Synthesis of the antifungal macrolide antibiotic (+)-roxaticin

Article abstract of DOI:10.1021/ja027638q

The total synthesis of the antifungal macrolide antibiotic roxaticin has been accomplished. The synthesis relies principally on aldol and directed reduction steps to construct the extended 1,3-polyol array present in the natural product. Three principal n

Full text of DOI:10.1021/ja027638q

Published on Web 08/14/2003
Synthesis of the Antifungal Macrolide Antibiotic (+)-Roxaticin
David A. Evans* and Brian T. Connell
Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
Received July 10, 2002; E-mail: Evans@Chemistry.Harvard.edu
Abstract: The total synthesis of the antifungal macrolide antibiotic roxaticin has been accomplished. The
synthesis relies principally on aldol and directed reduction steps to construct the extended 1,3-polyol array
present in the natural product. Three principal nonpolyene containing fragments were assembled and then
coupled using Julia olefination and methyl ketone aldol addition reactions. A series of functionalization
reactions incorporated the sensitive polyene and provided the protected roxaticin seco-acid, which was
lactonized in good yield. Acidic deprotection completed this convergent synthesis of roxaticin.
Introduction
Polyene macrolide antibiotics are an important class of natural
products currently used in the treatment of systemic fungal
infections. Although there are over 200 known members of this
family, most of which are produced by soil actinomycetes
belonging to the genus Streptomyces, only about a dozen
representatives have been fully characterized. The unique
stereochemical and structural characteristics of these natural
products, along with the increasing incidence and severity of
acute fungal infections, have stimulated the interest of organic
chemists.1-3 A variety of methods have been developed to
independently address the synthesis of both the polyol and
polyene portions of these potent fungicides. The first approach
to any member of this class was Nicolaou’s relay synthesis of
amphotericin B in 1988.4 At that time, amphotericin B was the
only polyene macrolide whose structure had been unequivocally
established. The strategies and methodologies developed by
other groups over the past decade have resulted in the successful
total syntheses of five other polyene macrolide antibiotics:
mycoticin A,5 roxaticin (1),6,7 filipin III,8 roflamycoin,9 and
dermostatin.10
Figure 1. Roxaticin heptaacetate X-ray structure (acetates omitted for
clarity).
were established by X-ray analysis of the derived heptaacetate
(Figure 1). The absolute stereochemistry was proposed on the
basis of the observed optical rotation of the C28-C30 degradation
product, 2,4-dimethyl-1,3-pentanediol, obtained by treatment of
the natural product with ozone followed by H2-Pd/C and
sodium borohydride. As is often the case with polyene mac-
rolides, roxaticin exhibits antifungal but not antibacterial activity.
Our interest in the synthesis of roxaticin was motivated by
the fact that the strict polyacetate origin of the roxaticin offered
an ideal forum for the extension of our aldol-based approach
to polypropionate natural product synthesis. In this paper, we
report a highly convergent total synthesis of the naturally
occurring (+)-enantiomer of roxaticin, which relies heavily on
recent advances in aldol methodologies developed in this
laboratory.
Structure. Roxaticin is a pentaene macrolide isolated from
streptomycete X-14994, found in the soil near Escalante, Utah,
USA.11 The structure and relative stereochemistry of roxaticin
(1) Rychnovsky, S. D. Chem. ReV. 1995, 95, 2021-2040.
(2) Sternberg, S. Science 1994, 266, 1632-1634.
(3) Omura, S.; Tanaka, H. Macrolide Antibiot. 1984, 351-404.
(4) Nicolaou, K. C.; Daines, R. A.; Uenishi, J.; Li, W. S.; Papahatjis, D. P.;
Chakraborty, T. K. J. Am. Chem. Soc. 1988, 110, 4672-4685.
(5) Poss, C. S.; Rychnovsky, S. D.; Schreiber, S. L. J. Am. Chem. Soc. 1993,
115, 3360-3361. Dreher, S. D.; Leighton, J. L. J. Am. Chem. Soc. 2001,
123, 341-342.
