Jan. 20, 1960
COMMUNICATIONS TO THE EDITOR
505
+
the 16~-bromo-17~~,20:20,21-bismethylenedioxy-A~111” (c 3.1),4 and VI1 to 1GP-fluorocortexolone, pregnene-3-one (I, 16@-bromocortexoloneBMD), m.p. 178-179”, [ a I z 7 D $89”, (C 1.0).4 m.p. 208-209’ dec., [cY:]’~D- 45’ (c 2.0).4 Additional reactions of some of these products including the preparation of l60-fluoro and 160methoxy derivatives related to cortisol will be the subject of forthcoming publications. RESEARCH LABOXATORIES WALTERT. MORELAKD CHAS.PFIZERA N D Co., INC. RUDOLPH G. BERG P. CAMERON GROTON,CONNECTICUT DONALD RECEIVEDKOVEMBER 21, 1959
Excess silver acetate in refluxing glacial acetic acid in the presence of sodium acetate smoothly converted I to 16P-acetoxycortexolone BMD (II), m.p. 197-198’, [ C Y ] ~ ~ D 4.5’ (c 3.6).4 Similarly, silver perchlorate in aqueous acetone gave the 160-hydroxy derivative (111), m.p. 231-233.8’, [ o c ] * ~ D - 12.5’ (c 3.8).4 Anhydrous methanolic silver perchlorate analogously formed 1GP-methoxycortexolone BMD (IV), map. 173-175, 184-187.(i0, [OCIz6D 5.9’ (c 1.0).4 Treatment of I with silver fluoride in 2-propanol gave, in addition to a minor amount of the isopropoxy compound (V), m.p. 149-152°, [ c Y ] ’ ~ D 5.9’ (c a 75% yield of 16@-fluorocortexoloneBMD (VI), m.p. 228-229’, [aIz6D 5.9’ (C Brief treatment of the acetate I1 with refluxing 607, formic acid3 and then by methanolic sulfuric acid to cleave residual formates and acetylation of the resulting mono-acetate gave a major product which was identical in all respects with an authentic sample of 16/3-hydroxycortexolone 16,21-diacetate.6n6s7 The alcohol I11 could be converted to the acetate I1 with acetic anhydride in pyridine a t room temperature ; similarly, methylation of the alcohol I11 gave the methyl ether IV. Thus unequivocal evidence is provided for the structures of 11, I11 and IV and, in view of the similarity in preparation, it is assumed that V and VI are analogous. It had been anticipated that the @orientation of the leaving bromine coupled with the general tendency for attacking species to approach the rear of the steroid molecules would strongly favor inversion of configuration a t C-16 leading to 16a-substituents. These results clearly show that this was not the case and that the displacements occurred predominantly with net retention of configuration, possibly as a result of interaction between one of the ether oxygens in the side-chain and C-16. Using the conditions for removal of the side-chain protection noted above for 11, IV was converted to 1B0-rnethoxycortexolone1m.p. 149.4-150°, [a]2 6 ~
+
+
+
+
(4) Satisfactory analyses have been obtained for all compounds herein described. Rotations are in dioxane. T h e infrared spectra (KBr) for I-VI are all characterized by absorption near 5.98, 6.16, 9.15, 10.10, 10.87 and 11.58 p ; infrared and ultraviolet absorption for all compounds is consistent with structures assigned. (6) K. Heusler and A. Wettstein, Chern. Ber., 87, 1301 (1954); c j . B. Ellis, F. Hartley, V. Petrow and D. Wedlake, J . Chem. SOL.,4383 (1055); J. Rorno and A. R . De Vivar, J . Org. Chem., 21, 902 (1956). (6) We are indebted to Dr. Seymour Bernstein fur samples of 10aand 168-hydroxycortexolune 16,2l-diacetates a n d 10a-hydroxycortexoione. (7) lfia-IIydroxycortexolone 16,21-diacetate was unchanged when suhjected t o t h e same sequence of reactions. (8) Ref. 1, p. 14.
