COMMUNICATIONS TO THE EDITOR
5368
Vol. 86
complex was treated with anhydrous HC1 to give transchloro - p -fluorophenylbis (triethylphosphine) platinum(11), m.p. 103-104". The other halogen and pseudohalogen complexes (2: X = Br, I, SCK, and CN) were prepared by simple metathetical reactions of the chloro compound. The aryl and methyl compounds (2: X = C6Hs, m-FC6H4, and CH3) were obtained by reaction of the appropriate Grignard reagents with the chloride. The m-fluorophenylplatinum compounds were prepared by similar procedures. The trans configurations of all the compounds studied were confirmed by the proton magnetic resonance technique of Jenkins and Shaw.' The CH3 resonance of the triethylphosphine ligands appeared as a characteristic five-line structure ( J = 7.S + 0.3 c.P.s.) owing to coupling with the CH2 protons and with the two trans P31nuclei equivalently.
hydrochloride and the hydrobromide in the present example).j The partial structure (22 atoms plus bromine) was then augmented, in steps, using more conventional techniques until 58 atoms (excluding hydrogen) were finally fixed. The two salts, crystallized from dioxane-water, were found to be orthorhombic, with space group P212121, and their cell dimensions agreed within 0.20.3%. The density of the hydrobromide gave an Xray molecular weight of 886 which, from elemental analysis, placed the asymmetric unit a t C40-441H4:-63T\;&-&-. Considering the empirical formula favored for the antibioticj2the high carbon and oxygen content suggested that dioxane and possibly water of solvation were present in the crystals. The positions of the chloride and bromide ions were determined from three principal difference Patterson projections employing coefficients of the type ( F i Br( 7 ) J . hI. Jenkins and B. L. Shaw, Proc. Chem. Soc., 273 (1963). G. W . PARSHALL ~ F ~ Cwhere COXTRIBUTIOS S o . 1021 I ) ~ ,IF B~ and ' F i c l are the structure ampliCENTRAL RESEARCH DEPARTMEST tudes of the hydrobromide and hydrochloride salts, EXPERIMENTAL STATION respectively. Their x-coordinate proved to be special E. I. D U PONTDE XEMOURS A N D COMPANY (x 0) and, consequently, generated an unL~ILMISGTOS, DELAWARE wanted mirror plane of symmetry a t x = in the RECEIVED OCTOBER 13, 1964 weighted electron density. The relative and absolute scales of the two isomorphs, requisite to the weighted isomorphous replacement method, were first obtained The Structure of Isoquinocycline A. An X-Ray by Wilson's method and then verified and improved Crystallographic Determination' upon from two-dimensional structure factor and elecSir: tron density considerations. The zero contribution Recently, a partial structure has been proposed for of the replaceable electrons to certain classes of twoisoquinocycline A based on spectroscopic, degradative, dimensional reflections and a comparison of peak and chemical studies.2 Two major and several heights (excluding bromide and chloride) in projected minor features of the antibiotic have been established electron densities served to obtain a self-consistent or suggested. These results are summarized as follows relative scale which was then fixed on an absolute basis by comparing the observed projected peak heights and peak shapes of the ions with those based on computation, systematically varying an isotropic thermal parameter for the ions. The weights for the observed structure amplitudes, functions of IF B r , F ~ 1 and , the OH replacable electron contribution, were computed but they proved to be, in general, inapplicable in a direct manner. A very large number of them had an absolute value greater than unity and these usually ranged between *lo. Such behavior was assumed to be the result of observational errors in intensity so their effect on the weight computation was assessed by approximating the standard error of the weight in terms of the estimated standard error of the structure amplitudes. Then, somewhat arbitrary but rather