Synthesis Plan. Our synthesis plan relied upon the three
stereoselective aldol bond constructions illustrated in Scheme
1 (eqs 1-3).12,13 These transformations provide rapid access to
chiral â-hydroxy ketones and ultimately 1,3-dioxygen relation-
ships, the prevalent stereochemical motif found in these natural
(6) Rychnovsky, S. D.; Hoye, R. C. J. Am. Chem. Soc. 1994, 116, 1753-65.
(7) Mori, Y.; Asai, M.; Okumura, A.; Furukawa, H. Tetrahedron 1995, 51,
5299-5314. Mori, Y.; Asai, M.; Kawade, J.-i.; Furukawa, H. Tetrahedron
1995, 51, 5315-5330.
(8) Richardson, T. I.; Rychnovsky, S. D. Tetrahedron 1999, 55, 8977-8996.
(9) Rychnovsky, S. D.; Khire, U. R.; Yang, G. J. Am. Chem. Soc. 1997, 119,
2058-2059.
(10) Sinz, C. J.; Rychnovsky, S. D. Angew. Chem., Int. Ed. 2001, 40, 3224-
3227.
(12) Nelson, S. G. Tetrahedron: Asymmetry 1998, 9, 357-389.
(13) Cowden, C. J.; Paterson, I. Organic Reactions; Paquette, L. A., Ed.;
Wiley: 1997; Vol. 51, Chapter 1.
(11) Maehr, H.; Yang, R.; Hong, L. N.; Liu, C. M.; Hatada, M. H.; Todaro, L.
J. J. Org. Chem. 1989, 54, 3816-3819.
9
10.1021/ja027638q CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 10899-10905
10899

A R T I C L E S
Evans and Connell
Scheme 1
Scheme 2
product targets. Stereochemical control in these aldol reactions
may be derived from the addition of enol nucleophiles to
â-alkoxyaldehydes (eq 1)14 or from remote stereocenters on the
enolate coupling partner (eq 2). The remarkable 1,5-anti
induction observed in this reaction independently by us15 and
Paterson16 is particularly well suited to polylene macrolide
applications. Silylketene acetal additions to benzyloxyacetal-
dehyde in the presence of a chiral Cu(II) catalyst provide
â-hydroxyketones with excellent enantioselectivity (eq 3).17
Stereoselective hydride reductions of these nonracemic â-hy-
droxyketones would provide access to either syn or anti 1,3-
diol subunits. We have developed two different methods for
the diastereoselective internally directed reduction of â-hydroxy-
ketones, as shown in eqs 4 and 5.18,19 The highly syn-selective
hydride reduction popularized by Prasad20 (eq 6) allows access
to syn diol subunits. With the proper choice of aldol reaction
conditions and hydride reduction reagents, any diol subunit can
be constructed with control over both relative and absolute
stereochemistry.
With these reactions in mind, our synthesis plan for roxaticin
(1) is outlined in Scheme 2. Given the sensitivity of the pentaene
moiety to light and air, this fragment would be incorporated
into the fully functionalized polyacetate precursor at a late stage
of the synthesis, potentially via a Horner-Wadsworth-Emmons
coupling with a C12 polyol-derived aldehyde. With this require-
ment in mind, we find a convergent synthesis relies upon a
penultimate seco-acid cyclization21-23 onto the hindered C30
oxygen. The C12-C30 fragment contains all the stereocenters
present in the natural product, and its synthesis may be
considered as the major subgoal of this project.
The 1,5-syn dioxygen relationship displayed on the polyac-
etate backbone provides the retron for the aldol reaction
illustrated in eq 2. The identification of the two such pairwise
relationships in the C12-C30 synthon is highlighted in Scheme
2, as are the two obligatory ketonic precursors 3 and 4. Each of
these intermediates may be derived from either of two potential
(14) Evans, D. A.; Dart, M. J.; Duffy, J. L.; Yang, M. G. J. Am. Chem. Soc.
1996, 118, 4322-4343.
(15) Evans, D. A.; Coleman, P. J.; Cote, B. J. Org. Chem. 1997, 62, 788-789.
(16) Paterson, I.; Gibson, K. R.; Oballa, R. M. Tetrahedron Lett. 1996, 37, 8585-
85888.
