STEREOCONTROL IN THE SYNTHESIS OF NONACTIN
A thesis presented by Hiten Sheth in partial fulfilment of the requirements for the award of the degree of Doctor of Philosophy of the University of London
Department of Chemistry, Imperial College, London SW7 2AY October 1982 ACKNOWLEDGEMENTS
I wish to thank Dr A.G.M. Barrett for his advice, guidance and friendship during this project. I am grateful to Ashley Fenwick,
Barrett Kalindjian and Mark Russell for friendship and help.
A special thanks to Mrs Denise Elliott for typing this thesis and to Tara, Gunvant, Bharat and Sonal for encouragement and support.
Thanks are also due to the SERC for a studentship, and to
W.R. Grace and Co., Research Division, Columbia, Maryland for support.
A special thanks to Mrs. Maria Serrano-Widdowson for helping out at times of crises. ABSTRACT
The racemic and chiral syntheses, of nonactic acid, in the
literature are reviewed.
Our approaches to the synthesis of racemic nonactic acid
utilized activated derivatives of pentane-1, 2R(S), 4S(R) triol
(101); 2R(S), 4S(R) dihydroxypentanoic acid (133) and 2R(S),
4S(R) dihydroxypentanal (119).
The key intermediate, 3R(S) acetoxy-5S(R)-methyltetra- hydrofuran-2-one (99), was prepared from the double elimination of
2,3,5-tri-0-acetyl-D-ribonolactone to give the desired 3-acetoxy-
5-methylene-2,5-dihydrofuran-2-one. Steric-approach controlled hydrogenation yielded the tetrahydrofuran-2-one (99). The lactol
(119), derived from the di-isobutylaluminiumhydride reduction of the tetrahydrofuran-2-one (99) was condensed with the ylide, ethoxycarbonylmethylenetriphenylphosphorane, the resultant enoate being hydrogenated and acidified to yield the lactone, 5S(R) -
[2S(R) hydroxypropyl ] tetrahydrofuran-2-one (120). Protection of the hydroxyl group as the t-butyldimethylsilyl ether and condensation with t-butyl 2-lithiopropanoate, on acidification gave the E- enoate, 2S(R)-[2S(R)-(t-butyldimethylsilyloxy) propyl]-5-[l-(t- butyloxycarbonyl) ethylidene] tetrahydrofuran (122). Steric-approach controlled hydrogenation gave the racemic nonactic acid derivative, t-butyl 8-0-t-butyl-dimethylsilylnonactate (123).
i. As proof of structure, compound (123) was converted to the known diol, 2S(R)-[2S(R) hydroxypropijl]-5R(S)-[2-hydroxy-lS(R)-
(methyl)ethyl]tetrahydrofuran, which was identical with authentic material obtained from the reduction of nonactin.
Other routes which were explored, but proved to be unsuccessful, included the reduction of the compound (99) to the triol (101) and after activation, gave the derivatives 2S(R), 4R(S)- bis(t-butyldimethylsilyloxy)-l-iodopentane (105) and 2S(R) -
[2R(S )-(t-butyldimethylsilyloxy)propyl] oxirane (112(a)). Neither
(105) nor (112(a)) reacted with the dianions of methyl 2-methyl-
3-oxo-butanoate (106(a)), and methyl 2-methyl-3-(methylamino)but-
2-enoate (107(a)).
Condensation of the dianion of t-butyl 2-methyl-3-oxo- butanoate (108(a)) with the lactone (99) and elaboration of the adduct gave the desired compound, 2-[3-(t-butyloxycarbonyl)-2- oxobutyl]-3R(S) acetoxy-5S(R)-methyltetrahydrofuran (117(a). Ring fragmentation, via its dianion, of the tetrahydrofuran-4-butanoic acid (117(a)) failed.
ii. CONTENTS
ABSTRACT i - ii
INTRODUCTION 1-2
SYNTHETIC APPROACHES TO THE NACTINS 3-16
CHIRAL SYNTHESES OF (+) AND (-) METHYL NONACTATE 17 - 23
CONSTRUCTION OF NONACTIN FROM NONACTIC ACID 24 - 27
RESULTS AND DISCUSSION 28 - 75
EXPERIMENTAL 76 - 143
REFERENCES 144 - 148
APPENDIX "A Concise Synthesis of (±)-t-Butyl 8-O-t-Butyldimethyl- silylnonactate" A.G.M. Barrett and Hiten G. Sheth J.C.S. Chem. Commun., 1982, 170. INTRODUCTION
The polyether antibiotics 1 are a group of compounds possessing
the ability to form lipid soluble complexes that mediate cation
transport across lipid barriers. This characteristic ion-bearing
property led to their being named ionophores. The macrotetrolides (1)
are all constructed from four tetrahydrofuranyl hydroxyacids linked
together as lactones. The key feature of the stereochemistry in this
particular class of ionophores is that two of the hydroxyacids are of
opposite configuration to the alternating pairs of acids. Thus
nonactin is a meso compound comprising of alternating enantiomers
of nonactic acid.
Nonactin and its homologues, the macrotetrolides, have been
isolated^,20 fr0m several Strentomvces cultures. These are
neutral ionophores which mediate an active uptake of cations
including potassium ions into mitochondria. The passive
diffusion of excess potassium ions across membranes, is probably the primary effect of the many different biological activities reported
for the polyethers. The actins afford one to one complexes with many alkali/ alkaline earth metal ions. Specifically, nonactin
+ + + + + exhibits a selectivity sequence of NH4 >K ^Rb >Cs >Na > Ba2 + 21.
1. R1 = R2=R3=R4= Me
Nonactin R1 = R2 = R3 = R4 = Me Monactin r2 = Et, RL = R3 = R4 = Me Dinactin r2 = R4 = Et, Rx - R3 = Me Trinactin r2 = R3 = R4 = Et, Ri = Me Tetranactin Rx = R2 = R3 = R4 = Et
2. Synthetic Approaches to the Nactins*^>^
Degradation^ of nonactin (and tetranactin) by base hydrolysis
yielded racemic nonactic acid (and tetranactic acid respectively).
Thus routes to the nactins require the synthesis of (it) or (+) and (-)
nonactic acid (or tetranactic acid) followed by oligomerization in the
correct sequence and lactonization to the macrolide.
The synthetic precursor to nonactin, the linear tetramer of subunits has been assembled both with^*^ and without ^ controlling the alternating chirality required. A number of syntheses (including the synthesis of each separate enantiomer^>^>^»®) of the nonactic acid subunit have
been reported.^,18# None, with the exception of Bartletts'^ is to any great degree stereoselective at each and every stage.
2-epi
OH OH
8-epi
3. Controlling the configurations at C-2 and C-8 are the real
synthetic problems, since the cis stereochemistry of the tetrahydrofuran
ring (C-3 and C-6) could be established by hydrogenation of a suitably
functionalized 2,5 disubstituted furan.
In the first synthesis of nonactic acid to be reported*0,
(Scheme 1), no control of the two extracyclic'chiral centres was attempted. Beck and Henseleit*** used methyl furan-2-acetate (6) which after monomethylation, was condensed with 3-methyl-3-buten-2- one under Lewis acid conditions to give a 2,5 disubstituted furan
(7). Catalytic hydrogenation of (7) over rhodium on alumina yielded the cis substituted 2,5 tetrahydrofuran ring as a mixture of 4 stereoisomers (9). Subsequent conversion of the acetyl group into a hydroxyl group afforded methyl nonactate (2) accompanied by the three diastereoisomers (3), (4) and (5).
Several of the other routes intersect at various intermediates, and take advantage of two observations for controlling stereochemistry at the C-2 and C-8 centres. Firstly, base-catalyzed epimerization with a cis tetrahydrofuran ring, favours the threo relationship between C-2 and C-3 positions. (This equilibration proceeds without ring opening since the cis stereochemistry of the tetrahydrofuran ring is preserved.)
Gerlach** reported that the equilibration of the 8-keto derivatives
(14) and (15) favoured the threo isomer by a ratio of 4:1. Schmidt
et al.2>5 and white et al.*^ were able to enrich methyl nonactate (2) using sodium methoxide in methanol over the C-2 epiraer (3) by only
3:2. Secondly, catalytic hydrogenation or L-selectride reduction*1»
4. Scheme 1 Beck and Henseleit 197110
0
•o OMe OMe (0
60%
•v
Me * OMe
iv 73%
UAC OH 0 CO,Me m OMe H °H i H °H (0+(j0+00+(s)
(i) NaH/Mel; o (ii) 3-methyl-3-buten-2-one, BF3.0Et2/60 C;
(iii) H2/Rh-Al203, MeOH;
(iv) CF3CO3H,CH2CI2,A ;
(v) OH " ;
(vi) H + ; (vii) CH2N2 5. of the 8-keto-derivative (14) gave 8-epi-methyl nonactate (4) as the
predominant diastereomer (9:1). The natural isomer was then obtained
by Walden inversion of configuration. White*2 suggested that the
stereoselectivity observed resulted from the coordination of boron with both the ketone carboxyl and ether oxygens as shown; followed
by hydride attack from the least hindered side in the complex.
OMe
8-epi product
The strategy adopted by Gerlach and Wetter** suffered from a non stereoselective introduction of the two chiral centres of the side chains. They condensed acetonylfuran (10) with the chloronitrone
(11) to yield a furan nitrone which on acid hydrolysis gave the aldehyde (12), in 79% yield. Oxidation and esterification gave the ketoester (13). Catalytic hydrogenation over rhodium on alumina gave the 8-keto-tetrahydrofurans (14) and (15). Using base catalyzed epimerization they enriched the threo composition of their product to 4:1. Subsequent borohydride reduction afforded methyl nonactate (2) and the 8-epi compound (4) (1:1), without any selectivity at C-8.
6. Scheme 2 Gerlach and Wetter*!
OMe 1:1
(i) AgBF4, CH2C12, -20°C; (ii) HC1/H20; (ill) Cr03;
(iv) CH2N2; (v) Rh/Al203, MeOH; (vi) NaOMe/MeOH;
(vii) NaBH4/H20/Me0H
7. Schmidt et al.2>5,6 synthesized methyl nonactate as shown in
Scheme 3. By combining both the selectivity of the ketone hydrogenation and the ability to epimerize the C-2 centre, they then developed a method for enriching in methyl nonactate (2), a mixture of equal amounts of all four C-2 and C-8 diastereoisomers. Propylene oxide was condensed with 2-lithiofuran (16) to give the alcohol (17).
Acetylation of (17) followed by a Vilsmeier reaction afforded the aldehyde (19). This was converted to the vinylfuran (20) by a
Wittig reaction. Hydroformylation employing a rhodium catalyst gave the aldehyde (21). Subsequent oxidation with the silver oxide, base hydrolysis, esterification and hydrogenation over rhodium on alumina gave a mixture of the four diastereoisomers (2), (3), (4) and (5).
After a lengthy process involving oxidation, reduction, Walden inversion and esterification; they enriched their sample in methyl nonactate (2).
8. 2 5 6 Scheme 3 Schmidt et al- 1975 > »
OH V 63% 0 (14) («)
ii 60% OAc OAc ill X-CACHO 0 81 % (19) (18)
57% OAc 5 V CHO 0 81% (20)
C.0 Me 2 OMe
(0 + (3 )+(it )+(5) XI,XII,XIII, xiv, xv 25 : 25 : 25 : 25
(2)+(3) + (4) + (5) xvi 37: 33: 21 : 10 (0 (3) 1-5:1
(i) 2-methyloxiran; (viii) CH2N2;
(ii) AC20; (ix) Rh/Al203, H2; (iii) DMF, POCI3; (x) Cr03; (iv) Ph3P « CH2; (xi) H2-Ra/Ni, 100 atm; (v) H2QL10 atm)/C0(90 atm), (xii) TsCl/py; (xiii) KOAc/DMSO, 70°; (PhjP) 33 RhCl (vi) Ag20; (xiv) KOH, MeOH (vii) KOH/MeOH; (xv) CH2N2; (xvi) NaOMe/MeOH
9. A further synthetic variant on the Schmidt route was reported by White et al.*2 (Scheme 4) 2-lithiofuran was condensed with propylene oxide to give the alcohol (17). Reaction with acetic- anhydride yielded the ketoacetate (22). Hydrogenation over rhodium on carbon gave the cis tetrahydrofuran (23) and established the correct stereochemistry at C-3 and C-6. Oxidation followed by a
Wittig reaction gave the methylene acetate (24). Hydroboration followed by chromium oxidation, gave without any selectivity, the C-2 epimeric acids (25) and (26) in a 1:1 ratio. Lithium tri(sec-butyl)borohydride reduction of the 8-keto compound
(25) followed by esterification afforded predominantly the 8- epiester(4) in a 9:1 ratio. Inversion of configuration yielded the methyl nonactate derivative (27).
10. Scheme 4 White §£ al.12 1976
OH , 1 (") 0 -—"LI 98% XJJ
83%
-I'
III
96%
VI, IV CO2H 90% H + H * 1:1 0 C0,H 0 H H = vii, viii
OCOPh OH IX C02Me * •q'i v OO^Me H °H 90% H H = (27) 00:(0 9:1
(i) 2-me thy1oxi ran; (v) Ph3P=CH2; (ii) (CH3CO)20, BF3.OEt2; (vi) B2H6, THF; (iii) Rh/C, H2; (vii) L-Selectride; (viii) MeOH, BF3. OEt2;
(iv) Cr03/H /Me2CO; (ix) Et02CN=NC02Et, Ph3P, PhC02H.
11. The second synthesis of methyl nonactate by Gerlach and Wetter**
(Scheme 5) utilized the diketone (28) obtained from the condensation of the dianion of 2,4-pentanedione with allyl bromide. Reduction of
(28), using sodium borohydride, gave the diols (29) and (30) in the ratio 3:2. The configuration of the major product (29) was inverted as shown, to give (30) with the stereochemistry required for C-6 and C-8 in methyl nonactate. Acetylation, ozonolysis and a
Wittig-Horner condensation on (30), gave the enoate (31). Base hydrolysis of the acetate moiety and subsequent cyclization afforded methyl nonactate (2) as the major product (60%).
The second synthesis of methyl nonactate by White et al.*2
(Scheme 6) utilizes the oxyallyl intermediate produced by means of a zinc/copper couple on the dibromoketone (32), and quenching with furan. The cycloadduct (33), after hydrogenation over palladium on carbon, underwent a Baeyer-Villiger oxidation to give the lactone (34).
Methanolysis of the lactone (34) produced the hydroxyester (35), with the desired cis stereochemistry at C-3 and C-6 but with configuration at C-2 corresponding to the epinonactate series. Homologation of the ethyl chain of (35) to 2-hydroxypropyl group of (2) was achieved by converting the hydroxyl group to the xanthate. Pyrolysis yielded the olefin (36). Hydroboration with disiamylborane followed by an oxidative work-up gave the alcohol (37). Loss of configurational homogeniety at C-2 of (37) was presumed to have occurred concurrent with the hydroxide-promoted oxidation of the alkylborane from (36).
12. Scheme 5 Gerlach and Wetter11 1974
70%
C0 Me •CO,Me 2
(0 Major product
(I) KNH2/ NH3; (vii) KOH/MeOH; (ii) Allyl Bromide (viii) AC20, pyridine;
(iii) NaBH4, MeOH, H20; (ix) 03; (iv) Separate , xC02Me (v) TsCl; (x) (EtO)2 P = C N (vi) NaOAc; CH3 (xi) KOH/MeOH/CH3CN
13. Scheme 6 White ££ al-12 1976
CO?Me
C0 Me OHC^ A n^ 2 H H - (38+ ) separate
C02Me H " H (39)
(i) Zn/Cu; (vii) A , 220°; (ii) Furan; (viii) (Sia)2 BH; (iii) H2, Pd/C; (iv) CF3CO3H; (ix) H202/0H " ; (v) NaOMe; (x) Cr03/py; (vi) NaH, CS2, Mel;' (xi) MeMgl, Et20
14. Collins oxidation of (37) yielded an easily separable mixture of epimeric aldehydes (38) and (39) in the ratio 1:1. Treatment of (38) with methyl-magnesium iodide produced methyl nonactate (2) and the 8-epi product (4) without any stereoselectivity.
A highly efficient and stereoselective synthesis of racemic methyl nonactate (2) was reported by Bartlett^ (Scheme 7). Dimethyl 1,
7-octadien-4-yl phosphate (41) was epoxidized stereo-and regio- specifically to give (42) in 63% yield. Lithium aluminium hydride reduction provided the erythro diol (43), which was acylated and oZonolyzed to give the aldehyde (44). A titanium tetrachloride catalyzed aldol condensation and subsequent Jones oxidation afforded the ketoester (45). The tetrahydrofuran ring was generated by acetate cleavage and dehydration to give the 8-epi-E-dehydro- nonactate (46). Hydrogenation afforded the 8-epimethyl nonactate (4).
Inversion of configuration at C-8 using White1s*2 procedure afforded
(+) methyl nonactate (2). The key feature in control of stereochemistry was the use of the cyclic phosphate ester intermediate to enable 1,3 asymmetric induction, in the formation of epoxide (42). This reaction, though setting up the stereochemistry at the would be centre C-6, left an inherent weakness - an erythro configuration and therefore the wrong stereochemistry (later to be inverted) at the would-be centre C-8 in methyl nonactate (2). Secondly, steric-approach controlled hydrogenation of enoate (46) established the stereochemistry at centres C-2 and C-3. Bartlett thus showed good overall control in introducing the four chiral centres.
15. Scheme 7 Bartlett and Jernstedt9 1980
C02Me
C02Me H H 85:15 at C-2 95% xiii.xiv
(0
(i) KH; (viii) MeCH=C(OTMS)OMe, TiCl4;
(ii) (Me0)2 P0C1; + (iii) I2, CH3CN; (ix) Cr03/H /Me2CO; (iv) NaOMe/TttF; (x) K2C03/Me0H;
(v) LiAlH4, Et20; (xi) (C02H)2/A; (vi) (CH3C0)20, py; (xii) Rh/A1203,H2, 3-5 atm; (vii) 03, -78°C; (xiii) Et02CNNC02Et, Ph3P, PhC02H (xiv) NaOMe/MeOH
16. Chiral Syntheses of (+) and (-) Methyl Nonactate
Schmidt et al.5>6. carried out the first chiral synthesis,
of both (-) (2R, 3R, 6S, 8S) and (+) (2S, 3S, 6R, 8R) methyl nonactates
from (-) (S) propylene oxide. Using the reaction sequence outlined
in Scheme 3, and preserving the original chiral center, they managed
to obtain (-) methyl nonactate, isolated by column chromatography in
25% yield. The mixture of the other three diastereoisomers (2S,
3R, 6S, 8S))(2R, 3S, 6R, 8S) and (2S, 3S, 6R, 8S) were converted by Walden inversion (Scheme 3) into a mixture from which (+) methyl nonactate (2S, 3S, 6R, 8R) was isolated in approximately 25% yield by chromatography. Thus both the laevo and dextro rotatory nonactates were accessible by asymmetric synthesis from a single optically active starting material.
The second chiral synthesis of both (+) and (-) methyl nonactate, was carried out by Fraser-Reid and Sun^, starting from the same derivative of D-ribose (47). They took advantage of the fact that
C-glycofuranosides with an activated methylene group at C-l of the glycosyl residue, were amenable to epimerization via a retro-Michael,
Michael addition sequence under base treatment (Schemes 8(a) and 8(b).
The known aldehyde (47) was converted into the ketone (48) via a Wittig reaction followed by acid hydrolysis. Raney-nickel hydrogenation gave predominantly the (S) alcohol (49) in a ratio of 9:1; the minor isomer being chromatographed out. Acid hydrolysis and acetonide formation gave the lactol (50). Subjecting the aldose (50) to a
17. Scheme 8(a) Fraser-Reid and Sun7 1980
OHC QMe I," 62%
0 0
(47)
HQ, SjtH OMe IV, Y
0 0
(49)
vin, ix,x, H Me xi,xii,xiii
H OH (-)2 (2R,3R,6S,8S)
(i) Ph3P=CH(OMe)CH3; (vii) NaOMe; (ii) HCI; (viii) PhCOCl; (iii) H2/Ni; + (ix) H30 ; + (iv) H30 j (x) Me2NC(OMe)2;
(v) Me2C(OMe)2; (xi) Ac90/A; (xii) H /Pd; (vi) Ph3P = C(C02Me)Me; 2 (xiii) NaOMe/MeOH;
18. Scheme 8(b) Fraser-Reid and Sun7 1980
0 H OMe
(M>) 90%
OMe
LLL,IV
HO H o 0 3:1 X (53)
N VII
84% l/Yle HvFO H 0 0 r(55) HO" 'H 0 0 (SO X VIII , I X ,X ,XI, xii,xiii,xiv H. OH CCUMe M7 (2S,3S,6R,8R) H Me
(i) NaOMe/MeOH; (viii) PhCOCl;
(ii) H2/Ni; + (ix) H30 + (iii) H30 (x) Me2NC(0Me)2;
(iv) Me2C(OMe)2; (xi) AC20/A;
(v) Ph3P=C(CN)Me; (xii) H2/Pd; (vi) NaOMe; (xiii) H202, NOC1, CH2N2; (vii) NaOEt/EtOH; (xiv) NaOMe/MeOH
19. Wittig reaction followed by base epimerization yielded the hydroxyester
(51). The final steps leading from (51) to (-) methyl nonactate
(2) employed the Eastwood^ deoxygenation procedure and subsequent hydrogenation to give (2). (Scheme 8(a)).
The dextrorotatory enantiomer (+) (2), (Scheme 8(b)) used thermodynamic epimerization to give ketone (52) in 90% yield.
Raney-nickel hydrogenation, followed by acid hydrolysis and acetonide formation gave the lactol (54). The addition of the
Wittig followed by prolonged treatment with base, gave the 2,5 cis tetrahydrofuran (56). Eastwood 15 deoxygenation of the C-8 benzoate of (56) was carried out as detailed above, and the nitrile converted to (+) methyl nonactate by treatment with hydrogen peroxide and nitrosyl chloride.
The third chiral synthesis of (-) and (+) methyl nonactate was achieved by starting from D-mannose (57) and D-gluono-y-lactone (64) respectively. The key step in this synthesis by Ireland and Vevert 8,13 was the [3,3] sigmatropic rearrangement of silylated ketene acetals via a boat transition state, leading to control of the C-2 configuration of the nonactate. -See Scheme 9(a) and 9(b).
The furanoid glycal (61) was made in ten steps from D-mannose (57) in 36% yield, for elaboration to (-) methyl nonactate (2).
