THE JOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 246, No. 8, Issue of April 25, pp. 2484 2501, 1071

Printed in U.S.A.

Mechanism of Squalene Cyclization

THE BIOSYNTHESIS OF FUSIDIC ACID*

(Received for publication, November 2, 1970)

L. J. MULHEIRN$ AND ELIAHU CASPI

From the Worcester Foundation for Experimental Biology, Xhrewsbury, Massachusetts 01545

SUMMARY

It is generally accepted that the C-20 cation or its stabi- lized equivalent is an intermediate in the biosynthesis of certain triterpenes and sterols. To evaluate this hypothesis the biosynthesis of the protosterol fusidic acid was investi- gated.

Fusidic acid was biosynthesized by incubating the fungus Fusidium coccineum on a medium containing (3RS,4R)- (2-14C,4-3H)-meva10nic acid (MVA). The obtained fusidic acid was shown to contain 6 14C atoms and 4 tritium atoms. It was proven that tritium atoms were located at the critically important 90 and 13a! positions. Evidence was also adduced for the location of the remaining 2 tritium atoms at the Sa! and 24 positions. The results show that the tritium atoms derived from 4-pro-R position of MVA are located at the theoretically predicted positions. Also this provides sup- port for the view that fusidic acid is biosynthesized by stabilization of the C-20 cation (protosterol numbering) or its stabilized equivalent by elimination of the C-17 proton with- out backbone rearrangement.

The retention of six 14C-atoms in the antibiotic confirms the view that in this fungus the transformation of the parent 4- gem dimethyl protosterol to fusidic acid involves the loss of the 4@methyl derived from 3’- of MVA. The absence of isotopic hydrogen at C-3, proven by us, indicates that the removal of the 4fi-methyl probably proceeds through a 3-keto intermediate.

The possible modes of stabilization of the C-20 cation and the formation of the C-17-C-20 double bond are discussed.

The reactions involved in the biosynthesis of sterols and t,riterpenes from mevalonic acid in plants and animals have been elucidated in considerable detail. In particular, it has been

*This work was supported by Grants AM 12156, HE 10566, and CA-K3-16614 from the National Institutes of Health, by Grants GB 5382 and GB 8277 from the National Science Founda- tion, by Grants P-500-g and P-500-h from the American Cancer Society, and by Grant 1377-C-1 from the Massachusetts Division of the American Cancer Society.

i Postdoctoral fellow 1968 to 1970. Present address, Shell Research Ltd., Milstead Laboratory of Chemical Enzymology, Sittingbourne, Kent, England.

shown (I) that conversion of MVA’ to squalene proceeds with stereospecific retention of all six 4-pro-R protons and loss of the 4-pro-S protons of MVA. Conversion of squalene to triterpenes can proceed either by proton-initiated cyclization (2, 3) or, more frequently, by conversion to squalene-2,3-oxide (4, 5)) which in turn undergoes cyclization.

According to present views the product of the cyclization is determined by the cyclase enzyme systems and by stereochemical interactions generated within the substrate (6). In the biosyn- thesis of cholesterol and the phytosterols squalene-2, a-oxide cyclization is thought to lead to the cation (Ia) (7, 8) or its stabilized equivalent (Ib) (9). Subsequent backbone rear- rangement of this intermediate could account for the forma- tion of lanosterol (II) cycloartenol (III) and other triterpenes. No direct evidence for this pathway has been obtained, al- though indirect evidence has been provided by examination of the elimination and migration of hydrogen atoms and methyl groups in the conversion of squalene to lanosterol and cho- lesterol mainly in rat liver (10-15). The results obtained are consistent with the initial formation of I followed by four con- certed 1,2-hydrogen and methyl migrations and terminating in the loss of a proton from C-9.

In order to provide more direct evidence for the role of I as an intermediate in the cylization of squalene-2,3-oxide we have examined the biosynthesis of fusidic acid2 (JVa), a member of the protosterol group of triterpenes produced by cert,ain fungi (16, 17). The structure of fusidic acid was considered (16) to be formed by direct stabilization of the cation (Ia) or its equivalent (Ib) without backbone rearrangement. We have tested the hypothesis by investigation of the pattern of incorporation of (3R,4R)-[2-14C ,4-3H]-MVA into fusidic acid (IVa) in Fusidium coccineum. This work has been the subject of a preliminary com- munication (18).

EXPERIMENTAL PROCEDURE

Materials-Samples of fusidic acid and F. coccineum were gifts from Dr. P. Diassi, Squibb Institute for Medical Research, New Brunswick, New Jersey, and Dr. W. 0. Godtfredsen, Leo Phar- maceutical Products, Ballerup, Denmark.

Physical Measurements-Melting points were taken on a hot

1 The abbreviations used are MVA, mevalonic acid; TLC, thin layer chromatography; NMR, nuclear magnetic resonance.

2 The rational nomenclature proposed for fusidic acid (15) is 3~,1l~-dihydroxy-16~-acetoxy-fusida-[17,20(16,21-cis); 24]-diene- 21-oic acid. The numbering system is shown in structure IV.

2494

by guest on February 9, 2020http://w

ww

.jbc.org/D

ownloaded from

Issue of April 25,197l L. J. Mulheirn and E. Caspi 2495

stage apparatus and are corrected. Infrared spectra were re- corded on a Perkin-Elmer 237 spectrophotometer as KBr discs. Proton magnetic resonance spectra were recorded on a Varian DA60 spectrometer at 60 MHz. Peaks are quoted in Hertz downfield from the tetramethylsilane internal standard. Mass spectra were measured on a Varian Associates M-66 instrument or on a Consolidated Electrodynamics Corporation 21-491 instru- ment.

(II)

(I) a. x = @ h. X = Nucleophilic group

(IV)a.R:O=R*=Rs=H b. R, = 111 = H R3 = CH, c. R, = R2 = &I&O, R, = CH, d. R, = CHICO, Rt = H, Rs = CH,

Ix T

u OOCH, COOCHI

(V) a. R, = a-OAc; P-H, Its = 0 b. R, = Rt = 0

Chromatography-Silica gel (Merck HF254+~,) was used for thin layer chromatography in the indicated solvent systems. Compounds were detected under ultraviolet light and by color reactions with phosphomolybdic acid. Radiochromatograms were scanned on a Vanguard automatic chromatogram scanner (model 885).