THERMAL REARRANGEMENT OF FEIST’S ESTER. A NEW TYPE OF INTERMEDIATE
Sir: The recent structure proof’ of the pyrolysis product I of Feist’s ester (IIa) has brought into question the mechanism of this remarkable rearrangement. Two routes appear possible: viz., (a) rotation of both ester-bearing carbons with scission of their common bond to give the resonance-stabilized “zwitterion” 111, and then a related transformation of I11 to product, or (b) a “valence-tautomeriIIa, R = H; R’ = COOCH3 IIIa, R IIb, R = COOCHa; R’ = H IIIb, R
,COOCHq
=
H ; R’ = COOCH COOCH; R’ = H
CH300C,
,H
%
H+R H
I H
K’
\R‘ Ia, R = H ; R’ = COOCHa Ib, R = COOCHB; R’ = H
H
I1
I\’, R = - C - or 0
zation” involving simultaneous cleavage and reformation of the ring bonds with concomitant rotation of one ester-bearing carbon and the terixinal methylene to give I directly. Pathway a is of particular interest in view of LCAO-RIO calculations suggesting considerable resonance stabilization in the related intermediates IV.2 Evidence supporting intermediate I11 in the pyrolysis reaction is given. I t is seen that rearrangement of optically active trans-ester I I a via the planar intermediate (path a) would lead to racemic product, while rearrange(1) E. F. Ullman, THISJOURNAL. 81, 5386 (1920). ( 2 ) J. G.Burr, Jr., and M. J. S. Dewar, J. Chem. Soc., 1201 (19S4).
506
COMMUNICATIONS TO THE EDITOR
Vol. 82
ment through path b should lead to retention of of the possible relationship of I to the tripyrrylconfiguration (IIa, starred atom). Accordingly, py- methane type of intermediate in the biosynthesis of rolyses of Feist’s ester (IIa), [aIz5o- 119°,3were p ~ r p h y r i n . ~We now present new evidence which carried to roughly 25 and 50% completion. The makes I untenable, and which shows that prodichromatographically separated products were Ia, giosin is 11, or 111. [CY]~’D4-28’, +loo6; Ib4, [ a ] “ +18’, ~ +So; and IIa, -93O, -34’, respectively. No cisFeist’s ester (IIb) was detected. The isolation of active product provides convincing evidence for the “valence-tautomerization” pathway (b), but the observed racemization of starting material requires additional comment. Two racemization mechanisms not involving the “zwitterion” I11 appear possible. (1) The transester (Ha) might equilibrate rapidly with product which has partially racemized either by equilibration with the internally-compensated cis-ester (IIb) or by some other unspecified route. (2.) The trans-ester might equilibrate directly with its cisi-f‘ isomer. IV The first alternative is untenable because the N H experimental conditions effectively inhibit reconversion of product to IIa or I I b by virtue of the demonstrably favorable equilibria between the product and these esters.6 The second was tested by partial pyrolysis of the cis-ester’ (IIb) using conJXTredeisolated 2-methyl-3-amylpyrrole, ClaHlin’, ditions under which the tram-isomer was substan(-1) from prodigio~in,~ while Santer and Voge15 tially unaffected.3 The resulting inixture was shown by n.m.r. to be pyrolysis product (I), trans-ester showed recently that a mutant strain (9-3-3) of (IIa) and cis-ester (IIb) in the approximate ratios Serratia accumulated a CloHloOzh’zprecursor (B) 75: 10: 15. Thus the cis-ester (IIb) rearranges which could be converted to prodigiosin by another about 7.5 times faster than it isomerizes to the mutant (W-l).