Correspondingly, they made the glycal (68), derived from D-gluono-Y- lactone (64) in 11% yield in 11 steps, required for the synthesis of
(+) methyl nonactate (2).
20. When the intermediate propionates of these enantiomeric glycals (61) and (68) were enolized with lithium diisopropylamide in THF, the Z,-enolate was the predominant isomer formed, thereby defining the rearrangement substrate. After trapping the enolate with chlorotrimethylsilane, and allowing the Claisen rearrangement to occur, the product was esterified and hydrogenated. The predominant hydroxyesters of (62) and (69) [in ratios 86:14 and 89:11 respectively] were separated by column chromatography and oxidized to the aldehydes
(63) and (70). No stereochemical control was possible in the organometallic addition step, thus giving the C-8 epimers as well as the desired products (+) 2 and (-) 2.
21. Scheme 9(a) Ireland and Vevert8>13 1981
OH HO*. *v„,0H Hv7 \"'H ,11,III,IV s * PhH2C0 "
V ,VI ,VII
0X0 •V|| I, IX H*7/ \'"H OCHjOMe 0CH20Me 2 H (59)
Me02C 0CH,0Me H 2 (62) xi,xii,xiii,xiv
xv MeO Me02C CHO
(~)2 + 00 1:1
(i) CH3COCH3, H (ix) Ph3P, CC14; (ii) NaH, PhCH2Cl, DMF; (x) n-Buli, C2H5C0C1, LDA, THF (iii) conc. HC1, MeOH; (iv) Me2NCH(OMe)2, CH2C12; SiMe3Cl, RT, H20, OH", CH2N2 (v) (CH3C0)20, 130°C; (xi) H2, Rh/C; (vi) 9-BBN, THF, H202, aq. NaOH; (xii) separate; (vii) KH, ClCH2OMe; (xiii) 2% HCl/MeOH; (viii) Li, NH3; (xiv) (COCl)2/DMSO, Et3N; (xv) LiCu(Me)2, pentane.
22. Scheme 9(b) Ireland and Vevert8'13 1981
HO OH 0^0 OH l,il,m, (65) OH
|V,V
0X0 0^0 VI ,VII,VIII, • ix.x OMe 0 NMe.
H (67) (66)
ix ,,i*0H
0 (68) (69) xii,xiii, xiv,xv
XVI Me09Q MeOjCv^ ^CHO (vo) 2 _ OH U = H H A H H (+)2 +(4) ' 4:3
(i) CH3COCH3, H (x) Ph3P, CC14;
(ii) i-Bu2AlH, Et20; (xi) n-Buli, C2H5C0C1, LDA, (ill) NaH, PhCHoCl, DMF; (iv) conc. HCl/MeOH; SiMe3Cl, RT, H20, OH "
(v) Me2NCH(OHe)2, CH2C12; (xii) H2, Rh/C; (vi) (CH3C0)20, 130°; (xiii) separate; (vii) 9-BBN, THF, H202, aq. (xiv) 2%HCl/Me0H; (viii) CH2(OMe)2, P205; (xv) (C0C12)/DMS0, Et3N;
(ix) Li/NH3; (xvi) LiCu(Me)2, pentane
23. Construction of Nonactin2>3,4,5,6,11 From Nonactic Acid
4 11
Gerlach et al. " built the macrotetrolide from nonactic acid
(71), by stepwise formation of the ester linkages (Scheme 10). Thus
the benzyl ether (72) and the t-butyl ester (73) were prepared following
conventional methods and condensed, after activating the carboxylic acid
using 2,4,6-trimethylbenzenesulphonylchloride and pyridine. The
ester (74) was transformed to the acid (75) by acid hydrolysis and to
alcohol (76) by hydrogenolysis. Coupling of (75) and (76) employing
the same esterification technique afforded the linear tetraraer (77) which
on deprotection furnished the hydroxyacid (78). Closure of the 32- membered ring was achieved via the 2-pyridinethiol ester (79) by
treatment with silver perchlorate. In benzene solution at 25° (0.5h), a 20% yield of tetramers was obtained, the yield rising to 35-40% in acetonitrile at 80° (lh). From the four possible tetrameric diastereomers
(starting with racemic nonactic acid) only three were observed. Nonactin
(1) was isolated by chromatography, comprising 25% of the mixture.
Schmidt et al.2?3,5,6 synthesized nonactin (1) with alternate
(+) and (-) nonactic acid units, stepwise from (-) and (+)-8-epinonactic acids (80 and 81 respectively) by suitable protecting and coupling operations. Thus the benzyl ester tosylate (81) of (+)-8-epinonactic acid was coupled with the potassium carboxylate of (-) nonactic acid
(80) in DMS0 to furnish the (-) nonactinyl (+) nonactic acid derivative (83) [Note inversion of configuration at C-8 in the Sn2 reaction.] Similar coupling of the tosylate (81) with the potassium Scheme 10 Gerlach etc al.4,11
ho2c
R=BZ,R'=BU .yl (75) R=Bz,R=H
viiyiiijx > 0) H H 9-10%' (77) RsBz.RVBu (78) RrR'=H (79 ) R=H,R = -S (i) NaH, Dioxan; (vi) H2, Pd - MeOH; (ii) Benzylbromide; (vii) <( )>~S}2, PPh3; (iii) (viii) AgC104, NaCN, CH3CN; CH3C02 ^Bu, CH3SO3H; (iv) (ix) separation. (v) CF3C02H; 25. salt (82) of 8-epi (-) nonactic acid gave 8-epi-(-) - nonactinyl (+) nonactic acid benzyl ester (84). Conversion of (84) to the tosylate (85) and of (83) to the carboxylate (86) followed by coupling of these two fragments accompanied by inversion of configuration of C-8, afforded the (-) - (+) - (-) - (+) hydroxybenzyl ester (87) which was transformed to the immediate precursor of nonactin, hydroxyacid (88). Formation of the 32 membered ring was accomplished in 20% yield, via the thioester (89) which was cyclized under silver ion catalysis to give nonactin (1). See Scheme 11. Scheme 11 Schmidt et al. 2 > 3,5,6 (83) R=H,R=Bz (86) R=H,R=K (8*0 R=H,R=Bz (as) R=Ts,R'-Bz (as)+ (86) R- / AgClOt nc —— H H = UK PhH 20% (87) R=H,R'=Bz (88) R=R= H (89) R=H,R'= -: 27. Results and Discussion A careful review of the literature revealed that although methyl nonactate had been synthesized by several groups in racemic form,11,12 or as dextro-or laevo-rotary enantiomers,5>6,7,8,13 an Qf these syntheses involved numerous stages. In addition, with one exception,9 these syntheses involved at least one step of low stereoselectivity. We were convinced that an efficient and stereocontrolled synthesis of (+) methyl nonactate (2) was feasible. As a preliminary study of tetrahydrofuranyl compounds, we undertook the synthesis of a nonactic acid analogue, 2R(S)-[2-hydroxyethyl]-5S(R) [(ethoxycarbonyl) methyl] tetrahydrofuran (138). The strategy envisaged, is as shown in Scheme 12. The strategy was based on the literature precedents of Bryson^, both for epoxide opening with the dianion (92(c)) and for the E-geometry of the tetrahydrofurylidene (139) thus formed. 3-Buten-l-o 124 was reacted with sodium hydride and benzyl bromide, to give the desired 4-benzyloxy-l-butene (90) in 61% yield. Treatment of this with peracetic acid afforded, after distillation, (2-benzyloxyethyl) oxiran (91) in 53% yield. The condensation of the dianion (92(c)) of ethyl 3-oxobutanoate with the oxiran (91) was attempted. Ethyl 3-oxobutanoate was reacted with sodium hydride at 0°C, followed by n-butyllithium at -78°C. The dianion^ (92(c)) was warmed up to room temperature for i hour and the 28. Scheme 21 (90) 111 0. IV 0& * •0' •Ph (91) PhCH20 OH PhCH20 ——»H0 H OEt (92) diastereomer (i) NaH, THF; / (ii) PhCH2Br; (iv) CH2 = C - CH = C(0~)0Et; (iii) CH3CO3H, CH2C12; (v) EtO*" H2; (vi) Rh/Al203. 29. Scheme 12(b) o- o- IXA OEt (92 C) OH 60 %/n CO Et 0 H (139) solution recooled to -78°C. The epoxide (91) was added and the reaction mixture was subsequently stirred for 5 days at room temperature. On quenching the reaction mixture, only the starting epoxide (91) (80%) was isolated. In a similar experiment, the dianion (92(c)) and epoxide (91) were refluxed for 72 hours in THF. On work-up, the epoxide (91) was recovered (56%). We speculated whether preliminary treatment of the epoxide (91) with iodotrimethylsilane2^>26 to form 4-benzyloxy-l-iodo-2-trimethyl- silyloxybutane IfVSItu , would enhance the reaction with dianion (92(c)). To a solution of hexamethyldisilane2^ in chloroform was added iodine, to form iodotrimethylsilane.After i hour, 30. the oxiran (91) was added, and the solution stirred for 2 hours. The solvent was evaporated, and the remaining solution was added to the previously formed dianion (92(c)) at -78°C. The solution was allowed to warm up to ambient temperature overnight and stirred for 7 days. Work-up afforded only the starting epoxide (91) (59%). This strategy and series of reactions were abandoned. Suitably chastised by the failure of model reactions, we now concentrated on the synthesis of the target molecule, (*) methyl nonactate (2). Racemic nonactic acid should be readily available by the condensation of activated derivatives of threo-pentane-l,2,4-triol (101), threo-2,4-dihydroxypentanoic acid (133) or threo-2,4-dihydroxypentanal (119), with the dianion (106(a)). Our approaches to nonactic acid utilized all three derivatives (101), (133) and (119). Initially, we speculated as to whether a mixture of erythro, threo-pentane-1,2,4-triol (93) would be adequate since reaction of the diastereoisomeric mixture of triol (93) with benzaldehyde, would give the protected erythro and threo-benzylidene acetals^* (94) and (94(b)). Chromatography should permit the separation of the threo from the erythro isomers.(Scheme 14) This regioselectivity in acetal formation was predicted from the literature^* of standard carbohydrate chemistry. Ketones normally show a preference for the formation of five-membered 31. Retrosynthetic Analysis Diagram (13) MeO,C- 4 H Me02C OMe 4 (106a) MeO,C HO V-H v-0H or I r or J - HO Hi •OH H0^-ftQH H OH ('01) (119) 0») 32. rings to give 1,3-dioxalans and therefore react preferentially with vicinal diols. Aldehydes show a preference for six-membered 1,3-dioxan rings. The reason for the preferences lies in the conformational stability of the acetal ring. When ketones form six-membered 1,3-dioxans, one of the alkyl groups must occupy an axial position, destabilizing it relative to the 1,3-dioxalan ring. Under acidic conditions, aldehydes condense with diols to give 1,3 dioxans with the large substituent equatorial. Six-membered acetal rings formed from two secondary hydroxyl groups with an erythro relationship are preferred to six or seven-membered rings from threo diols. The relative stabilities of these ring systems are nevertheless, difficult to predict and vary according to the conditions and the aldehyde used. Having separated the isomers, subsequent reactions, as outlined in Scheme 14, with the threo-acetal (94(b)) would afford the desired stereochemistry at the end. Thus, penten-4-ol28 was reacted with aqueous potassium permanganate to yield, after distillation, pentane -1, 2, 4- triol^?,29(93) (33% yield). Reaction of the triol (93) with a, c^-dimethoxytoluene^O under catalysis by boron trifluoride etherate (0.1 eq) gave 6S(R)-methyl-4S(R)-hydroxymethyl-2S(R)- phenyl-1,3-dioxan (94) as the predominant product (7:1, 58%); the minor product not being characterized. The erythro 33. Scheme 14 HO (93) HO OH OH Ph ^0 0 (94b) erythro threo 11 iv, v, VI Ph Ph ok ok OMe 4-1! 0 0 VIII Me 0,0 IX H H MeCUC , N (0 (i) KMn04, H20; (vii) H2C=C(0")-C(CH3)=C(0~)OMe; (ii) PhCHO; + (iii) Chromatography; (viii) H30 ; (iv) NaH, THF, CS2; (ix) H2-Rh/Al203. (v) Pyrrolidine; (vi) Mel; 34. configuration of the dioxan (94) was authenticated by its conversion into 4S(R), 6S(R)-dimethyl-2S(R)-pheny1-1,3-dioxan (97). The dioxan derivative (94) was reacted with toluene-4- sulphonylchloride to yield the tosylate (96) (84%). Lithium aluminium hydride reduction afforded the dimethyl dioxan (97). Similarly, treatment of the hydroxymethyl-dioxan (94) with n-butyllithium3*, carbon disulphide and methyl iodide afforded 6S(R)-methyl-4S(R)-(methylthio-thiocarbonyloxy) methyl-2S(R)-phenyl- 1,3-dioxan (95) (95%). This xanthate31 (95) was reduced with tri-n-buty1s tannane32 to give the 4S(R), 6S(R)-dimethyl-2S(R)- phenyl-1,3-dioxan (97) (27%). Examination of both the 'H and *3C high resolution n.ra.r. spectra of (97) confirmed the symmetrical erythro configuration. (See Scheme 14b). A sample of the acetal (95) was treated with pyrrolidine^2 followed by iodomethane to afford 4S(R)-iodomethyl-6S(R)-methyl-2S(R)- pheny1-1,3-dioxan (98) as a low melting solid, (85%). Since the predominant product isolated from the erythro, threo- triol mixture was the undesired erythro-dioxan (94),we proceeded to synthesize threo-pentane-1,2,4-triol (101) from the known 33 lactone 3-acetoxy-5-methyl-2,5-dihydrofuran-2-one (134). 35. Scheme 14(b) Ph Ph 0rA n 0 o^o (») OH (96) OTs Ph Ph 0 X IV o^o (95) S 0-C- SMe V,VI Ph 0^0 (i) TsCl, py; (iv) n-Bu3SnH; (ii) LAH, THF; (v) pyrrolidine; (iii) n-BuLi, CS2, Mel; (vi) Mel. 36. The literature^ preparation (Scheme 15) involves the condensation of diethyl oxalate with ethyl sodioacetate to afford the salt (135) in quantitative yield. Reaction of the salt (135) with acetaldehyde followed by a strongly acidic work-up, gave the lactone (136). Hydrolysis, followed by decarboxylation afforded lactone (100(c)). Acetylation with acetyl chloride gave the 3-acetoxy-5-methyl-2,5-dihydrofuran-2-one (134). The dihydrofuran-2-one-^-1 (134) was hydrogenated at atmospheric pressure over palladium on calcium carbonate.^ The hydrogenation of the lactone (134) had been reported^ as being stereospecific, but the stereochemistry of the product had not been elucidated. The high resolution fH n.m.r. spectrum clearly showed that the sole product formed (99) had a cis vicinal coupling (C-3, C-4) constant of 8.4 Hz and a trans vicinal (C-3, C-4) coupling constant of 10.6 Hz. Comparison with authentic^ n.ra.r. spectral data confirmed the cis^3 stereochemistry. Ironically, Daremon, Rambaud and Verniette^ had speculated that the hydrogenation of lactone (134) would give the trans stereochemistry. Their own spectral data helped to refute this suggestion. (i) Pd - CaC03, H2 37. Scheme 12(b) 0"NQ+ CK jott (135) 100 % Cft^OEt 90% II,HI (Tft) (134) (i) NaOEt, EtOAc; (iv) conc. HCl/H20/g. AcOH, 100°C; (ii) CH3CHO, EtOH; (v) AcCl, py. (iii) conc. HCI; 38. Since 3R(S)-acetoxy-5S(R)-methyltetrahydrofuran-2-one (99) was a crucial intermediate, we investigated alternative routes to it, the literature33 route being inelegant and low yielding. Of all the various steps, the hydrolysis and decarboxylation of (136) needed a major improvement, due to the atrocious yield and the drastic conditions involved. The literature yield claimed for this step was 75%. We therefore decided to synthesize the analogous t-butyl ester of lactone (136), lactone (100(b)). We speculated that the hydrolysis of the t-butyloxycarbonyl moiety with trifluoracetic acid would be milder and would give a higher yield of lactone (100(c)) on decarboxylation. Reaction of di-t-butyl oxalate3^ with t-butyl potassioacetate followed by an acidic work-up afforded di-t-butyl 2-oxobutan-l, 4-dioate (100) (78%). Treatment of this with sodium hydride and acetaldehyde followed by an acidic work-up gave 4-(t-butyl- oxycarbonyl )-3-hydroxy-5-methy1-2,5-dihydrofuran-2-one (100(b)) in 95% yield. Hydrolysis and decarboxylation in aqueous THF with trifluoracetic acid, after 60 hours, afforded cleanly, the lactone (100(c)) in an acceptable 72% yield. The overall modification is summarized in Scheme 16. The degree of improvement achieved is quite significant. The yield from diethyl oxalate to the hydroxylactone (100(c)) for 3 steps was 15%. Changing to the t-butyl oxalate and using milder 39. Scheme 21 J0+ , .. _U] 78% 95% in OH BuO?C OH IV 72% O'^O 0 (100 c) (100 b) (i) KOBu*, t-BuOAc; (iii) NaH, THF, CH3CHO; (ii) conc. HC1/H20; (iv) H20/THF, CF3C02H. 40. hydrolytic and decarboxylation conditions gave the hydroxylactone (100(c)), in 3 steps in an overall 53% yield. However, a cheaper, simpler and highly efficient route was found, starting with tri-O-acetyl ribonic acid "/-lactone (99(c)). Treatment of lactone (99(c)), at -20°C in THF, with 1,5-diazabicyclo [5.4.0] undec-5-ene (DBU). gave the doubly eliminated 3-acetoxy- 5-methylene-2,5-dihydrofuran-2-one (99(b)) as a white crystalline solid (94%). Steric-approach controlled*^ hydrogenation over palladium on calcium carbonate-^, afforded the desired 3R(S)- acetoxy-5S(R)-methyltetrahydrofuran-2-one (99) (90%). Comparison of the high resolution fH n.m.r. spectrum with that of authentic (99), confirmed that the hydrogenation proceeded with a cis disastereoselectivity of 97:3, showing the ease of control of stereochemistry in relatively flat 5-membered rings. 94% AcO OAc (99) (i) 1 eq. DBU; (ii) H2 - Pd/CaC03 41. The structure of 3-acetoxy-5-methylener-2,5-dihydrofuran-2-one (99(b)) was unambiguously assigned from its simple n.m.r. spectrum. Three olefinic resonances, a singlet at 7.3 ppm. (C-4) and two doublets at 4.95 and 5.25 ppra with a J^ value of 3 Hz indicated the presence of vinylic proton and an exocyclic methylene moiety, respectively. This lactone (99(b)) is very stable in stark contrast to protoanemonin^^ which readily dimerlzes to anemonin, an antibiotic agent present in Anemone Pulsatilla. 0 J^K protoanemonin anemonin Interestingly, the synthesis of butenolides from carbohydrate- y-lactones have featured prominently in the recent literature.50,65 Having optimized the synthesis of the acetoxylactone (99) to give an overall yield of 84% over two steps, we proceeded to prepare threo-pentane-l,2,4-triol (101). Lithium aluminium hydride reduction of the lactone (99) proceeded satisfactorily on a small scale to afford the triol (101) as a viscous oil (72%). On large scale reductions, problems were generally encountered from aluminium salt residues. To overcome this problem, the lactone (100(c)) was treated with t-butylchlorodimethylsilane^9 in 42. DMF to afford 3-(t-butyldimethylsilyloxy)-5-methyl-2r5-dihydrofuran -2-one (110(a)) (53%). Hydrogenation over palladium on calcium carbonate afforded the desired lactone (110) (82%) as a clear liquid. We reasoned that if the polarity of the triol (101) was creating the isolation problems, then, bis-1, 4S(R)-hydroxy-2R(S)- (t-butyldimethylsilyloxy) pentane obtained from the reduction of (110) would be a more accessible intermediate. OH II. Ill (100c) (i) LiAlH4 (ii) t-BuSiMe2Cl, DMF; (iii) H2/Pd-CaC03 However, the reduction of lactone (110) proved to be unnecessary. Reduction of the lactone (99) using sodium borohydride33 in water at pH 9-10 proceeded splendidly to afford the triol (101) (100%), identical to the previous sample by n.m.r. spectroscopy. 43. Treatment of the threo-pentane-l,2,4-triol (101) with dimethoxytoluene3^ in the presence of zinc bromide, slowly gave a mixture of several products and not the desired 4R(S)- hydroxymethyl-6S(R)-methyl-2S(R)-phenyl-l,3-dioxan (102). The difference in reactivity*^ during acetal formation between erythro triols and threo triols is quite noticeable. Erythro triols react rapidly whilst threo triols are remarkably sluggish. The desired benzylidene (102) may not have readily formed, because the threo configuration would necessarily require either the phenyl moiety (of (of the triol (101)) to be axial in the transition state and therefore in Che final acetal formed. We now explored the possibility of using other hydroxyl protecting groups, but leaving the overall strategy (Strategy 1) unchanged. The triol (101) was reacted regioselectively^'^with toluene-4- sulphonyl chloride to afford the desired l-(toluene-4-sulphonyloxy)- 2R(S), 4S(R)-diol (103) (60%). The two remaining hydroxyl groups 3 were protected by reaction with t-butylchlorodiraethylsilane ^ in DMF to give the required 2R(S), 4S(R)-bis-(t-butyldiraethylsilyloxy)- l-(toluene-4-sulphonyloxy) pentane (104) in excellent yield (93%). 