Counting-Samples were counted on a Nuclear-Chicago Mark I automatic liquid scintillation counter. The samples were dis- solved in 15 ml of a scintillation solution of toluene containing 4 g of 2,5-diphenyloxazole and 100 mg of 1,4-bis[2-(5-phenyl- oxazolyl)]benzene per 1000 ml.

Incubation of (SR ,4R)-[2-W,4-3H]-Mevalonic Acid with F. coccineum-A4 preliminary inoculum of F. cocci?zeum was obtained by incubation of the organism in four 500-ml Erlenmeyer flasks

each containing 100 ml of a medium composed of glucose (5 g per liter), yeast extract (2 g per liter), and micronutrients (1 ml) at pH 6.5. The flasks were shaken at 26” for 5 days.

A second medium (19) was prepared consisting of sucrose (60 g per liter), corn steep liquor (20 g per liter), KH2PO( (10 g per liter), MgS04.7 Hz0 (0.5 g per liter) at pH 6.5. Four l-liter flasks, each containing 250 ml of this medium, were sterilized and the medium was supplemented with (3R,4R)-[2-14C,4-3H]-meva- ionic acid dibenzylethylenediamine salt (50 &i of 1% distributed equally among the flasks). The preliminary inocula were then added and the cultures shaken for 7 days at 26”

The mycelium was filtered off and the liquors were acidified to pH 5.5 and continuously extracted with ethyl acetate for 12 hours. The extract was washed, dried, and, after the addition of fusidic acid, concentrated to a residue. The residue was frac- tionated by TLC (methanol-dichloromethane (1: 12))) and the area corresponding to fusidic acid was removed and extracted. One aliquot of this extract was further diluted with nonradioac- tive fusidic acid (IVa) (400 mg) and crystallized twice from ben- zene to give the solvate (276 mg), m.p. 189-191”. A further 60 mg of the solvate were obtained from the mother liquors.

3H4-%‘s-MethyZ Fusidate (ZV6)-Fusidic acid-benzene solvate (IVa) (20 mg) was methylated with diazomethane in ether- methanol. The product was purified twice by thin layer chroma- tography in systems ethyl acetate-benzene (1: 20) and ethyl ace- tate-benzene (1:7). The major band was extracted with ethyl acetate and crystallized from ether-hexane to constant 3H:W- ratio and r4C-specific activity giving IVb (15 mg) ; m.p. 151-153” (literature (20), 153.5-154”) (specific activity 3.95 x lo5 dpm of i4C! per mmole; 3H :W ratio 18.0).

3H4-14C6-Sa-Acetoxy Methyl F&date (ZVd)-Fusidic acid-ben- zene solvate (IVa) (40 mg) was diluted with nonradioactive (IVa) (60 mg) and esterified with diazomethane in ether-metha- nol. The crude product was acetylated by treatment with acetic anhydride (0.3 ml) and pyridine (0.3 ml) at 25” for 18 hours. The solution was poured into water and the product was extracted with ethyl acetate. The organic phase was washed with water and dried. After removal of solvents and drying under reduced pressure the product (IVd) (55 mg) was purified by repeated preparative TLC (ethyl acetate-benzene (1: 10)). This prepara- tion could not be induced to crystallize (specific activity 1.55 x lo5 dpm of r4C per mmole; 3H : 14C ratio 17.6) ; vE:i 3500 (OH), 1740,1730,1710 (sh) (C=O), 1250 cm-i.

3H4-14C6-Soc-Acetoxy-ll-keto Methyl Fusidate (Va)---Fusidic acid benzene solvate (IVa) (25 mg) was converted to 3a-ace- toxy methyl fusidate (IVd) described above and the product was treated with Jones reagent (3 drops) in acetone (5 ml) at 0” for 15 min. The reaction was terminated by addition of metha- nol and water. The product was extracted with dichlorometh- ane, washed, and dried. After removal of solvent the product was purified by repeated TLC (ethyl acetate-benzene, (1: 19)) to yield 15 mg (Va) (specific activity 3.88 x lo5 dpm of r4C per mmole; 3H:14C ratio 18.1); YE: 1740, 1735, 1720, 1700 (C=O), 1240 cm-l; mass spectrum, m/e 526 (M-44), 511,510 (M-60, base peak), 497, 478, 467, 451,450.

3H4-W6-S ,1 I-Diketo Methyl Fusidate (Vb)-Fusidic acid-ben- zene solvate (IVa) (25 mg) was methylated as described above. The crude product was dissolved in analytical grade acetone

3 Micronutrient solution: FeS04.7 Hz0 (0.5 g per liter), ZnClz, (0.21 g per liter), MnSOa.HzO (0.15 g per liter) and CaC12.2 Hz0 (18.5 g per liter).

by guest on February 9, 2020http://w

ww

.jbc.org/D

ownloaded from

2496 Biosynthesis of Fusidic Acid Vol. 246, No. 8

(5 ml) and treated with Jones reagent (5 drops) at 0” for 10 min. Conventional work-up followed by TLC (ethyl acetate-benzene, (1: 19)) gave a single product which was crystallized from ether- hexane to give 13 mg of Vb; m.p. 129” (specific activity 4.17 X

lo5 dpm of 14C per mmole; 3H : 14C ratio 18.1) ; YE:: 1745 (-0. CO * CH,), 1725 (-COe0.CH3), 1715 (C=O) cm-l; NMR 62.5 (doublet, 3H (30~CHa, 6.0 Hz)), 63.0, 69.0, 72.0 (singlets, each 3H (19-, 32-, and 1%CH3)), 95.5, 100 (6H (26- and 27-CHs,)), 120 (singlet, 3H(-0 .CO. CH3)), 218 (singlet, 3H(-CO .O. CHa)), 302 (multiplet, lH(C-16)); mass spectrum, m/e 466 (M-60), 431 (M-92), 420, 407, 397, 365, 246, 243, 219 (base peak), 177.

