THE GIBBERELLINS By JOHN FREDERICK GROVE (IMPERIAL CHEMICAL INDUSTRIES LIMITED AKERS RESEARCH LABORATORIES THE FRYTHE WELWYN HERTS.) 1. Fungal gibberellins THE discovery of the gibberellins originated from an investigation of a soil-borne disease of rice caused by the fungus Gibberella fujikuroi. Infected plants eventually wilt and die but at an early stage of the disease called “bakanae” in Japan the leaves and stems of some seedlings elongate more rapidly than those of healthy plants. In 1926 Kurosawal showed that cell-free filtrates from cultures of the fungus produced in healthy seedlings the elongation symptons characteristic of the disease and eventually in 1938 Yabuta and his collaborators2 succeeded in isolating from such culture fluids a crystalline active material which they named gibberellin A.The chemistry and plant-growth promoting properties of this material were described in a series of papers from the University of Tokyo these papers have been re-assembled and re~iewed.~ More recently in an attempt to repeat the isolation of gibberellin A from cultures of G. fujikuroi,* a group of workers at the Akers Research Laboratories of Imperial Chemical Industries Limited obtained a new active metabolite gibberellic acid,4 which differed from gibberellin A in its physical and chemical properties. The same compound was also isolated at the North Regional Research Laboratories Peoria U.S.A. where5 it was named gibberellin X until the identity with gibberellic acid was established.6 The American workers obtained a mixture of two compounds gibberellic acid and gibberellin A, which was similar to gibberellin A.These results led to a re-examination of the crude gibberellin produced by various Tokyo University strains of G. fujikuroi and gibberellic acid (gibberellin A3) gibberellin A, and a third active metabolite gibberellin A, were eventually obtained.’ Gibberellin A, another component of the mixture was described later,8 and the exact composition of the gibberellin A obtained in 1938 remains obscure. Modification of the fermentation conditions used to produce gibberellic acid has led to the isolation by the Kurosawa Trans. Nat. Hist. SOC. Formosa 1926 16 213. 2Yabuta and Sumiki J. Agric. Chem. Soc. Japan 1938 14 1526; Yabuta and Hayashi J. Agric. Chem. SOC. Japan 1939 15 257. Stodola “Source Book on Gibberellin 1828-1957,” U.S. Dept. of Agriculture 1958.* Curtis and Cross Chem. and Ind. 1954 1066. Stodola Raper Fennell Conway Sohns Langford and Jackson Arch. Biochem. Cross J. 1954,4670. Takahashi Kitamura Kawarada Seta Takai Tamura and Sumiki Bull. Agric. * Takahashi Seta Kitamura and Sumiki Bull. Agric. Chem. SOC. Japan 1957,21,396 1955,54 240; Stodola Nelson and Spence Arch. Biochem. 1957,66,438. Chem. Soc. Japan 1955 19 267. 56 GROVE THE GIBBERELLINS 57 group at Imperial Chemical Industries Limited of gibberellin A,9 and gibberellin A9.10 Gibberellic acid has been produced in much greater yieldll than the other gibberellins and a more extensive investigation of its chemistry and biological activity has been carried out than was possible with gibberellin A. The structure and some of the stereochemistry of gibberellic acid has been elucidated and the structures of the other gibberellins have been related to gibberellic acid.This contribution from the group at Imperial Chemical Industries Limited has been reviewed in a comprehensive treatise.12* Despite widespread searches there is no well-authenticated example of the production of a gibberellin by a fungus other than G..fujikuroi. 2. Occurrence of gibberellins in higher plants The fungal gibberellins were found to promote many normal processes of plant-growth development and this led to the discovery that compounds similar in chemical structure and physiological properties were widely distributed in higher plants. Mitchell Skaggs and Anderson13 had shown in 1951 that ether extracts of immature dwarf-bean seeds contained a substance which stimulated the elongation of seedling epicotyls to a degree much exceeding that observed after auxin application-an effect now recognised as gibberellin-like.More recently many w o r k e r ~ ~ * - ~ ~ have obtained similar extracts from the seed roots and shoots of a wide variety of plants.7 From immature seed of the runner bean Pl~aseolus multiJEorus MacMillan and his collaborator^^^ isolated gibberellin A in a yield of 2 mg./kg. fresh weight of seed together with three new compounds gibberellins As,,' A6,I8 and West and Phinney19 