44. Strategy 1 (a) HO' HO (102) HO OH OyO Ph 0- 0' 0 0 I OMe Me X' (106a) 0 0 vPh Y X-leaving group Ph Me020 Explore (b) Hi HO OH ,0 0SiMe2Bu (101) 0- Q- OMe 0 0 OMe Me020 ^ i I .0 0SiMe2Bu —f~Sr 45. The conversion of the tosylate (104) to the iodide (105) by treatment with 10% sodium iodide4® in acetone was most unsatisfactory (4%). However, the method of Place, Roumestant and Gore4*- whereby freshly prepared magnesium iodide etherate4*- was reacted with the tosylate (104), afforded the iodide (105) as a pale oil (70%). (Scheme 18) The key step in our strategy, required the formation of a new C-C bond envisaged by us as the alkylation of dianion (106(a)). This type of reaction had been demonstrated by both Weiler44 and Harris60. 0 0' Ml ^Nwi T)Me OMe R-I Yield of (140) Mel 81% EtBr 84% i-Prl 73% PhCH2Cl 81% + (i) NaH, THF; (iii) R-I; H30 (ii) n-Buli; Weiler44, using sodium hydride and then n-butyllithium to form the dianion of methyl 3-oxobutanoate, succeeded in his alkylations in respectable yield, as shown above. 46. Scheme 12(b) HO' 'sO HO OH 60 % HO OH (101) (103) 93% S' i : 111 i 0 6siMe,Bu « 0 0SiMe2Bu 70% TsO I (10 5) (104) X4/ 0 0 MeN" 0" OMe (106) (l07Q) I ,0 0SiMe2Bu —Si 1 (i) TsCl, py; (iv) H2C=C(0 )-CMe=C-OMe (ii) t-BuMe2 SiCl, DMF; (ili) MgI2.OEt2; 47. Harris,3® using an amide base was less successful in his dianion alkylations: OEt (1M) R-I Yield of (141) Mel 37% EtBr 29% (i) 2.2 KNH2; (ii) R-I Using Weiler's methodology, we attempted many coupling reactions to form an adduct between the dianion (106(a)) of methyl 2-methyl-3- oxobutanoate,^3 and the iodide (105). Reaction of the dianion (106(a)) with the iodide (105) at 0°C in THF for 6 hours, on work- up afforded only the starting materials. Changing the solvent to DME and extending the reaction time to 3 days at room temperature, still only gave unreacted starting materials on work-up. Similarly, using hexamethylphosphorictriamide/DME as a solvent, after 14 days at room temperature, showed no reaction. Reactions involving the addition of copper (I) iodide and phenylthiocopper (I) to the 48. dianion (106(a)) and subsequent addition of the iodide (105), after 7 days, showed no condensation. We eliminated incomplete formation of dianion (106(a)) as a possible cause of failure by reacting it with iodopropane. Thus methyl 2-methyl-3-oxobutanoate was treated with sodium hydride and then n-butyllithium at 0°C. After i hour, the dianion44 (106(a)) was quenched with iodopropane. After 15 hours at room temperature, the reaction mixture was quenched with dilute hydrochloric acid and worked up to give methyl 2-raethyl-3- oxoheptanoate (109) (64%). Clearly, the formation of dianion (106(a)) was not in question. These series of reactions were abandoned. We speculated whether the sluggish reactivity of the dianion (106(a)) could be overcome by using methyl 3-(methylamino)-2-methylbut- 2-enoate42 (107(a)). The dianion of inline (107(a)) would be electron rich and would therefore be very reactive. Thus, the dianion of the but-2-enoate (107(a)) was formed at 0°C in THF, with n- butyllithium. The iodide (105) was then added. On quenching the reaction after 4 days, only the starting iodide (105) and methyl 2-methyl-3-oxobutanoate were detected by n.m.r. and i.r. spectroscopy and t.l.c. analysis. t-Butyl 3-oxobutanoate4^ was treated with sodium hydride and methyl iodide to give t-butyl 2-methyl-3-oxobutanoate (108(a)) (85%). Treatment of the B-ketoester (108(a)) with 2.2 equivalents of n-butyllithium at 0°C gave the dianion. The solution was 49. cooled to -78°C and the iodide (105) was added. The reaction mixture was allowed to warm slowly to ambient temperature. On quenching the reaction after 15 hours, only the two starting materials were detected by n.m.r. and i.r. spectroscopy, and t.l.c. analysis. This approach was subsequently abandoned. We now re-examined Bryson's^ studies (Scheme 12(b)) on the condensation of the dianion (92(c)) with epoxides to give E-tetrahydrofurylidenes (139). With this reaction in mind, we considered the use of 2S(R)-[2R(S)-hydroxypropyl] oxiran (111(a)) as a suitable activated derivative of the triol (101). To (toluene-4-sulphonyloxy) pentane-2R(S), 4S(R)-diol (103) was added 1,5-diazabicyclo [5.4.0.] undec-5-ene (DBU). The desired epoxide (111(a)) was isolated after chromatography in moderate yield (46%). We also prepared a mixture of the erythro/threo (2-hydroxypropyl) oxiran- (111(b)) from 2-hydroxy-4-pentene23 by peracid oxidation. The erythro/threo epoxide (111(b)) was synthesized to serve as a model in any condensation reactions, for the threo epoxide (111(a)). The epoxides (111(a)) and (111(b)) were very similar by their n.m.r. and i.r. spectra and identical by t.l.c. analysis. The structure of (111(a)) was assigned as the epoxide, and not the corresponding 2R(S)-hydroxy-4S(R)-raethyltetrahydrofuran (141) by careful consideration of the n.m.r. and i.r. spectra. The C-l and C-2 protons of the oxiranehad a chemical shift of 2.35-3.3 p.p.m., whereas the C-2 and C-5 protons of 50. a tetrahydrofuran ring would be expected at 3.6 p.p.m. The tertiary proton of the 2-hydroxypropyl moiety had a chemical shift of 4.1 p.p.ra., which in the tetrahydrofuran (141) would have a chemical shift of 3.6 p.p.m. The i.r. spectrum showed the strong characteristic epoxide C-0 stretch at 1260 cm"r l Of course, the similarity in the i.r. and n.m.r. spectra of (111(a)) and (111(b)) was the final confirmation. Both hydroxyepoxides (111(a)) and (111(b)) were protected as t-butyldimethylsilyl ethers (112(a)) and (112(b)) respectively, 2 by reaction with t-butylchlorodimethylsilane ^ in DMF. Scheme 19 OH OH 0SiMe2Bu H H in II TsO 46 °/o 0' 92% 0 OH OH OH 0SiMe2Bu (i) CH3CO3H, CH2C12; (iii) 1,5-diazabicyclo [5.4.0.] (ii) t-BuMe2SiCl, DMF; undec-5-ene, THF. 51. Condensations44 were now attempted between the epoxides (112(a)) and (112(b)) with the dianion (106(a)) of methyl 2-methyl-3-oxobutanoate. Therefore methyl 2-methyl-3-oxobutanoate was treated with sodium hydride and then n-butyllithium in THF to form the dianion (106(a)). The threo-epoxide (112(a)) was added at 0°C and the reaction mixture was stirred for 15 hours at room temperature. On quenching the reaction and isolating the components, only the two starting materials, the epoxide (112(a)) (60%) and 8-ketoester (40%) were recovered. Due to the sluggish reactivity of the epoxide (112(a)), a reaction was attempted using Schlessingerfs46 conditions with 10 equivalents of dianion (106(a)) in DME at 40°C. During the synthesis of dl-bisnorvernolepin, Schlessinger et. al,.46 opened epoxide (142) with 10 equivalents of t-butyl dilithioacetoacetate to give rise to the ketoester (143). OMe H 52. The dianion of methyl 2-methyl-3-oxobutanoate was formed with lithium di-isopropylamide in DME at 0°C. The erythro/ threo epoxide (112(b)) was added and the reaction mixture was heated at 40°C for 4 days. No reaction was seen by either t.l.c. or n.m.r. spectroscopy and the reaction was abandoned. Scheme 20 0" 0' OMe OMe (113) HQ 0SiMe2Bu (i) NaH; (ii) n-BuLi; (iii) 0 OSi —|— A titanium tetrachloride33 mediated reaction was tried. After the dianion (106(a)) had been formed at 0°C, the solution had been formed at 0°C, the solution was cooled to -78°C and titanium tetrachloride (1 eq.) was added to form the titanium 53. enolate. The black suspension was stirred for 2 hours at -78°C, when the epoxide (112(b)) was added. The reaction mixture was allowed to warm up to room temperature overnight, and quenched with dilute hydrochloric acid. The crude reaction mixture was shown to contain only the two starting materials by n.m.r. spectroscopy, and t.l.c. analysis. This series of reactions was subsequently abandoned. Having failed to utilize the pentane derivative (101), we now considered a strategy based on the threo-2,4-dihydro- xypentanoic acid (133). We now investigated the reaction of enolates with lactones. Meinwald et al^ had reported the condensation of fi-lactone (144) with various enolates: 0 0 R (144) R R' Yield of (145) H t-Bu 97% O-t-Bu t-Bu 87% 0-CH(Me)0Et Me 85% 54. Trost and Runge®>9 had reported the condensation of t-butyl lithloacetate with the lactone (146) to afford on work-up, the alkylidenetetrahydrofuran (147) in 91% yield. BuOjCSv 1,11 111 91% 2:1 (146) (147) 1 (i) LiCH^O^u* , -78°C; (ii) CH3S02C1, DBU We attempted a condensation of methyl lithioacetate*^ at -78°C with the acetoxylactone (99). After 48 hours at ambient temperature, work-up afforded an intractable orange oil. However, reacting the lactone (99) with t-butyl lithioacetate at -78°C for 1 hour and then at room temperature overnight gave, 2-hydroxy- 3R(S)-hydroxy-2-[(t-butyloxycarbonyl)methyl]-5S(R)-raethyl- tetrahydrofuran (114(b)) as a yellow oil (63%). Treatment of the adduct (114(b)) with trifluoracetic acid in methanol afforded 2R(S), 6-dioxabicyclo [3.3.0.]-7S(R)-methyl-5R(S)-methoxy- octan-3-one (115) as a clear oil (54%). 55. Scheme 21 OAc X ^OMe 0 '0 OH (99) 63% OH o 0. Ji 54% 0+ OH OMe (lub) (115) (i) H2C = C(OLi) OMe: (iii) MeOH, CF3C02H fc (ii) H9C = C(OLi) OBu ; This unusual bicyclo [3.3.0.] octan-3-one (115) was immediately recognised from its high resolution n.m.r. spectrum. A double doublet resonance at 4.76 p.p.m. with coupling constants, 2.3 Hz and 7.0 Hz was identified as the C-l proton. An A-B quartet with a coupling constant of 17 Hz at 2.85 p.p.m. was assigned to the C-4 protons. All other resonances were assigned without any difficulty. 56. Encouraged by this successful condensation, we attempted to react the lactone (99) with dianion (106(a)). To the dianion (106(a)) at -78°C was added the lactone (99). After 1 hour at -78°C, the solution was allowed to warm up to room temperature and stirred there for 48 hours. On quenching the reaction mixture, an intractable polymer resulted. Repeating the experiment using lithium di-isopropylamide as the base for dianion (106(a)) formation only afforded the two starting materials after 72 hours at room temperature. These series of reactions with methyl 2-methyl-3- oxobutanoate were therefore abandoned. Dianion acylations are quite well documented due to the work of Hauser^® and Harris^*-. Light and Hauser^® acylated pentane-2,4-dione with methyl benzoate to give the trione (148): i,ii (148) (i) KNH2 (2 eq.); (ii) PhC02Me 57. Similarly, Harris and Hauser?1 acylated the dione (149) o Ph 58% (1A9) (150) I, III,IV 28% (151) (i) 2 eq KNH2; (iii) PhCHO; (ii) PhC02Me; (iv) TsOH, heat We now attempted a dianion condensation with t-butyl 2-methyl-3- oxobutanoate (108(a)) and the lactone (99). After 72 hours at room temperature, quenching the reaction with glacial acetic acid and subsequent purification by column chromatography afforded the desired 2-[3-(t-butyloxycarbonyl)-2-oxobutyl]-2-hydroxy-3R(S)-acetoxy- 5S(R)-methyltetrahydrofuran (117(a)) as a clear oil (35%). This was immediately treated with trifluoracetic acid and the ensuing furylidene (117(b)) was hydrogenated in situ to afford the tetrahydrofuran (117) as a clear oil (93%). 58. Van der Gen et had reported the fragmentation of esters of tetrahydrofuran-l-acetic acid such as that shown in Scheme 23. Reaction of tetrahydrofuran (152) with lithium di-isopropylamide afforded the anion a to the ester. Subsequent 3-elimination fragmented the tetrahydrofuran to afford the E-olefin (153) on acidic work-up. With this precedent in hand, we attempted the analogous fragmentation of 2-[3-(t-butyloxycarbonyl)-2-oxobutyl]-3R(S)- acetoxy-5S(R)-methyltetrahydrofuran (117), by double metallation giving the intermediate (154), (Scheme 24). To lithium di- isopropylamide at -78°C was added the tetrahydrofuran (117). The mixture was allowed to warm up to room temperature over 6 hours and stirred there for a further 15 hours. Work-up afforded only decomposition residues. Due to the failure of the 3-elimination, the route that we had envisaged (Scheme 24) had to be abandoned. Incidentally, Seebach et al.^9 were unable to fragment ketone enolates or the type shown in Scheme 25. The attempted syntheses of methyl nonactate (2) using Synthons (101) and (133) having failed, the aldehyde (119) was examined. 59. Scheme 22 (117Q) (117) (i) H2C=C(0")-CMe=C0Me; (iii) CF3C02H, THF; t (ii) H2C=C(0")-CMe=C(0")OBu' ; (iv) Pd/C,H2 Scheme 23 CO,Et C02Et 4 y N/ nx' —2 o 65% HO R R (152) (153) R=H,Me E- only + + (i) Li N~ (i-Pr)2, -78°C; (ii) H30 60. Scheme 22 (154) R=C02Me (i) Li+ N~ (i-Pr), -78°C; (iv) Na/MeOH; + + (ii) H30 ; (V) IR - 120H , THF (iii) H2 - Pd/C; Scheme 25 R' - t-Bu R2- Et or Ph R'= Ph R2= Et orPh + (i) Li N~ (i-Pr)2; (ii) H30" 61. Synthesis of (+)-t-Butyl 8-£-t-Butyldimethylsilyl Nonactate (123) The strategy we now considered, was focussed on the lactol (119) as the key intermediate. The lactol (119) was seen as a latent aldehyde and was therefore capable of undergoing a Wittig reaction. The resultant enoate would then be hydrogenated immediately and acidified to afford lactone (120). Using the knowledge acquired earlier on enolate condensations with lactones would lead us to the furylidene (155). Conversion of this to t-butyl nonactate would then be a cosmetic step. Clearly the Wittig reaction of the lactol (119) with any ylide would be an important step in our strategy. An important consideration would be choice of reagent and conditions so as to minimise any conconL itant cyclization of the enoate formed (See Strategy 2). 72 Corey in his pioneering syntheses of prostaglandins F2 and E2 utilized the following reaction: OH 0H Ph3P=CH(CH2)£0- C55 H1n1i DMSO 80% OTHP OTHP OTHP OTHP 62. Strategy 2 ^OH (120 a) An0 ^ OH (119) Cr^NDEt 0' "1. Bu02C ±t-butyl nonactate 63. The reactions of the ylide ethoxycarbonylmethylenetrlphenyl- phosphorane, which we were going to use, were also well documented. Ruchardt, Eichler and Panse" had condensed It with various ketones, as shown, In high yield. Ph3P=CHC02Et Ketone product Yield OEt 80% OEt 66% OEt 95% Wittig reactions followed by conjugate additions are also known. Kato et al.^6 utilized the following reaction in the synthesis of pyrazofurins (see Scheme 25b). 64. Scheme 25 (b) LU 3 > t~i 95% i—( o 0 V6 V3 o (i) Ph P=CHCCH C0 Et, CH CN, A 3 2 2 3 With all the precedents clearly established, we proceeded as shown in Scheme 26. Di-isobutylaluminiumhydride reduction74 of the acetoxy lactone (99) afforded 2,3 R(S)-dihydroxy-5S(R)-methyltetrahydrofuran (119) as a clear oil (85%). A Wittig condensation between the lactol (119) and ethoxycarbonylmethylenetriphenylphosphorane^S in THF gave the enoate (120(a) which was hydrogenated immediately to give ethyl 4S(R), 6S(R)-dihydroxyheptanoate. Acidification afforded the required 5S(R)-[2S(R)-hydroxypropyl] tetrahydrofuran-2-one (120) (60%), after chromatography. A lot of difficulty was encountered in isolating lactone (120), due to contamination by triphenylphosphine oxide. The best procedure involved partitioning the ethyl 4S(R), 6S(R)-dihydroxyheptanoate and triphenylphosphine oxide mixture between water and ether. 65. Separation of the aqueous phase and evaporation afforded nearly pure heptanoate. Acidification then gave the lactone (120). Treatment of 5S(R)-[2S(R)-(t-butyldimethylsilyloxy)propyl] tetrahydrofuran-2- one (121) with t-butyl lithiopropanoate (10 eq.) at -78°C and then at room temperature overnight, afforded a yellow oil which was probably Che condensed hydroxyester (70%). This material was refluxed for 8 hours in THF with IR-120H+, to afford 2S(R)- [2S(R)- (t-butyldimethylsilyloxy)propyl]-5-E-[1-(t-butyloxy- carboxyl)-ethylidene] tetrahydrofuran (122)^7,50 as a clear oil (91%). Generally, a one-pot reaction involving condensation of t-butyl 2-lithiopropanoate with lactone (121) and subsequent treatment with IR-120H+ afforded the furylidene (122) in 75% yield. The furylidene had a characteristic u.v. spectrum with \ 246 n.m., e 13900; values which are unmistak able and comparable to those in the literature.^ Having secured this key compound (122), the production of the nonactic acid derivative (123), was merely a cosmetic operation.® However, hydrogenation of the a, 8 unsaturated ester at atmospheric pressure using rhodium on alumina, palladium on carbon or platinum oxide as catalysts were extremely sluggish and only the unreacted furylidene (122) was recovered. When the furylidene (122) was hydrogenated at 70 p.s.i. over rhodium on alumina,® for 72 hours, chromatography afforded the starting furylidene (122) and the hydrogenated derivative (123) (89%). High resolution fH n.m.r. spectroscopy showed this to be a 85:15 mixture of the desired nonactic acid derivative 66. Scheme 27 JDAc >OH * rr 85% ^oNi60% (99) II ,III,IV ,v 55% Bu02C 89% ± . H H Bu02C Bu 02C (123) (137) 85: 15 t (i) Di-isobutylaluminium hydride, (vi) H3CH=C(OLi)OBu , THF; PhCH3,-78°C; (vii) g. AcOH: (ii) Ph3P=CHC02Et, (viii) IR-120H ; (iii) H2-Rh/Al203; (ix) Rh/Al203-H2 70 p.s.i, (iv) CF3C02H, CH2C12; (v) t-BuMe2SiCl, DMF; 67. (123) and the diastereoisomer (137) (Scheme 26). A similar hydrogenation carried out at 1000 p.s.i. in a Parr hydrogenator afforded a 1:1 mixture of the desired (123) and undesirable diastereoisomer (137) by 250 MHz *H n.m.r. spectroscopy. Hydrogenation of the furylidene (122) at 1000 p.s.i. over rhodium on alumina using anhydrous magnesium bromide (1 eq.) as a coordination source, gave a 2:1 mixture of the desired (123) and its diastereoisomer (137). Expected Mechanism t Rh from below However, from the results, the magnesium bromide did not coordinate as expected thus giving a 2:1 mixture of (123) to (137) instead of a higher diastereoselectivity. The synthetic (+)-t-butyl 8-0-t-butyldimethylsilylnon- actate (123) was correlated by conversion to the known diol12*22 2S(R)-[2S(R)-hydroxypropyl]-5R(S)-[2-hydroxy-lS(R)- (methyl)ethyl] tetrahydrofuran (124). The nonactate derivative (123) was reduced with lithium aluminium hydride to give 68. 