3H3-14Cs-SmAcetoxy-A9(11) Methyl Fusidate (VI)-To a stirred and cooled solution (-30”) of 3a-acetoxy methyl fusidate (IVd)

(VII)a.R,=Rz=Rs=H b. RI = Rz = H, R, = CR,

0 OAc

(XVII)

o,# o/&b +OOR

; + OH

(XI) (XII) (Xl a. R = CHs b. R = CH,COOpBr

(50 mg) in pyridine (3 ml), a solution of thionyl chloride (0.2 ml) in pyridine (3 ml) was added dropwise over a period of 15 min. The mixture was stored for 1 hour at O”, and the reaction was terminated by pouring onto ice. The product was extracted with dichloromethane and fractionated by TLC (ethyl acetate- benzene, (1: 9)) to give 25 mg of VI. The homogeneity of VI was assayed by repeated TLC (specific activity 1.76 x lo5 dpm of 14C per mmole, aH:14C ratio 14.9); v:z 1745, 1735, 1720 (C=O), 1240, 1010, 900, 810, 790 cm-*; NMR 49.5 (doublet, 5 Hz, C-30-CHs), 52.5,59.5,71.0 (9H (18-, 19., and32-CH3)) 95.5, 100.0 (singlets, 6H (26- and 27-CH3)), 119, 122 (singlets, 3H each (3a and 16fi O.CO.CHI)), 220 (singlet, 3H (CO.O.CH,)), 294 (multiplet, 1H (3/3H)), 305 (multiple& 1H (C-16 proton)), 328 (triplet, 1H (11-olefinic proton, 3.5 Hz)), 356.5 (doublet, 1H

(24H)) ; mass spectrum, m/e 494 (M-60), 462, 425, 419, 365 (base peak).

3H4-‘4CB-Dihydrojusidic Acid (VIIa)-Fusidic acid-benzene solvate (IVa) (215 mg) was diluted with cold material (105 mg) and dissolved in ethanol (10 ml) and 5% palladium-calcium car- bonate (60 mg) was added (20). The mixture was stirred vigor- ously and hydrogenated until 1.1 moles of hydrogen had been absorbed. The catalyst was filtered off and the solution was diluted with water. The solid was filtered off and dried. The essentially quantitative conversion to VIIa was demonstrated by mass spectrometry (m/e 518),

(XVI) a. R = H b. R = CH,

(XXIV) a. It = CHs b. R = CH,OH

aH4-14Cs-Methyl Dihydrojusidate (VIIb)-Dihydrofusidic acid (VIIa) (25 mg) was esterified with diazomethane in ether-metha- nol. The product was purified by TLC (ethyl acetate-benzene (1: 9)) and crystallized from ether-hexane to give 17 mg of VIIb; m.p. 150” (literature (20), 150-150.5”) (specific activity 2.72 x lo5 dpm of 14C per mmole; aH :14C ratio 16.8).

Deacetoxylation of 3H4-Ws-Dihydrofusidic Acid (VIIa)-To a solution of VIIa (60 mg) in dimethylformamide (0.2 ml), lithium chloride (15 mg) was added, and the mixture was heated at 140- 150” for 3 hours (21). The reaction mixture was cooled and diluted with water, and the product was recovered with ethyl acetate. The ethyl acetate solution was washed with water and

by guest on February 9, 2020http://w

ww

.jbc.org/D

ownloaded from

Issue of April 25, 1971 L. J. Mulheirn and E. Caspi 2457

dried, and the solvent was removed under reduced pressure. The residue was dissolved in methanol and esterified by addition of diazomethane in ether. Separation of the products was achieved by TLC (ethyl acetate-benzene (1:19), developed four times). The major product was identified as the Ala(r7)~20(22)-diene (VIII) (25 mg) which was purified by repeated chromatography (specific activity 2.55 x lo5 dpm of 14C per mmole; 3H:14C ratio 13.7); YE:: 3450 (3~y and liar-OH), 1725 (-CO.OMe) cm-i; X,,, 252 nm NMR 51.0, 57.0, 59.0, 66.0, 77.0 (singlets, each 3H (19, 32, and WCHZ) and doublet, 3H (30CH3), 6H 26- and 27-CHs)), 224 (multiplet, IH (3@H)), 226 (singlet, 3H, (-CO .O.CHa)), 256 (multiplet, lH, (11&H), 351 (triplet, 1H (22-H, 7.5 Hz)).

A second product from this reaction was identified as the isomeric A15~17(20)-diene (IX) (10 mg) and was purified as described above (specific activity 2.52 x 10” dpm of i4C per mmole; 3H :14C ratio 17.1); X,,, 271 nm (~16,000).

Oxonolysis of 3H4-14CG-Methyl Dihydrojusidate (vIIb)-Dihy- drofusidic acid (170 mg) was esterified with diasomethane and the resulting product was dissolved in dichloromethane (25 ml) con- taining pyridine (0.025 ml). The mixture was cooled to -75’ and a stream of ozonized oxygen was bubbled through the solu- tion for 15 min. The excess ozone was removed in a stream of nitrogen, zinc (200 mg) and acetic acid (2.5 ml) were added, and the mixture was stirred briefly at 0” and then at 25” for 90 min. The filtered solution was washed with a dilute sodium bicarbon- ate solution and with water and then dried, and the solvent was removed under vacuum. The resulting gum was dissolved in a little ether and excess hexane was added. The supernatant liquid was decanted from the precipitated solid steroidal material. The solution was diluted further with hexane to ensure complete removal of the steroidal product. This process gave a crude steroidal fraction (98 mg) while the hexane solution gave, on evaporation, methyl 2-hydroxy-6-methylheptanoate (Xa) (33

md; vE2 3470,1740,1385,1365,1230,1140,1090,1020 cm-l. 3H1-Y3-p-Bromophenacyl d-Hydroxy-6-methyl Heptanoate (Xb)