have also obtained gibberellins A and A (bean factor 11) from seed of the French bean Phaseolus vulgaris and gibberellin Al has been isolated from young shoots of Citrus reticulata Blanco var. unshiu.20 The presence of gibberellic acid * When this treatise was written all the available evidence favoured a trans-fusion of rings A/B antipodal to that normally found in diterpenes and an a-orientation for the ring A lactone bridge the latter is now believed to be fl-oriented (see Section 6) but no ring A stereochemistry has been adopted here since no definite conclusions have been reached.t For a more extensive bibliography see references 12 and 23. Cross. Galt. and Hanson. Tetrahedron Letters. 1960. No. 15. 18. lo Cross; Galt and Hanson; Tetrahedron Letters 1960 No. 23 22. l1 Brit. P. 838033. l2 Brian Grove and MacMillan Progr. Chem. Org. Nat. Prod. 1960,18,350. l3 Mitchell Skaggs and Anderson Science 195 1 114 159. l4 Phinney West Ritzel and Neely Proc. Nat. Acad. Sci. U.S.A. 1957 43 398. l5 Radley Nature 1956,178 1070; Ann.Bot. 1958,22 297. l6 Lona L'Alteneo Purmense 1957,28 11 1. l7 (a) MacMillan and Suter Natiirwiss. 1958 45 46; (b) MacMillan Seaton and Suter Proc. Chem. SOC. 1959,323; (c) Tetrahedron 1960,11 60. l8 MacMillan Seaton and Suter unpublished work. l9 West and Phinney J. Amer. Chem. Soc. 1959 81 2424. 2o Kawarada and Sumiki Bull. Agric. Chem. SOC. Japan 1959 23 343. 58 QUARTERLY REVIEWS in green maltz1 and in several higher plantsz2 has also been claimed on the basis of chromatographic evidence but the isolation of gibberellic acid from these sources has still to be achieved. 3. The significance of the gibberellins in plant physiology The discovery of the gibberellins in plant tissues implies that they are natural hormones with a regulating function in many aspects of plant growth and development.Only very small quantities of some gibberellins are available and little biological work has been done with them; but so far as is known they all have a qualitatively similar action. In most circumstances gibberellic acid is the most active followed by gibberellin A,.24 Exceptionally members of the Cucurbitaceae respond most readily to gibberellins A and A,.25 In the simple case of plants whose growth is not rigorously determined by day-length or temperature a gibberellin increases the length of the stem internodes without altering their number and only those internodes actually extending at the time of application are affected.26 In bushy plants a gibberellin enhances apical dominan~e.~' In general genetically dwarf plants show the greatest response and treated plants are similar to the tall varieties.Increases in both cell-division2* and cell-size2' are involved in gi b berellin-induced s t em-ex t ension. In the more complex case of plants whose development is determined by day-length or temperature or both the effect of a gibberellin is more intricate. Many biennials which normally require a period of cold treat- ment can be induced to bolt and flower by gibberellin applicati~n.~~ A gibberellin will similarly replace the long-day requirements of some plants30 and will terminate the dormancy induced by short-days in deciduous shrubs and trees;31 and the development of autumn foliage colours and leaf-fall can be Although the gibberellins promote cell elongation they differ from the auxins in many ways. For example they have little or no action on root which is inhibited by auxins; neither do they initiate roots on 22 Adler Medwick and Johl 138th Meeting Amer.Chem. SOC. New York Sept. 23 Phinney and West Ann. Rev. Plant Physiol. 1960 11 411. 24 Bukovac and Wittwer Nature 1958 181 1484. 25 Lockhart and Deal Naturwiss. 1960 47 141 ; Brian and Hemming Nature 1961 26 Brian and Hemming Physiol. Plant. 1955 8 669. 27 Brian Elson Hemmmg and Radley J. Sci. Food Agric. 1954 5 602; Brian 28 Sachs Bretz and Lang. Amer. J. Bot. 1959 56 376. 2 9 Lang Naturwiss. 1956 43 257 284. 3O Bukovac and Wittwer Quart. Bull. Mich. Agric. Exptl. Stn. 1957 39 650. 31 Lockhart and Bonner Plant Physiol. 1957 32,492. 32 Brian Petty and Richmond Nature 1959 183 58. 33 Brian Hemming and Radley Physiol. Plant. 1955 8 899. Lazar and Dahlstrom personal communication.ll-l6th 1960 Abs. p. 25A. 189 74 Hemming and Lowe Physiol. Plant. 