2S(R)-[2S(R)-(t-butyldimethylsilyoxy)propyl]-5R(S)-[2-hydroxy-lS(R)- (methyl)ethyl]-tetrahydrofuran (124(a)) (92%). This was treated with potassium fluoride and trifluoracetic acid to afford the desired diol (124) (89%), (Scheme 27). Nonactin was reduced with lithium aluminium hydride to afford authentic diol (124) (95%). The synthetic diol was identical to the authentic diol by i.r. spectroscopy and t.l.c. analysis. The 250 MHz fH n.m.r. of the synthetic material, was identical to that of the authentic material, except for a doublet at 0.96 p.p.m. due to the C-2 methyl of the minor diastereoisomer (137(a)). We now decided to investigate other propanoate esters in the condensation of t-butyl 2-lithiopropanoate with the lactones (120) and (121). We wanted to select an ester which would undergo hydrolysis in high yield under very mild conditions. We also speculated whether the S^t-butyl thioester of (123) could be similarly synthesized and Masamune^^ copper (I) trifluoromethane sulfonate lactonization, effected on it, to give nonactin (1). Thus, to lithium di-isopropylamide solution at -78°C was added Sr-t-butylthiopropanoate^l. After allowing metallation to occur for 40 minutes, the hydroxylactone (120) was added. After allowing the solution to warm up to room temperature overnight, and quenching, column chromatography afforded the starting lactone (120) and j3-t-butyl 2-methyl-3-oxothiolpentanoate. 69. Scheme 27 C 0 s OSi-j- H H BU02C (123) /H 82% BU02C HCK (137) (137a) 0) 95% H H HO (124) (i) Lithium aluminium hydride, Et20; (ii) KF, CF3CO2H. 70. Similarly using JS-t-butyl 2-lithiothiopropanoate on the lactone (121) afforded j3-t-butyl 2-methyl-3-oxothiolpentanoate and the silyloxylactone (121). N- + (120) (120) (121) (121) Similarly, when a condensation of lactone (121) with 2-(trimethylsilyl)ethyl 2-lithio-propanoate32 was attempted, no adducts were detected and only the starting lactone (121) and some degradation products were detected. Due to the sluggish reactivity of these other propanoate enolates, we now abandoned the idea of using them and speculated whether we could utilize the ylides of B-ketoesters with lactol (119) (Scheme 28). 71. Scheme 28 JJH Cj + X3-0H Ar3P=CHCCHCD2MeJA,Me0?c 0 Me (119) (128) Ar = Ph OMe (131) Ar=Meo \\ // + (i) A/THF; (ii) H2-Rh/Al203; (iii) IR 120H The advantage of these ylides would be to reduce the number of steps required to synthesize methyl nonactate from 7 to 4. The overall yield would presumably be higher; and, forming the required carbon skeleton in one reaction, would remove the need for using protecting groups. The preparation and reactions of ethoxycarbonyl- acetylmethylenetriphenylphosphorane^ (156) are well documented. (i) PhH, a, 24 hours 72. (Methoxycarbonyl-2-propanoyl)methylene-triphenylphosphorane55 (128) was prepared by reacting triphenylphosphine with methyl 4-bromo-2- methyl-3-oxobutanoate^2 and treating the resultant gum with aqueous sodium hydrogen carbonate. Chromatography afforded the phosphorane (128) as a white crystalline solid. The synthesis of methyl 4-bromo-2-methyl-3-oxobutanoate^2 deserves mention. Bromination of the ^-ketoester, methyl 2-methyl-3-oxobutanoate at room temperature affords methyl 2-bromo- 2-methyl-3-oxobutanoate. Aeration of this acidic solution, causes the bromine atom to migrate and gives the desired methyl 4-bromo- 2-methyl-3-oxobutanoate. Interestingly, the bromination and isomerization of methyl 3-oxobutanoate is clean and high yielding. This is not the case for the 2-methyl ester, which is messy and low yielding.^2 We also decided to make a highly electron rich ylide, thereby increasing its reactivity. Therefore, we attempted to prepare (methoxycarbonylacetyl) methylene-tris-(2,4-dimethoxyphenyl) phosphorane (131) by reacting tris-(2,4-dimethoxyphenylphosphine)^4 with methyl 4-bromo-3-oxobutanoate, and treated the resultant salt with aqueous sodium hydrogen carbonate. Work-up afforded an atrociously low recovery of a yellow solid, identified as a complex mixture of products by n.m.r. spectroscopy and t.l.c. This preparation was subsequently abandoned. The results of the attempted condensations between the ylide (128) and lactol (119) 73. are summarized in the table below. Ylide Solvent Reaction Conditions Products 128 THF 23°C, 96 hours Ylide and decomposition 128 DMF 100°C, 96 hours No reaction, only starting materials Interestingly, condensation of the lactol (119) with ylide (128) in refluxing acetonitrile^^ after 5 days afforded the cyclized product, 2-[3-(methoxycarbonyl)-2-oxobutyl]-3R(S)- hydroxy-5S(R)-methyltetrahydrofuran (129) as a clear oil (24%). This has precedent in the reaction performed by Kato et. al.56> cited previously. (Scheme 25(b)). (i) CH3CN,A We now considered whether we could resolve 3R-hydroxy-5S methyltetrahydrofuran-2-one (138) and 3S-hydroxy-5R-methyl- tetrahydrofuran-2-one (138) from 5S(R)-methyl-3R(S)-[2S(R)- methoxy-2-phenylacetoxy] tetrahydrofuran-2-one (132) by HPLC. 74. If the two enantiomers of (138) could be readily separated, we were in a position to synthesize the two enantiomers of t-butyl nonactate, using the route in Scheme 26. Thus the acetoxy-lactone (99) was treated with potassium carbonate in methanol. After 48 hours, the reaction was quenched with trifluoracetic acid to afford the hydroxylactone (138) (57%). Treatment of this with (S)-methyl-O-mandelic acid57 in the presence of 4—dimethylamino— pyridine and dicyclohexylcarbodiimide53 afforded the desired mandelate (132) (25%). We were unable to separate the two enantiomers by analytical HPLC, even though the ^"H 250 MHz n.m.r. indicated that the sample was enriched (2.6:1) in one enantiomer. This partial resolution may have occurred due to a kinetic reactivity process.75 In conclusion, racemic t-butyl 8-0-t-butyldimethyl- silylnonactate (123) has been prepared in seven steps from 2,3,4- tri-O^-acetyl-D-ribonolactone (99(c)) via the highly stereoselective hydrogenations of 3-acetoxy-5-methylene-2,5-dihydrofuran-2-one (99(b)) and subsequently, 2S(R)-[2S(R)-(t-butyldimethylsilyloxy) propyl]-5- [l-(t-butyloxy-carbonyl)ethylidene] tetrahydrofuran (122). The (+) nonactic acid derivative (123) is now available in seven steps from readily available starting materials in an overall yield of 24%. Samat and Bibout7^ have stated, and I quote, "the total synthesis of these macrotetrolide nactins is long and uneasy"' We have gone some way in refuting this statement. 75. EXPERIMENTAL Reactions were performed under a dry N2 (argon where specified) atmosphere at room temperature unless otherwise stated. Temperatures were measured in degrees Celcius (°C); low reaction temperatures were recorded as bath temperatures. n-Butyllithium (in hexane) or diisobutylaluminium hydride (in toluene) were added dropwise over ten minutes. Reaction times are recorded in minutes (min.), hours (h) or days (d). Reactions were studied by t.l.c., i.r., n.m.r. or u.v. analysis prior to work-up. Ultra-violet inactive compounds on t.l.c. were visualized either with iodine or by concentrated sulphuric acid charring. Reaction mixtures were evaporated at 50° or below on a Buchi Rotavapor (unless otherwise specified); involatile compounds were further evaporated ( mm Hg)• Organic extracts were dried over anhydrous sodium or magnesium sulphate. T.l.c. was carried out on Merck Kieselgel GF 254, column chromatography on Kieselgel H, developing solvents are given in parentheses. M.p.Ts were determined on a Kofler hot stage. I.r. spectra were recorded as nujol mulls (solids) or films (oils) on a Perkin-Elmer 257 instrument. Broad (br.), medium (m), strong (s) or weak (w) bands were reported. U.v. spectra were recorded on a Unicam SP-800A spectrophotometer. N.m.r. spectra were recorded as deuteriochloroform solutions (unless otherwise stated) on a Varian T60, Varian EM-360A or Brucker WM 250, instruments using tetramethylsilane as internal 76. reference unless otherwise stated. Multiplicities were recorded as br. (broad) peak, s. (singlet), d. (doublet), t. (triplet), q. (quartet), and m. (multiplet). .Mass spectra were recorded on a Micromass 4040 A instrument; only molecular ions, fragments from molecular ions, and major peaks were reported. Microanalyses were recorded by the Microanalytical Laboratory, Imperial College. Solids were recrystallized and oils purified by column chromatography prior to microanalysis. All starting materials and reagents were purified (t.l.c., n.m.r., i.r.) and dried unless otherwise stated. Common solvents, dichloromethane, diethylether and petroleum (b.p. 40-60°) benzene and ethyl acetate were redistilled prior to use. For organometallic reactions, the following solvent drying techniques were employed: THF was dried by refluxing with potassium and benzophenone and distillation of the blue solution prior to use; DMF was refluxed with activated 4A° molecular sieves and distilled onto 4A° molecular sieves. Di-isopropylamine and chlorotrimethylsilane were refluxed over calcium hydride and distilled; the former stored over 4A sieves, the latter, used 23 immediately. Reagents were purified according to Perrin unless otherwise stated. Abbreviations THF - tetrahydrofuran DME - dimethoxyethane DMF - N,N - dimethyIformamide 77. Preparation of 4-Benzyloxy-l-Butene (90) To 3-buten-l-ol24 (12.8g) in THF (20 ml) at 0°C under nitrogen, were added sodium hydride (4.7 g, 1.1 eq.) and imidazole (0.4 g) in small portions. To the sodium salt, was added benzyl bromide (28.8 g, 0.95 eq.). When the reaction was complete, the suspension was filtered off, the THF was evaporated and the benzyl ether purified by distillation (17.56 g, 61%)., b.p. 80-82° at 4 mm mercury, V max (neat) 3070 (m, olefinic C-H stretch), 3050 (m), 1645 (m, C=C stretch), 1500(m), 1470(m), 1365(s), 1105 (s, ether C-0 stretch), 910(s), 740(s), and 705 (s, aromatic G-H out of plane deformation) cm"1, 6 2.4 (2H, t, vJ=6.0 Hz, C-3), 3.5 (2H, t, J=6 Hz, C-4) , 4.5 (2H, s, C6H5 CH9), 4.9 (1H, m, C-1ED, 5.1-5.2(1H, m, C-jZ), 5.5-6.1(1H, m, C-2), and 7.35 (5H, br. s, C6H5), m/e 162 (Mt), 161, 107, 105, 92, 91 (100), 65, 55 and 51. 78. Preparation of (2-Benzyloxyethyl)oxiran (91) To dichloromethane (30 ml) at 0°C, was added 4-benzyloxy- 1-butene (90). Peracetic acid (40.4 ml, 3.95 M, 1.5 eq.) which had been neutralized with excess sodium acetate trihydrate (10 g) was then added. After 60 hours, when the reaction was complete, aqueous potassium iodide solution was added. To the brown solution was added 10% aqueous sodium thiosulphate. The organic layer was separated, dried and evaporated. Distillation afforded the title compound (91) as a clear liquid (10.1 g, 53%), b.p. 120-122° at 4 mm mercury, V max (neat) 3040 (m, C-H stretch), 1600 (w), 1460(m), 1370 (s), 1265 (m, epoxy C-0 stretch), 1100 (s, ether C-0 stretch), 1035 (m), 910 (m), 840 (m), 740 (s), and 705 (s, aromatic C-H out of plane deformation) cm"*-, X max (e) 242 (100), 248 (117), 253 (144), 258 (170), 264 (141), and 286 nm (98), 5 1.8-2.20 (2H, m, C-l1), 2.45-3.0 (2H, m, C-2), 3.0-3.30 (1H, br. ra, C-l), 3.75 (2H, t, J=6.0 Hz, C-2'), 4.70 (2H, s, Ph + CH2) and 7.45 (5H, s, Ph), m/e 178 (M"t), 160, 150, 107 (Ph CH20 ), 91 (100), (Found: C, 74.07, H, 8.06.Cn H^ 02 requires C, 74.13; H, 7.92%). 79. Attempted Condensation of Ethyl 3-0xobutanoate and (2-Benzyloxyethyl) Oxiran (91) (92) To ethyl 3-oxobutanoate (0.26 g, 2 mmole, 0.254 ml) in THF (5 ml) under nitrogen at 0°C, were added sodium hydride (2.2 mmole, 52.8 mg.) and imidazole (5 mg.). The reaction was left overnight (15 hours). The suspension was cooled to -78°C and n-butyllithium (2.4 ramoles, 1.34 M, 1.80 ml) was added. The solution was allowed to warm to room temperature (for i hour) and the solution recooled to -78°C. The epoxide (91) (0.356 g, 2 mmole) was added. The reaction was quenched with water after 5 days at room temperature. Dichloromethane was added, and the organic phase was separated, dried and evaporated. The crude material was shown to be the starting epoxide (91) (0.28 g, 80%) by t.l.c., and n.ra.r. spectroscopy. The reaction was repeated on a 1 mmole and 20 mmole scale, but again there was no reaction. In one experiment, the THF was refluxed for 3 days, after the addition of the epoxide (91) to the dianion (92(c)), but to no avail. The epoxide (91) was recovered (56%), the rest having polymerized. To a solution of hexamethyldisilane2^ (0.146 g, 1 mmole) in chloroform (5 ml) was added iodine (0.254 g,. 1 ramole), to form iodotrimethylsilane2^ jn-situ. After \ hour under nitrogen, the oxiran (91) (0.356 g, 2 mmole) was added. 80. The solution was stirred for 2 hours and the chloroform was evaporated. The dark solution remaining, was added to the dianion of ethyl 3-oxobutanoate (2 mmole) in THF (10 ml) at -78°C (the dianion (92(c)) having been made as in [92(a)] above). The dark solution was allowed to warm to room temperature overnight and the solution stirred for 1 week. The reaction was quenched with water and the organic material extracted with dichloromethane. The dichloromethane phase was separated, dried and evaporated. The residue was found to be the starting epoxide (91) (0.21 g, 59%) from its n.m.r. spectrum. This method and series of reactions were subsequently abandoned. 81. Preparation of Pentane -1, 2, 4-triol27 (93) 2 To 2-hydroxy-4-pentene 8 (16 g) in water (533 ml) was added aqueous potassium permanganate (29.3 g, 800 ml) at 5°C with vigorous stirring. After the addition was complete, the suspension was stirred for 1 hour. The manganese dioxide was filtered-off, the water evaporated and the residue distilled, to give the title triol as a yellow viscous oil (7.41 g, 33%) b.p. 126-130°C at 5 x 10"4 mm mercury (lit.27, b.p. 150-155°C at 0.2 mm mercury; lit.2^ b.p. 149-153°C at 0.03 mm mercury); V max (neat) 3500-3200 (s.br., 0-H stretch), 1460 (s), 1420 (s), 1380 (s), 1325 (s), 1100 (br. m, C-0 stretch), 1060 (s), 1 1025 (s), 940 (m), and 890 (w) cm" , 5 (D20), 1.35 (3H, d, J= 6.0 Hz, C-5), 1.60-1.80 (2H, m, C-3), 3.55-3.80 (2H, m, C-l), 3.90-4.05 (1H, m, C-4), and 4.10-4.20 (1H, m, C-2), ra/e 119 (Mt - H), 101, 89, 71 (100). 82. Preparation of 6S(R)-Methyl-4S(R)-hydroxymethyl-(R)-Phenyl-l, 3-Dioxan(94) (a) To pentane -1, 2, 4-triol (0.48 g, 4 mmol) in dichloromethane was added benzaldehyde (0.406 ml; 1 eq.) and anhydrous zinc chloride (0.54 g, 1 eq.). After 48 hours at room temperature, more zinc chloride (0.27 g, 0.5 eq.) was added. After a further 5 days, the dichloromethane was evaporated and the product isolated by column chromatography (dichloromethane, ether gradient 1:0 to 1:1), to give the title benzylidene derivative (94) as a clear oil (93.4 mg. 11.5%), Vmax (neat) 3430 (s. br. 0-H stretch), 1455 (s), 1410 (s), 1385 (s), 1350 (s), 1220 (m), 1160 (s), 1120 (s), 1060 (s), 1025 (s), 910 (m), and 765 (sh. s, aromatic C-H out of plane deformation) cm"*, 5 1.25 (3H, d, £=6.0 Hz, C-6 methyl), 1.40-1.60 (2H, m, C-5) 2.1) (1H, br., OH), 3.40-4.10 (4H, m, C-6, C-4, CH9OH) 5.50 (0.875H, s, C-2), 5.80 (0.125H, s, C-2), and 7.25-7.50 (5H, m, Ph), m/e 208 (M+), 177 (M*"-CH20H), 122, 105 + + (100, PhC0 ), 77 (Ph ) (Found: C, 69.06; H, 7.88. C12H1603 requires C, 69.21; H, 7.74%). (b) To pentane -1, 2, 4-triol (1.95 g) in dichloromethane (35 ml) was added -dimethoxytoluene3^ (2.80 g, 1.1 eq.). To the solution was added boron trifluoride etherate (0.1 eq., 0.21 ml.) and the reaction left at 25°C under nitrogen, for 10 d. Excess triethylamine was added and the solution evaporated. The crude material was purified by column chromatography (tichloromethane ether gradient, 1:0 to 1:1) to give the 83. title benzylidene derivative (5(a) (1.96 g, 58%) as the predominant isomer (7:1) formed in the reaction, the minor isomer was not characterised. 84. Preparation of 6S(R)-Methyl-4S(R)-[Methylthio-thiocarbonyloxy]methyl- 2S(R)-Phenyl-1, 3-DioxanJ1 (95) To a solution of the pure 1,3-dioxan derivative (5(a)), (75 mg, 0.361 mmole) in THF (10 ml), at -78°C under nitrogen was added n-butyllithium and the solution was stirred for i hour. The solution was warmed to room temperature and carbon disulphide (5 eq., 0.108 ml) was added. After 2 hours stirring methyl iodide (5 eq., 0.113 ml) was added. After 15 hours at room temperature, the solution was evaporated and the crude material column chromatographed (benzene/dichloromethane gradient, 1:0 to 8:2) to give the title compound (95) as a clear oil (102 mg 95%), V max (neat) 1710 (m), 1460 (m), 1420 (w), 1385 (w), 1350 (w), 1225 (s), 1070 (s), 845 (m), 770 (m)and 710 (s, aromatic C-H out of plane deformation) cm"~l, 6 1.30 (3H, d, J=6 Hz, C-6 Me), 1.40-1.60 (2H, m, C-5), 2.50 (3H, s, S-Me), 4.0-4.4 (2H, m, C-4, C-6), 4.60 (2H, d, J=5 HZj-CI^O CSSMe), 5.50 (1H, s, C-2), and 7.20-7.55 (5H, m, Ph), m/e 298 (Mt), 282 + + + (M -CH3), 190, 177, 160, 105 (100%, PhC0 ), 91, 77 (Ph ) (Found: C, 56.14, H, 6.09. C14 H18 O3 S2 requires C, 56.35, H, 6.08%). Preparation of 6S(R)-Methyl-4S(R)-(toluene-4-sulphonyloxymethyl)-2S(R)- phenyl-l,3-dioxan (96) To a solution of the acetal (94), (1.13 g, 5.4 nnnole) in pyridine (10 ml) at 0°C, was added a solution of toluene-4-sulphonyl chloride (1.5 g, 1.5 eq.) in pyridine (10 ml). After 48 hours the solution was poured into iced water and extracted with dichloromethane. The organic phase was washed with dilute hydrochloric acid (2M), separated and dried. Evaporation of the solvent afforded the title tosylate (96) as a white solid (1.65 g, 84%) m.p. 72-73.5 °C; Vmax (nujol) 3040 (w), 3020 (w), 1605 (s, C=C stretch), 1500 (w), 1460 (m), 1370 (s), 1220 (w), 1190 (s), 1130 (w), 1115 (w), 1025 (m), 980 (m), 930 (w), 710 (s) and 675 (s, aromatic C-H out of plane deformation) cm"1, 6 1.30 (3H, d, J=6.6 Hz, C-6 Me), 1.55-1.70 (2H, m, C-5), 2.40 (3H, s, C^CHO, 4.0-4.15 (2H, m, C-4, C-6) 4.30 (2H, t, J=6.6 Hz, superimposed dd, Clh? 0 S02 Ar), 5.50 (1H, s, C-2) and 7.30-7.80 (9H, m, aromatic protons), m/e 362 (Mt), 208, 177, 155, 105 (100%, PhCOt), 91 (Found C, 62.67; H, 6.10. c19h22°5s requires C, 62.96; H, 6.12%). 