(17) Evans, D. A.; Kozlowski, M. C.; Murry, J. A.; Burgey, C. S.; Campos, K.
R.; Connell, B. T.; Staples, R. J. J. Am. Chem. Soc. 1999, 121, 669-685.
(18) Evans, D. A.; Hoveyda, A. H. J. Am. Chem. Soc. 1990, 112, 6447-6449.
(19) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am. Chem. Soc. 1988,
110, 3560-3578.
(20) (a) Chen, K.-M.; Hardtmann, G. E.; Prasad, K.; Repic, O.; Shapiro, M. J.
Tetrahedron Lett 1987, 28, 155-158. (b) Narasaka, K.; Pai, H. C. Chem.
Lett. 1980, 1415-1418.
(21) Paterson, I. M. Tetrahedron 1985, 41, 3569-3624.
(22) Boeckman, R. K.; Goldstein, S. W. The Total Synthesis of Natural Products;
Apsimon, J., Ed.; Wiley: New York, 1988; Vol. 7, pp 1-139.
(23) Bartra, M. U.; Vilarrasa, J. Recent Progress in the Chemical Synthesis of
Antibiotics and Related Microbial Products; Lukacs, G., Ed.; Springer-
Verlag: Berlin, 1993; Vol. 2.
9
10900 J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003

Antifungal Macrolide Antibiotic (+)-Roxaticin
A R T I C L E S
Scheme 3 a
Scheme 4
a Key: (a) Hg(OAc)2, AcOH, ∆; (b) LiAlH4, THF, 0 °C to rt, then O2;
(c) triethylphosphonoacetate, NaH, -78 °C to rt, THF; (d) NaBH4, EtOH;
(e) SOBr2, 2,6-di-tert-butylpyridine, -20 °C, THF; (f) (EtO)3P, toluene,
110 °C.
aldol reactions, depending upon which fragments serve as the
nucleophilic and electrophilic reaction components. This family
of disconnections provides several options for construction of
this fragment. The aldol option that was selected was based on
the C19-C20 aldol bond construction from the two fragments 5
and 6 of most similar complexity. In principle, stereochemical
control of this proposed aldol reaction could also arise from a
Lewis acid-mediated 1,3-syn aldol union of 5 with the enolsilane
of ketone 6 (eq 1), providing additional flexibility to the
synthetic plan.
salt of triethylphosphonoacetate to deliver a monoester-
monoaldehyde. All subsequent reactions and manipulations of
compounds containing this polyene structure were carried out
with careful exclusion of light to maximize product yield and
stability. Treatment of the unpurified reaction mixture with an
excess of sodium borohydride resulted in formation of alcohol
10, which could be quickly purified by flash chromatography
and isolated in 60% yield. Exposure to thionyl bromide at -20
°C provided the corresponding allylic bromide. When this
unpurified bromide was heated to reflux in toluene containing
an excess of triethyl phosphite, the desired phosphonate 11 was
obtained in 60% yield over the two steps. Fortunately, this
polyene was the most stable of any intermediate in this sequence
and typically could be stored frozen in argon saturated benzene
at -20 °C for several months with less than 20% decomposition.
Flash chromatography purification was employed to purify this
phosphonate immediately before its use (vide infra).
C12-C19 Subunit. The plan for the synthesis of methyl ketone
fragment 6 (Scheme 4) evolved from a chelate-controlled
allylmetal addition to the readily available â-alkoxy aldehyde
12, a subsequent homologation, and a stereocontrolled, hemi-
acetal-mediated heteroconjugate addition (eq 7) to introduce the
C16-syn oxygen heteroatom through methodology previously
reported by us.28
Synthesis of methyl ketone fragment 6 began with aldehyde
12 (Scheme 5). While this aldehyde is available by several
routes,29 we choose to produce it in quantity by asymmetric
alkylation of the titanium enolate derived from propionyl
oxazolidinone 13 with BOM-Cl30 followed by reductive removal
(LiBH4) of the chiral auxiliary and Parikh-Doering oxidation
to the aldehyde. This procedure affords large quantities of
enantiopure aldehyde 12 in 85% yield for the three steps.