-Methyl 2-hydroxy-6-methylheptanoate (Xa) (32 mg) was hydrolyzed with potassium hydroxide (13 mg, 1.2 moles) in aqueous methanol (I :2, 5 ml) at 25” for 4 hours, after which sol- vents were removed under reduced pressure. The brown gum was dissolved in dimethyl formamide (2 ml) and treated with triethylamine (0.2 ml) and p-bromophenacyl bromide (130 mg, 2.5 moles) at 25” for 18 hours. The mixture was poured into water and extracted with ether. The organic phase was washed with water and dried. Following removal of the solvent the product was purified by TLC (ether-hexane, (1: 3)) and crystal- lized from ether-hexane to give 30 mg of p-bromophenacyl2-hy- droxy-6-methylheptanoate (Xb); m.p. 88-91” (specific activity 0.88 X lo5 dpm of 14C per mmole; 3H:14C ratio 9.8); vz:i 3420 (OH), 1735 (-CH*.CO.O-), 1690 (-CO.+--), 1585 (aromatic absorption), 1290, 1225, 1185, 1130, 1095, 1065, 1000, 970, 810 cm+; NMR 50.0, 55.0 (doublet, 6H(-CH(CH,)2) 127 (multi- plet, 1H; ---CH-(CH3)2), 138 (multiplet 2H, CH2) 311 (singlet, 2H; &-CH2) 442,450,454,462 (aromatic, 4H) ; mass spectrum, m/e 358/356 (M+), 327/325 (M-31), 274/272,217/215, 200/198, 185/183 (base peaks), 157/155.

Isolation of the 3Hz-‘4C4 9@ ,IS&Triketone (XI)-The steroidal portion from the ozonolysis described above (98 mg) was dis- solved in acetone and treated with Jones reagent at 25” for 15 min. After conventional work-up the crude product was treated with zinc (600 mg) and glacial acetic acid (15 ml) under reflux for 30 min. Following filtration and washing of the zinc with chloro-

form the solvents were removed under vacuum, and the product was purified by TLC (ethyl acetate-benzene (1:lO)). Crystal- lization from ether-hexane gave 60 mg of XI; m.p. 174-175” (liter- ature (22), 171-174”) (specific activity 1.77 X lo5 of W per mmole; 3H:14C ratio 14.2); vzi 1735, 1710, 1700 (C=O), 1220, 1190, 1150, 1120, 990 cm-l; NMR 49.5, 74.0, 88.0 (singlets, each 3H; 18, 19, 32-CH3); 62.0 (doublet, 3H, 30-CH3, 6Hz); mass spec- trum, m/e 330 (M+), 302 (M-28), 273, 234, 219, 205, 191, 179, 137, 97 (base peak).

Epimerization of XI to 3Hl-14C4-9cr, lS&Triketone (XII)-The triketone (XI) (35 mg) was dissolved in dioxane (8 ml) previously saturated with dry hydrogen chloride. After 18 hours at 25” the solution was diluted with water and concentrated by evaporation at 60” under vacuum. The cooled solution was extracted with dichloromethane which was then washed with water and dried. After removal of solvents under vacuum, the crude product was purified by TLC (ethyl acetate-benzene (1:9)) and crystallized from ether-hexane to give 24 mg of XII; m.p. 253-254” (literature (22), 255-257”) (specific activity 1.89 X lo5 dpm of i4C per mmole; 3H:14C ratio 7.0); vE:G 1740, 1710, 1690 (C=O), 1230, 1195, 1130, 1100 cm-‘; NlMR 62.5, 68.0, 70.5 (singlets, each 31-I; 18, 19, and 32-CH3); 62.0 (doublet 3H, 30-CH3, 6Hz); mass spectrum, m/e 330, 179 (base peak), 135, 123, 110.

Xidechain Degradation of 3H4-14Cs-Fusidic Acid (I Vu)--The mother liquors from crystallization of the 3H-r4C-fusidic acid were combined with the remaining extract of the culture broth in which the biosynthesis was carried out. This was diluted with non- radioactive fusidic acid (220 mg), and crystallized from benzene to give the solvate (IVa) (230 mg). To a solution of IVa in aqueous methanol (1:4; 5 ml), 30% aqueous sodium hydroxide (0.6 ml) was added, and the mixture was heated for 15 min at 75” (22). The methanol was partially removed under a stream of nitrogen, and the aqueous solution was treated with glacial acetic acid (2.5 ml) on a water bath for 90 min. The cooled solution was diluted with water, and the product (XIII) was collected (aH:14C ratio 17.2). The lactone (XIII) was dissolved in metha- nol (10 ml), and treated with 5% aqueous sodium borohydride (1 ml) (26). After 30 min the reaction mixture was carefully acidified with acetic acid and water was added. The crude satu- rated lactone (XIV) was collected (150 mg) and was purified by TLC (ethyl acetate-benzene (1:9)); m.p. 173-174” (literature (21), 174-176”) (3H:14C ratio 16.9); vzti 3450, 1755, 1625, 1315, 1190, 1060, 1010, 965, 925 cm-i.

The homogenous XIV was treated with osmium tetroxide (120 mg, 1.2 moles) in ether-pyridine (5:4, 6 ml) with stirring at 25” for 36 hours. TLC indicated the absence of starting material. A solution of sodium bisulphite (250 mg) in pyridine-water (1: I,9 ml) was added, and the mixture was stirred for 30 min. The resulting product was extracted with dichloromethane, and the organic phase was washed with water and dried. The solvent and traces of pyridine were removed under reduced pressure. The product was purified by repeated preparative TLC (ethyl acetate) to give (XV) (80 mg) (specific activity 0.81 x lo5 dpm of r4C per mmole; 3H:14C ratio 16.7) ; v”,fi 3450, 1750, 1180, 1075, 1050, 1015, 970 cm-l; mass spectrum, m/e 438 (M3HsO), 423, 420, 405, 370, 340, 325, 281, 265.