86. Preparation of 4S(R), 6S(R)-Dimethyl-2S(R)-phenyl-l,3-dioxan (97) (a) The benylidene tosylate (96), (0.4 g) was dissolved in THF (10 ml) and added dropwise to a suspension of lithium aluminium hydride (46.0 mg, 1.1 eq.) in THF (10 ml) at 0°C. When the addition was complete, the solution was stirred at ambient temperature for 3 hours and refluxed under nitrogen for a further 9 hours. The reaction was quenched with saturated sodium sulphate solution and anhydrous sodium sulphate was added. The THF was evaporated and the product extracted with benzene. Evaporation and column chromatography (benzene) afforded the title compound (97) as a clear liquid. (0.126 g, 59%), V max (neat) 1600 (w, C=C stretch), 1470 (s), 1350 (w), 1200 (w), 1190 (w), 1170 (w), 1060 (m, C-0 stretch), 990 (s), 860 (m), 830 (w), 775 (m), 715 (m), and 680 (m, aromatic C-H out of plane deformation) cm"1, 6 H 1.30 (6H, d, J,6,4=6.5 Hz, C-4 Me & C-6 Me), 1.42 =12 2 Hz 1H dt (1H, t, C-5 axial, J 5eq,5ax « )> < » » =2 2 Hz =12 2 Hz) l5seq,4 ' ' i5seq,5sax * ' 3.90-4.0 (2H, m, C-4, C-6), 5.50 (1H, s, C-2), and 7.30 and 7.55 (5H, m, C6H5), 6 c 139 (Ar C-l!), 128.6 (Ar C-4»), 128.2 (Ar C-2f), 126.3(Ar C-31), 101.0 (C-2), 73.0 (C-6), 40.5 (C-5), and 21.7 (C-4 Me), m/e 192 (Mt), 105 (100, PhC0+), 77 (Pht). (Found: C, 74.86; H, 8.28%. C12H1G02 requires C, 74.97; H, 8.39%). 87. The benzylidene xanthate^l (95), (100 mg., 0.33 mmole), was dissolved in toluene (5 ml) and added dropwise over 1 hour to a solution of tri-n-butylstannane (1.05 eq., 102.5 mg) in toluene (5 ml). The mixture was refluxed for 18 hours and the reaction quenched with aqueous potassium fluoride2^. The toluene layer was separated, dried and evaporated. Column chromatography (petrol/benzene gradient, 1:0 to 0:1) afforded the starting xanthate (95), (12 mg.) and the desired dioxan (97) (15 mg., 27%). This was identical with the previous sample (97(a)) by t.l.c. and n.m.r. and mass spectroscopy. / 88. Preparation of 4S(R) - Iodomethyl-6S(R) methyl-2S(R)-phenyl-l, 3-dioxan (98) A sample of the acetal (95), (0.44 g, mmole) was dissolved in pyrrolidine (5 ml) and stirred for 6 d. The pyrrolidine was evaporated off and the residue dissolved in methyl iodide (4 ml) and stirred at room temperature for 36 hours. The solution was subsequently refluxed for 45 hours and the methyl iodide then evaporated. Chromatography (benzene) of the crude material afforded the title iodide (98) as a pale yellow solid (0.4 g, 85%) m.p. 27-29°C, V max"(Nujol) 3030 (w), 1450 (s), 1400 (m), 1380 (s), 1335 (s), 1140 (m), 1060 (m), 1040 (s, C-0 stretch), 760, and 705 (s, aromatic C-H out of plane 1 deformation) cm" , 6 R 1.3 (3H, d, = 6.5 Hz, C-6 Me), 1.4 (1H, dt, £ 5ax>4 = 2.2 Hz, J5ax,5eq = 15.7 Hz, C-5 ax.), 1.9 (1H, dt, 5315 7 Hz c 5 3 25 2 J5eqj4=2.2 Hz, Jseq^ax - > " - ( H, oct, J = 7 Hz, Jgem = 15.7 Hz, CH2I), 3.8 - 4.0 (2H, m, C-4, C-6), C 1 5.5 (lH,s, C-2), and 7.3-7.55 (5H, m, C6Hq ; 6 138.3 (Ar C-l ), 128.8 (Ar C-4'), 128.2 (Ar C-31), 126.2 (Ar C-2'), 100.9 (C-2), 76.2 (C-4), 72.8 (C-6), 38.5 (C-5), and 7.9 (CH2D, m/e 318 (M+), 195, 145, 105 (PhCOt), 98 (100%), and 77 (Ph+) (Found: C, 45.70; H, 4.9; C12H1502I requires C, 45.30; H, 4.75%). 89. Preparation 33,34 0f 3R(s)-Acetoxy-5S(R)-methyl-tetrahydrofuran- 2-one (99) (a) 3-Acetoxy-5-raethyl-2,5-dihydrofuran-2-one33 (134) (20.34 g) was dissolved in ethanol (330 ml) and palladium on calcium carbonate33 catalyst (8.7 g, 2% Pd) was added. The compound was hydrogenated at atmospheric pressure for 24 hours. The suspension was then filtered off through celite, the catalyst was extracted with more ethanol and the ethanol evaporated to give the title compound (10) as a clear liquid (20.2 g, 98%), b.p. 80-81° at 0.1 mm mercury, V max (neat) 1790 (s.br, lactone C=0), 1750 (s. br, ester C=0), 1460 (m), 1380 (s), 1240 (s), 1200 (s), 1130 (m), 1110 (m), 1060, 1030 (m, C-0 stretch), 950 (w), 910 (w), and 825 (w) cm-1, 6 1.5 (3H, d, J = 6.5 Hz, C-5 = 13 X Hz = 8 6 Hz Me), 1.90 (1H, q, J4(gem) ' » £4 ,3 ' » > 2.15 (3H, s, COCH3), 2.85 (1H, o, £3^4 = 10.8 Hz, J4(gem) = 13.1 Hz, J4 >5 = 6.8 Hz, C-^B), 4.6 (l.H, o, J = 6.5 Hz, J4 j5 = 12.2 Hz, C-5), and 5.55 (1H, 2d, J3)4 = 8.6 ^ 0.2 Hz, J3>4 = 10.8 ^ 0.2 Hz, C-3), m/e 158 (M"t"), 142 (H^" - Me), 115 (M+ - COCH3), 86, 72 (100). The exclusive cis stereochemistry was confirmed by comparison with authentic nrar38 spectral data. 90. Preparation of 3-Acetoxy-5-methylene-2.5-dihydrofuran-2-one (99(b)) Tri-O-acetyl-ribonic acid y-lactone (99(c)) (8 g) was dissolved in THF (100 ml) and cooled to -20°C under nitrogen. 1, 5- Diazabicyclo [5.4.0] undec-5-ene (4.9 ml, 1.1 eq.) was added dropwise over 20 minutes. After 3 hours at -20°C, the solution was warmed to room temperature and stirred for a further 2 hours. The solution was cooled to 0°C and quenched by pouring into a hydrochloric acid, diethyl ether and ice mixture. The ethereal phase was separated, dried and evaporated. The crude material was column chroraatographed (petrol/dichloromethane gradient, 1.0:0, 0:1) to give the title compound (99(b)) as a while crystalline solid (4.3 g, 94%) m.p. 75-76°C, Vmax (chloroform) 3040 (m, olefinic C-H stretch), 1785 (s, lactone C=0 stretch), 1740 (S, ester C=0 stretch), 1645 (m, C=C stretch), 760 (s), and 680 (w) cm-1, 6 2.32 (3H, s, 0 COCH3), 4.95 (1H, d, Jgem = 3 Hz, C-l'E), 5.25 (1H, d, Jgem = 3 Hz, C-l'Z), and 7.30 (1H, s, C-4); m/e 154 (Mt), 112, 88, 85, 83, 70, 69, 67, 61, 60, 43 (100, COCH3+) (Found: C, 54.27; H, 3.92. CyHgO^ requires C, 54.55; H, 3.92%). 91. Hydrogenation of 3-Acetoxy-5-methylene-2,5-djhydrofuran-2-one (99(b)) to 3R(S)-Acetoxy-5S(R)-methyltetrahydrafuran-2-one(99) The methylene-lactone (99(b)), (0.344 g), was dissolved in ethanol (25 ml) and palladium on calcium carbonate (2% Pd, 0.2 g) was added. The lactone was hydrogenated at atmospheric pressure for 6 hours. The catalyst was filtered off through celite, the celite was extracted with ethanol and the ethanol evaporated. The crude material was column chromatographed (petrol- dichlororaethane gradient, 9:1 to 0:1) to give a clear liquid (0.316 g, 90%) which was identical with authentic (99) by t.l.c. and n.m.r. spectroscopy. This confirmed that the hydrogenation proceeded with a cis-diastereoselectivity of 97:3. 92. Preparation33 of Di-t-butyl 2-oxobutan-l,4-dioate (100) (a) To a stirred suspension of potassium t-butoxide (7.06 g, 1.05 eq.) in ether (50 ml), under nitrogen at 0°C, was added a solution of di-t-butyl oxalate37 (12.12 g, 0.06 mole) and t-butyl acetate (6.96 g, 8.07 ml, 0.06 mole) in ether (50 ml). When the addition was complete, the ether was refluxed for 3 hours. The reaction was quenched with ether/ice/aq. hydrochloric acid. The ether layer was separated, dried and evaporated to yield the desired di-t-butyl 2-oxobutan-l, 4-dioate (100) as a white solid (11.5 g, 78.5%) m.p. 74-76°C , Vmax (nujol) 3400- 3300 (w, br. 0-H stretch), 1725 (s, C=0 stretch), 1650 (s, C=C stretch), 1460 (s), 1370 (s), 1285 (s), 1245 (s), 1150 (s, C-0 stetch), and 1140 (m) cm-1, 6 1.50, 1.55 (18H, br. 2s, OBut); m/e 245, 244 (Mt), 215, 189, 173, 159, 133 (M+ - COjBu*), 115, 87, and 57 (100). (Found: C, 59.13; H, 8.29. C12H2o05 requires C, 59.00; H, 8.25%. (b) Preparation of 4-(t-Butyloxycarboxyl)-3-hydroxy-5- methy1-2,5-dihydrofuran-2-one (100(b)) The foregoing diester (100) (10.5 g), was dissolved in THF (50 ml) and sodium hydride (100%, 1 eq., 1.03 g) was added. To this was added acetaldehyde (1.5 eq., 2.85 g) and the solution heated at 50°C for 6 hours. The solution was evaporated and the yellow solid decomposed with a 93. mixture of conc. hydrochloric acid, ice and diethyl ether. The ether layer was separated, dried and evaporated to give the title compound (100(b)) as a yellow oil (8.7 g, 95%) Vmax (neat) 3400 (s.br., 0-H stretch), 1780 (s, lactone C=0 stretch), 1715 (s, ester C=0 stretch), 1655 (s, C=C stretch), 1400 (m), 1380 (m), 1250 (br. ra), 1150 (br.m, C-0 stretch), 1060 (m), 930 (w), 850 (m), and 770 (m) cm"1, 5 1.45-1.50 (12H, m, C-5 Me and t-Bu), 5.35 (1H, br.q, C-5), and 6.8 (1H, br.s, OH); m/e 214 (Mt), 199, 188, 175, 169, 159, 141, 113, 73, 57 (lOOfeBut). Hydrolysis and Decarboxylation of 4-(t-Butyl-oxycarbonyl)- 3-hydroxy-5-methyl-2.5-dihydrofuran-2-one (100(b)) The lactone (100(b)) (1 g), was dissolved in a solution of water and THF (1:1, 10 ml), and trifluoracetic acid (5 ml) and Amberlite 1R 120 H+ resin (0.1 g) were added. The reaction mixture was stirred for 2 days and then refluxed for 12 hours. The resin was filtered off and the solution evaporated. The crude material was column chromatographed (petrol, dichloromethane gradient, 1:1 to 0:1) to give 3-hydroxy-5-methyl-2,5-dihydrofuran-2-one one11 100(c)) as a pale yellow solid (0.38 g, 72%) m.p. 68-70°. This material was identical with authentic33 material by both t.l.c. and n.m.r. spectroscopy. 94. Preparation of Pentane -1, 2R(S), 4S(R)-triol (101) (a) To lithium aluminium hydride (2.2 eq., 0.475 g), in THF (20 ml), under nitrogen, at 0°C was added a solution of the acetoxylactone (99), (0.84 g) in THF (30 ml) over 20 minutes. On completion of reduction, saturated aqueous sodium sulphate was added. Excess anhydrous sodium sulphate was added and the suspension was filtered off. Evaporation of the filtrate gave the crude triol, which was further purified by filtration through a plug of silica, to give the title compound (101), as a clear viscous oil (0.48 g, 72%), V max (nujol) 3400 (br. s, 0-H stretch), 1450 (m) 1 and 1380 (m) cm" , 6 R (D20) 1.35 (3H, d, J5>4 = 6.5 Hz, C-5), 2.62-2.75 (2H, m, C-3), 3.55-3.80 (2H, m, C-l), 3.92-4.05 (1H, m, C-2), and 4.15 (1H, q, J4j5 = 6.5 Hz, C-4), 6 c 71.6 (C-2), 68.6 (C-4), 67.2 (C-l), 44.0 (C-3), and 25.5 (C-5), m/e 120 (Mt), 83 (100) and 45. Benzoylation of an aliquot (1.2 g) using benzoyl chloride (4.6 ml) and pyridine (4 ml) in dichloromethane (15 ml) gave the derived 1, 2R(S), 4S(R)-tribenzoyloxypentane as an oil (1.28 g, 30%), (Found: C, 72.11; H, 5.68. C26 H24 06 requires C, 72.21; H, 5.59%). 95. Reduction of 3R(S)-Acetoxy-5S(R)-methyltetrahydrQfuran-2-one (99) GJith sodium borohydride38 To a mixture of acetoxylactone (99), (10 g) in water (120 ml) was added boric acid (1.96 g) and Amberlite 1R 120 H+ (25 ml). To the stirred suspension, was added sodium borohydride (14.4 g, 6 eq.) in four portions, stirring for i hour in between additions. After addition was complete, the pH was adjusted to 9-10. The reaction was stirred overnight under nitrogen at room temperature. The reaction was quenched with Amberlite 1R 120 H+ in methanol, until the solution was neutral. The resin was filtered off, and the filtrate evaporated. The resultant crude syrup was dissolved in methanol and the solution was re-evaporated. This was repeated to constant weight to leave the triol as a yellow oil (7.6 g, 100%). This was identical with the foregoing sample by both t.l.c. and n.m.r. spectroscopy. 96. Attempted Preparation of 4R(S)-Hydroxymethyl-6S(R)-methyl-2S(R) phenyl-1, 3-dioxan (102) The threo-pentanetriol (101), (0.32 g, 2.67 mmole ) was dissolved in methanol (3.5 ml) and a, a'-dimethoxytoluene (5.5 eq., 2.225 g) was added. Zinc bromide (25 mg.) and p- toluene sulphonic acid (50 mg) were added as catalysts. The solution was stirred at ambient temperature for 6 days. The reaction was quenched with aqueous sodium hydrogen carbonate solution, and dichloromethane was added. The organic layer was separated, dried and evaporated. The crude material was purified by column chromatography (petrol, dichloromethane gradient 9:1 to 1:1) to yield a mixture of several benzylidene adducts (0.2 g) as an oil. N.m.r. spectroscopy revealed a this oil contained. a complex mixture, 6 1.15 (d, J = 6.5 Hz), 1.20 (d, J = 6.5 Hz), 1.25 (d, J = 6.5 Hz) [combined integral 3H], 1.48 (1H, d, J ® 7.8 Hz), 1.52 - 2.10 (2H, m, C-5), 2.40 (1H, br.s, OH), 3.55-3.84 (2H, m, CH7OH), 3.95-4.5 (3H, m, C-4, C-6), 5.77 (s), 5.85 (s), 5.9 (s) [C-2 combined integral 1H], and 7.3-7.5 (5H, m, Ph). 97. Preparation of l-(Toluene-4-sulphonyloxy) pentane-2R(S), 4S(R)-diol (103) The triol (101), (1.14 g) was dissolved in pyridine (5 ml) and chloroform (1.5 ml), at 0°C. To this was added toluene- 4-sulphonyl chloride (1 eq., 1.81 g) and reaction mixture was warmed up to room temperature. On completion of the tosylation, the reaction was quenched with ice, dilute hydrochloric acid and diethyl ether. The ether layer was washed with dilute hydrochloric acid and then with water. The ether layer was separated, dried and evaporated to give a pale yellow oil. Column chromatography (petrol, ether gradient 1:0 to 1:1) yielded the pure mono-toluene- 4-sulphonate (103) as a white solid (1.5 g, 60%) m.p. 36-39°C, Vmax (ne*t) 3370 (br., 0-H stretch), 1605 (m, C=C stretch), 1350 (m.br.), 1180 (m), 1100 (m), 970 (m), 825 (s, aromatic C-H out of plane deformation), and 670 (s) cm-*, 6 1.35 (3H, d, Jj5>4 = 6.5 Hz, C-5), 1.45-1.65 (2H, m, C-3), 2.45 (3H, m, C6 H4 CH3), 3.5 (2H, q, Jj^ = 6.5 Hz, C-l), 3.80 - 4.0 (2H, m, C-2, C-4), 5.65 (2H, br., OH), 7.35 (2H, d, J^ = 8.6 Hz), and 7.8 (2H, d, J^ = 8.6 Hz), m/e 275 (M+l^) 257, 244, 229, 215, 200, 176 (100), 155, 107, 91. (Found: C, 52.25; H, 6.68. c12 h18 so5 requires C, 52.54; H, 6.61%). 98. Preparation39 of 2R(S), 4S(R)-Bis-(t-butyldimethylsilyloxy)-l- (toluene-4-sulphoxyloxy) pentane (104) The dihydroxytosylate (103), (4.0 g) was dissolved in DMF (20 ml)' and the solution cooled to 0°C, under nitrogen. Imidazole (5.0 eq. 4.97 g) and t-butylchlorodiraethylsilane39 (2.1 eq., 4.62 g) were added. After 3 hours, the reaction was quenched by pouring into water and diethyl ether. The ether layer was separated, dried and evaporated, to yield the title compound (104) (6.8 g, 93%) as a clear liquid, V max (neat) 1605 (s, C=C stretch), 1460 (s), 1370 (s, S=0 stretch), 1255 (s), 1200 (m), 970 (s), 905 (w), 850 (s, aromatic C-H out of plane deformation) 775 (s), and 670 (s) cm"1, <5 0.02 (12H, 2s, 4SiMe), 0.82 (18H, 2S, 2_t Bu), 1.1 (3H, d, = 6.5 Hz, C-5), 1.25-1.33 (1H, m, C-3), 1.40-1.60 (1H, m, C-3), 2.42 (3H, s, C6H4 CH3), 3.70-3.98 (4H, m, C-l, C-2, C-4), 7.34 (2H, d, JAB=8.6 HZ), and 7.75 (2H, d, J^ = 8.6 Hz), m/e 501 (M+ - H), 487 (Mt - Me), 445 (M+ - tBu), 415, 403, 361, 273, 229, 199 (100), 189, 159, 147, 117, 91, and 73 (Found: C, 57.08; H, 9.29. C24 H46 05 S Si2 requires C, 57.32; H, 9.22 %). Preparation of l-Iodo-2R(S), 4S(R)-bis-(t-butyldimethylsilyloxy) pentane (105) To magnesium turnings41 (0.6 g, 2.5 eq.) in diethyl ether (37.5 ml), at 0°C under nitrogen, was added iodine (5.08 g, 2. eq.) in small portions. After the addition was complete and the ether was decolourized, the solution was filtered to remove the excess magnesium. The filtrate was added to the tosylate (104) (5 g) in diethyl ether (15 ml). The mixture was refluxed under nitrogen for 3 hours. The reaction was quenched with aqueous sodium thiosulphate solution. The ether layer was separated, washed with water, dried and evaporated to give the title iodide (105) as a clear liquid (3.2 g, 70%), Vmax (neat) 1470 (s), 1390 (m), 1260 (s), 1080 (br. s), 830 (br. s), 785 (m) and 680 1 (m) cm" , 6 0.02 (12H, 2s, 4 SiMe3), 0.85 (9H, s, t-Bu), 0.88 (9H, s, t-Bu), 1.28 (3H, d, J=6.5 Hz, C-5), 1.39-1.52 (1H, m, C-3), 2.21 (1H, pentet, C-3), 3.70 (1H, d, J = 4.6 Hz, C-l), 3.72 (1H, d, J = 4.6 Hz, C-l), 3.98 (1H, sextet, C-2), and 4.35-4.45 (1H, m, C-4), m/e No M+, 399, 271, 173, 139, 129, 117, 91, 95, 75 (100) and 73. 100. The tosylate (104 (0.54) was dissolved in acetone (2 ml) (0.54 g) was dissolved in acetone (2 ml) and a 10% solution of sodium iodide in acetone (1.5 eq., 0.24 g in 2.4 ml) was added. The solution was refluxed for 9 hours. The inorganic salts were filtered off, and the acetone was evaporated. The crude material was chromatographed (benzene) repeatedly to give the iodide as a clear liquid (20 mg., 4%). This was identical with authentic material (105(a)) by n.m.r. spectroscopy and t.l.c. 101. Attempted Coupling Reactions of Iodo-2R(S), 4S(R),-bis-(t- butyldimethylsilyloxy) pentane (105) with the Dianion of Methyl 2-Methyl-3-Oxobutanoate42 (106) (a) To methyl 2-methyl-3-oxobutanoate (0.125 ml, 1 mmole) in THF (5 ml) was added sodium hydride (1.2 eq., 25 mg.), at 0°C under nitrogen. After the salt formation was complete, n-butyllithium (1.1 mmole, 1.69 M, 0.6 5 ml) was added. After i hour at 0°C, the dianion44 (106(a)) quenched with the iodide (105), (0.52 g, 1.1 eq.) in THF (5 ml). After 6 hours, the reaction was quenched with dilute hydrochloric acid and diethyl ether. The ether layer was separated, dried and evaporated. Column chromatography (petrol, dichloromethane gradient, 1:0, 0:1) only gave the two starting materials (confirmed by t.l.c. and n.m.r. spectroscopy). (b) To methy3r-2-methy1-3-oxobutanoate (0.8 mmole, 0.10 ml) in DME (5 ml) was added sodium hydride (1.2 eq., 23 mg.) at 0°C under nitrogen. After the salt formation was complete, n-butyllithium (1.1 eq., 1.4 M, 0.63 ml) was added. After i hour, the iodide (105) (0.4 g, 0.873 mmole) in DME (5 ml) was added. The reaction mixture was stirred at room temperature for 3 days and then quenched with dilute hydrochloric acid and diethyl ether. Separation of the ethereal layer, drying and evaporation, gave only starting materials (t.l.c. and n.m.r. spectroscopy), iodide (105) (0.14 g, 35%) and ester (40 mg, 40%). 102. (c) To methyl 2-methyl-3-oxobutanoate (0.25 ml, 2 mmole) in DME (5 ml) was added, at 0°C, sodium hydride (1.2 eq., 60 mg.). When the salt formation was complete, n-butyllithium (1.4 M, 2.2 mmole, 1.6 ml) was added followed by HMPA (2.2 mmole, 0.348 ml) and the iodide (105), (2 mmole, 0.92 g) in DME (5 ml). After 14 days, there was still no reaction. This reaction was therefore abandonded. (d) To the methyl 2-methyl-3-oxobutanoate (0.25 ml, 2 ramole) in DME (5 ml) was added, at 0°C, sodium hydride (1.2 eq., 60 mg.). When salt formation was complete, n-butyllithium (1.4 M, 2.2 mmole, 1.6 ml) was then added. Cuprous iodide (2.1 mmole, 0.38 g) was then added. The solution was stirred at 0°C, for 1 hour and the bissilyloxyiodide (105) (1 g, 2.18 mmole) in DME (5 ml) was added. After 1 week stirring at room temperature, the usual work-up afforded only starting materials, the iodide (105) (0.48 g, 48%) and ester (0.1 g, 38%). (e) The dianion of methyl 2-methyl-3-oxobutanoate (0.25 ml, 2 mmole) was formed as above, with sodium hydride (60 mg, 1.2 eq.) and n-butyllithium (1.4 M, 2.2 mmole, 1.6 ml). After i hour, phenylthiocopper (I) (2 mmole, 0.345 g) was added. The yellow suspension was stirred for 1 hour and then the iodide, (105) (1 g, 2.18 mmole) in DME (5 ml) was added. After 7 d, there was no reaction and the experiment was abandoned. 