958
COMMUNICATIONS TO THE EDITOR
tion was characterized not by being the best a t the early stages of the study, b u t rather by showing a very slow increase in R as the amount of data was increased. Further studies are planned as to the generality of this behavior. The unusual features of the phytolaccagenin structure are its high degree of oxidation and the appearance of a carbomethoxy group. Other triterpene esters have been i ~ o l a t e d ,but ~ their formation has been attributed to esterification during treatment with acidic alcohol. Since phytolaccagenin can be prepared without any exposure to methanol its ester function is apparently of natural origin. Phytolaccagenin represents one of the largest natural products to be solved without the use of a heavy atom, and its solution exemplifies the growing possibilities for the use of chemical information in the early stages of an X-ray structural analysis. Acknowledgments.-This work was supported in part by funds from the U. S.Public Health Service, Grant CA-04087, and by generous gifts of computer time from the Research Computer Laboratory, University of Washington. We also wish to thank Prof. C. E. Nordman for making available to us his model rotation program.
Vol. 86 TABLEI CRYSTALLIZATIOS DATA
Sample
Specific activity with resp. to SH 35s in c p m /mg
1st Crystallization Mother liquor 2nd Crystallization Mother liquor 3rd Crystallization Mother liquor
1134 970 1154 1090 1141 1155
212 174 210 200 204 213
3H-"S ratio
5.4 5.6 5.5 5.5 5.6 5.4
5 : 5 ) >which separates the monosulfates from other urinary conjugates. The sulfate fraction was distributed between 0.3 Jd pyridinium sulfate solution and chloroform.6 Evaporation of the CHC13 afforded the pyridinium salts of the steroid sulfates (264,000 c.p.m. 3H and 40,000 c.p.m. 35S)which were purified further by partition chromatography on Celite using the system : isooctane-chloroform-n-butyl alcohol-methanol-0 3 M pyridinium sulfate-pyridine (4 : 2 : 0.7 : 2 : 2 : 0.1). A t least two radioactive metabolites were eluted, the more polar of which was rechromatographed on Celite using the system : isooctane-ethyl acetate-n-butyl alcohol-methanol-1 M N H 4 0 H (2 : 4 : 1 : 2 : 3 ) . One crystalline radioactive product was eluted in the (5) C Djerassi and J. S . Mills, J . A m . Chem. Soc., 8 0 , 1236 (19.581, W. A . third to fifth hold-back volume, and this was identified Jacobs and 0.Isler. J. Bioi. Chem , 119, 155 (1937). (6) National Science Foundation Cooperative Fellow, 1960-1961 by its infrared spectrum as the ammonium salt of DS. DEPARTMENT OF CHEMISTRY GEORGE H . STOUT It weighed S7 mg. and contained 92,900 c.p.m. 3H UNIVERSITYOF WASHINGTON BERNARDM. MALOFSKY~ and 17,10035S. The product was recrystallized from 5, WASHIXGTON VIRCISIA F. STOUT a mixture of hleOH and acetone to constant specific SEATTLE RECEIVED DECEMBER 5, 1963 activity with respect to both 3H and 35Sin both crystals and mother liquor residues during three successive crystallizations The data are presented in Table I , Evidence that Cholesterol Sulfate is a Precursor of where the 3H-35Sratios are also shown. The latter values were essentially identical with the ratio present Steroid Hormones in the injected CS. I n addition, the specific activities Sir: with respect to both isotopes remained constant followSince the initial observation' that conjugation of ing conversion of the DS to its oxime and two crystalsteroids does not necessarily destine the hormone for lizations of that derivative from methanol-ether. The excretion, two important findings have supported the yield of radioactivity associated with DS in the urine of idea that intact steroid sulfates may serve as hiosynthe first day was approximately 0 . 5 ~ ~ . thetic intermediates. In a recent study, Calvin, et a1.,2 found that urinary dehydroisoandrosterone sulfate SCHEME I (DS)may be synthesized in vivo from pregnenolone PROPOSED BIOSYSTHETIC PATHWAY INVOLVING sulfate (PS) by a pathway involving intact steroid SCLFATES AS ISTERMEDIATES STEROID sulfates as intermediates. They have also shown3 CH3 t h a t a homogenate of hyperplastic tissue is able to hydroxylate PS a t C-17. To extend these findings, c h ~ l e s t e r o l - i a - ~sulfateH l 35S(CS) (24.9 X lo6 C.P.S. 3H and 4.35 X 106 c.p.m. 35S 3H--35S ratio = 5.7) was injected into the left s p l k i c artery of a female subject with an inoperable -0sso adrenal carcinoma. This vessel supplied 90yo of the blood going to the neoplasm. The CS was injected Pregnenolone Sulfate (PS) Cholesterol Sulfate (CS) through a small catheter introduced into the artery under radiographic visualization prior to the selective 1 perfusion of the neoplasni with an antimetabolite. The conjugates from the first 24-hr. urine collection CH, were extracted into ethanol-ether ( 3 :l ) . 4 After I evaporation of the organic solvents, the remaining aqueous phase was extracted with n-butyl alcohol. The dry residue left after removal of the butyl alcohol contained 2.5y6 of the injected radioactivity. I t was purified by partition chromatography on Celite using the -080 -03s isooctane-t-butyl alcohol-1 31 NHlOH ( 2 : systenl" :
w -w
&
(1) K . D. Roberts, R . L. VandeWiele, and S Lieberman. J . B m l . C h e m . , 236, 2213 (1961). (2) H I Calvin, R L L'andeWiele, and S Lieberman, Riochem, 2 , 648 (1963). (3) II. I. Calvin and S Lieberman, z b d , 3 , 2.59 (1964). (4) I< U' J Edwards, A . E Kelly, and A P. Wade, Mem. Soc Endocrinol , 2 , 53 (1862)