Chelate-controlled allylation (SnCl4, CH2dCHCH2SnBu3,
CH2Cl2, -78 °C) by the method of Keck delivered the homo-
allylic alcohol 14 in 90% yield and 35:1 diastereoselectivity.31
Temporary protection of the alcohol as its TES ether (99%)
was required to efficiently execute the following homologation.
Results and Discussion
C2-C11 Polyene Subunit. Much effort has been devoted to
the stereocontrolled synthesis of polyenes.24,25 They are present
not only in antifungal macrolides but also in anticancer retinoids,
and they are increasingly being used to explore the concept of
molecular electronics. We preferred, for reasons of efficiency
and convergency, to incorporate this fragment into the polyol
chain in a single stereocontrolled step, rather than to build up
the pentaene in a repetitive, stepwise manner. This plan required
a practical synthesis of a pentaene fragment coupling precursor.
We decided to use methodology developed by Anet26 to
construct a suitable polyene precursor, due to the relative ease
with which this segment may be stereoselectively constructed
from readily available starting materials. The polyene fragment
required would be a phosphonate such as 2, which would be
coupled to polyol C12 aldehyde 3 in a trans-selective Horner-
Emmons reaction. The synthesis of phosphonate 2 began with
commerically available cyclooctatetraene (7) (Scheme 3).
As initially reported by Cope, a one-pot electrocyclic ring
closure and Hg(II)-catalyzed trans-acetoxyation delivered com-
pound 8 in 88% yield (Scheme 3).27 Reductive removal of the
acetates using LiAlH4, followed by stirring an ethyl acetate
solution of the transient diol in air led to 2-electron oxidative
ring fragmentation and olefin isomerization to afford the highly
unstable dialdehyde 9 in 60% yield. Since 9 was highly sensitive
to both air and light and was prone to spontaneous polymeri-
zation, it was immediately treated with 1 equiv of the sodium
(24) Lipshutz, B. H.; Lindsley, C. J. Am. Chem. Soc. 1997, 119, 4555-4556.
(25) Babudri, F.; Cicciomessere, A. R.; Farinola, G. M.; Fiandanese, V.;
Marchese, G.; Musio, R.; Naso, F.; Sciacovelli, O. J. Org. Chem. 1997,
62, 3291-3298.
(28) Evans, D. A.; Gauchet-Prunet, J. A. J. Org. Chem. 1993, 58, 2446-2453.
(29) Romo, D.; Johnson, D. D.; Plamondon, L.; Miwa, T.; Schreiber, S. L. J.
Org. Chem. 1992, 57, 5060-5063.
(26) (a) Anet, R. Tetrahedron Lett. 1961, 720-723. (b) Poss, C. S. Ph.D. Thesis,
Harvard University, Cambridge, MA, 1993.
(30) Evans, D. A.; Urpi, F. P.; Somers, T. C.; Clark, J. S.; Bilodeau, M. T. J.
Am. Chem. Soc. 1990, 112, 8215-8216.
(27) Cope, A. C.; Nelson, N. A.; Smith, D. S. J. Am. Chem. Soc. 1954, 76,
1100-1104.
(31) Keck, G. E.; Abbott, D. E. Tetrahedron Lett. 1984, 25, 1883-1886.
9
J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003 10901

A R T I C L E S
Evans and Connell
Scheme 5 a
a Key: (a) TiCl4, NEt3, BnOCH2Cl, 0 °C, CH2Cl2; (b) LiBH4, 0 °C, THF; (c) SO3‚pyr, DMSO, -10 °C, CH2Cl2; (d) allyltributyltin, SnCl4, -78 °C,
CH2Cl2; (e) TESCl, imidazole, CH2Cl2; (f) O3; Ph3P; then 15; TsOH, CH2Cl2; (g) cat. KHMDS, PhCHO, THF, 0 °C; (h) Zn(OTf)2, EtSH, NaHCO3, CH2Cl2;
(i) cyclopentylidene dimethyl ketal, PPTS, CH2Cl2; (j) MeLi, THF, -78 °C.
