2 Physical Methods Part (iii)X-ray Crystallography By M. B. HURSTHOUSE ChemistryDepartment Queen Mary College Mile End Road London El 4NS S. NEIDLE BiophysicsDepartment King’s College 2629Drury Lane London WC2B5RL Since the end of 1973 over 2000 organic structural papers have appeared. Some of these structures are described in other chapters of this and the 1974 Report; references to virtually all structures published can be found in ‘Molecular Structures by Diffraction Methods’’ and ‘Molecular Structures and Dimensions’.2 It is therefore our intention in this short contribution to report on recent developments in the processes of structure determination by X-ray methods and also on work which can best be described as ‘applied structure analysis’ i.e.use of crystal structure data for purposes other than molecular structure identification. One development which has arisen out of the availability of accurate X-ray data is the study of electron-density distributions in molecules. Coppens has outlined3 the implication of combined X-ray and neutron diffraction studies of bonding and lone-pair electron distributions and the results of such a study have been described for the molecule 1,3,5tria~etylbenzene,~ and compared with charge populations determined from semi-empirical calculations. Suitable modifications to the struc- ture refinement process can lead to the same information just with the use of X-ray data as has been shown for 3,6-dimethyl-S-thioformyl pyrrolo[:,l-b ]thiazole+ and tetracyanocyclobutane;b in the latter the polarities of the C-H bond and C=N group have been determined.The development of direct methods of structure analysis in which the phases needed for the correct synthesis of the crystal structure image are determined by building up relationships between them so that assignment of a few key phases leads to estimates for many has been of great importance for organic structures almost eliminating the need to make heavy-atom derivatives. ‘Molecular Structures By Diffraction Methods’ ed. L. E. Sutton and G. A. Sim (Specialist Periodical Reports),The Chemical Society London,1975 Vol. 3. ‘Molecular Structures and Dimensions’ Crystallographic Data Centre Cambridge University. 3 P.Coppens Acta Cryst. 1974 B30,255. B. H. O’Connor and E.N. Maslen Am Cryst. 1974 B30,383. A. Sharrna and R.C. G. Killean Acca *st 1974 B30,2869. 6 M. Hare1 and F.L. Hirshfield Actu Cryst. 1975 B31 162. 24 Part (iii) X-Ray Crystallography The solution of small and medium-sized (i.e.up to 50 atoms per asymmetric unit) ‘equal-atom’ (i-e. atoms of similar atomic number) crystal structures by well-established direct methods is now often a routine procedure. Indeed the ‘symbolic- addition’ method of phase determination which can be applied without the use of a computer has had some remarkable successesin the hands of skilled practitioners; for example the analysis’ of the cyclic dodecadepsipeptide valinomycin involved 156independent non-hydrogen atoms and is probably the largest structure solved by direct methods to date.However many structure analyses still pose problems and improvements and new developments are still being sought. Hauptmann,’ Schenk,’ and others have evolved a promising approach based on quartets of related reflec- tions rather than the triplets of ‘classical’ direct methods. Quartets seem to be a more reliable means of phase determination judging from the few applications which have been made so far. They have also been shown to be especially useful when in the form of negative quartets,” in which very weak reflections are utilized being particularly powerful in discriminating between correct and incorrect phase sek1 Woolfson and his collaborators have recently developed the concept of ‘magic integers’12” by means of which several unknown phases may be represented by a single symbol thus enabling a large number of phase sets to be developed.This technique has been successfully employed in the analysis12b of cephalotaxine which had previously resisted attempts at solution by conventional direct methods. Over the past few years an increasing amount of attention has been paid to the idea ofstructure-solving via crystal-packing considerations. In part these techniques have been developed as