The above tetrol (XV) (80 mg) was oxidized with Jones reagent in the usual manner to give a polar compound which was isolated by TLC (methanol dichloromethane (1: 12)). Mass spectrom- etry confirmed its structure as the 24-carboxylic acid (XVIIa) (25 mg). The product was purified by TLC (specific activity

by guest on February 9, 2020http://w

ww

.jbc.org/D

ownloaded from

2498 Biosynthesis of Fusidic Acid Vol. 246, No. 8

0.65 x lo5 dpm of 14C per mmole, 3H : 14C ratio 16.8) : vE:G 3500, 2900, 1760, 1700, 1175, 1090, 950 cm-l; NMR 61.0, 67.5, 74.5 (doublet and singlets (1%, 19-, 30., and 32-CHs)), 305 (multiplet, 1H (16~H)) ; mass spectrum, m/e 444 (N+, base peak), 398,384, 372, 265, 123.

The acid (XVIa) was esterified with diazomethane in ether- methanol, and the product was purified by TLC (ethyl acetate- benzene (1: 1)) to yield 8 mg of the ester (XVIb) (specific activity 0.64 x lo5 dpm of 1% per mmole, 3H:14C ratio 16.5) ; vE:z 1760, 1740, 1700, 1690, 1215, 1175, 1075, 1000 cm-l; mass spectrum, m/e 458 (M+, base peak), 427 @l-31), 336, 307, 304, 275, 225, 219, 206, 193, 123.

RESULTS AND DISCUSSION

Fusidic acid has been shown to retain six C-2 carbon atoms of MVA (24, 25) and to be derived from squalene via squalene-2,3- oxide (26). The enzymatic cyclization of the epoxide to the cat- ion (Ia) or its stabilized equivalent (Ib) should proceed with the retention of six 4-pro-R protons of MVA at C-3, C-5, C-9, C-13, C-17, and C-24 as shown in I (1). Conversion of I to fusidic acid would be expected to result in retention of the protons at C-5a, C-9& C-13a, C-24, and possibly C-3fi while that at C-17 should be lost (16).

In order to evaluate this hypothesis F. coccineum was grown in a medium (19) containing (3R,4R)-[2-14C,4-3H] MVA (50 PC1 of 14C) and the fusidic acid isolated showed 0.2% incorporation of 1% radioactivity. The metabolite was characterized as the methyl ester (IVb) (3H:14C ratio 18.0). Conversion of methyl fusidate (20) to the 3,16-diacetate (IVc) (3H:14C ratio 17.6), fol- lowed by oxidation, gave the 11-mono ketone (structure Va) which had an unchanged 3H:14C ratio (l&l), indicating that no loss of tritium from C-9 occurred during oxidation with Jones reagent. Oxidation of methyl fusidate under similar conditions gave the diketone (Vb) (3H:14C ratio l&l), without loss of tritium, demonstrating the absence of tritium from the 3p posi- tion of fusidic acid.

The location of tritium atoms at C-9/3 and 13cr was then exam- ined. The diacetate (IVd) was dehydrated (16) to give the triene (VI). The decreased 3H :14C ratio (14.9) revealed the presence of tritium at C-9 of fusidic acid, which is removed on introduction of

the A9(11) double bond. However, the change in 3H:14C rat,io of VI was less than predicted for removal of 1 atom of tritium. This may have been due to the presence of small amounts of contaminants (e.g. 11-chloro or All analogues), which could have cochromatographed with the major product. The presence of 1 atom of tritium at C-9/3 was confirmed by later results (see below).

For investigation of the presence of tritium at C-13a! fusidic acid was hydrogenated over a deactivated catalyst (20) to give dihydrofusidic acid (VIIa) characterized as the methyl ester (VIIb) (3H :14C ratio 16.8). The decreased ratio is consistent with the loss of some tritium from C-24 during hydrogenation. Deacetoxylation of (VIIa) with lithium chloride-dimethyl form- amide (21) gave the A~a(~7)~20(22)-dier~e (VIII) (3H :14C ratio 13.7). The lower ratio indicates the removal of a tritium atom from C-13. It is noteworthy that the isomeric Als,l7@“-diene (IX) (3H:14C ratio 17.1) showed an unchanged tritium content.

Additional evidence for the presence of tritium atoms at C-9 and C-13 was obtained by a different route. Ozonolysis of VIIb gave a mixture of products from which the steroidal fragments were separated by precipitation with hexane. The crude solid material was oxidized with Jones reagent and resulting acetoxy- triketone (XVII) (16) was treated with zinc-acetic acid to give the 9fi, 13/3 (H)-triketone (XI) (22). The aH:14C ratio of the product (14.2) is consistent with the loss of 2 tritium atoms from C-13 and 24 and two l4C atoms from C-22 and 26 of methyl dihy- drofusidate (VIIb). The tritium atom originally present at C-13a! would be expected to be lost during the epimerization of (XVII) on treatment with zinc-acetic acid. Exposure of the 9fi, 13/Striketone (XI) to hydrogen chloride in dioxane resulted in epimerization (22) at C-9 and concomitant loss of tritium to give the 9a(H), 13/3(H)-triketone (XII) (3H:14C ratio 7.0). This con- stitutes added evidence for the presence of a tritium atom at C-9p of fusidic acid. The location of a third tritium atom at a nonenolizable position in the nucleus of the triketone (XII) is also demonstrated by this result. Biogenetic considerations suggest its location at C-5a! in fusidic acid (16, 17). The results are summarized in Table I.

The actual presence of the 4th tritium atom in the side chain of the antibiotic was proven by isolation of the hydroxy-ester

TABLE I

Summary of results of degradation of fusidic acids biosynthesized from (SR,4R)-(214C,43H)-mevalonic acid

Product

Methyl fusidate (IVb). ................ Methyl fusidate-3ol-acetate (IVd). ...... 11.Ket,o-3ol-acetoxy ester (Va) ......... Methyl fusidate-3, ll-diketone (Vb). .... 9(11), 17(20) ,24-Triene (VI). ............ Methyl dihydrofusidate (VIIb). ....... 13(17),20(22)-I>iene (VIII) ............. p-Bromophenacyl ester (Xb). ......... 9p, la@-Triketone (XI). ........... .... 9a, 13p-Triketone (XII). ............... Lactone (XIV). ....... .......... Diketo ester (XVIb) .............