1959,12,15. GROVE THE GIBBERELLINS 59 cuttings an effect which is promoted by auxins. Exogenous auxins have a large effect on the extension of isolated plant tissues but little effect on intact plants; the gibberellins on the other hand induce large responses in intact plants but have relatively little effect on shoot or coleoptile sections unless auxins are added. This synergism between the gibberellins and auxin^,^^-^^ and the nature of the interaction between the gibberel- lins and other growth regulators such as the kin in^,^' is not fully under- stood. The failure of the gibberellins to simulate the effects of cold treatment or long days in a number of cases (reviewed elsewhere12) may be a consequence of such hormonal interactions in which hormones other than the gibberellins are limiting.The practical uses of the gibberellins in agriculture have been reviewed.38 The gibberellins are used to stimulate swelling of fruits e.g. grapes and tomato; and they are more effective than auxins in inducing partheno- carpic fruit setting in tomato. They can also be used to break dormancy of seeds of e.g. lettuce peach and Douglas fir and to accelerate the germination of barley. Growth of bacteria and fungi is unaffected2' by gibberellic acid which has negligible mammalian 4. Nomenclature The systematic nomenclat~re~~ is based on the trivial name gibbane for the fully saturated tetracarbocyclic system (I) numbered as shown. The 8,9-bridge in (1) is /3 (absolute configuration) and the ring system derived from gibbane by inversion at positions 7 and 9a is called 7a-gibbane.Gibberellic acid (2 ; R = OH) is thus 2,4a,7-trihydroxy-l-methyl-8- methylenegibb-3-ene- 1,10/3-dicarboxylic acid 1 +4a-lactone. Catalytic reduction (steric control) of an 8-methylene group gives epimeric pairs of methylgibbanes in which the absolute configuration at position 8 is not known with certainty. Only one epimer is obtained by chemical reduction (thermodynamic control) of gibberellic acid and is arbitrarily called an 8-methylgibbane the epimer is called an 8-epimethyl- gibbane. The 8-methyl compounds may be expected to have the less 34 Brian and Hemming Nature 1957 179 417; Ann. But. 1958 22 1. 36 Purves and Hillman Physiul. Plant. 1958 11 29. 36 Galston and Warburg Plant Physiul.1959 34 16. 37 Wickson and Thimann Physiul. Plant. 1958 11 62. 38 Wittwer and Bukovac Ecun. But. 1958 12 213. 39 Peck McKinney Tytell and Byham Science 1957 126 1064. 40 Grove and Mulholland J. 1960 3007. 60 QUARTERLY REVIEWS hindered configuration and in this event will be 8cc-methyl compounds. The term gibberellic acid nor-ketone is used to describe the 8-ketone obtained by oxidative removal of the 8-methylene substituent. Trivial names e.g. allogibberic acid (18) and gibberic acid (8) are retained for degradation products in which ring A is aromatic. Within this class the prefix epi is reserved for those compounds in which the 4b- hydrogen atom is /3-oriented. Degradation products in which ring D of gibbane has been opened are named as derivatives of fluorene e.g.9p-carboxy-4ba,5,6,7,8,8a-hexa- hydro- 1 -methyl-7-oxofluorenyl-8a~-acetic acid (24; R = H). The gibberellins are defined as a group of naturally occurring plant hormones containing the tetracyclic system (1). As new gibberellins are isolated they are allotted trivial names in the series gibberellin A . . . . A,. This procedure is adopted because the names gibberellin B and gibberellin C were given by the Japanese workers to allogibberic acid (18) and the 7a-gibbane (34; X = H OH; R = H) respectively of which the former has no plant-growth promoting proper tie^.^^ 5. The gibberellins and their structural relationships The structures and physical properties of the gibberellins are listed in the Table. All are gibbane-10P-carboxylic acids with a 1+4a lactone bridge. Gibberellic acid and gibberellin A have an 8-methylene substituent Compound Structure Formula M .P . ~ [a]; SourceC Gibberellic acid 2; R=OH C1gH2206 233-235” - T92 F Gibberellin A 4; R=R’=OH R”=H c,QH,o,( :::-258 +36 F Gibberellin A2 5 Gibberellin A4 4; R=R”=H R’=OH C,QH,O~ 21k215 - 3 Gibberellin 6 Gibberellin A 2 ; R = H C19H2205 202 f20 F Gibberellin A8 4; R=R=R=OH C1gH240 210-215 +30 P c1QH2606[:&237 + 12 F Gibberellin A 3 C ~ Q H & ~ 260-261 -77 P 222-225 Gibberellin A 4; R=R’=R=H CigH2404 208-211 -12 F aDecomposition point. The alternative values given are for polymorphic forms. dLater work shows this to be a 2,3-epoxide CleH220,. ethanol or methanol. cisolated from fungus (F) or higher plant (P). and d3 double bond in gibberellin A the latter is replaced by a A 2 double bond. Gibberellins A, A, A, and A have the 8-methylene substituent as the only unsaturated centre in gibberellin A 2 this has been saturated by the addition of the elements of water.With the exception of 41 Brian Grove Hemming Mulholland and Radley Plant Physiol. 