* In a formal way, the condensatrans-ester (IIa) and, contrary to fact, racemization tion of these two C-10 fragments leads to prodigioCloHloOzN2 + C20H250N3 HaO. of I I a via I I b could occulr only less than 1/7.5 sin, CIoH1,N Lye have now, in fact, shown that pure A and B retimes the rate of product appearance. It is thus concluded that racemization must occur act readily under conditions of dipyrrylmethene by reversible equilibration of trans-ester with the synthesis to form prodigiosin, indistinguishable “zwitterion” IIIa (and/or IIIb) and that rear- (infrared spectra of zinc salt and hydrochloride) rangement proceeds competitively by both mech- from the natural material. The n.m.r. spectrum of U (60 111.c. in dimethyl anisms a and b. sulfoxide) contains two broad peaks a t T = -1.35 The author thanks JV.voii E. Doering and A. S. and 1.52 (two pyrrolic XH protons) and a sharp Kende for helpful suggestions and J. E. Lancaster singlet a t T = 0.57 (aldehyde hydrogen). The for the n.m.r. spectra. a,a’-linkage of the pyrrole rings is shown by the (3) Pyrolyses were carried o u t i n 17; b e n z e n e solutions in a bomb isolation of pyrrolecarboxamide from the alkalinepartially immersed in a 210’ bath for intervals of 30 and 45 minutes peroxide oxidation of both B and prodigiosin. for I I a and 23 minutes for 111,. The above evidence, and other properties of B5 (4) Stereochemistry nut implied. ( a l T h e + I O ” sample wits coiitaniinuted niL!i Ib. (one OCH3, ultraviolet absorption a t XlCz2’3 254 (6) An equilibrium ratio of 1 I a : I I b : I was found by n.m.r. analysis m p , E 13000; 363 nip, E 33000), now can be acconitii be ( j : < 0 . 3 : 0 4 (6. ref. 1). iiiodated by expression IIr. This can be refined fur(7) Cf,AI G. E l l l i n g e r and 17. Kennedy. C h e f i i i s t i y & I?idu,tiy, 891 ther : (i) the strong ultraviolet absorption requircs L 1937). conjugation of both pyrrnle rings with the formyl ORGANIC CmmCar, RESEARCII Sscrro~ group and excludes ally structure with -CHO a t I,EDERLE LABORATORIES DIVISION EDWIS1’. ULLM.4N AMERICANCYANAMID COMPAXY C-3. (ii) In view of the isolation of methoxyirialePEARLRIVER,h7mv YORK imide from prodigiosin7 the methoxyl must be a t RECEIVED DECEMBER 3, 1959 C-3 or C-4. (iii) In the 1i.m.r. spectrum of B, the doublet a t T = 3.69, which changes to a singlet in base, corresponds to the lone C-H proton on ring PRODIGIOSIN. STRUCTURE AND PARTIAL A (IV). Were this proton located a t (2-3, adjacent SYNTHESIS’ to CHO a t C-2, it would be expected to show Sir : resonance closer to T = 2.7-2.8, as in 2-pyrroleProdigiosin, the blood-red pigment2 of Serratzk aldehyde. B thus contains OCH3a t (2-3, and CHO marcescens, assigned the tripyrrylmethene structure D. Shemin, C. S . Russel and T. Abramsky, J . B i d . Chenz., 215, I by Wrede,3 has attracted recent interest because 613(4)(1955).
c
+
( I ) This investigation was supported by Grants P-64 and P-64A from t h e American Cancer Society. (2) E. N. Morgan and E. h.1. Tanner, J. Chem. S O L . 3805 , (1955). (3) F. Wrede and A. Rothhaas, Z . phrsiol. Chem., 226, 95 ( 1 9 3 4 ) .
0
+
(5) U. V. Santer and H. J . Vo!3l. Biochim. Bi (1956). (6) hf. T.hl. Rizki, P r o c . S u l l . A t a d . Sca., 40, 1057 (1954). (7) F. Wrede a n d A. Rothhaas. Z . physiol. Chetn., 215, 67 (1933).
