examined' with consistent-force-field calculations. Ab inifio molecular orbital calculations of hydrogen bonds in carbohydrates6 suggest that for 0-H -* 0systems the energy minimum is not at a 180" bond angle but at -163"; neutron-diffraction data gives an average of 166".The Cambridge data files have been used for an alternative approach to the problems of conformational variability in crystals based on the examination of all structures determined and employing standard statistical methods to analyse the data.7 In one instance the geometry of the p-1'-aminofuranoside ring (in nucleic acids and their constituents) was examined in some 100 structures and statistically meaningful correlations among several geometric parameters were noted.' The geometry of substituent- induced deformation of bond length in cyclopropane-containing structures has recently been investigated;' the trends found are not large but support the predic- tions of molecular orbital theory.The structure analysis of 10,10-dimethyl-3,4- dioxatricyclo[5.2.1 .0'*5]decane-2-spiro-2'-adamantane(1)has led to a comparison of C-0 and 0-0 bond lengths in small rings as a function of ring size." The exploration of reaction pathways via examination of relevant crystal struc- tures is proving to be most instructive. Dunitz and his colleagues have approached the problem by examining related compounds that all react in the same manner but say with varying rates. A crystallographic study of four compounds that display aspects of the keto-acid-hydroxy-lactoneisomerization (2) has shown two to be on OQ OH the ring-opened side but with short 0 --C=O distances.The two ring-closed compounds have lengthened C-0 bonds.' ' Explorations of solid-state reactions continue to be a fruitful area; for example a plausible explanation in terms of differences in molecular packing has been advanced for the racemic +chiral thermal conversion of 2,6-dimethyl-4-(a,a-diphenylmethylene)cyclohexa-2,5-dien-1-one.12 X-Ray data on eight derivatives of cis-4a,5,8,8a-tetrahydro-l,4-naphtho-quinone have been correlated with their solid-state phot~chemistry.'~ J. Bernstein Actu Cryst. 1979 B35 360. 'S. Lifson A. T. Hagler and P. Dauber J. Amer. Chem. Soc. 1979 101,5111. M. D. Newton G. A. Jeffrey and S. Takagi J. Amer. Chem. Soc. 1979,101,1997. 'P.Murray-Rust and W. D. S. Motherwell Acfu Cryst.,1978 B34 2518 2527.a P.Murray-Rust and W. D. S. Motherwell Actu Cryst. 1978 B34,2534. F. H. Allen Actu Cryst. 1980 B36 81. lo P. B. Hitchcock and I. Beheshti J.C.S. Perkin IZ 1979 126. l1 D.J. Chadwick and J. D. Dunitz J.C.S.Perkin ZZ 1979 276. l2 T.W.Lewis I. C. Paul and D. Y. Curtin Actu Cryst.,1980 B36,70. l3 J. R. Scheffer and A. A. Dzakpasu J. Amer. Chem. Soc. 1978,100,2163. Physical Methods and Techniques -Part (ii) X-Ray Crystallography 15 It is not uncommon for crystal structures to provide the theoretician with new phenomena to explain. This is well illustrated by for example [3]-and [4]-rotane (3)and (4) both of which have abnormally short mean central C-C bond distances of 1.470 A suggesting that there are v-bonding contributions in these structure~.'~ The central C-C bond length in 9,9-bitriptycyl has been found to be shorter than predicted by force-field calculations even though these correctly predicted other aspects of the ge~metry.'~ An authentic example of an intramolecular 0-H * -v hydrogen bond has been reported in the crystal structure of 2,6-diphenylphenol.l6 2 Natural Products and Biological Molecules The natural world continues to cause surprises with the bizarre nature of the secondary metabolites of plants. Noteworthy among these crystal structures are mintsulphide (9 a sulphur-containing sesquiterpene from peppermint oil," and solanscone (6) a novel sesquiterpene ketone from Nicotiana tabacum (as the oxime).18 DL-Bi-( 0-trimethyl-cis- brazilane) (7) a derivative of brazilin has an H (8) (7) l4 C.Pascard T. Prange A. de Meijer W. Weber and J.-P. Barnier J.C.S. Chem. Comm. 1979,425. lS M. H. P. Ardebili D. A. Dougherty K. Mislow L. H. Schwartz and J. G. White J. Amer. Chem. SOC. 1978,100,7994. l6 K. Nakatsu H. Yoshioka K. Kunimoto T. Kinugasa and S. Ueji AcfaCryst. 1978 B34 2357. '' T. Yoshida S. Muraki K. Takahashi T. Kato C. Kabuto,T. Suzuki T. Uyehara andT. Ohnumen J.C.S. Chem. Comm. 1979,512. 18 T. Fujimori R. Kasuga H. Kaneko S. Sakamura M. Noguchi A. Furusaki N. Hashiba and T. Matsumoto J.C.S. Chem. Comm. 1978 563. 16 S. Neidle expected staggered conformation; however the three central bonds C-10-C-11 C-11-C-1 l’ and C-ll’-C-lO‘ are all longer than normal C(sp3-sp3) bonds with the central one of length 1.61 This is most likely a result of the steric overcrowding in the molecule.An important code of recommended practice and standardized procedures for the assignment of absolute configurations has been established.20 The use of the principles laid down in this code has led to the revision of the chirality originally assigned for clerodin,2’ one of two errors found so far among the ca. 900 absolute configurational assignments derived by X-ray crystallography that are in the lit- erature. The analysis of steroid structures continues to be an active topic. The two independent molecules of 3p-acetoxy-l6/3-methylpregn-5-en-20-one have their side-chains in very different conformations;22 energy calculations show that the minimum is a broad one. The careful analysis of the structures of four oestranes in the light of conformational data on other 1,3,5(10)-oestratriene structures has shown the role of conformational transmission in that exocyclic non-bonded interactions are crucial factors in determining conf~rmation.