103. Attempted Condensation of Methyl 3-(Methylamino)-2-methylbut-2- enoate (107(a)) with Iodo-2R(S), 4S(R)-bis-(t-butyldimethylsilyloxy) pentane (105) (107) To methyl 3-(methylamino)-2-methylbut-2-enoate42 (2 mmole, 0.286 g) in THF (5 ml) was added n-butyllithium (1.4 M, 2.2 eq., 1.6 ml) at 0°C whilst stirring under nitrogen. After i hour, the iodide (105), (1 g, 2.18 mmole) in THF (5 ml) was added. After 4 days, stirring at room temperature, the reaction was quenched with dilute hydrochloric acid and. diethyl ether. The ether layer was separated, dried and evaporated. T.l.c. and n.m.r. spectroscopy showed the presence of methyl 2-methyl-3-oxobutanoate and the unreacted iodide (105). This reaction was abandoned. 104. Preparation and Attempted Condensation of t-Butyl 2-Methyl-3- oxobutanoate with Iodo-2R(S), 4S(R)-bis(t-butyldimethylsilyloxy) pentane (105) (108) To t-butyl 3-oxobutanoate45 (20 g) in THF (100 ml) at 0°C, under nitrogen, was added sodium hydride (1.1 eq., 3.34 g) in small portions. When the salt formation was complete, methyl iodide (1.1 eq., 8.7 ml) was added. The reaction was stirred overnight at room temperature. The inorganic salts were filtered off and the THF evaporated. Distillation afforded the desired t-butyl 2-methyl-3-oxobutanoate (108(a)) as a clear liquid (18.4 g, 85%), b.p. 66-67° at 6 mm mercury, Vmax (neat) 3550 (w. br., enol 0-H stretch), 1730 (s. br., C=0 stretch), 1455 (m), 1395 (w), 1370 (m), 1270 (m), 1170 (m), 960 (m), and 850(s) cm"1, 6 1.25 (3H, d, J = 6.5 Hz, C-2 Me), 1.50 (9H, s, t-Bu), 2.20 (3H, s, C-4) and 3.30 (1H, q, J = 6.5 Hz, C-2), m/e No Mt, 115 (Mt-t-Bu), 99 (Mt - OBu*), + 57 (100, t-Bu ) (Found: C, 62.88; H, 9.51. C9 H16 03 requires C, 62.77; H, 9.36%). To the t-butyl ester (108(a)) (0.35 g, 2 mmole) in THF (30 ml) at 0°C under nitrogen, was added n-butyllithium (2.2 eq., 1.6 M, 2.79 ml). The dianion was stirred at 0°C for i hour and cooled to -78°C. The iodide (105), (0.93 g, 2.03 mmole) was added and the temperature maintained at -78°C for 4 hours. The reaction mixture was allowed to warm up to room temperature, overnight. The solution was recooled to 0°C and the reaction mixture was quenched with glacial acetic acid (0.5 ml). The solvent was evaporated off and the residual organic material was extracted 105. with dichloromethane. Evaporation and chromatography afforded the two starting materials, the ester (108(a)) (0.2 g, 57%) and the iodide (105) (0.5 g, 54%) by n.m.r. and t.l.c. This reaction was subsequently abandoned. 106. Preparation of Methyl 2-Methyl-3-oxoheptanoate (109) To methyl 2-methyl-3-oxobutanoate (0.254 ml, 2 mmole) in THF (6 ml) at 0°C under nitrogen was added sodium hydride (1.2 eq., 58 mg.) When the salt formation was complete, n- butyllithium (1.69 M, 1.3 ml, 2.2 mmole) was added. The dianion44 was stirred for i hour. This was then quenched with iodopropane (2.2 mmole, 0.215 ml) and allowed to warm to room temperature and stirred for a further 15 hours. The reaction was quenched with dilute hydrochloric acid and diethyl ether. The ether layer was separated, dried and evaporated, the crude material was purified by column- chromatography (petrol, dichloromethane gradient, 3:1 to 0:1) to give the title compound (109) as a clear liquid (0.22 g, 64%), V max (neat) 1755 (s, ester C=0 stretch), 1710 (s, kectone C=0 stretch), (1450 (m), and 960 (w) cm-1, 6 0.8-1.70 (10H, m, C-5, C-6, C-7, C-2 Me), 2.6 (2H, br. t, J = 7.0 Hz, C-4), 3.6 (1H, q, J = 7.0 Hz, C-2), and 3.8 (3H, s, OMe), m/e 172 (Mt), 158, 141 (Mt -OMe), 130, 116, 101,and 85 (100) (Found: Mt 172.1097. C9 H16 03 requires Mt 172.1099 ). 107. Preparation39 of 3R(S)-(t-Butyldimethylsilyloxy)-5S(R)- methyltetrahydrofuran-2-one (110) (a) To the hydroxy "/-lactone (100(c)), (0.8 g)m in DMF (10 ml) was added imidazole (2.5 eq., 1.2 g) and t-butylchlorodimethylsilane (1.1 g). The solution was stirred at room temperature under nitrogen. When the silylation had gone to completion (15h) the solution was poured into diethyl ether and water. The ether layer was separated, dried and evaporated. Column chromatography (petrol, dichloromethane gradient 1:0, 0:1) afforded the desired 3-(t- butyldimethylsilyloxy)-5-methyl-2,5-dihydrofuran-2-one as a pale-yellow oil (0.85 g, 53%), V max (neat) 1780 (s, lactone C=0 stretch), 1665 (s, C=C stretch), 1470 (s), 1310 (s), 1260 (s), 1130 (m), 1120 (m), 1030 (m, C-0 stretch), 910 (s), 860 (s), and 800 (s), 1 cm" , 5 0.2 (6H, s, SiMe2), 1.0 (9H, s, t-Bu), 1.4 (3H, d, J5 4 = 6.0 Hz, C-5 Me), 4-85-5.15 (1H, m, C-5), 6.2 and (1H, d, J^ 5 = 3 Hz, C-4), m/e No M+, 171 (M+- - tBu), 132, 75, 73. (b) Hydrogenation of 3-(t-Butyldimethylsilyloxy)-5- methyl-2,5-dihydrofuran-2-one The foregoing silyl enol ether (110(a)) (0.7 g) was dissolved in ethanol (25 ml) and palladium on calcium carbonate (2% Pd, 0.5 g) catalyst was added. The 108. suspension was hydrogenated at atmospheric pressure. On completion of hydrogenation (48 hours), the catalyst was filtered off, and the ethanol evaporated to give 3R(S)-(t-butvldimethylsilvloxv)-5S(R)- methyltetrahydrofuran-2-one (110) as a clear liquid (0.58 g, 82%), Vmax (neat) 1780 (lactone C=0 stretch), 1455 (s), 1390 (w), 1150 (br.s, C-0 stretch), 1050 (m), 1010 (s), 950 (m), 900 (m), 850 (s), and 780 (s) cm"1, 6 0.10 (6H, s, SiMe7), 0.85 (9H, s, t-Bu), 1.37 (3H, d, J = 4 Hz, C-5 Me), 1.60-2.04 (1H, m, C-4), 2.24-2.74 (1H, m, C-4): 3.18-3.64 (1H, m, C-5), and 4.10- 4.64 (1H, m, C-3), m/e No Mt, 215 (Mt - Me), 173 (M+-t-Bu), 129, and 75 (Found: C, 57.59; H, 9.70. C11 H22 °3 Si requires C, 57.35; H, 9.63%). 109. Preparation of 2S(R)-[2R(S)-Hydroxypropyl]oxirane (111(a)) To l-(toluene-4-sulphonyloxy)-pentane-2R(S), 4S(R)-diol (103), (0.65 g) in THF (25 ml) at 0°C under nitrogen, was added 1,5-diazabicyclo [5.4.0] undec-5-ene (1 eq., 0.355 ml). When the reaction was complete, ice-water and diethyl ether were added. The ether layer was separated and the aqueous phase re-extracted with ether. The combined ether extracts were washed rapidly with dilute hydrochloric acid and then with aqueous sodium hydrogencarbonate solution. The ethereal layer was separated, dried and evaporated. Column chromatography (dichloromethane, ether gradient 1:0, 8:2) yielded the title epoxide (111(a)) as a clear liquid (0.11 g, 46%), Vmax (neat) 3420 (br. 0-H stretch), 3050 (w), 1450 (m), 1410 (s), 1370 (s), 1260 (s, epoxide C-0 stretch), 1140 (s, C-0 stretch), 1060 (m), 1020 (m), 960 (m), 920 (m), 850 (s), 830 (s) and 760 (m) cm"1, 6 1.25 (3H, d, J = 6 Hz, C-31), 1.55-2.25 (3H, m, C-lf,-0H), 2.35-3.35 (3H, m, C-l, C-2), and 4.10 (1H, br. q., C-2'), m/e 101 (Mt-H), 87 (M* - Me), 69, 58 (100), 57, 45, 43. 110. Preparation4^ of (2-Hydroxypropyl) oxirane (111(b)) 2-Hydroxy-4-pentene2B (15.2 g) was dissolved in dichloromethane (100 ml) and peracetlc acid (40% w/v, 1.2 eq., 39.8 ml) buffered with sodium acetate trihydrate (3 g) was gradually added. When the epoxidation was complete, the organic phase was separated, washed with aqueous sodium hydrogencarbonate solution and evaporated. Distillation afforded the title epoxide (111(b)) as a clear liquid (3.68 g, 21% b.p. 88-90° at 22 mm. mercury, Vmax as above, 6 1.22 (3H, d,. J = 6 Hz, C-3f), 1.45-1.95 (2H, m, C-l?), 2.35-2.90 (3H, m, C-l, C-2), 3.15 (1H, br. s, -OH), 4.0 (1H, br. q., J = 6 Hz, C-2f), m/e as above. 111. Preparation39 of 2S(R)-[2R(S)-t-Butyldimethylsilyloxypropyl ] oxirane (112(a)) The epoxide (111(a)) (50 mg.) was dissolved in DMF (3 ml) and imidazole (2.5 eq., 85 mg.) was added at 0°C, under nitrogen. t-Butylchlorodimethylsilane (75 mg., 1 eq.) was added. On completion of silylation, ice-water and diethyl ether were added. The organic phase was separated, washed with water, dried and evaporated to give the title epoxide (112(a)) as a clear liquid (0.1 g, 92%), Vmax (neat) 1460 (m), 1380 (w), 1260 (s, epoxide C-0 stretch), 1220 (m), 1090 (s, Si-0 stretch), 1065 (w), 1040 (w), 1000 (w), 910 (w), 840 (m), 770 (s), 740 (m), and 670 (m) cm"1, 11 6 0.05 (6H, s, SiMe2), 0.85 (9H, s, SiBu ), 1.15 (3H, d, J = 6 Hz, C-31), 1.40-1.80 (2H, m, C-l'), 2.30-3.15 (3H, m, C-l, C-2), and 4.0 (1H, br. q., C-2'), m/e 215 (Mt-H), 201 (Mt-Me), 183, 173, 159 + + + (M -t-Bu), 141, 127, 115, 101 (M - SiBu Me2), 85 (100), 57 (tBut). (Found: C, 60.90; H, 11.45. Cn H24 02 Si requires C, 61.05; H, 11.18). 112. Preparation3® of (2-t-Butyldimethylsilyloxypropyl)oxirane (112(b)) To the mixture of the erythro and threo epoxyalcohols (111(b)), (3.6 g) in DMF (10 ml) at 0°C, under nitrogen, were added imidazole (2.5 eq., 6 g) and t-butylchlorodimethylsilane. (1.0 eq., 5.3 g). On completion of silylation (15 hours), the reaction was quenched with ice-water and diethyl ether. The ethereal layer was separated, washed with water, dried and evaporated. Distillation afforded the silyloxyepoxide as a clear liquid (2.82 g, 37%) b.p. 92-95° at 17 mm mercury, V^x as above, 6 0.05 (6H, br. s. SiMe?) 1 0.85 (9H, br. s, Si But), 1b05 (d, J = 6 Hz, C-3 ), 1.15 (d, J = 6 Hz, C-3f), [combined integral 3H], 1.40-1.80 (2H, m, C-l'), 2.30-3.15 (3H, m, C-l, C-2), 4.0 (1H, br. q., C-2f), m/e as above. 113. (a) Attempted Condensation47 of the Epoxide (112(a)) with Methyl 2-Methy1-3-oxobutanolate4z (113) Methyl 2-methyl-3-oxobutanoate (0.064 ml, 0.5 mmole), was dissolved in THF (5 ml) at 0°C, and stirred under nitrogen and sodium hydride (1.1 eq., 13.2 mg) was added. When the salt formation was complete, n- butyllithium (1.4 M, 1.1 eq., 0.4 ml) was added. After i hour, the silyloxyepoxide (112(a)), (0.1 g) in THF (5 ml) was added to the dianion. The reaction mixture was stirred at room temperature for 15 hours. The mixture was quenched with glacial acetic acid (0.25 ml), the solvent was evaporated and the constituents isolated by column chromatography (petrol, dichloromethane gradient 5:1, 0:1). The two isolated components were the epoxide (112 (a)) (60 mg., 60%) and methyl 2-methyl 3-oxobutanoate (25 mg, 40%), (confirmed by n.m.r. and t.l.c.) indicating that no reaction had occurred. (b) Attempted Condensation48 of Epoxide (112(b)) with Methyl 2-Methyl-3-oxobutanoate To di-isopropylamine (2.06 ml, 14.7 mmole) in DME (6 ml) at 0°C under nitrogen, was added n-butyllithium (1.55 M, 8.66 ml, 13.4 mmole). After 20 minutes, methyl 2-methyl- 3-oxobutanoate (1.7 ml, 13.4 mmole) was added to the lithium di-isopropylamide solution and the mixture was stirred at 0°C for | hour. The dianion was warmed up 114. to room temperature and a solution of the silyloxy epoxide (112(b)), (0.29 g, 1.34 mmole) in DME (3 ml) was added After 4 days at 40°C, no reaction (t.l.c. and n.m.r.) was noticed and the reaction was abandoned. Attempted Condensation of the Epoxide (112(b)) with Methyl 2-Methyl-3-oxobutanoate To methyl 2-methyl-3-oxobutanoate (0.254 ml. 2 mmole), in THF (15 ml) at 0°C, under nitrogen, was added sodium hydride (1.1 eq., 52 mg.). After the salt formation was complete, n-butyllithium (1.6 M, 2.2 mmole, 1.375 ml.) was added. After i hour, the solution was cooled to -78°C and titanium tetrachloride (1 eq., 0.22 ml) was added. A black suspension was immediately formed. The suspension was stirred vigorously for 2 hours and the epoxide (112(b)), (0.43 g, 2 mmole) was added. The reaction was allowed to warm up from -78°C to room temperature over 15 hours. Quenching with dilute hydrochloric acid and diethyl ether and examination of the ethereal phase showed that the two starting materials (n.m.r., t.l.c. and mass spectroscopy). This reaction was subsequently abandoned. 115. (a) Attempted Condensation of 3R(S)-Acetoxy-5S(R)-methyltetrahydrofuran- 2-one (99) with Methyl Lithioacetate (114) To di-isopropylamine (4.4 eq., 1.95 ml), in THF (40 ml) at 0°C under nitrogen, was added n-butyllithium (1.6 M, 4.0 eq., 8.7 ml). The lithium di-isopropylamide solution was stirred for i hour and methyl acetate (4.0 eq., 0.965 ml) was added. After ± hour at 0°C, the solution was cooled to -78°C. The acetoxy-lactone (99), (0.5 g, 3.16 mmole, 1 eq.), was added. Thfe reaction was stirred for 1 hour at -78°C and warmed to 0°C. The reaction mixture was allowed to warm up to ambient temperature and stirred for a further 48 hours. The solution was recooled to 0°C and quenched with glacial acetic acid (0.5 ml). The THF was evaporated and the residue partitioned between water and ether. The ether layer was separated, dried and evaporated to yield an intractable orange liquid. This reaction was abandoned. 69 (b) Condensation of Lactone (99) with t-Butyl lithioacetate To di-isopropylamine (4.4 eq., 7.8 ml) in THF (50 ml) at 0°C under nitrogen, was added n-butyllithium (4.0 eq., 1.55 M, 36 ml). The lithium di-isopropylamide solution was stirred for i hour at 0°C and cooled to -78°C. t-Butyl acetate (4.0 eq., 6.84 ml) was added, and the solution was stirred for a further i hour. The acetoxy-lactone (99) (2 g) was added and the solution maintained at -78°C for a further 1 hour. The reaction mixture was allowed 116. to warm to room temperature overnight. The solution was recooled to 0°C and the reaction was quenched with glacial acetic acid (3.2 ml). The solvent was evaporated, and the residue extracted with diethyl ether. The ether was filtered and evaporated to give an oil. This was column chromatographed (petrol, dichloromethane gradient 1:1, 1:3), to give t-butyl-3-oxobutanoate (0.98 g, 49%) (t.l.c. and n.m.r.) and 2-hvdroxy-3R(S)-hydroxy-2-[(t-butyloxy carbony1) methyl1-5S(R)-methyltetrahydrofuran as a yellow oil (1.86 g, 63%), Vmax (neat) 3450 (s. br., 0-H stretch), 1720 (s, ester C = 0 stretch), 1460 (m), 1370 (w), 1250 (m), 1160 (m), 1060 (m) and 740 (w), cm"1, 6 1.25 (3H, d, J = 6.0 Hz, C-5 Me), 1.45 (9H, s, t-Bu), 1.75-2.50 (1H, m, C-4 a), 2.65(s), 2.80 (br.s) [combined integral 2H, C-l1], and 3.80 - 4.55 (3H, m, C-3, C-5, C-4 B), m/e No Mt, 215 (Mt - OH), 176 (Mt-t-Bu), 159 (100, Mt - OBut), 141, 132, 117, 105, 99, 89, 87, 86, and 72. (Found: C, 56.85; H, 8.74. C^ H20 O5 requires C, 56.88; H, 8.68%). 117. Attempted Conversion of the Dihydroxyester (114(b)) to 3R(S)- Hydroxy-5S(R)-methyl-2-[(Methoxycarbonyl) methylene]- tetrahydrofuran (115) The dihydroxyester (114(b)), (0.25 g) was dissolved in methanol (9.5 ml) at 0°C under nitrogen and trifluoracetic acid (0.5 ml) was added. After 1 hour, the solution was evaporated and the resultant oil chromatographed (petrol, dichloromethane gradient 4:1 to 0:1) to give 2R(S), 6- dioxabicvclo [3.3.0] -7S(R)-methyl-5R(S)-methoxvoctan-3- one as a clear oil (0.10 g, 54%), V ^^ (neat) 1790 (s, lactone C=0 stretch), 1445 (m), 1410 (w), 1390 (m), 1310 (w), 1260 (m), 1220 (m), 1150 (m), 1110 (m), 1070 (m), 1005 (m), 980 (m), 940 (w), 920 (w), 880 (w), 840 (w) and 720 (w) cm"1, 6 1.34 (3H, d, J = 6.0 Hz, C-7 Me), 1.74 (1H, dq, J = 2.3 Hz, J = 8.0 Hz, J = 13.7 Hz, C-8) , 2.55 (1H, p, C-8), 2.85 (2H, q, J » 17 Hz, C-4), 3.35 (3H, s, OMe), 4.32 (1H, m, C-7), and 4.76 (1H, dd, J = 2.3 Hz, J = 7.0 Hz, C-l), m/e 172 (Mt), 157 (Mt - Me), 141 (Mt - OMe), 130, 113 and 101 (100). (Found: C, 55.58; H, 7.27. Cg H12 O4 requires C, 55.81; H, 7,02%). 118. Attempted Condensation of Lactone (99) with Methyl 2-Methyl-3-oxobutanoate^z (116) To methyl 2-methyl-3-oxobutanoate (0.823 g, 0.8 ml) in THF (10 ml) at 0°C under nitrogen, was added sodium hydride (1.2 eq., 0.182 g). When salt formation was complete, n-butyllithium (1.2 eq., 1.6 M, 4.75 ml) was added. The solution was stirred at 0°C for i hour and then cooled to -78°C. The preformed dianion was transferred into a solution of the acetoxylactone (99), (1 g) in THF (10 ml) at -78°C. The solution was allowed to warm up to room temperature and stirred for 48 hours. The reaction mixture was recooled to 0°C and quenched with glacial acetic acid (0.5 ml). The THF was evaporated and the residue partitioned between water and diethyl ether. The ether layer was separated, dried and evaporated. Column chromatography of the residue, (petrol, dichloromethane gradient, 3:1 to 0:1), afforded the starting materials methyl 2-methyl-3-oxobutanoate (0.1 g, 12%) and lactone (99) (0.1 g, 20%) and intractable polymer. To di-isopropylamine (1.07 ml, 2.42 eq.) in THF (40 ml) at 0°C under nitrogen was added n-butyllithium (1.5 M, 2.2 eq., 4.66 ml). The lithium di-isopropylamide solution was stirred for i hour and methyl 2-methyl-3-oxobutanoate (0.4 ml, 3.15 mmole) was added. After 1 hour at 0°C, 119. the solution was cooled to -78°C and the acetoxy lactone (99), (0.5 g) was added. The reaction mixture was allowed to warm-up to 0°C and stirred for 1 hour. The reaction mixture was stirred at room temperature for 72 hours. The solution was then recooled to 0°C and the reaction quenched with glacial acetic acid (0.5 ml). The THF was evaporated, and the residue partitioned between water and ether. The ether layer was separated, dried and evaporated. The residue contained only the two starting materials (t.l.c. and n.m.r.). This set of reactions were abandoned• 120. Preparation of 2R(S)-[3-(t-Butyloxycarbonyl)-2-oxobutyl]-3R(S) acetoxy-5S(R)-methyltetrahydrofuran (117) (a) To t-butyl 2-methyl-3-oxobutanoate (108(a)), (1.72 g, 10 mmole) in THF (70 ml) under nitrogen at 0°C, was added sodium hydride (1.1 eq., 0.26 g). When salt formation was complete, n-butyllithium (1.55 M, 1.1 eq., 7.1 ml) was added. The dianion was stirred at 0°C for 40 minutes, and then cooled to -78°C. After | hour, the acetoxy lactone (99), (1.58 g, 10 mmole) was added, and the solution allowed to warm to 0°C. The reaction was then stirred for 3 days at room temperature and recooled to 0°C when glacial acetic acid (0.75 ml) was added. The solvent was evaporated and the residue extracted with diethyl ether. The ether layer was filtered and evaporated to give an oil which was purified by column chromatography (petrol, dichloromethane gradient, 3:1 to 0:1). This afforded the desired 2-[3-(t- butyloxycarbonyl)-2-oxobutyl] -2-hydroxy-3R(S)-acetoxy-5S(R) -methyltetrahydrofuran as a yellow oil (1.15 g, 35%), which was used directly in the next stage. (b) The hydroxy adduct (117(a)), (1.0 g) was immediately dissolved in THF (50 ml) at 0°C and trifluoracetic acid (0.2 ml) was added. On completion of the elimination reaction, palladium on carbon (10% Pd, 0.5 g) was added. The mixture was hydrogenated at atmospheric pressure. The catalyst was 121. then filtered through celite (5 g), washing through with more THF and the solvent was then evaporated. Purification by column chromatography (petrol, dichloromethane gradient 3:1 to 0:1) afforded the title compound (117) as a clear oil (0.89 g, 93%), V^x (neat) 3450 (w, enol 0-H stretch), 1730 (br. s, C=0 stretch), 1600 (w), 1460 (w), 1370 (m), 1240 (w), 1160 (s, C-0 stretch), and 1080 (m) cm"1, 6 1.35 (3H, d, J = 6.6 Hz), 1.47 (12 H, overlapping s and d, OBut, C-5 Me), 1.9 - 2.1 (4H, overlapping s and m, OCOCH3 and C-4), 2.61 (2H, dq, J = 15.5 Hz,J = 6.6 Hz), 3.0-3.15 (1H, m, C-4), 3.47 (1H, q, J = 6.6 Hz), 4.50-4.60 (1H, m, C-5), 5.1-5.40 (2H, m, C-l, C-2), m/e 313 (M+ - H), 299, 257, 252, 240, 213, 196, 159, 147, 143, 129, 103, 87, 71, 57 (100) (Found: Mt-H 313 .1649. Clfi H2fi 0fi requires M+-H 313.1651). 