I Isotopic atoms removed”

- SH

9cY

13a 5a, 9p, 13CY 13~u, 24 9/3, 13or, 24

24

‘4C

1, 7, 15, 31 22, 24 22, 24

26

dfim 18.0 17.6 18.1 18.1 14.9 16.8 13.7 9.8

14.2 7.0

16.9 16.5

3H:W (atomic)

Observed

3.9:6 4.0:6 4.0:6 3.3:6 3.7:6 3.0:6 0.7:2 2.1:4 1.0:4 3.8:6 3.0:5

Calculated

4: 6 3 : 6 4:6 4:(i 3:G 4:6 3:f.i 1:2 2:4 1:4 4: 6 3:5

a With reference to methyl fusidate (IVb)

by guest on February 9, 2020http://w

ww

.jbc.org/D

ownloaded from

Isshe of April 25, 1971 L. J. Mulheirn and E. Caspi 2499

(Xa) obtained by osonolysis of methyl dihydrofusidate (VIIb) (zinc-acetic acid work-up (20)). The ester (Xa) was hydrolyzed and the resulting hydroxy acid was characterized as the p-bromo- phenacyl ester (Xb) (311:14C ratio 9.8). The isotopic ratio and 14C specific activity of this product were compatible with the presence of 1 tritium and two 14C atoms in the side chain of fusidic acid. For a more precise definition of the location of the tritium atom a degradation of the side chain was performed. Fusidic acid (IVa) was transformed to the lactone (XIII) which was re- duced with sodium borohydride (23) to the saturated lactone (XIV) (3H :14C ratio 16.9). Exposure of XIV to osmium tetrox- ide provided the glycol (XV) which was converted to the diketo acid (XVIa) (3H:14C ratio 16.8) by treatment with Jones reagent. The 3H:1*C ratios of the acid (XVIa) and its methyl ester (XVIb) (3H:14C ratio 16.5) showed the loss of one tritium and one 14C atom, a result consistent with the predicted presence of a tritium atom at C-24.

In summary it may be concluded that the biosynthesis of fusi- die acid from (3R,4R)-(2-14C,4-3H) MVA proceeds with the retention of 4 tritium atoms. The location of the important isotopic hydrogen atoms at C-9@ and 13a has been demonstrated, and evidence provided in support of the presence of the remaining tritium atoms at C-5a and 24, while no tritium was detected at the C-3/I position. This demonstrates that the 4-pro-R protons of MVA retained in IVa occupy their original positions of attach- ment to the carbon skeleton, as predicted for the conversion of squalene, via the cation (I) to fusidic acid. Consequently, the protosterol skeleton of IVa seems to be formed by elimination of a proton from C-17 of I without rearrangement. In principle, our results on the mode of formation of fusidic acid can be interpreted as further evidence in favor of the biogenetic scheme proposed by Eschenmoser et al. (7) for the formation of lanosterol.

It has been shown (25) that the protosterol (XVIII) can be re- arranged in the presence of acid to lanosterol (II). This suggests that lanosterol is the most energetically favored product of the carbonium ion formed by protonation of the 13(17) double bond of the protosterol. As a corollary to this, the idea was forwarded (6) that, in lanosterol biosynthesis, the rearrangement of the cat- ion (I) to lanosterol may be governed largely by thermodynamic factors, with the enzyme having only a peripheral role. In addi- tion the cyclization of the 10,15-di-nor-squalene (XIXa) by rat liver enzymes to the protosterol (XX) (27,29) was demonstrated. When only one of the methyl groups of squalene is removed, as in XIXb and XIXc, the lanosterol analogues (XXIa and b) were formed, respectively (30). These results were viewed (6) as evi- dence for the importance of intramolecular steric repulsion as the driving force for the backbone rearrangement leading to lanosterol. However, alternative routes of stabilization may overcome the internal repulsion. This is exemplified by the cy- clization of the conjugated analogue (XXII) which gives the protosterol (XXIII) when cyclized with rat liver enzymes (28). It is important to note, however, that the stereochemistry of the products (XX) and (XXIII), particularly at C-17, was not proven rigorously. The saturated analogue of XXIII obtained by catalytic hydrogenation, was recently converted by acid treat- ment (24, 25) to dihydrolanosterol. This was interpreted as evidence in support of the C-20 location of the hydroxyl group and of the indicated stereochemistry of (XXIII) (29).

The current ideas on the cyclization of squalene fail to provide a complete rationalization for the formation of the protosterols. In F. coccineum two different pathways for stabilization of an

intermediate of type I appear to function in parallel (16). Our results are consistent with the postulate that fusidic acid, and hence the protosterols (XVIII and XXIV) are formed by direct elimination of a proton from C-17 of I. On the other hand the presence of ergosterol in the mold is clear evidence for the opera- tion of the backbone rearrangement process (7) which terminates in lanosterol (or cycloartenol).

The question arises as to why the C-20 cation (or its stabilized analogue) should collapse to different end products in the same species. The problem is mechanistically interesting and several explanations for this divergence can be visualized. For instance, it has been proposed (8) that the formation of lanosterol may re- quire two enzymes, a cyclase which generates the intermediate

(XIX) a. R, = R, = H b. R, = H, RZ = CH, c. R, = CH,, RI = H

Ho,& o& % (XXI) a. RI = H, Rf = CH,

b. R, = CH,, Rt = H (XXII)

(XXIII) -OH at one C*

(I) and an isomerase which promotes rearrangement. Suppres- sion of the latter enzyme might result in the activation of a shunt mechanism leading to protosterols. Alternatively, cyclization and isomerization may be a single enzyme process. The two groups of compounds might then be formed on completely sepa- rate enzyme systems, which would control the fate of the inter- mediate (I) and hence the structure of the products. One exten- sion of this idea leads to the postulate that the two cyclases could produce different intermediates (XXVa) and (XXVIa) having opposite stereochemistry about C-17. The trans relationship of the 13a and 17/3 protons of XXVa would facilitate backbone rear- rangement to lanosterol. The cis relationship of the 13a! and

17a protons of XXVIa could prevent rearrangement and the most

by guest on February 9, 2020http://w

ww

.jbc.org/D

ownloaded from

2500 Biosynthesis of Fusidic Acid Vol. 246, No. 8

favorable stabilization pathway under these circumstances may be elimination of the 17a proton.