1958 33 329. GROVE THE GIBBERELLINS 61 gibberellins A, A, and A all have a 2(ax)-hydroxyl group while gib- berellic acid and gibberellins Al A, A, and A8 have a 7(eq)-hydroxyl substituent. Gibberellin A is unique in having a hydroxyl group at posi- tion 3. The position of one hydroxyl group in gibberellin A has still to be located (see however the Table). The relations between the gibberellins were elucidated as follows (for practical convenience the methyl esters were frequently used). Reduction of the 3,4- double bond in gibberellic acidg2 and gibberellin A,g gave gibberellin A and with simultaneous reduction of the 8- methylene substituent dihydrogibberellin A respectively.Reduction with zinc and acetic anhydrideg3 of the 7-hydroxy-8-ketone obtained from gib- berellin A by ozonolysis gave gibberellin A nor-ketone (as its acetyl derivative). The relation between gibberellins A and A has been con- firmedlO by an alternative route. Treatment of gibberellin A with dilute mineral acidg4 gave gibberellin A,; and the action of collidine on the 2-toluene-p-sulphonyl derivative of gibberellin Al gave gibberellin The 7-methyl-8-0~0-7a-gibb-2-ene derived from gibberellin A by the acid-induced rearrangement of rings C/D (see p. 65) proved to be a key compound in relating gibberellin A to gibberellin A and to A,.'' Hydroxy- lation with osmium tetroxide of the 2,3- double bond gave the correspond- ing 7-methyl-8-ketone obtained from gibberellin A,; and catalytic reduction of the 2,3- double bond gave a 7-methyl-8-0~0-7a-gibbane obtained from gibberellin A6 by a reaction sequence in which after rearrangement of rings C/D the remaining hydroxyl group was replaced by halogen and the halogen removed by treatment with Raney nickel.Collidine treatment of the toluene-p-sulphonyl derivative of gibberellin A nor-ketone methyl ester followed by catalytic reduction gave gibberellin A nor-ketone methyl ester.1° 6. The chemistry of gibberellic acid The evidence for the structure and stereochemistry of gibberellic acid is discussed first in paragraphs (a)-(d) and some of the reactions commonly encountered with the gibberellins are mentioned in paragraph (e).Gibberellic acid (2; 'R = OH) formed mono- and di-acetyl derivatives each of which yielded a monomethyl ester as did the acid itself.6 One of the hydroxyl groups is secondary since in the reduction products of gibberellic acid it was oxidised to a ketone :45 the other was considered to be tertiary from the difficulty of acylation. When gibberellic acid was kept with excess of alkali at room temperature a second equivalent was con- sumed this together with the fact that the infrared spectra of the acid and 4 2 Grove Jeffs and Mulholland J. 1958 1236; Takahashi Seta Kitamura and Sumiki Bull. Agric. Chem. SOC. Japan 1957 21 327. 43 Kitamura Takahashi Seta Kawarada and Sumiki Bull. Agric. Chem. SOC. Japan 1959,23 344. 44 Grove unpublished work.45 Cross J. 1960 3022. 62 QUARTERLY REVIEWS its derivatives showed a strong band near 1780 cm.-l indicated the presence of a y-lactone ring. Microhydrogenation revealed the presence of two ethylenic double bonds. This evidence6 showed that gibberellic acid was a tetracarbocyclic dihydroxylactonic carboxylic acid. Further information about the ring HO m~ W O H -OH 1 M . O H CH2 H02C CH2 H02 (2) R'j$-&. Hoc$+oH Me H0* CH2 CH2 H 02C R H02C (4) (5) (See however the Table p. 60.) system was derived largely from the study of two products of acid hydro- lysis allogibberic acid (18) and gibberic acid (8). Selenium dehydrogenation of both allo- gibberic and gibberic acid has earlier given46 a hydrocarbon gibberene which from its ultraviolet spectrum was regarded correctly as a substituted fluorene but was incorrectly given the formula C16H16 and 4-ethyl-5- methylfluorene was suggested as a possible structure.By degradation to the known fluorene- 1,7-dicarboxylic acid Mulholland and Ward47 showed that gibberene was 1,7-dimethylfluorene (11; R = Me) Cl5HI4 and this was confirmed by unambiguous synthesis47 from 2-amino-5-methylbenzoic acid and o-tolylmagnesium bromide via the intermediate benzophenone (b) Structure of gibberic acid. Treatment of gibberellic acid or allo- gibberic acid with boiling dilute mineral acid gave6 gibberic acid (8) C18H2003 m.p. 153-154" or 174-175" [ M I D -7" together with an isomer epigibberic acid (23) m.p. 227-230" or 252-255" [RID + 131". The formation of an ester an oxime and an oxime ester showed that gibberic acid was a keto-acid and a band at 1741 cm.