122. Attempted Fragmentation49 of 2R(S)-[3-(t-Butylo*ycarboxyl)-2- oxobutyl]-3R(S)-acetoxy-5S(R)-methyltetrahydrofuran (117) via its dianion (118) To di-isopropylamine (0.6 ml) in THF (5 ml) at 0°C under nitrogen, was added n-butyllithium (1.6 M, 2.41 ml). After i hour, the lithium di-isopropylamide solution was cooled to -78°C and the tetrahydrofuran-4-butanoate ((117), 0.55 g) in THF (5 ml) was added. The reaction mixture was allowed to warm to 0°C over 6 hours and stirred at room temperature for 15 hours. The solution was cooled to 0°C and the reaction mixture was quenched with glacial acetic acid (0.4 ml). The solution was evaporated and the residue was extracted with diethyl ether. The ethereal extract was filtered and evaporated. Column chromatography (petrol, dichlororaethane gradient 1:1 to 0:1) afforded only decomposition products. This reaction was therefore abandoned. 123. Preparation of 2,3R(S)-Dihydroxy-5S(R)-methyltetrahydrofuran (119) Di-isobutylaluminium hydride (3.1 eq., 33% w/w, 30 ml), was added to toluene (25 ml) at -78°C, under nitrogen. To the solution was added 3R(S)-acetoxy-5S(R)-methyltetrahydrofuran-2- one (99), (3.1 g). After 2 hours at -78°C, the reaction mixture was quenched with glacial acetic acid (3.5 ml, 3.1 eq.) over 30 minutes. The oil was allowed to warm up to 0°C, and water was gradually added. The precipitate was filtered off and soxhlet extracted with ethyl acetate for 15 hours. The filtrate and soxhlet extracts were combined and evaporated to give the title compound (119) as a clear oil (1.97 g, 85% yield), Vmax (neat) 3400 (br. s., 0-H stretch), 1720 (m, C=0 stretch), 1450 (s), 1390 (s), 1260 (s), and 1060 (s) cm"1, 6 (CDCI3) 1.45 (3H, d, Me, J = 7.0 Hz), 1.85-2.75 (2H, m, C-4), 3.85-4.55 (2H, m, C-3, C-5), 4.75 (2H, br.s., -OH), 5.15-5.55 (1H, m, C-2), m/e 118 (Mt), 83 (100) and 47. Benzoylation of an aliquot (0.2 g) using benzoyl chloride (0.4 ml) and pyridine (0.5 ml) in benzene (5 ml) gave the derived 2. 3R(S)- dibenzoyloxy-5S(R)-methyltetrahydrofuran (0.4 g, 73%) as an oil. (Found: C, 69.98; H, 5.83. C19 H18 05 requires C, 69.93; H, 5.56%). 124. Preparation of 5S(R)-[2S(R) Hydroxypropyl] tetrahydrofuran-2-one (120) Ethoxycarbonylmethylenetriphenylphosphorane55 (7.58 g 1 eq.) was added to the lactol (119), (2.5 g) in THF (50 ml). The reaction mixture was stirred at room temperature for 14 hours and the mixture refluxed for three hours. The solution was filtered through Kieselgel H (5 g) washing through with more THF and 5% rhodium on alumina (1 g) was added. The enoate (120(a)) was hydrogenated at atmospheric pressure for 15 hours. The catalyst was filtered through celite (10 g) washing through with more THF and the THF was evaporated. The crude product was partitioned between water and diethyl ether and the aqueous phase separated. Evaporation of the water gave mainly, ethyl 4S(R), 6S(R) dihydroxyheptanoate. The crude material was dissolved in dichloromethane (40 ml) and refluxed with trifluoracetic acid (1 ml) for 6 hours. Evaporation of the solvent and subsequent distillation afforded the title compound (120) as a clear oil (1.84 g, 60%), b.p. 120° at 0.35 mm mercury, V max (neat) 3420 (br. s, OH stretch), 1780 (s, lactone C=0 stretch) 1470 (ra), 1425 (m), 1390 (m), 1345 (m) and 1160 (s., C-0 stretch) cm"1, 6 1.6 (3H, d, J2«3»= 6 Hz, C-3f), 1.66 (2H, m, C-lf), 1.73- 1.90 (1H, m, C-4) 2.24-2.36 (1H, m, C-4), 2.45-2.52 (2H, m, C-3) 2.54 (1H, br. s, OH), 3.93-4.04 (1H, m, C-2'), and 4.65-4.76 (1H, m, C-5), m/e 144 (M+), 127, 126 (11+ - H20, 100%), 116, 100, 85, 71, 67, 55, and 42. The lactone was microanalysed as the following t-butyldimethylsilyl ether (121). 125. Preparation of 5S(R)-[2S(R)-(t-Butyldimethysilyloxy)propyl] tetrahydrofuran-2-one (121) To the hydroxylactone (120), (60 mg.) in DMF (5 ml) was added imidazole (2.5 eq., 71 mg.) and t-butylchlorodimethyl- silane (1.05 eq., 66 mg.) at 0°C under nitrogen. After 15 hours at room temperature, the reaction mixture was quenched with water, and diethyl ether was added. The ether layer was separated, dried and evaporated to give the title compound as a clear oil (90 mg., 84%), V max (neat) 1780 (s, lactone C=0 stretch), 1460 (m), 1360 (m), 1250 (s), 1160 (s, C-0 stretch), 1100 (m), 1050 (s, C-0 stretch), 1000 (s), 910 (m), 830 (m) 1 88 9H s and 770 (m) cm" , 6 0.05 (6H, s, SiMe2), °* ( > > SiBu*), 1.16 (3H, d, J = 6.5 Hz, C-3'), 1.64-1.72 (2H, m, C-l'), 1.76-1.93 (1H, m, C-4), 2.28-2.41 (1H, m, C-4), 2.49-2.56 (2H, m, C-3), 3.96-4.12 (1H, ra, C-21) and 4.62-4.72 (1H, m, C-5), m/e 259 (ttt + H), 201 (100), 183, 159, 158, 157, 105, 101, 85, 75, 73 and 57. (Found: C, 60.61; H, 10.26. C13 H2$ 03 Si requires C, 60.42; H, 10.14%). 126. Due to isolation problems involved in the preparation of lactone (120), caused by triphenylphosphine oxide, reactions were carried out as above on the crude hydroxylactone (120). The silyl ether (121) was then easily purified by column chromatography (petrol, dichlororaethane gradient, 1:0 to 0:1) to remove the triphenylphosphine oxide. Generally, the conversion of lactol (119) (0.67 g) to the isolated silyl ether (121) (0.8 g) went in 55% yield. 127. Preparation of 2S(R)-[2S(R)-(t-Butyldimethylsilyloxy)propyl3-5-E- [l-t-butyloxycarbonyl)ethylidene3tetrahydrofuran (122)9>^ To di-isopropylamine (11 eq., 2.7 ml) at 0°C in THF (30 ml) under nitrogen was added n-butyllithium (1.52 M, 10 eq., 11.47 ml). After 0.5 hours at 0°C, the lithium di-isopropylamide solution was cooled to -78°C. To this was added t-butyl propanoate (2.27 g, 10 eq.). After 40 minutes at -78°C, the silyloxy lactone (121), (0.45 g, 1.74 mmole, 1 eq.) was added. The reaction mixture was allowed to warm up to room temperature, overnight. The solution was recooled to 0°C and quenched with glacial acetic acid (1.2 ml). The solvent was evaporated and the crude product partitioned between water and diethyl ether. The ether layer was separated, dried and evaporated. The crude material was column chromatographed (petrol, dichloromethane gradient, 4:1 to 0:1) to yield a pale yellow oil (0.477 g, 70.5%) which was probably the condensed hydroxy ester (122(a)). This material (0.1 g), was dissolved in THF (10 ml) and 1R-120H+ (0.2 g) resin was added. The mixture was refluxed for 8 hours under nitrogen, the resin was filtered and the THF evaporated. Column chromatography of the residue (petrol, dichloromethane gradient, 1:0 to 0:1), gave the title compound (122) as a clear oil (87 mg., 91% yield), Vmax (neat) 1690 (s. sh., ester C=0 stretch), 1638 (s, C=C stretch) 1640 (m), 1365 (s), 1305 (s), 1255 (s), 1110 (s), 1080 (m), 910 (m), 840 (m), 780 (m) and 740 (s) cm"1, + X (ethanol) 246 mm, ( 13900 — 700), 6 0.07 (6H, s, max e 1 SiMe2), 0.88 (9H, br. s., Si Bu^, 1.17 (3H, d, C-3 , J2t3» = 128. 5.8 Hz), 1.47 (9H, s, OBut), 1.55-1.70 (3H, m, C-l1 and C-3), 1.77 (3H, s, Me), 2.10-2.23 (1H, m, C-3), 2.83-2.97 (1H, m, C-4), 3.07-3.23 (1H, m, C-4), 4.0-4.1 (1H, m, C-2'), and 4.42-4.54 (1H, m, C-2), m/e 371 (MT), 313 (Mf - tBu) 297, (MT - OBut), 257, 213, 165, 159, 130, 113, 75, 73 and 57 (100) (Found: C, 64.58; H, 10.59. C2Q H38 O4 Si requires C, 64.82; H, 10.33%). Generally, it was found that a 1 pot reaction gave comparable yields. The silyloxylactone (121), (0.4 g) was condensed with t-butyl-2-lithiopropanoate. The reaction was quenched with glacial acetic acid, the solvent evaporated, and the crude adduct partitioned between water and ether. The ether layer was separated, dried and evaporated. The crude product was dissolved in THF (15 ml) and refluxed under nitrogen with 112-120H4" (1 g.). The resin was filtered off and the solvent was evaporated. The crude material was column chromatographed to give the title compound (122) as a clear oil (0.43 g, 75% yield). 129. Preparation of (±) t-Butyl 8-0-t-Butyldimethylsilylnonactate9 (123) (a) To the furylidene (122) (88 mg.) in THF (20 ml) in a pressure bottle was added rhodium on alumina (5%, 0.1 g). The pressure bottle was charged with hydrogen at 65 psi and the mixture vigorously shaken. After 72 hours, the catalyst was filtered off and the THF was evaporated. The crude material was column chromatographed (petrol, dichlororaethane gradient, 1:0 to 1:3) to separate starting material (122) (24 mg recovered) and the title compound (123) (58 mg, 89%) which was isolated as a clear oil, Vmax (neat) 1735 (s, ester C=0 stretch), 1460 (s), 1390 (m), 1365 (s), 1260 (s), 1150 (s, C-0 stretch), 1060 (s, C-0 stretch), 900 (m), 840 (s), 810 (m) 1 88 9H s t and 780 (s) cm" , 6 0.05 (6H, s, SiMe2), °« ( > > Si Bu ), 0.98 (3H, d), 1.06 (3H, d, J = 7.0 Hz), C-2 Me), 1.13 (3H, d, J= 6.5 Hz, C-8 Me), 1.34 (9H, s, OBu/), 1.50-1.60 (4H, br.t, C-4, C-5), 1.87-1.97 (2H, m, C-7), 2.34-2.45 (1H, pentet, C-2) and 3.85-4.10 (3H, m, C-3, C-6, C-8), m/e 373 (M+ + H), 300 (M+ - OBut), 258 (100), 243, 217, 215, 203, 184, 167, 159, 145, 143, 125, 119, 113, 111, 101, 75, 73 and 57 (Found: C, 64.74; H, 11.02. C20 H40 04 Si requires C, 64.47; H, 10.82%). Attempted hydrogenations of the unsaturated ester (122) in the presence of rhodium on alumina (5%), palladium on carbon (10%) and platinum oxide at one atmosphere were extremely sluggish; only unreacted ester (122) was recovered. The furylidene derivative (122) (85 mg) was dissolved in THF (15 ml) and anhydrous magnesium bromide (42 mg., 1 eq.) was added. The reaction mixture was pressurized under hydrogen (1000 psi) in a Parr hydrogenator. After 24 hours, the catalyst was filtered off and the solvent was evaporated. The crude product was separated by column chromatography ( petrol, dichloromethane gradient, 1:0 to 1:3) to give the starting material (122) (55 mg.) and the hydrogenated material (20 mg., 65%). The 250 MHz *H n.m.r. spectrum showed this to be a 2:1 mixture of the desired (123) and the diastereoisomer (137). The furylidene derivative (122) (25 mg.) was dissolved in THF (15 ml) and rhodium on alumina (10 mg.) was added. The Parr hydrogenator was charged with hydrogen (1000 psi) and the mixture vigorously stirred. After 24 hours, the catalyst was filtered off and the solvent was evaporated. Column chromatography of the residue (petrol, dichloromethane gradient, 1:0 to 1:3) gave the starting material (122) (2 mg.) and the hydrogenated product (17 mg., 74%). The 250 MHz fH n.m.r. spectrum this to be a 1:1 • mixture of the desired (123) and its diastereoisomer (137). 131. Preparation of 2S(R)-[2S(R)-Hydroxypropyl]-5R(S)-[2-hydroxy-lS( R)- (methyl) ethyl] tetrahydrofuran^z»zz (124) (a) The synthetic derivative (123) (73 mg.) was dissolved in diethyl ether (15 ml) and lithium aluminium hydride (1.25 eq., 10 mg.) was added. After 24 hours at room temperature, the reaction mixture was quenched with saturated aqueous sodium sulphate solution and anhydrous sodium sulphate was added. The inorganic salts were filtered off and the solvent was evaporated to leave the desired 2S(R)-[2S(R)-(t-butyldimethylsilyloxy)- propyl]-5R(S)- [2-hydroxy-lS(R)-(methyl) ethyl]- tetrahydrofuran as a clear oil (54.5 mg., 92%). This was dissolved in aqueous THF (1:1, 10 ml) and potassium fluoride (3 eq., 30 mg.) and trifluoroacetic acid (0.2 ml) were added. The solution was refluxed under nitrogen. On completion of deprotection, the solution was evaporated and the crude diol column chromatographed (dichloromethane, ethyl acetate gradient, 1:0 to 0:1) to afford the title diol (124) as a clear oil (30 mg., 89%). This was identical with the diol obtained from the reduction of nonactin (by i.r. and t.l.c.), Vmax (neat) 3620 (s, 0-H stretch), 3500 (s, 0-H stretch), 1430 (m), 1375 (ra), 1080 (s, C-0 stretch) and 1035 (s, C-0 stretch) cm"1, 5 0.85 (2.55H, d, J = 6.5 Hz, CH^CHCH?0H), 0.95 (0.45H, d, J = 6.5 Hz), 1.21 (5H, br. d, J = 6.5 Hz, CH3 CHOH, 2xOH), 132. 1.50-1.80 (5H, m, C-3, C-4, CH3CHCH2OH), 1.90-2.10 (2H, m, CH2 CH(0H)CH3),' 3.57-3.72 (3H, m, CH2OH, CHOH), and 3.95-4.18 (2H, m, C-2, C-5). To nonactin (20 mg.) in ether (20 ml) at 0°C under nitrogen was added lithium aluminium hydride (20 eq., 20 mg). After 15 hours, the suspension was quenched with saturated aqueous sodium sulphate and anhydrous sodium sulphate was added. The inorganic materials were filtered off and the ether was evaporated to yield the diol (124) as a clear oil (19.5 mg, 95%). 133. Attempted Condensation of 5S(R)-[2S(R)-Hydroxypropyl]- tetrahydrofuran-2-one (120) with S-t-Butyl 2-Lithiothiopropanoate (125) To di-isopropylamine (2.67 ml) in THF (40 ml) at 0°C under nitrogen, was added n-butyllithium (1.5 M, 11.6 ml). After hour, the lithium di-isopropylamide solution was cooled to -78°C and j5-t-butyl thiopropanoate^1 (2.5 g, 5 eq.) was added. After 40 minutes at -78°C, the hydroxylactone (120), (0.5 g, 1 eq.) in THF (1 ml) was added. The reaction mixture was kept at -78°C and allowed to warm to room temperature, overnight. The solution was recooled to 0°C, and quenched with glacial acetic acid (1.1 ml). The solvent was evaporated and the crude residue partitioned between water and diethyl ether. The ethereal phase was separated, dried and evaporated. Column chromatography (dichloromethane, diethyl ether gradient, 1:0 to 0:1) afforded the unreacted hydroxy lactone (120) (0.35 g, 70%) and S-t-butyl 2-methyl-3-oxothiolpentanoate (0.6 g, 34%), vmax (neat) 1725 (s, C=0 stretch), 1675 (s, C=0 stretch), 1600 (w, C=C stretch), 1450 (s), 1365 (s), 1160 (m), 1125 (w), 1110 (w) and 950 (s) cm'1, 6 1.05 (3H, t, J = 7.0 Hz, C-5),-1.30 (3H, d, J = 7.0 Hz, C-2 Me), 1.50 (9H, s, S-Bu*1), 2.55 (2H, q, J = 7.0 Hz, C-4) and 3.65 (1H, q, J = 7.0 Hz, C-2), m/e 202 (Mt), 146 (Mt-t-Bu), 113, 111, 109, 97, 95, 86, 69, 57 (100, t-Bu). (Found: Mt, 202.1028. C1Q Hlg 02 S requires Mt 202.1027). This reaction has abandoned. 134. Attempted Condensation of 5S(R)-[2S(R)-(t-Butyldimethylsilyloxy)- propyl] tetrahydrofuran-2-one (121) with S-t-Butyl 2- Lithiothiopropanoate (126) To di-isopropylamine (3.0 ml) in THF (40 ml) at 0°C, under nitrogen was added n-butyllithium (1.5 M, 12.92 ml). After i hour, the lithium di-iospropylamide solution was cooled to -78°C and S-t-butyl thiopropanoate (2.83 g, 10 eq.) was added. After 40 minutes, the lactone (121) (0.5 g), was added. The reaction mixture was kept at -78°C and then allowed to warm up to room temperature overnight. The solution was recooled to 0°C and the reaction quenched with glacial acetic acid (1.22 ml). The solvent was evaporated and the crude material was separated by column chromatography (petrol, dichloromethane gradient, 2:1 to 0:1) to afford the silyloxylactone (121) (0.2 g, 40%) and £-t-butyl 2-methyl- 3-oxothiolpentanoate (0.72 g, 37%). This reaction was subsequently abandoned. 135. Attempted Condensation of 5S(R)-[2S(R)-(t-Butyl dimethylsilyloxy)- propyl] tetrahydrofuran-2-one (121) with 2-(Trlmethylsllyl) ethyl 2-Lithio propanoate3Z (127") To di-isopropylamine (2.15 ml) in THF (40 ml) at 0°C, under nitrogen, was added n-butyllithium (1.5M, 9.3 ml). After hour at 0°C, the lithium di-isopropylamide solution was cooled to -78°C and 2-(trimethylsilyl)ethyl propanoate (10 eq., 2.43 g) was added. The metallation was allowed to proceed for 1 hour. The silyloxylactone (121), (0.36 g, 1.4 mmole, 1 eq.) was added, and the temperature maintained at -78°C. The reaction mixture was allowed to warm up to room temperature overnight. The solution was recooled to 0°C, and the reaction quenched with glacial acetic acid (0.9 ml). The solvent was evaporated, the residue partitioned between water and ether, and the ethereal layer separated, dried and evaporated. Column chromatography (petrol, dichloromethane gradient) afforded the starting silyloxylactone (121) and degradation products but no adducts were detected. This reaction was subsequently abandoned. 136. Preparation of (Methoxycarbonyl-2-propanoyl)methylenetrisphenyl- phosphorane (128) To methyl 4-bromo-2-methyl-3-oxobutanoate33 (3.48 g) in benzene (10 ml) was added a solution of triphenylphosphine (4.3 g) in benzene (10 ml). The solution was stirred for 15 hours at room temperature. The supernatant liquid was decanted off, from an oily deposit. This was leached with more benzene, to leave a thick gum. The gum was dissolved in dichloromethane (50 ml) and aqueous sodium hydrogencarbonate (2.8% w/v) was added. The organic phase was separated, dried and evaporated. Chromatography (dichloromethane, methanol gradient) afforded the title compound (128) as a crystalline solid (1.9 g, 29%) m.p. 115-117°C, V max (nujol) 1730 (s, C=0 stretch), 1540 (m), 1480 (s), 1430 (s), 1380 (m), 1190 (s), 1100 (m), 740 (s), 700 (s) and 685 (s) cm""1, 6 1.40 (3H, d, J = 7 Hz, CH-CH3), 3.2-4.0 (2H, m, Ph3P=CH, CH-CH3), 3.7 (3H, s, OMe), and 7.20-7.80 (15H, m, Ph), m/e 390, 303 , 277, 262, 183, and 77 (100) (Found: C, 73.60; H, 6.02. C24 H23 03P requires C, 73.83; H, 5.94%.) 