Scheme 6
Scheme 7 a
Ozonolysis followed by in situ reduction (Ph3P) and treatment
with N-methoxy-N-methyl-2-(triphenylphosphoranylidene)acet-
amide (15)32 afforded, after acidic workup, amide 16 in 86%
yield. Failure to protect alcohol 14 during this reaction resulted
in retro-aldol cleavage of the intermediate â-hydroxy aldehyde
under the reaction conditions. Amide 16 was induced to undergo
intramolecular hemiacetal conjugate addition with catalytic
quantities of potassium hexamethyldisilazide (KHMDS) in the
presence of excess benzaldehyde (eq 7) to give the thermody-
namically favored syn-protected acetal 17 (76%) along with 15%
recovered 16. Since the benzylidene acetal proved to be
incompatible with later reductive reaction conditions (vide infra),
it was replaced with a cyclopentylidene ketal by treatment with
bicarbonate buffered Zn(OTf)2 and EtSH in CH2Cl2 followed
by acid-catalyzed reprotection to afford amide 18 in 97% yield
over the two steps. Finally, treatment of amide 18 with
methyllithium at -78 °C gave the desired C12-C19 subunit 19
in 92% yield.33
C20-C30 Subunit. The assembly of this fragment is outlined
below (Scheme 6). The trans-selective Julia-Lythgoe coup-
ling34-36 of the C28-C30 propionate subunit to the extended
C20-C28 polyacetate fragment, the catalytic asymmetric synthe-
sis of the C22-C27 fragment and its acetate extension highlight
the important bond constructions in the assemblage of this
fragment.
The synthesis was initiated with a boron-mediated aldol
reaction of the (R)-N-propionyl imide 13 with isobutyraldehyde
to afford adduct 21 in 99% yield and >200:1 diastereoselectivity
(Scheme 7).37 Reductive removal of the auxiliary with lithium
borohydride38 and direct acetalization with p-anisaldehyde
a Key: (a) Bu2BOTf, NEt3, i-PrCHO, CH2Cl2, -78 °C; (b) LiBH4,
MeOH, THF, -78 °C; (c) cat. TsOH, p-MeOPhCH(OMe)2, CH2Cl2; (d)
DIBAl-H, CH2Cl2, -78 °C; (e) MsCl, NEt3, CH2Cl2; (f) PhSLi, THF, -78
°C f 23 °C; (g) m-CPBA, CH2Cl2.
Scheme 8 a
a Key: (a) ref 17; (b) Et2BOMe, NaBH4, MeOH, THF, -78 °C; (c)
TBSCl, imidazole, CH2Cl2; (d) 2000 psi H2, 10% Pd/C, EtOAc; (e) Dess-
Martin, CH2Cl2.
dimethylacetal delivered 93% of the protected diol 22, along
with 99% of recovered oxazolidinone. The mixture of 22 and
oxazolidinone could be efficiently separated by trituration with
hexanes, filtration to collect the crystallized oxazolidinone, and
flash chromatography of the mother liquor to deliver pure 23.
Regioselective reductive acetal cleavage at the less-hindered
oxygen with DIBAl-H at 0 °C afforded the primary alcohol 23
in 98% yield.39,40 Alcohol 23 was converted to phenyl sulfone
24 by mesylation, displacement with lithium thiophenylate, and
oxidation to the sulfone with MCPBA. This simple three-step
(32) Evans, D. A.; Kaldor, S. W.; Jones, T. K.; Clardy, J.; Stout, T. J. J. Am.
Chem. Soc. 1990, 112, 7001-7031.
(33) (a) Sibi, M. P. Org. Prep. Proced. Int. 1993, 25, 15-40. (b) Nahm, S.
Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815-3818.
(34) Julia, M. P. Tetrahedron Lett. 1973, 4833-4836.
(35) Kocienski, P. Phosphorus Sulfur Relat. Elem. 1985, 97-127.
(36) Kocienski, P. J.; Ruston, S. J. Chem. Soc., Perkin Trans. 1 1978, 829-
837.