-l in the infrared spectrum indicated that the ketone group was present in a five-membered ring.Microhydrogenation showed the absence of ethylenic double bonds but the ultraviolet spectrum (Amax265 274 mp; log E 2-56 2-47) revealed the presence of a benzenoid ring. The presence of the hexahydrofluorene nucleus in the tetracarbocyclic system (8) was established by stepwise d e g r a d a t i ~ n ~ ~ ~ ~ via the a-diketone (9) and the tricarboxylic acid (12) to 1,7-dimethylfluorene and by oxidation of gibberic acid to benzene- 1,2,3- tricarboxylic The substitution pattern of the five-membered ring (a) Structure of gibberene. (17). 46 Yabuta Sumiki Aso Tamura Igarashi and Tamari J. Agric. Chem. SOC. Japan 47 Mulholland and Ward J. 1954 4676. 48 Cross Grove MacMillan and Mulholland J.1958 2520. 1941 17 975. GROVE THE GIBBERELLINS 63 containing the ketone group was deduced from ultraviolet absorption studies on the cc-diketone (9) which was found to have no enolisable hydrogen atom. The position of the carboxyl group was establi~hed~~ by dehydrogenation of the methyl ester of (9) to methyl 1,7-dimethyl- fluorene-9-carboxylate identical with a synthetic specimen prepared by carboxylation of the 9-lithium derivative of 1,7-dimethylfluorene followed by methylation. 1. at? Me 2 __c 3 c- h - - - I (91 0 0 { 3 M e H02C C0,H 0 2) AcHN 07) Me CH,-Ci.H CMe I C q M e C0,Me (16) Me CHART 1 Reagents 1 KMnO,. 2 SeO,. 3 Pd-C. 4 Se. 5 H,O,-NaOH. 6 CrO,. 7 KMn0,-Mg (NO,),. The position of the -CH,CO- bridge followed from a second stepwise degradati~n~~ in which the key compound was a ketone gibberone C1,Hl,O (10) obtained directly from gibberic acid by dehydrogenation over palladium-charcoal or indirectly by decarboxylation of dehydro- gibberic acid (7) (Amax.260 269 mp; log E 4.14 4-09) a permanganate oxidation product of gibberic acid. Oxidation of gibberone with chromic oxide gave the 1-oxoindanespirocyclopentanone (13) which was stable to hydrolysis and was not therefore a 1,2’-diketone. Further oxidation of the indanone (1 3) gave 3-methylphthalic acid (14) and /3-methyltricarballylic acid (15); and opening of both five-membered rings by second-order Beckmann rearrangement ofthe 2’-oximino-compoundgave after hydrolysis 64 QUARTERLY REVIEWS and methylation two diastereoisomeric tetramethyl esters (16) m.p.83- 84" ( a ) D -6' and m.p. 47-48" (a)D +12". The structure (16) of the esters was confirmed by the unambiguous synthesis of their race mate^.^^ This synthesis completed the elucidation of the structure of the indanone (13) and consequently of gibberone (10). It followed that the methylene carbonyl bridge in gibberic acid must be attached as in (8). Racemic gibberone has recently been synthe~ised.~~ (c) Structure and stereochemistry of allogibberic acid. With cold dilute mineral acid both gibberellic acid6v41 and the intermediate gibberel- lenic acid (19)51 gave allogibberic acid (18) C1SH2003 m.p. 201-203" ( a ) D -84" and 1 mol. of carbon dioxide With hydrazine hydrate gibberellenic acid gave both allogibberic acid and an isomer epiallo- gibberic acid (20),40 m.p. 244" ( a ) D +87".Allogibberic acid (Amax- 266 274 mp; log 6 2.50 2.35) contained a benzenoid ring and an ethylenic double bond present in an exocyclic methylene grouping since ozon01ysis~~ gave formaldehyde and a nor-ketone Cl7Hl8O4 (21). The carboxyl group was attached in the same position as in gibberic acid since the methyl ester was isomerised with acid to methyl gibberate. The third oxygen atom was present as a hydroxyl group (vmax. 3460 cm.-l) considered to be tertiary because of the difficulty of acylation and the failure to oxidise dihydro- allogibberic acid to a ketone. These facts together with the following evidence established allogibberic acid as a tetracarbocyclic hydroxy-acid (18) the presence of an a-ketol system in a five-membered ring in the nor-ketone (21) was shown by the infrared spectrum (vmax.1742 cm.-l) and by oxidation with sodium bi~muthate~~ to a tricyclic dibasic keto-acid C17Hls05 (24; R = H) in which the carbonyl group was contained in a saturated six-membered ring. The position of this carbonyl group and hence of one point of attachment of the five-membered ring in allogibberic acid was ascertained by selenium dehydr~genation~~ to 8-methylfluoren-2- 01 (1 1 ; R = OH) whose structure was proved by ~ynthesis.