alternatives to direct methods and have been particularly useful in cases where special features of the molecular geometry have upset the phasing processes in the direct method approach. In addition however the depen- dence of these techniques on calculations of lattice energies using semi-empirical potential energy functions for the various ‘non-bonded’ interatomic interactions present has also led to the development of methods applicable to the calculation of conformations of individual molecules.For example in the structure analysis of guanosine- 3’,5’-cytidine monophosphate the most probable molecular conf orma- tion was first computed and then used as the basis for the packing search. This was also restricted by known constraints from the X-ray data (e.g. unit cell parameters space group position of the phosphorus atom orientations of the bases) and by inclusion of a geometric function which imposed Watson-Crick base-pairing. During the caIculations an interactive computer graphics system was used for visualizing the calculated structures.The structure of the rigid molecule spirodienone (1)was also determined by a visual packing analysis coupled with the minimization of repulsive energies.l4 I. L. Karie J. Amer. Chem. Sac. 1975,97,4379. H. Hauptmann Acra Crysf.,1974 AM 822. H. Schenk Actu Cryst. 1974 AM,477. lo H. Hauptmann Acra Cryst. 1974 AM,472. G. T. De Titta J. W. Edmonds D. A. Langs and H. Hauptmann Acru Cryst. 1975 A31,472. l2 (a) P. S. White and M. M. WooIfson,Acta Crysf.,1975 A31,53;(b)J. f.Declerq G. Germain and M. M. Woolfson ibid. p. 367. l3 S. D. Stelfman,B. Hingerty,S. B. Broyde,E. Subramanian T. Sato and R. Langridge,Biopolymers 1973 12 2731. l4 B. S. Hass T. V. Willoughby C. N. Morimoto D. L. CuIlen and E. F. Meyer Acru Cryst. 1975 B31 1225. M.B.Hursthouse and S.Neidle In the structure determination of coumarin by potential energy calculations coupled with minimum residual analysis attempts were made to assess the relative importance of van der Waals and dipole-dipole interactions in deciding the actual packing minima.” The results suggest that in spite of the high (ca. 4.50 D) moment for the molecule dipole-dipole interactions do not seem to determine the packing. The success of the application of these studies to unknown structures depends on the availability of accurate energy parameters and of course these can be determined from known structures. Lifson et al. have outlined16” a procedure for determination of such parameters and have shown it to be applicable to other fields of conforma- tional analysis.The derivation of a consistent force field for amides by the fitting of parameters of various trial functions to such experimental data as heats of sublima- tion dipole moments and crystal-structure parameters has also been described.16’ A further ‘by-product’ of these studies has been the development of methods for representing molecular geometries and effecting changes in them. For example Mackay has described” how arbitrary combinations of bond lengths bond angles and torsion angles can be used as generalized co-ordinates for describing molecular models and how these and conventional Cartesian or fractional unit cell co-ordinates can be interconverted. In the ‘method of local change’ the computer manipulation of moleCular conformation is effected by movement of only one atom at a time giving considerable simplification of the programming required.l8 Also relevant to the study of molecular geometries and conformations derived from crystal structure analyses is the descriptionlg of a new computer program for the comparison of these features by a least-squares-fitting procedure for equivalent molecules or fragments found in different structural environments. Energy considerations are also involved in the study of strained molecules and a number of structure analyses have been reported in this area. The analysiszo of the two isomers of 2,3:6,7:2’,3’:6’7-tetrabenzoheptafulvalene(2)confirms an earlier assignment of the trans and syn isomers. Both are severely overcrowded and the strain is to some extent relieved by buckling of the non-alternant heptafulvalene ring as well as significant non-planarities of the phenylene rings.The overcrowded polycyclic aromatic hydrocarbon tetrabenzo[ a cd j lmlperylene (3)assumes a non- planar shape largely in order to relieve the overcrowding between C-2and C-30 and C-13 and C-19,which would otherwise be some 2.4 A apart instead of the 3 A separations observed.2’ In spite of the severe deformations the bond lengths still suggest a large measure of aromatic character for the molecule. *5 E. GavBzzo F. Mazza and E. Giglio Actu Cryst. 