(37) Evans, D. A.; Bartroli, J.; Shih, T. L. J. Am. Chem. Soc. 1981, 103, 2127-
2129. Gage J. R.; Evans, D. A. Org. Synth. 1989, 66, 83-91.
(38) Penning, T. D.; Haack, R. A.; Kalish, V. J.; Miyashiro, J. M.; Rowell, B.
W.; Yu, S. S. Synth. Commun. 1990, 20, 307-312.
(39) Johansson, R. S. J. Chem. Soc., Perkin Trans. 1 1984, 2371-2374.
(40) Johansson, R. S. J. Chem. Soc., Chem. Commun. 1984, 201-202.
9
10902 J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003

Antifungal Macrolide Antibiotic (+)-Roxaticin
A R T I C L E S
Scheme 9 a
a Key: (a) n-BuLi, BF3‚OEt2, -78 °C, THF; (b) Na/Hg, Na2HPO4, -40 °Cf 23 °C, MeOH; (c) HF‚pyr, THF; (d) cyclopentylidene dimethyl ketal,
PPTS, CH2Cl2; (e) LiAlH4, THF; (f) Dess-Martin, CH2Cl2; (g) BF3‚OEt2, 31, -90 °C, toluene; (h) TBSOTf, 2,6-lutidine, -78°C, CH2Cl2; (i) DIBAl-H,
-78 °C, toluene.
sequence proceeded in 96% yield without the need for purifica-
tion of either intermediate.
OEt2, rather than BF3‚OEt2, afforded equivalent results, but
freshly distilled BF3‚OEt2 was chosen for ease of use on larger
scale reactions. Alternatively, aluminum-based Lewis acids led
only to decomposition of starting materials. Functionalization
of the â-hydroxysulfone for subsequent elimination was unsuc-
cessful under a variety of conditions. All acetylation and
mesylation attempts yielded only recovered starting material.
Fortunately, it was eventually found that direct treatment of the
hydroxysulfone mixture with 5% Na/Hg amalgam in buffered
methanol at -20 °C provided the desired E-olefin 30 in 81%
yield as a single isomer by 1H NMR spectroscopy (Scheme 9).
Synthesis of the C20-C27 fragment began with the Lewis acid-
catalyzed addition of the Chan’s diene analogue 26 to (benzyl-
oxy)acetaldehyde (25) (Scheme 8). Utilizing 2 mol % of our
previously reported chiral Lewis acid catalyst, [Cu(S,S)-Ph-
pybox](SbF6)2, (eq 3), 500 mg of recoverable ligand afforded
20 g of 27 in 85% yield and >99% ee from 25 and 26.17 The
protected syn diol 28 was prepared via Et2BOMe-mediated
NaBH4 reduction of 27, followed by protection of the diol as a
bis(TBS) ether 37 in 87% yield for the two steps.20 A cyclic
ketal protecting group for this diol would have been ideal for
the synthesis, as it would promote a stereoselective aldol addition
in a later reaction. Unfortunately, the proposed Julia fragment
coupling onto a C27 aldehyde failed with all cyclic protecting
groups installed at this position (vide infra). For maximal
efficiency and protecting group compatibility, 28 was prepared
for immediate coupling to sulfone fragment 24 before further
elaboration of the polyacetate region from the ester terminus.
Catalytic hydrogenation of 28 proceeded smoothly under forcing
conditions (2000 psi H2, 50 h) to reveal the primary alcohol
that was efficiently oxidized using the Dess-Martin periodinane
to afford aldehyde 29 in 97% yield for the two steps.41
The desire for a synthetically useful level of diastereoselec-
tivity in the projected aldol reaction onto the C22 aldehyde
necessitated a protecting group exchange, since â-silyloxyal-
dehydes usually yield poor diastereofacial control.14 Indeed,
aldol reaction of C24,C26-bis(TBS) protected aldehyde equivalent
of 33 with trimethylsilylketene acetal of S-tert-butyl thioacetate
(31) mediated by BF3‚OEt2 yielded only 2:1 syn diastereose-
lectivity. Accordingly, the TBS ethers were cleanly removed
(HF‚pyridine, THF), and the cyclopentylidene ketal was intro-
duced under mildly acidic conditions (PPTS) to deliver 32 in
97% yield for the two-step sequence. The ester was transformed
to the aldehyde in a two-step process by LAH reduction to the
alcohol and Dess-Martin oxidation to afford aldehyde 33 in
90% yield. Treatment of 33 in toluene at -110 °C with 31 in
the presence of BF3‚OEt2 afforded a 5-7:1 inseparable mixture
of C22-diastereomers which were immediately silylated with
TBSOTf to afford protected thioester 34 in 83% yield. Con-
trolled DIBAl-H reduction of the thioester proceeded quanti-
tatively to deliver the desired aldehyde 35 as a 5-7:1 mixture
of diastereomers.