137. Condensation of (Methoxycarbonyl-2-propanoyl)methylenetrisphenyl- phosphorane (128) with 2,3R(S)-Dihydroxy-5S(R)-methyltetrahydrofuran (119) (129) (a) To the lactol (119)(0.4 g) in acetonitrile56 (5 ml) was added the ylide (128), (1.32 g). The reaction mixture was refluxed under nitrogen for 5 days. Evaporation of the solvent and column chromatography afforded 2-[3- (methoxycarbony1)-2-oxobuty1]-3R(S)-hydroxy-5 S(R)-methy1- tetrahydrofuran as a clear oil (0.18 g, 24%) together with decomposition products, Vmax (neat) 3600-3400 (w, br. 0-H stretch), 1730 (s, C=0 stretch), 1680 (m), 1450 (m), 1430 (w), 1380 (m), 1260 (m), 1100 (s), and 920 (s) cm"1, 6 1.0-1.15 (1H, m), 1.20 (3H, d, J = 6.2 Hz, Me), 1.40 (3H, d, J = 7.0 Hz, Me), 1.58-2.90 (5H, m, C-3, C-4, CH?C0), 3.05-3.45 (1H, m), 3.60-4.35 (2H, m), 3.65 (3H, s, OMe), and 5.50 (1H, br. m), m/e No M+, 212, 201, 199, 194, 183, 177, 167, 149, 135, 122, 115, 95 (100), + 91 and 83. (Found: M -H20 212.1049, CnH1604 requires 212.1049). (b) To the ylide (128) (0.73 g), in DMF (10 ml) was added the lactol (119) (0.25 g), at 0°C under nitrogen. The mixture was then heated at 100°C for 96 hours. The solution was poured into water and the organic materials extracted with diethyl ether. Drying and evaporation afforded 138. the starting ylide (128) (0.4 g, 53%). Evaporation of the aqueous phase afforded the starting lactol (119) (0.1 g, 40%), indicating there had been no reaction. To di-isopropylamine (0.17 ml) in THF (5 ml) at 0°C under nitrogen, was added n-butyllithium (1.6 M, 0.68 ml). After i hour at 0°C, the lithium di-isopropylamide solution was cooled to -78°C and the ylide (128) (0.39 g), was added. After 1 hour at -78°C, the lactol (119) (0.12 g) was added, and the solution allowed to warm up to room temperature. The reaction mixture was further stirred at room temperature for 96 hours. The solution was poured into water and the organic residues extracted with diethyl ether. Drying and evaporation afforded a brown oil. Column chromatography (dichloromethane, methanol gradient, 1:0 to 3:1) gave the starting ylide (128), (0.18 g) apart from decomposition products. These series of reactions were abandoned. 139. Preparation of Tris-(2,4-dimethoxyphenyl)phosphine54(i3Q) To 1,3-dimethoxybenzene (12 g) and anhydrous zinc chloride (3.9 g) was added phosphorus trichloride (2.6 ml). The mixture was heated from room temperature to 175°C over 20 minutes and the reaction was maintained there for hours. The solution was cooled and added to aqueous sodium hydrogencarbonate (25 g) and ice. The emulsion was extracted several times with benzene. The benzene extracts were combined, washed with water, dried and evaporated. The resultant oil was dissolved in benzene (100 ml) and extracted with conc. hydrochloric acid (3 x 50 ml). The aqueous phase was neutralized with 880 ammonia and then reextracted with dichloromethane. The dichloromethane was dried and evaporated, and the residue was recrystallized from ethanol, to give the title compound (130). The benzene layer was neutralized with 880 ammonia and the aqueous phase was further extracted with chloroform. The chloroform layer was dried and evaporated. The residue was recrystallized from ethanol to give the title compound (130). as white needles ( 1.03 g from 5A both extractions, 8%), m.p. 200.5-202°C (lit. 187-188°C,) Vmax (nujol) 1590 (m), 1570 (m), 1460 (s), 1440 (s), 1410 (m), 1305 (m), 1282 (m), 1250 (m), 1210 (m), 1160 (s), 1080 (s), 1030 (s), 920 (s) and 845 (s) cm"1, 6 3.72 (3H, s, OMe), 3.80 (3H, s, OMe), and 6.25-6.80 (3H, m, Ar), m/e 442 (Ml), 411, 291, 150, 121 and 78. (Found: C, 64.93; H, 6.15. C24 H27 Og P requires C, 65.13; H, 6.15%. 140. Attempted Preparation33 of (Methoxycarbonylacetyl)methylenetris- (2,4-dimethoxyphenyl)phosphorane (131) To methyl 4-bromo-3-oxobutanoate33 (0.44 g) in benzene (10 ml) was added, 2,4-dimethoxyphosphine (130) (1.0 g). The solution was refluxed for 24 hours, under nitrogen. The salt was filtered off and dissolved in water (15 ml). A solution of sodium hydrogencarbonate (10% w/v, 4 ml) was added drop- wise and an oil settled out. The oil was extracted with diethyl ether and ethylacetate (1:1, 3 x 10 ml). Evaporation of the extract afforded an oil. Trituration of the residue with petroleum afforded a yellow solid (68 mg) which was found to be a complex mixture (by n.m.r. and t.l.c.) of products. This reaction was subsequently abandoned. 141. Preparation of 5S(R)-methyl-3R(S)-[2S(R)-miethoxy-2-phenylacetoxy] tetrahydrofuran-2-one (132) To 3R(S)-acetoxy-5S(R)-methyltetrahydrofuran-2-one (99), (2.4 g) in methanol (20 ml) was added potassium carbonate (2.1 g, 1 eq.). The hydrolysis was allowed to continue at room temperature for 48 hours. The solution was cooled to 0°C and trifluoracetic acid (3 eq., 3.5 ml) was slowly added. The solution was evaporated and the desired 3R(S)-hydroxy-5S(R)- methyltetrahydrofuran-2-one (138) extracted with ether- ethylacetate (1:1). Evaporation afforded the hydroxy lactone (138) as a clear oil (1 g, 57%). To 3R(S)-hydroxy-5S(R)- methyltetrahydrofuran (0.20 g) in dichlororaethane (5 ml) at 0°C, were added 4-dimethylaminopyridine (84 mg) and (S) methyl-O-mandelic acid^7 (0.33 g). To the solution was added dicyclohexyl carbodiimide^® (0.54 g). The reaction mixture was kept at 0°C for 5 minutes, and then at room temperature for 6 hours. The If, N'-dicyclohexyl urea that formed was filtered off and the filtrate was evaporated. The crude material was column chromatographed (dichlororaethane) to give the title compound (132) as a white, low melting solid (0.13 g, 25%), m.p. 47-49°C, Vmax (neat) 1785 (s, lactone C=0 stretch), 1750 (s, ester C=0 stretch), 1665 (w, C=C stretch), 1455 (w), 1390 (w), 1200 (m), 1170 (m), 1110 (m, C-0 stretch), and 905 (s) cm"1, 6 1.35 (d, J= 6.6 Hz), 1.43 142. (combined integral 3H, d, J = 6.6 Hz, C-5 Me), 1.81- 1.96 (1H, m, C-4 a), 2.65-2.80 (1H, m, C-4 3), 3.45 (3H, s, OMe), 4.42-4.58 (1H, m, C-5), 4.88, 4.89 (1H, 2s, CHPh), 5.50 (1H, 2d, = 8.3 Hz, J^^ = 10.6 Hz, C-3), and 7.30-7.50 (5H, m, Ph), m/e 264 (M+), 249, 233, 220, 207, 122, 121 (100), 105, 91, 84 and 77. (Found: C, 63.71; H, 6.24. C14 H18 O5 requires C, 63.63; H, 6.10%). REFERENCES 1. J.W. Westley, Adv. Appl. Microbiol., 1977, 22, 177. 2. U. Schmidt, J. Gombos, E. Haslinger, and H. Zak Chem. Ber., 1976, 109, 2628. 3. U. Schmidt, J. Gombos, E. Haslinger,and H. Zak, Tetrahedron Lett., 1975, 3391. 4. H. Gerlach, K-Oertle, A. Thalman,and S. Servi, Helv. Chim. Acta., 1975, 58, 2036. 5. H. Zak and U. Schmidt, Angew. Chem. Int. Ed. Engl., 1975, 14, 432. 6. U. Schmidt, J. Gombos, E. Haslinger, and H. Zak, Monatsh. Chem., 1975, 106, 219. 7. K.M. Sun and B. Fraser-Reid, Can. J. Chem., 1980, 58, 2732. 8. R.E. Ireland and J.P. Vevert, Can. J. Chem., 1981, 59, 572. 9. P.A. Bartlett and K.K. Jemstedt, Tetrahedron Lett., 1980, 1607. 10. G. Beck and E. Henseleit, Chem. Ber., 1971, 104, 21. 11. H. Gerlach and H. Wetter, Helv. Chim. Acta., 1974, 57, 2306. 12. M.J. Arco, M.H. Trammel,and J.D. White, J. Org. Chem., 1976, 41, 2075. 13. R.E. Ireland and J.P. Vevert, J. Org. Chem., 1980, 45, 4259. 14. P.A. Bartlett, Tetrahedron, 1980, 36, 3. 15. F.W. Eastwood, K.J. Harrington, J.S. Josan,and J.L. Pura, Tetrahedron Lett., 1970, 5223. 16. K.C. Nicolaou, Tetrahedron, 1977, 33, 683. 17. S. Masamune, G.S. Bates, and J.W. Corcoran, Angew. Chem. Int. Ed. Engl., 1977, 16, 585. 18. "The Total Synthesis of Natural Products", Ed. J. ApSimon, Wiley, New York, 1981, Vol IV, 287. 144. 19. R. Corbaz, L. Ettlinger, E. Gaumann, W. Keller-Schierlein, F. Kradoffer, L. Neipp, V. Prelog,and H. Zahner, Helv. Chim. Acta., 1955, 38, 1445. 20. W. Keller-Schierlein and H. Gerlach, Fort. Chem. Org. Naturst., 1968, 26, 161. 21. D.H. Haynes and B.C. Pressman, J. Membr. Biol., 1974, _18, 1. 22. J. Dominguez, J.D. Dunitz, H. Gerlach,and V. Prelog, Helv. Chim. Acta., 1962, 45, 129. 23. D.D. Perrin, W.L.F. Armarego and D.R. Perrin, "Purification of Laboratory Chemicals", Pergamon Press, London, 1966 First edition. 24. M.G. Ettlinger and J.E. Hodgkins, J. Am. Chem. Soc., 1955 77, 1831. (b) G.E. Gream, Aust. J. Chem. 1972, 25, 1051. 25. H. Sakurai and A. Okada, J. Orgmet. Chem. 1972, 36C-13. (b) H. Gilman, K. Shiina, D. Aoki, B. Gaj, D. Wittenberg,and T. Brennan, J. Organomet. Chem., 1968, 13, 323. 26. G.A. Olah, Angew. Chem. Int. Ed. Engl., 1979, 18, 612. 27. M.J. Anteunis and R. Camerlynck, J. Chem. Soc., Perkin Trans. 2, 1975, 1434. 28. W.H. Yanko, H.S. Mosher,and F.C. Whitmore, J. Am. Chem. Soc., 1945, J37, 666. 29. F.C. Hartman and R. Barker, J. Org. Chem., 1964, 29, 873. 30. M.E. Evans, Carbohydr. Res., 1972, 21, 473. 31. D.H.R. Barton and S.W. McCombie, J. Chem. Soc. Perkin Trans 1., 1975, 1574. 32. J.E. Leibner and J. Jacobus, J. Org. Chem., 1979, 44, 449. 33. A. Rossi and H. Schinz, Helv. Chim. Acta., 1948, 31, 473. 34. von F. Fleck, A. Rossi, M. Hinder, and H. Schinz, Helv. Chim. Acta., 1950, 33, 130. 35. A.I. Vogel, "Practical Organic Chemistry", Longmans, London, 1961, 891 (3rd Edition). 145. 36. C. Daremon, R. Rambaud,and M. Verniette, C.R. Acad. Sci. Paris., 1971, 273, 1273. 37. G.J. Karabatsos, J.M. Corbett,and K.L. Krumel, J. Org. Chem., 1965, 30, 689. 38. R.L. Whistler and M.L. Wolfrom, "Methods in Carbohydrate Chemistry", Vol. II, Academic Press, London, 1963, 65. 39. M. Fieser and L. Fieser, "Reagents for Organic Synthesis", Vol. 4, Wiley, 1974, 57. 40. J. Brennan, G. Richardson,and R.J. Stoodley, J.C.S. Chem. Commun., 1980, 49. M. Fieser and L. Fieser, "Reagents for Organic Synthesis", Vol I, Wiley, 1967, 1088. 41. P. Place, M. Roumestant,and J. Gore, Bull. Soc. Chim. Fr., 1976, 169. 42. "Dictionary of Organic Compounds", Eyre and Spottiswoode, London, 1965, 2115. 43. P.C. Anderson and B. Staskun, J. Org. Chem., 1965, 30, 3033. 44. L.F. Weiler, J. Am. Chem. Soc., 1970, 6702. 45. S.O. Lawessen, Org. Synth., 1962, 42, 28. 46. V.I. Pansevich-Kolyada and T.A. Galysheva, Vests! Akad. Navuk. Belaruss. SSR Ser. Khim. Navuk, 1971, 1_, 89. V.I. Pansevich-Kolyada and T.A. Galysheva, Chem. Abstr., 1971, 74, 140817 z. 47. T.A. Bryson, J. Org. Chem., 1973, 38, 3428. 48. R.H. Schlessinger, G.R. Kieczykowski,and M.L. Quesada, J. Am. Chem. Soc., 1980, 102, 782. 49. C.G. Kruse, A.C.V. Janse, V. Dert and A. van der Gen, J. Org. Chem., 1979, 44, 2916. D. Seebach, M. Pohmakotr, C. Schregenberger, B. Weidmann, R.S. Mali, and S. Pahmakotr, Helv. Chim. Acta., 1982, 65_, 419. 50. P. Camps, J. Font, and 0. Ponsati, Tetrahedron Lett., 1981, 1471. 146. 51. U. Schmidt and M. Schwochau, Tetrahedron Lett., 1967, 4491. 52. R.J. Fessenden and J.S. Fessenden, J. Org. Chem., 1967, 32, 3535. (b) H. Gerlach, Helv. Chim. Acta., 1977, 60, 3039. 53. A. Svendson and P.M. Boll, Tetrahedron, 1973, _29, 4251. 54. I.S. Protopov and M. Yakraft, J. Gen. Chem. USSR, 1964, 34, 1451. 55. H. Muxfeldt, G. Grethe, K. Uhlig,and H. Zeugner, Chem. Ber., 1963, 96, 2943. 56. N. Katagiri, K. Takashima,and T. Kato, J. Chem. Soc. Chem. Commun., 1982, 664. 57. W.A. Bonner, J. Am. Chem. Soc., 1951, 73_t 3126. 58. R.E. Ireland, W.J. Thompson, G.H. Srouji,and R. Etter, J. Org. Chem., 1981, 46, 4863. 59. S. Masamune, M. Hirama, S. Mori, Sk. A. Ali, and D.S. Garvey, J. Am. Chem. Soc., 1981, 103, 1568. 60. T.M. Harris and C.M. Harris, Org. React., 1969, 17, 177. 61. "Comprehensive Organic Chemistry", Ed. D.H.R. Barton and W.D. Ollis, Pergamon Press, Oxford, 1979, _5, 724. 62. D.H.R. Barton, R.V. Stick,and R. Subramanian, J. Chem. Soc., Perkin Trans. 1, 1976, 2112. 63. S.A.M.T. Hussain, W.D. Ollis,and J.F. Stoddart, J. Chem. Soc., Perkin Trans. 1, 1975, 1480. 64. Y. Asahina and A. Fujita, Chem. Abstr., 1920, 14, 1384. 65. D.H.R. Barton, M. Benecbie, F. Khuong-Huu, P. Potier, and V. Reyna-Pinedo, Tetrahedron Lett., 1982, 651. 66. W.S. Johnson, J.C. Collins, R. Pappo, M.B. Rubin, P.J. Kropp, W.F. Johns, J.E. Pike,and W. Bartmann, J. Am. Chem. Soc., 1963, 85, 1409. 67. M. Fieser and L. Fieser, "Reagents for Organic Synthesis", Vol I, Wiley, 1967, 1181. 147. 68. M.T. Reetz and R. Peter, Tetrahedron Lett., 1971, 4691 and references cited therein. 69. A.J. Duggan, M.A. Adams, P.J. Brynes,and J. Meinwald, Tetrahedon Lett., 1978, 4323. (b) B.M. Trost and T.A. Runge, J. Am. Chem. Soc., 1981, 103, 7559. (c) R.J. Cregge, J.L. Herrmann, C.S. Lee, J.E. Richman,and R.H. Schlessinger, Tetrahedron Lett., 1973, 2425. 70. R.J. Light and C.R. Hauser, J. Org. Chem., 1960, ^5, 538. 71. C.R. Hauser and T.M. Harris, J. Am. Chem. Soc., 1958, 80, 6360. 72. E.J. Corey, N.M. Weinshenker, T.K. Schaaf,and W. Huber, J. Am. Chem. Soc., 1969, 91, 5675. 73. C. Ruchardt, S. Eichler,and P. Panse, Angew. Chem. Int. Ed. Engl., 1963, 2, 619. 74. J. Schmidlin and A. Wettstein, Helv. Chim. Acta., 1963, 46, 2799. 75. E.L. Eliel, "Stereochemistry of Carbon Compounds", McGraw-Hill, London, 1962, 65. 76. A. Samat and M. Bibout, Heterocycles, 1982, 19, 469. 148. A Concise Synthesis of (±)-t-Butyl 8-O-t-Butyldimethylsilylnonactate Anthony G. M. Barrett* and Hiten G. Sheth Department of Chemistry, Imperial College, London SW7 2AY, U.K. The racemic title compound was prepared in seven steps from 2,3,4-tri-O-acetyl-D-ribonolactone (3) via the highly stereoselective hydrogenations of 3-acetoxy-5-methylenefuran-2(5//)-one (4) and, subsequently, 2S(/?)-[2S(/?)-(t-butyldimethylsilyloxy) propyl]-5-[1 -(t-butyloxycarbonyl)ethylidene]tetra- hydrofuran (9). Reprinted from the Journal of The Chemical Society Chemical Communications 1982 170 J. CHEM. SOC., CHEM. COMMUN., 1982 A Concise Synthesis of (±)-t-Butyl 8-O-t-Butyldimethylsilylnonactate Anthony G. M. Barrett* and Hiten G. Sheth Department of Chemistry, Imperial College, London SW7 2AY, U.K. The racemic title compound was prepared in seven steps from 2,3,4-tri-O-acetyl-D-ribonolactone (3) via the highly stereoselective hydrogenations of 3-acetoxy-5-methylenefuran-2(5//)-one (4) and, subsequently, 2S(R)-[2S(fl)-(t-butyldimethylsilyloxy)propyl]-5-[1 -(t-butyloxycarbonyl)ethylidene]tetra- hydrofuran (9). Nonactin (1) is an antibiotic ionophore produced by Strepto- myces. It is notable for the ability to mediate cation trans- port. In particular potassium ions are readily complexed. Nonactin (1) is a tetrameric mesomolecule composed of alternating dextro- and laevo-rotatory nonactic acid (2a)f units. Although methyl nonactate (2b) has been synthesised by several groups in racemic form1'2 or as dextro- or laevo- rotatory enantiomers,3 all these syntheses involve numerous stages. In addition, with one excellent exception,1 these syntheses involve at least one step of low stereoselectivity. Herein, we report a seven-stage, highly stereoselective synthesis of (±)-t-butyl 8-O-t-butyldimethylsilylnonactate (10) which is clearly amenable to multigram synthesis t Structures (2) and (5)—(14) refer to racemic modifications. (Scheme 1). 170 J. CHEM. SOC., CHEM. COMMUN., 1982 AcO OAc 94V. OAc AcO OAc (4) (11) V(3) r 90'/. Jii OSiMe2Bul ^O^OH 85 V, CC^Bu1 (6) (5) (13) (14) phosphorane to give the unstable enoate (12). This was hydrogenated in situ to give, on acidification, the lactone (7). v % OSif l w ( 55V. 0' i " 'OSiMe2Bu "0 "^ ! 'OH [from (5)] Most conveniently, this was isolated as the lactone (8) in H 55% yield [based on (6)]. Both lactones (7) and (8) were (8) stereochemically pure. (7) t-Butyl 2-lithiopropanoate condensed with the lactone (8) to give, on acidification, the enoate (9) as a single geometric 75V. vi isomer. This was assigned the E-geometry on the basis of ample precedent.1-8 Hydrogenation of this product gave the (i)-nonactic acid derivative (10) and the isomer (13) (85:15). l In the final step the remaining two chiral centres (C-2 and Bu 02C l C-3) were introduced via steric-approach-controlled delivery OSiMe2Bu1 OSiMe2Bu of hydrogen. vii As proof of the structure, the (±)-nonactic acid derivative CO2BU 89 V. (10) was converted (LiAlH4, THF; CF3C02H, H20, THF) into the known diol (14) (82%). The product was identical (10) (9) (i.r., n.m.r., and t.l.c.) with authentic material obtained by the reduction of nonactin (l).7 Scheme 1. Reagents and conditions (THF = tetrahydrofuran; Clearly, the (±)-nonactic acid derivative (10) is now avail- DMF = dimethylformamide): i, DBU (1 equiv.), THF, —20 to able in seven steps from readily available starting materials 0 °C; dil. HCI; ii, H2, Pd-CaCO„ THF; iii, BU*2A1H (3.1 equiv.), PhMe, —78 °C; HOAc; iv, PhsP=CHC02Et, THF; H2, Rh-Al2Oa, in an unoptimised overall yield of > 24%. t THF; CFjC02H; v, BuMe2SiCl, imidazole, DMF; vi, MeCH= We thank the S.E.R.C. for a studentship (to H. G. S.), 4 C(OBu)OLi (10 equiv.), THF; HOAc; Amberlite 120 H; vii, H2, W. R. Grace and Co., Research Division, for most generous Rh-Al2Os, THF. support, and the University of London Central Research Fund for an equipment grant. We also thank David Neuhaus for the n.m.r. spectra. 2,3,5-Tri-O-acetyl-D-ribonolactone (3) is an inexpensive, readily available starting material. On reaction with 1,8- Received, 18th November 1981; Com. 1345 diazabicyclo[5.4.0]undec-7-ene (DBU), the double elimina- tion of acetic acid occurred giving the diene (4), m.p. 75— 76 °C.J Routinely, this was hydrogenated without purification References to give the lactone (5).4§ Clearly this steric-approach-controlled 1 P. A. Bartlett and K. K. Jernstedt, Tetrahedron Lett., 1980, 21, hydrogenation6 correctly established two of the four chiral 1607. centres of (±)-nonactic acid (C-6 and C-8). The isomeric 2 H. Gerlach and H. Wetter, Helv. Chim. Acta, 1974, 57, 2306; lactone (11) was not detected [(5): (11) >97:3]. The lactol M. J. Arco, M. H. Trammell, and J. D. White, J. Org. Chem., (6), derived from (5) and di-isobutylaluminium hydride, 1976, 41, 2075; G. Beck and E. Henseleit, Chem. Ber., 1971, condensed cleanly with ethoxycarbonylmethylenetriphenyl- 104, 21. 3 R. E. Ireland and J.-P. Vevert, Can. J. Chem., 1981, 59, 572; U. Schmidt, J. Gombos, E. Haslinger, and H. Zak, Chem. Ber., 1976, 109, 2628; H. Zak and U. Schmidt, Angew. Chem., Int. X Compounds (4)—(10) were fully characterised by spectral data, Ed. Engl., 1975, 14, 432; K. M. Sun and B. Fraser-Reid, Can. including 250 MHz n.m.r. Compounds (4), (6), (8), (9), and (10) J. Chem., 1980, 58, 2732. microanalysed correctly. 4 C. Daremon, R. Rambaud, and M. Verniette, C. R. Acad. Sci. Paris Sect. (Q, 1971, 273, 1273. § The known4 mixture of lactones (5) and (11) was readily obtained via 2,4-dihydroxypentanenitrile, acid-catalysed hydrolysis, and 5 S. A. M. T. Hussain, W. D. Ollis, C. Smith, and J. F. Stoddart, acetylation. N.m.r. spectra are detailed elsewhere.4 In addition J. Chem. Soc., Perkin Trans. 1, 1975, 1480. sodium borohydride reduction of the lactone (5) gave threo- 6 T. A. Bryson, J. Org. Chem., 1973, 38, 3428; S. A. Krueger pentane-l,2,4-triol (100%) (microanalysed as the tri-benzoate). -and T. A. Bryson, ibid., 1974, 39, 3167. None of the erythro-isomer was detected by *H or "C n.m.r. 7 J. Dominguez, J. D. Dunitz, H. Gerlach, and V. Prelog, Helv. spectroscopy. Chim. Acta., 1962, 45, 129.