This communication presents the complete structure stringent criteria were employed to select acceptable of the molecule and its relative stereochemistry as weights. These were used to compute a weighted determined by the X-ray method.3 electron density: The density contained 146 peaks X satisfactory partial structure for the antibiotic greater than 1 eA.? in the asymmetric unit, of which was obtained through the use of the weighted isomorabout 73 were related by mirror symmetry. A 22phous replacement m e t h ~ d . In ~ this way, it was atom partial structure (the tetracyclic nucleus and its demonstrated that structures can be solved in the hydroxyl and carbonyl substituents) was easily derived noncentrosyrnmetrical case employing isomorphous reand it served as a starting point for the use of more placement procedures with the information that is conventional methods to obtain the remainder of the available from only h o isomorphous derivatives (the structure. (1) This research has been supported by the National Institutes of Health, U. S. Public Health Service, and Lederle Laboratories Division, American The structure of isoquinocycline -4 is shown above, Cyanamid C o . from which most of the stereochemistry is self-evident. ( 2 ) D. B. Cosulich. J. H. Mowat, R . W. Broschard. J. B. Patrick, and The antibiotic consists of five fused rings joined to a \T. E Xeyer, T e l v a h e d r o n L e f f a r s , No, 7 , 453 (1963); i b i d . , No. 13, 750 (1964). pyrrolopyrrole liiu a spiro atom and also to a sugar-like
-
n
( 3 ) .5 preliminary account oi this work was presented a t the Annual American Crystallographic Association lleeting held in March, 1363, a t the Llassachusetts Institute oi Technology, Cambridge, Mass. (4) G. K a r t h a , Acta C v y s t . , 14, 680 (1361).
(5) T h e author wishes t o thank the Organic Chemical Research Section, Lederle Laboratories Division, American Cyanamid Co., for supplying these sa m p 1e 9.
COMMUNICATIONS TO THE EDITOR
Dec. 5, 1964
5369
be the 1,2-isomer by reaction with phenyl isocyanate followed by hydrolysis to give hydrazine-N,N'dicarboxylanilide, m.p. 245-248", lit. m . p . 245O.lh T h a t it was free of 1,l-isomer was shown by the n.m.r. spectrum which has a single peak in the Si-C-H region a t r 9.92. I n a typical experiment 4.1 g. (0.023 mole) of 1 , 2 bis(trimethylsily1) hydrazine, prepared by the method of Wannagat and Liehr,'"Zb was treated with 0.048 mole of n-butyllithium in hexane a t 0". After being stirred for about 30 min. to ensure complete reaction, the mixture was chilled to --70°, and 25 ml. of freshly HO OH distilled tetrahydrofuran was added, causing the preC33HaN20io .HX viously white suspension to become light yellow. Methyl iodide (0.48 mole) was then slowly injected. moiety through a glycosidic link, with the pyrrolopyrrole The mixture was then allowed to warm up slowly with nitrogen atoms away from the aromatic nucleus and stirring, filtered in a drybox, and fractionally distilled the furan-pyrrolopyrrole system trans to the sugar. yielding 2.8 g. (60%) of pure colorless product, b.p. The pyrrolopyrrole is planar within *0.04 A. and contains one carbon-carbon double bond. Both 67-71' (17 Torr). Anal. Calcd. for C8H&i2K2: C, 46.99; H , 11.83; K j 13.70. Found: C, 47.33; nitrogen-to-bridge-carbon distances are equal and H, 11.97: N, 13.62. The proton n.m.r. spectrum of the shorter than expected for a carbon-nitrogen single bond (1.33 and 1.31 f 0.03 A,). The two secondary product shows two peaks a t r 9.98 and 9.89, assigned hydroxyl groups of the sugar are hydrogen bonded with to trimethylsilyl protons, and two peaks a t r 7.43 dioxane molecules of solvation. The latter occupy and 7.37, assigned to the methyl protons in the two relatively large cavities in the solid between antiisomers. Relative peak areas are approximately 3 : 3 : biotic molecules and probably assist in maintaining the 1 : 1, respectively, indicating that