~~ With aqueous alkali the keto-ester (24; R = Me) gave the acid (24; R = H) together with a new dibasic keto-acid C17H1805 epimeric at position 9. With acetic anhydride both dibasic acids gave the same cis- anhydride but hydrolysis of the anhydride regenerated only the acid (24; R = H) derived directly from allogibberic acid.The 10-carboxylic acid substituent and 8,9-two-carbon bridge are therefore cis in allogibberic acid.40 Since catalytic redwtion of dehydrodihydroallogibberic acid (26) a 4b(5)-ene related to dehydrogibberic acid (7) will take place from the less-hindered side of the molecule opposite to the 10-carboxyl substituent *O Morrison and Mulholland J. 1958 2536. 5 0 Loewenthal Proc. Chem. SOC. 1960 355. 61 Moffatt J. 1960 3045. 5a Mulholland J. 1958 2693. 53 Morrison and Mulholland J. 1958 2702. GROVE THE GIBBERELLINS 65 and 8,9-bridge7 and since this process regenerated the original stereo- chemistry at position 4b,40 it followed that rings B/C were trans-fused in allogibberic acid. The absolute configuration (1 8) followed from measure- ments of optical rotatory dispersion on the keto-ester (24; R = Me).40754 In addition to racemisation at position 9 the ester (24; R = Me) on CHART 2 Reagents 1 H+.2 N2H4,H20. 3 0,. 4 NaBiO,. 5 OH-. treatment with base underwent*O an intramolecular Claisen-type condensa- tion at position 6 this was followed by fission of the 6,7-bond and hydro- lysis liberating the 4ba78aa,9cc-isomer (25) of the acid (24; R = €€). Epiallogibberic acid which gave epigibberic acid (23) on acid treatment was chemically similar to allogibberic acid and yielded the 4b/3,8a/3,9p- enantiomer (22) of (25) on ozonolysis followed by fission of the resulting cc-ketol. It followed that epiallogibberic acid differed from allogibberic acid only in configuration at position 4b.40 The acid-catalysed rearrangement of these acids takes place by a Wagner- Meerwein m e c h a n i ~ r n ~ ~ .~ ~ via the cation (27) and in the resulting gibberic acids the 8,9-bridge has the opposite configuration (a) to that which it occupied in the allogibberic acids. The racemate of the ester (28) a reduction product of the ketone (22) has been ~ynthesised.~~ 54 Stork and Newman J. Amer. Chem. Soc. 1959 81 3168. 55 Grove MacMillan Mulholland and Turner J. 1960 3049. 56 House Paragamian and Wluka J. Amer. Chem. Soc. in the press. 3 66 QUARTERLY REVIEWS (d) Structure of gibberellic acid. The position of the carboxyl sub- stituent in gibberellic acid was the same as in allogibberic acid since the methyl ester with acid gave methyl gibberate. An early suggestion5' that the two acids had the same B/C/D structure was by the ozono- lysis of methyl gibberellate which took the same course as the ozonolysis of methyl allogibberate giving formaldehyde and ultimately a keto-acid (32; R = H).The methyl ester (32; R = Me) of the latter gave the ester (24; R = Me) with acid and the formation of allogibberic acid from gibberellic acid therefore involves only the aromatisation of ring A. Ring A of gibberellic acid must accommodate the methyl group which appeared at position I in the fluorene degradation products the saturated y-lactone ring and the secondary hydroxyl group which was shown to be allylic by oxidation with manganese of methyl gibberellate to the d3-2-one (29) (Amax. 228mp; log E 3-99). Catalytic hydrogenation of the ketone (29) afforded the 2-ketones (3 l) epimeric at position 8 also obtained HO w methyl gi bberellate 0 MeCqR 4 (R=H) Me02C (30) 0 1 CHART 3 Reagents 1 MnOz.2 Hz/Pt. 3 H2/Pd. 4 Cr03. 5 03. by oxidation with chromic oxide of methyl tetrahydrogibberellate and its 8-epimer. The position of the allylic hydroxyl group was e ~ t a b l i s h e d ~ ~ ~ ~ by selenium dehydrogenation of the ketones (35; R = H X = H2 Y = 0) and (34; R = Me X = 0) both derived from gibberellin A, to l-methyl- fluoren-2-01 and 1,7-dimethylfluoren-2-01~~ respectively. The position of 57 Cross Grove MacMillan and Mulholland Chem. and Id. 1956 954. 58 (a) Seta Takahashi Kitamura Takai Tamura and Sumiki Bull. Agric. Chem. SOC. Japan 1958 22 61 ; (6) Seta Takahashi Kitamura and Sumiki Bull. Agric. Chenz. SOC. Japan 1958 22,429. 