1974,B30,1351. l6 (a)A.T. Hagler and S. Lifson Acra Oysr. 1974 B30 1336; (b)A.T.Hagler E. Huler and S. Lifson J. Amer. Chem. Soc. 1974,% 5319. A.L.Mackay Actu Cryst.1974 A30,440. l8 J. Hermans jun. and J. E. McQueen jun. Actu Cryst. 1974 A30,730. l9 S.C.Nyburg Acra Cryst. 1974 B30,251. 2o K. S.Dichman S. C. Nyburg F. H. Pickard and J. A. Potworvorski Actu Cryst. 1974 B30,27. 21 Y.Kohno M. Konno Y. Saito and H. Inokuchi Acru Oyst. 1975 B31,2076. Part (iii)X-Ray Crystallography co The structure determination of l-methoxycarbonyl-5,7,12,14,16(4)-tetracyclo[9,2,2 14311,08*’6]hexadecapentaene (4) is of interest in that a simultaneous analysis of both quantitative strain-energy calculations and ‘Hand I3Cn.m.r. spectra was carried out.22 The molecule exhibits many of the characteristics of strained molecules such as long C-C single bonds. In a series of nine papers,23 Dunitz and Winkler have described studies on amide deformations in medium (7-12) ringlactams,and Iactam salts.The results compare quite well with those from force-field calculations on cyclo-olefins. An attempt was also made with some success to derive the form of the potential energy surface for amide-group out-of -plane deformations. Similar deformations but in carbonyl groups which arise out of short N * * C=O and 0 * * C=O contacts have also been and the correlation between the short contacts and the out-of-plane deviations is discussed in terms of incipient chemical reactions involving addition of a nucleophile to a carbonyl group and the reverse breakdown of the tetrahedral species produced. As in previous years many structural studies of biologically active molecules have been reported and of particular interest are studies from which details of the intermolecdar interactions provide some insight into the structure-activity relation- ships.Interactions between drugs possessing planar aromatic groupings and nucleic acid are of particular current interest. Two structures which model such systems have been reported with dinucleotide monophosphates mimicking the natural nucleic acids. In the complex between 9-aminoacridine and adenylyb3’,5‘-uridine the drug molecules stack between non-Watson-Crick base-pair~.~’ However in the ethidium-bromide-uridylyl-3‘,5’-adenosine the drug molecules interca- ~ompIex,*~ late between the Watson-Crick base-pairs of an unwound RNA double-helical fragment. Also in the biological field the first structure analysis of a phospholipid the synthetic compound 1,Z-dilauryl-D-p hosphatidylethanolamine (5) has been de~cribed.~’ The analysis has provided much detailed information relevant to the 22 E.Maverick S. Smith L. Kozerski F. A. L.Anet and K. N. Trueblood Acfa Cryst. 1975 B31,805. 23 J. D. Dunitz and F. K.Winkler Acta Crysf. 1975 B31,251 et seq. 24 H.B. Burgi J. D. Dunitz and E. Shefter Acra Cryst. 1974 BM,1517. 25 N. C. Seeman R.0.Day and A. Rich Nature 1975,253,324. z6 C. C. Tsai S. C. Jain and H.M. Sobell Roc. Nat. Acad. Sci. U.S.A. 1975,72,628. 2’ P.B.Hitchcock R. Mason K.M. Thomas and G. G. ShipIey,Roc. Nat. Sci. U.S.A.,1974,71,3036. M. B.Hursthouseand S. Neidle ' 0 II CH2 -O-C-(CH2) o-CH2 CH,-O-P-O-(CH2)2 -NH3+ I 0- (5) packing and conformation of lipids in both artificial and natural membranes; in particular the observed packing of phospholipid molecules is exactly that of the classic bilayer model.Finally mention is made of a group of three papersz8 describing a study of the reaction of solid crystalline aromatic carboxylic acids and related compounds with ammonia and amines. Generally the reaction proceeds anisotropically and the reactivities along the principal directions through crystals have been correlated with the crystal structures of the substrates in particular the arrangement of the reactive functions. 28 R.S. Miller D. Y.Curtin and I. C. Paul J. Amer. Chem. Soc. 1974,% 6329 6334,6340.