A further observation of relevance to this discussion concerns the application of the nucleophilic stabilization (X group) mecha- nism (9) to the biosynthesis of the protosterols. It has been sug- gested that the involvement of an enzymic or chemical nucleo- phile in the conversion of squalene-2,3-oxide to lanosterol would lead to the transiently stabilized intermediate (XXVb). Rota- tion of the side chain of (XXVb) relative to the ring system would then permit backbone rearrangement to lanosterol via trans elimination of HX. This hypothesis circumvents some of the difficult,ies inherent in the nonstop carbonium ion theory, particu-

(XXV) a. x = 6 b. X = Nucleophilic

group

(XXVI) a. X = 8 b. X = Nucleophilic

group

(XXIX)

larly with regard to the spatial requirements of the enzyme. The

invoked “stable” intermediate also allows accommodation of the hypothesis that two enzymes may be involved in the production of lanosterol. However, this concept appears not to be directly applicable to the production of the protosterols, since, irrespective of whether the side chain of the nucleophile-stabilized inter- mediate has configuration (XXVb) or (XXVIb), a tram elimina- tion of the elements of HX (as required by the theory) from C-17 and 20 would lead to a protosterol having the wrong geometry about the A17(20) bond.

Several interpretations can be projected which would accom- modate this discrepancy. First it is possible that a free car- bonium ion is indeed involved in the biosynthesis of protosterols.

Unfortunately, however, acceptance of this view would leave unanswered the previously mentioned criticism of this hypothesis. On the other hand, to fit the facts, the X group theory would have to be modified or broadened. A basic tenent of the X group con- cept as it was formulated by Cornforth (9) is the trans elimination of the proton and of the prosthetic group. Obviously, a tram elimination will be thermodynamically favored over a cis elimina- tion. To reconcile this requirement, it could be postulated that the “X” group in XXV or XXVI is displaced by another nucleo- phile “Y” with inversion of configuration.4 Thus in the newly formed Y-stabilized intermediate (corresponding to XXV or XXVI, but with inverted configuration of the C-X(Y) bond) the C-17 proton and the prosthetic Y moiety are positioned for a facile trans elimination, leading to protosterols with the correct geometry at the C-17-20 double bond. Finally, a broadening of the concept of the “X” group may also be considered. For exam- ple two types of “X” group may be invoked, one allowing a truns elimination-migration from XXVb leading to a lanosterol typic side chain, and the other producing the protosterol (XXIVa) by cis elimination of HX from either XXVb or XXVIb. This mech- anism is illustrated in XXVII and XXVIII in which the pros- thetic group is shown as a phosphate ester. The indicated six- membered transition states would facilitate cis dehydration. It must be stressed, however, that the phosphate moiety is used only as an illustration and no experimental support for its par- ticipation is claimed.

Murata et al. (31-33) have recently isolated and determined the structure of several new protosterols which they named Alisols. Alisol A was shown by x-ray crystallography to have Structure XXIX and the indicated 20R, 23s) 24R, (23,24-threo)- configuration. In Alisol B the 24,25-diol of XXIX is repla,ced by a 24,25-epoxide; Alisol C has the structure of 16-keto-Alisol B. The A13(17)-20R-moiety common to the Alisols suggests their for- mation via an intermediate of type XXV. Indirectly therefore this provides some support for the two-enzyme concept of a squalene-oxide-cyclase and an isomerase in the biosynthesis of these triterpenes. In any event the actual nature of the factors governing the in vivo formation of the different triterpenes is not known and requires a detailed study.

Finally, the present results are of relevance to recent observa-

tions on the mechanism of elimination of the C-4 gem dimethyl group from triterpenes. It has been demonstrated that, in the

formation of cholesterol in rat livers, the sequence involves oxida- tive loss of the 4cy methyl group (via a 3-keto intermediate) and subsequent equilibration of the 4fi methyl group to the 4a! posi- tion (34-36). A similar sequence seems to be operative in some phytosterols (35). However, in fusidic acid the demethylation process appears to be different. Isolation of the triterpenes XVIII and XXIV from F. coccineum suggests that this group of

compounds is formed from (3s)squalene-2,3-oxide, as are lanos- terol and other triterpenes (38, 39). Hence, the 4cr-methyl

group of XXIV and its cogeners will originate from C-2 of MVA The reported (24, 25) retention of six C-2 carbon atoms of MVA by fusidic acid (confirmed by the present investigation) and the absence of tritium from C-3 on incorporation of (3R,4R)- (214C,4-3H) MVA suggest that the formation of fusidic acid

involves a direct loss of the 4@-methyl group, presumably via a 3-keto analogue.

4 The displacement idea was formulated by Dr. J. W. Cornforth to whom we are indebted for a stimulating exchange of views.

by guest on February 9, 2020http://w

ww

.jbc.org/D

ownloaded from

Issue of April 25,197l L. J. Mulheirn and E. Caspi

1.

2.

3.

4.

5.

6.

7.

8.

9. 10.

11.

12.

13. 14.

15.

16.

17. 18.

19.

20.

REFERENCES

CORNFORTH, J. W., CORNFORTH, R. H., DONNINGER, C., AND POPJAK, G., Proc. Roy. Sot. Ser. B Biol. Sci., 163, 492 (i966).

ZANDER, J. M., GREIG, J. B.. AND CASPI. E.. J. Biol. Chem.. 246, 1247 (1970).

I I

BURTON, D. H. R., GOSDEN, A. F., MELLOWS, G., AND WID- DOWSON, D. A., Chem. Commun., 184 (1969).

COREY, E. J., RUSSEY, W. E., AND ORTIZ DE MONTELLANO, P. R., J. Amer. Chem. Sot., 88, 4750 (1966).