59 Cross and Melvin J. 1960 3038. GROVE THE GIBBERELLINS 67 the lactone bridge was deduced from the high yield of acidic hydrogen- olysis products on catalytic reduction of methyl gibberellate which suggested an allylic lactone system and from the formation of a hetero- annular conjugated diene gibberellenic acid51 (19) (Amitx.253 mp; log E 4.35) from gibberellic acid in aqueous solution. It was concludedGo that 0 0 gibberellic acid has structure (2; R = OH) and this conclusion was supported by nuclear magnetic resonance studiesG1 An alternative ring A structure (36) has been s~ggested~~b but is inconsistent with much of the foregoing evidence. The quasi-axial nature of the 2-hydroxyl substituent in gibberellic acid was deduced from the base-catalysed isomerisation of gibberellin Al to the more stable 2(eq)-hydro~y-epimer.~~ The rotary dispersion curvc s for the keto-esters (35; R = Me X = 0 Y = H,OH) and (24; R = Me) were almost identicalG3 and significantly different from that for the methyl ester of the keto-acid (22) obtained from epiallogibberic acid.Gibberellic acid therefore probably has the same trans-^/^ ring fusion as allogibberic acid but more evidence on this point would be desirable. Oxidation and de- carboxylation of the 8-epimethyl acid (30; R = H) obtained by hydrogeno- lysis of methyl gibberellate gave the keto-ester (33) the rotatory dispersion curve of which showed a negative Cotton effect,63 but this fact is in itself in- sufficient to determine unequivocally the nature of the A/B fusion in the acid (30; R = H). The a~signment,~~ from nuclear magnetic resonance studies of a trans-orientation to the hydrogen atoms at positions 10 and 10a now appears to be unwarranted.61 Some controversy arose over the interpretation of the physical evidence for the orientation of the 1+4a lactone bridge.64 The chemistry of the ester (30; R = Me)65 indicated that the 2-hydroxyl substituent was equatorial and that consequently hydrogenolysis of methyl gibberellate involved inversion of configuration at position 4a.This chemical evidence is decisively in favour of a /%orientation for the lactone bridge but a rigorous proof of the configuration at position 10a is still required. Cross Grove MacMillan Moffatt Mulholland Seaton and Sheppard Proc. Chem. SOC. 1959,302. G1 Sheppard J. 1960 3040. G2 Cross Grove and Morrison J 1961 in the press. 63 Cross Grove McCloskey Mulholland and Klyne Chem. and Ind. 1959 1345. 64 (a) Stork and Newman J .Arner. Chern. SOC. 1959,81,5518; (6) Edwards Nicolson Apsimon and Whalley Chem. and Ind. 1960 624. G5 Grove and McCloskey unpublished work. 68 QUARTERLY REVIEWS (e) General chemical reactions of the gibberellins. Some of the more important reactions encountered with the gibberellins are indicated below. (i) 2(ax)-Hydroxygibbane 1 +4a-lactones are epimerised in dilute alkali to the more stable 2(eq)- hydroxy-compounds and a retroaldol mechanism via the intermediate (37) has been suggested.62 Under the same conditions62 2(ax)-hydroxygibb-3-ene 1+4a-lactones undergo an allylic type of rearrangement to gibb-4-ene 1 +3-lactones without concomitant epimerisation of the hydroxyl substituent. 0 (ii) In the presence of a l+4a-lactone bridge catalytic reduction of a 3-ene precedes that of an 8-methylene substituent ; but the 8-methylene group is reduced before a 4-ene group in gibb-4-ene 1+3-lactones.The latter are difficult to reduce and hydrogenolysis of the lactone pre- dominates. 62 (iii) 2-Hydroxyl groups are smoothly oxidised to 2-ketones only in gibbanes in which an 8-methylene substituent has been reduced or eliminated.45 The 2-ketones are reduced by alkali-metal hydrides to the 2(eq)-hydroxy-~ompounds.~~ (iv) In gibbanes 2(ax)-hydroxy-substituents are readily eliminated,17c either directly under the influence of nucleophilic reagents or via the toluene-p-sulphonate giving gibb-2-enes. (v) Compounds containing the C/D partial structure (38; R = OH) undergo Wagner-Meerwein rearrangement with acid to give ketones of partial structure (39).55 Under the same conditions the elements of water are added to the 8-methylene group in compounds (38; R = €3).(vi) Among degradation products in which ring A is aromatic compounds having a 10-methoxycarbonyl substituent in the less stable configuration undergo base-catalysed racemisation at this centre. Only compounds in which the 4b-hydrogen atom is trans to the 10-carboxylic acid substituent are oxidised by permanganate to 4b(5)-enes in some cases neighbouring groups cause steric inhibition of this r e a ~ t i o n . ~ ~ * ~ ~ The configuration at position 10 also determines the steric course of the catalytic reduction of 4b(5)-enes hydrogenation occurring trans to a 10-carboxyl or 10-methoxy- carbonyl substituent. 