~~ The structures of three 4,4,14a-trimethyl-19(10 -B 9)-abeo- 5P,9p,lOa-pregnane-6,1 l-diols [typified by (8)] have been dete~mined.~~ Their conformations are as predicted by force-field methods which have also been used to analyse puckering parameters and steric energies.Interest in structural studies of small peptides has been revitalized by the dis- covery and subsequent intensive biochemical study in many laboratories of the enkephalins. These are pentapeptides that occur naturally in animal brains that mimic the action of morphine and which bind to the opiate receptor.Several crystallographic analyses of enkephalin fragments have been reported. The N-terminal tripeptide residue Tyr-Gly-Gly exists as a zwitterion with an a-helix-like conformati~n,~~ although no intramolecular hydrogen bonds were observed. The structures of two tetrapeptide enkephalin fragments Tyr-Gly-Gly-Phe and Gly- Gly-Phe-Leu have been determined.26 Their conformations are quite distinct; the former has a Gly-Gly @-turn whereas the latter has a bent conformation without intramolecular head-to-tail interaction. As the p-turn has been suggested to occur for enkephalin itself in solution it has been proposed that Tyr-Gly-Gly-Phe is the biologically active sequence.This and other possible biologically important features have been reported in the crystal structure of [Leus]enkephalin itself .27a However it has now transpired2” that the original diffraction data recorded (from small thin crystals) corresponded to a small sub-cell of the true unit cell. Although it is now apparent that the crystal structure is in fact an average from four closely l9 M. F. MacKay and N. W. Isaacs Tetrahedron 1979,35 1893. ’O D.Rogers and F. H. Allen Acta Cryst. 1979 B35,2823. ’’ D. Rogers G. G. Unal D. J. Williams S. V. Ley G. A. Sim B. S. Joshi and K. R. Ravindranath,J.C.S. Chem. Comm. 1979,97. 22 H.Campsteyn 0.Diderberg L. Dupont and J. Lamotte Actu Cryst.,1979 B35 2971. 23 W. L. Duax D. C. Rohrer R. H. Blessing P.D. Strong and A. Segaloff Actu Cryst. 1979 B35,2656. 24 J. C.A. Boeyens J. R. Bull A. Tuinman and P. H. Van Rooyen J.C.S. Perkin 11 1979 1279. ’’ W. M.Carson and M. L. Hackert Acta Cryst. 1978 B34 1275. 26 T. PrangC and C. Pascard Actu Cryst. 1979 B35 1812. ” (a)G. D. Smith and J. F. Griffin Science 1978,199,1214;(b)T. L.Blundell L. Hearn I. J. Tickle R. A. Palmer B. A. Morgan G. D. Smith and J. F. Griffin Science 1979,205,220. Physical Methods and Techniques -Part (ii)X-Ray Crystallography related molecules it has not as yet been possible to solve the true structure and hence unequivocally to define all the conformational features especially those of the side-chains. Among the several structures of cyclic peptides published during the period under review two are especially worthy of note.The structure of the cyclic decapeptide gramicidin S (as a complex with urea) has finally been solved,28 after many years of effort. It has many of the features predicted in an early as well as some unexpected ones such as an intramolecular hydrogen bond. The conformation of the cyclic decapeptide antamanide is the same as that found in several solvents as well as in several derivatives with variants in the ~ide-chains.~' It has thus been concluded that the conformation found is an intrinsic property of the molecule. The crystal structures of several fragments of the nucleic acid DNA have attracted much attention. The tetranucleotide d(pApTpApT) forms Watson-Crick base-pairs in the solid but not as a continuous fragment of double-stranded DNA.Instead two short segments of base-pairing are formed by each tetranucleotide by means of hydrogen-bonding to two other molecules in the crystal structure. On the other hand the hexanucleotide d(CpGpCpGpCpG) crystallizes as a double-helical Watson-Crick base-paired duplex.32 This structure has a number of remarkable and unexpected features which may well have considerable biological implications; for example the helicity is left-handed rather than right-handed as in the classical model of DNA. ** S. E. Hull R. Karlsson P. Main M. M. Woolfson and E. J. Dodson Nature 1978 275 206. 29 D. C. Hodgkin and B. M. Oughton Biochem. I. 1957,65,752. 30 I. L. Karle,T. Wieland D. Schemer andH. C. Ottenhegm Proc. Nu?.Acad. Sci. U.S.A.,1979,76,1532. 31 M. A. Viswamitra,P. Kennard P.G. Jones G. M. Sheldrick S. Salisbury L. Falvello and Z. Shakked Nature 1978,273 687 32 A. H.-J. Wang G. J. Quigley F. J. Kolpak. J. L. Crawford J. H. van Boom,G. van der Marel and A. Rich Nature 1979,282,680.