VAN TAMELEN. E. E.. WILLETT. J. D.. CLAYTON. R. B.. AND LORD, K. E., J. Amer. Chem.’ Sot., 88, 4752 (1966). ’

VAN TAMELEN, E. E., HANZLIK, R. P., CLAYTON, R. B., AND BURLINGAME, A. L., J. Amer. Chem. SOL, 92, 2137 (1970).

ESCHENMOSER, A., RUZICKA, L., JEGER, O., AND ARIGONI, D., Helv. Chim. Acta, 38, 1890 (1955).

WOODWARD, R. B., AND BLOCH, K., J. Amer. Chem. Sot., 76, 2023 (1953).

CORNFORTH, J. W., Angew. Chem., 7, 903 (1968). MAUDGAL, It. K., TCHEN, T. T., AND BLOCH, K., J. Amer.

Chem. Sot., 80, 2589 (1958). CORNFORTH, J. W., CORNFORTH, R. H., PELTER, A., HORNING,

M. G., AND POPJAK, G., Tetrahedron, 6, 311 (1959). CORNFORTH, J. W., CORNFORTH, R. H., DONNINGER, C.,

POPJAK, G., SHIMIZU, Y., ICHII, S., FORCHIELLI, E., AND CASPI, E., J. Amer. Chem. Sot., 87, 3224 (1965).

CASPI, E., AND MULHEIRN, L. J., Chem. Commun., 1423 (1969). CASPI, E., VARM.~, K. R., AND GREIG, J. B., Chem. Commun.,

45 (1969). JAYME, M., SCHAEFER, P. C., AND RICHARDS, J. H., J. Amer.

Chem. Sot., 92, 2059 (1970). GODTFREDSEN, W. O., VON DAEHNE, W., VANGEDAL, S., MAR-

QUET, A., ARIGONI, D., AND MELERA, A., Tetrahedron, 21, 3505 (1965).

COOPER, A., AND HODGKIN, D. C., Tetrahedron, 24,909 (1968). CASPI, E., AND MULHEIRN, L. J., J. Amer. Chem. Sot., 92,

404 (1970). GODTFREDSEN, W. O., LORCH, H., AND JAHNSEN, S., U. S.

Patent 3,072,631 (1963). GODTFREDSEN, W. O., AND VANGEDAL, S., Tetrahedron, 18,

1029 (1962).

21.

22.

23.

24.

25.

26.

- \----I

27. COREY, E. J., ORTIZ DE MONTELLANO, P. R., AND YAMAMOTO, H., J. Amer. Chem. Sot., 80, 6254 (1968).

28. COREY, E. J., LIN, K., AND YAMAM~TO, ‘H., J. Amer. Chem. Sot., 91, 2132 (1969).

29. COREY, E. J., AND YAMAMOTO, H., Tetrahedron, 27,2385 (1970). 30. VAN TAMELEN, E. E., HANZLIK, R. P., SHARPLESS, K. B..

31.

32.

33.

34. SHARPL&S, I?..h., SNYDER, T.‘E., SPENCER, T. A., MAHESH- WARI, K. K., GUHN, G., AND CLAYTON, R. B., J. Amer. Chem. Sot., SO, 6874 (1968).

35. SHARPLESS, K. B., SNYDER, T. E., SPENCER, T. A., MAHESH-

36.

37.

38.

39.

GODTFREDSEN, W. O., VON DAEHNE, W., AND VANGEDAL, S., Chem. Commun., 638 (1966).

DIASSI, P. A., KR~KOW~R, 6. W., BACSO, I., AND VAN DINE, H. A., Tetrahedron, 22, 3443 (1966).

GODTFREDSEN, W. O., VON DAEHNE, W., TYBRING, L. AND VANQEDAL, S., J. Med. Chem., 9, 15 (1966).

ARIGO’NI, D., Conference on the Biogenesis of Natural Products, Academia Nazionale dei Lincei, Rome, 1964.

VISCONTI, G., Ph.D. thesis, Eidg. Technische Hochschule, Zurich, 1968.

GODTFREDSEN, W. O., LORCK, H., VAN TAMELEN, E. E., WIL- LETT, J. D., AND CLAYTON, R. B., J. Amer. Chem. Sot., 90, 208 (1968).

CLAYTON, R. B., RICHTER, W. J., AND.BTJRLINGAME, A. L.; J. Amer. Chem. Sot., 96. 3284 (1968).

MURATA, T., AND MIY~MO~O, M.; Chem. Pharm. Bull (Tokyo), 18, 1354 (1970).

MURATA, T., SHINOHARA, M., AND MIYAMOTO, M., Chem. Pharm. Bull. (Tokyo), 18, 1369 (1970).

KAMIYA, K., MURATA~ T., AND NISHI~AWA, M., Chem. Pharm. Bull. (Tokuo). 18. 1362 (1970).

WARI, K.. K., NELSON, J. A., AN; CLAYTON, R. B., J. Amer. Chem. Sot.. 91. 3394 (1969).

RAHM~N, R.,‘SHARPLES& K. ‘B., SPENCER, T. A., AND CLAYTON, R. B., J. Biol. Chem., 246, 2667 (1970).

GHISALBERTI, E. L., DE SOUZA, N. J., REES, H. H., GOAD, L. J., AND GOODWIN, T. W., Chem. Commun., 1403 (1969).

Moss, G. P., AND NICOLAIDIS, S. A., Chem. Commun., 1072 (1969).

STONE, K. J., ROESKE, W. R., CLAYTON, R. B., AND VAN TAMELEN, E. E., Chem. Commun., 530 (1969).

by guest on February 9, 2020http://w

ww

.jbc.org/D

ownloaded from

L. J. Mulheirn and Eliahu CaspiMechanism of Squalene Cyclization: THE BIOSYNTHESIS OF FUSIDIC ACID

1971, 246:2494-2501.J. Biol. Chem. 

  http://www.jbc.org/content/246/8/2494Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/246/8/2494.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on February 9, 2020http://w

ww

.jbc.org/D

ownloaded from