7. Biogenesis of the gibberellins Inspection of structure (2; R = OH) showed that it could have arisen by a variant of the processes leading to the tricyclic diterpene skeleton (40) in which (i) the 17-carbon atom had been lost (ii) contraction of ring B to GROVE THE GIBBERELLINS 69 a five-membered ring had occurred with extrusion of a carboxyl group and (iii) formation of the phyllocladene-type of bridged-ring structure had occurred from ring c and its substituents according to the scheme suggested by Wenkert.6s The correctness of these speculations was e~tablished~~ by the degradation of gibberellic acid obtained from G.fujikuori grown on 0 0 [carboxy-14C] acetic acid [*indicates labelled atom in (40)] and [2-14C] mevalonic lactone (42) ; the results of the degradation were consistent with the labelling pattern (41) expected from the usual mode of incorporation (40) of these precursors in a tricyclic diterpene.The stereochemical implications of this important work have been the subject of much peculation.^^^^^ Two points are clear however first oxidation at position 7 in the gibberellins is subsidiary to the main biogenetic process and this may also be true of the hydroxylation of ring A. Secondly since the 17-carbon atom is lost and contraction of ring B is likely to involve an intermediate 10-0x0-derivative it is not necessary to postulate that gibberellic acid is derived from other than the normal (1 1 a) type of trans-anti-trans-hydrophenanthrene precursor (40). 8. The gibberellins in diterpene chemistry Many terpen~ids~~ and more commonly compounds with isoprenoid side chains70 are known among fungal metabolic products but few isolated.Of these only the groups related diterpenoids have been OH (4 4) to rosenon~lactone~~ (43) and gibberellic acid have been extensively investi- gated. Few diterpenes have the 2-hydroxy-substituent which occurs frequently 66 Wenkert Chem. and Ind. 1955 282. 67 Birch Ricjcards and Smith Proc. Chem. SOC. 1959 192; Birch Rickards Smith 68 Djerassi Cais and Mitcher J. Amer. Chem. SOC. 1959 81,2386. 69 (a) Haagen-Smit Progr. Chern. Org. Nat. Prod. 1955 12 1 ; (b) Jones and Halsall 'O Birch English Massy-Westropp and Smith J. 1958 369. 71 Harris Robertson and Whalley J. 1958 1799 1807; Freeman Morrison and Harris and Whalley Tetrahedron 1959 7 241. ibid. p. 44. Michael Biochem. J. 1949,45 191. 70 QUARTERLY REVIEWS in the gibberellins. Among those possessing this feature several e.g.dar~tigenol~~ (44) and andr~grapholide~~ (45) also have the abnormal antipodal trans-A/B stereochemistry which however is not invariably associated with- the presence of a 2-hydroxyl group e.g. eperuic (46) and cassaic (47). The axial configuration of the 2-hydroxyl substituent in the gibberellins is also unusual although similarly situated axial hydroxyl groups have been recorded in the triterpenoids e.g. in the mould product polyporenic acid A.69b The 8,9-two-carbon bridge in ring c carrying an 8-methylene substituent is found in the tetracarbocyclic diterpenes related to phyllocladene (48).76 The structure of steviol been amended to (49) and the steviol -+ isosteviol rearrangement is now recognised as analogous to the acid- induced conversion of allogibberic into gibberic acid.Contraction of ring B in a diterpene skeleton such as (40) may be assumed to involve a 9,lO-dioxygenated intermediate and xanthoperol (51)78 provides an example of a diterpenoid 9,lO-diketone. Benzilic acid rearrangement of the enantiomer of methyl-9,1 O-dioxopodocarpa- 5,7,13(14)-trien-16-oate has been shown to give the hydrofluorene- carboxylic acid (51).79 OH The gibberellins differ from all other tri- and tetra-cyclic diterpenoids in that selenium dehydrogenation gives substituted fluorenes instead of substituted phenanthrenes. Apart from this fact the more characteristic reactions (see Section 6e) are associated with the 2-hydroxygibb-3-ene 1 +4a-lactone and 2-hydroxygibbane 1 +4a-lactone systems for which there are no exact analogies among natural products. 72 Diara Asselineau and Lederer Bull. SOC. chim. France 1959 693. 73 Cava and Weinstein Chem. and Ind. 1959 851; Chan Haynes and Johnson 74 Barltrop and Bigley Chem. and Ind. 1959 1447. 76 Turner Herzog Morin and Riebel Tetrahedron Letters 1959 No. 2 7. 78 Briggs Cain Davis and Wilmshurst Tetrahedron Letters 1959 No. 8 13. 77 Dolder Lichti Mosettig and Quilt J. Amer. Chem. SOC. 1960 82 246. 78 Bredenberg Acta Chem. Scand. 1960,14,385. Chem. and Ind. 1960,22. Grove and Riley J. 1961 1105.