the two isomers must orientation of the sugar moiety. be present in nearly equal amounts. Partial separation of the mixture could be effected by fractional The details of this structure determination will be distillation. submitted to Acta Crystallographzca. The mixture was solvolyzed with 1-propanol t o STERLING CHEMISTRY LABORATORY A . TULINSKY produce the mixed dimethylhydrazines which were YALEUNIVERSITY SEW HAVE?;,CONNECTICUT treated with p-nitrobenzaldehyde yielding p-nitroRECEIVED OCTOBER15, 1964 benzaldehyde dimethylhydrazone, m.p. 111-1 12, lit.3 m.p. 111') infrared spectrum identical with that of authentic p-nitrobenzaldehyde dimethylhydrazone. A Novel Rearrangement of Isolation of this known hydrazone confirms the pres1,2-Bis(trimethylsily1) hydrazine ence of the rearranged hydrazine in the original reSir: action mixture. I n extending the work of Wannagat and co-workers The rearrangement reported here, though not on organosilylhydrazines,l we have found a novel previously noted, may have occurred in previous rearrangement involving migration of silicon from one reactions of silyl-substituted hydrazines with bases. nitrogen atom to another. If 1,2-bis(trimethylsilyl)- Thus, in the reaction of 1-phenyl-2-trimethylsilylhyhydrazine is treated with 2 equiv. of n-butyllithium in drazine with phenyllithium followed by bromine, a tetrahydrofuran and then with 2 equiv. of methyl large amount of N-trimethylsilylaniline was observed iodide, the reaction mixture is found t o contain both as a b y l p r ~ d u c tperhaps ,~ arising via a 1,2-silicon shift the expected product, 1,2-bis(trimethylsilyl) 1,2-diof the type described above. ,Many reactions of lithmethylhydrazine (I), and the rearranged product, ium derivatives of silylhydrazines have been reported l,l-bis(trimethylsilyl)-2,2-dimethylhydrazine (11). in which structures of products are assigned without considering the possibility of rearrangementle,g ; some 1 n-CHsLi of these structures may now have to be reconsidered (CHa)rSiNH--NHSi( CH,)r 2 CHDI in the light of our evidence. At present i t is not ( CH3),Si ( CHdSi CHa known whether the rearrangement occurs during or \ / \ N-N after treatment with n-butyllithium, or even conceivN-x / \ / \ ably during the reaction with methyl iodide. InCH, CHI CHI (CH3)aSi vestigations of these possibilities are under way, and I1 I studies of the limits and of the reaction in general are Pure 1,2-bis(trimethylsilyl)hydrazine, b.p. 1 4 7 O , planned for the near future. nz5D 1.4253, d2440.834, was prepared by the method of (2) The material originally supposed by Wannagat a n d Lieht-" t o be Wannagat and Liehr.1a'ba2 This material was shown to 1,2-bis(trimethylsilyl)hydrazinem a y have been a m i x t u r e of t h e 1 . 2 - a n d
7iCHa)s +
(1) (a) U. Wannagat and W. Liehr, Angew. Chem., 69, 783 (1957); ( b ) U. Wannagat a n d W . Liehr, 2. anorg. allgem. Chem.. 297, 133 11958); (c) U. Wannagat and H. Niederprum, A n g e w C h e m . , 70, 745 (1988); (d) U. W a n n a g a t a n d W. Liehr, Z . a n o r g . allgem. Chem., 299, 311 (1959), (e) U . Wannagat a n d H . Xiederprum, i b i d . , 310, 32 (19611, (1) H. Niederprum and U. W a n n a g a t , ibrd., 311, 270'(196l); (g) U. W a n n a g a t , C . Krueger, and H. Niederprum. i b i d . , 311, 198 (1963): (h) U. Wannagat a n d C. Krueger, Monalsh. C h e m . , 94, 63 (1963).
1,l-isomers (private communication from Professor U W a n n a g a t ) . T h e proportions of the t w o isomers formed in the reaction of trimethylchlorosilane w i t h h y d r a z i n e d e p e n d s o n t h e reaction conditions i n a w a y n o t yet f u l l y u n d e r s t o o d . When p r e s e n t , t h e l , l - i s o m e r c a n easily be d e t e c t e d h y n.m.r. spectroscopy (Si-C-H resonance at 7 9.99). ( 3 ) 0. L . Btady a n d G P. h I c H u g h , J . Chem. Soc., 121, 1618 11922) (4) U . Wannagat a n d C K r u e g e r , Z . a n o v g . n l i E e m ChPm , 326, 288 ( I 964).