&&
Dehydroisoandrosterone
17-hydroxy pregnenolone
Sulfate (DS)
Sulfate
( 5 ) P. K . Siiteri P h D . Dissertation, Columbia University, 1963. (6) J McKenna and J K Norymberski, B i o c h e p . J , 7 6 , Ix (1960)
March 5 , 19G4
COMMUNICATIONS TO THE EDITOR
959
distinguishcs them from these other radicals. One of McConnell and Robertson’s generalizations6 originally applied to aromatic radicals, i . e . , the g value of a radical increases with the spin orbit coupling parameter, j-, of the atoms that have significant spin density, applies also t o this case. To remove some spin density from the terminal oxygen atom and distribute i t to any extent on the R group requires the polarization of the 0-0 bond so ( 7 ) IZ 1- VandeU’iele, P C. hiacDonald, E. Gurpide, and S Lieberman, as to make the terminal oxygen more negatively Recent Progv. Hovmone R e s . . 19, 275 (1963) ( 8 j Predoctoral Research Fellow, National Institutes of Health charged. For this reason radicals with electron-with(9) Career Scientist. Health Research Council, City oi New York drawing groups in R , such as the radical of Lontz, KENNETH D. ROBERTS should have the greatest spin density on the terminal DEPARTMENTS OF BIOCHEMISTRY AND OBSTETRICS ASD GYNECOLOGY LAJOSBASDI oxygen atom and therefore the highest g value. In COLUMBIA UNIVERSITY HAROLD I . CALVIN^ solution this radical would have the averaged g value D E P A RM~E N T OF MEDICINE WILLIAM D. DRUCKERg A-E‘~vYORKUSIVERSITY SEYMOUR LIEBERMAN of 2.0187. SEIV YORK,XE\VYORK Our microwave frequency was measured by a HewRECEIVED J A N U A R Y 6, 1964 lett-Packard 524C counter with a Hewlett-Packard 540B transfer oscillator. The magnetic field was determined by proton magnetic resonance using a marginal oscillator whose frequency was measured on the same Identification of Peroxy Radicals by Electron 524C counter. The line broadening effect of the presParamagnetic Resonance ence of molecular oxygen was a principal source of Sir : error in the g value measurement. In e.p.r. studies free radicals are identified by their ( 6 ) H M McConnell and R . E. Robertson, J . P h y s . Chem., 61, 1018 characteristic hyperfine splittings, their g values (line (19571, ( 7 ) University of Toronto, Toronto, Canada. positions), and their relaxation times in a particular milieu. The last parameter is the least specific and CALIFORNIA RESEARCH CORPORATION MALCOLM BERSOHN’ the first parameter gives the most information. Peroxy RICHMOND, CALIFORNIA J . R. THOMAS RECEIVED JANUARY 8, 1964 radicals, ROz,, have no hyperfine structure detectable by present methods except that due to O17. Because of the difficulty of 0 1 7 labeling only one experiment on Studies of Reactions of Atomic Carbon by a Double the OI7 hyperfine splittings of free radicals has been reported,’ so that method is not yet broadly applicable. Tracer Technique We wish t o report evidence that the g values of perSir : oxy radicals easily distinguish them from RO. and Reaction mechanisms of atomic carbon have recently R.. We have measured the g value of the cumyl been studied using C” (20.5 min.) as produced by peroxy radical which is 2.()155 f 0.0005. Lontz2 nuclear transformation. 1-4 The C’l, being radioactive, found that the radical CF3CFCONH2 reacts in the serves as its own tracer. However, it would also be solid state with molecular oxygen to form a free radical greatly desirable t o trace the origin of the atoms without noticeable hyperfine splitting and with g with which the C” combines. Because the nuclear values measured along the three principal axes of 2.0102, recoil technique provides only a small number of carbon 2.0193, and 2.026’7. He assumed this is the correspondatoms ( ~ 1 0 and ~ ) hence of product molecules, this has ing peroxy radical CF3CFCONH2. Fessenden and not been practical thus far. Labeling the molecules with which the carbon reacts with another radioisotope 02 is not feasible because the amount of such a secondary Schuler3 found that alkanes irradiated with fast label which could be incorporated in the product is electrons in the absence of oxygen formed hydrogen too small to d e t e ~ t . ~ atoms plus readily identifiable alkyl radicals. In the The use of stable isotopes, such as deuterium, as presence of oxygen under the same conditions radicals secondary tracers has faltered on the formidable probwithout noticeable hyperfine splitting were produced. lem of product analysis. Mass spectrometric separaThe g values of these latter radicals ranged from 2.0155 tion of isotopic deuterated molecules followed by when pentane was used to 2.0148 for paraffin oil, the counting of their Cll content is possible in principle, g value decreasing with increasing molecular weight but very difficult in practice.6 The alternative, of the parent alkane. These were also presumed to gas chromatographic separation of isotopic molecules be peroxy radicals. followed by C” radioassay, has been restricted by the Our cumyl peroxy radicals were produced in benzene unavailability of columns of sufficient resolving power. solution as a result of the reaction of cumene with With the recent publication by Cvetanovic, et al. , 7 oxygen in the presence of azobisisobutyronitrile. N o of a technique for separation of deuterated light olefins other radicals were detected in this system by e.p.r. Evidence that the principal radical formed in this (1) C. M a c K a y , et al.. J . A m . Chem. So