Our initial fragment coupling study in the assemblage of the
C20-C30 subunit was disappointing. Deprotonation of sulfone
24, followed by addition of aldehyde 29 did not produce any
coupled products (Scheme 9). Regardless of the exact experi-
mental conditions, including choice of metal counterion, solvent,
order of addition, and temperature, the sulfone component was
always recovered unchanged along with aldehyde decomposition
products. Competing enolization of 29 led to substrate decom-
position in every instance examined. Similar observations were
made in an earlier study when the C24-C26 hydroxyls had been
protected as an acetonide. Since the problems associated with
the Julia olefination stemmed from the excessive basicity of
the sulfone anion, it was reasoned that buffering the reaction
mixture with a Lewis acid could overcome this problem.42
Indeed, deprotonation of 1.0 equiv of sulfone 24 with n-BuLi
followed by sequential addition of 1.0 equiv of BF3‚OEt2 and
aldehyde 29 provided the desired â-hydroxy sulfones as a
mixture of diastereomers. The use of freshly prepared MgBr2‚
Roxaticin. The final stages of the roxaticin synthesis are
summarized in Scheme 10. Formation of the dibutylboron
enolate of methyl ketone 19 (NEt3, Et2O, -78 °C), followed
by addition of aldehyde 35 at -110 °C delivered the aldol
adduct 36 in 79% yield as a single epimer at C20, as evidenced
1
by H NMR spectroscopy, along with 15% recovered methyl
ketone 19.15 It is interesting to note that a kinetic resolution
took place over the course of the reaction, affording the desired
product in preference to the adduct derived from reaction of
the undesired aldehyde epimer. The observed product contained
both the 1,5-anti and the 1,3-anti stereochemistry in the polyol
array. Therefore, an excess of the isomeric mixture of aldehydes
was employed in this reaction, and the recovered aldehyde
(41) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277-7287.
(42) Achmatowicz, B.; Baranowska, E.; Daniewski, A. R.; Pankowski, J.; Wicha,
J. Tetrahedron 1988, 44, 4989-4998.
9
J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003 10903

A R T I C L E S
Evans and Connell
Scheme 10 a
a Key: (a) 19 with Bu2BOTf and NEt3, -78 °C, then cool to -110 °C, then 35, Et2O; (b) Me4NBH(OAc)3, -25 °C, CH3CN, AcOH; (c) cyclopentylidene
dimethyl ketal, PPTS, CH2Cl2; (d) LiDBB, -78 °C, THF; (e) Dess-Martin, CH2Cl2; (f) DDQ, H2O, CH2Cl2; (g) 2, LiHMDS, -78 °C, THF; (h) LiOH,
THF, H2O, MeOH; (i) 2,4,6-trichlorobenzoyl chloride, NEt3, DMAP, toluene, 23 °C; (j) PPTS, rt, MeOH.
Scheme 11 a
typically was isolated as a 1:1 epimeric mixture at the â-position.
The subsequent hydroxyl-directed syn reduction of aldol 36
(Me4NHB(OAc)3, MeCN/AcOH, -35 °C19 proceeded smoothly,
affording the corresponding diol in good yield (85%) as a single
1
observable diastereomer by 500 MHz H NMR spectroscopy.
Ketalization afforded fragment 37 in quantitative yield. This
fully protected roxaticin precursor contains all the stereocenters
present in the natural product and is differentially protected in
a manner suitable for the introduction of the polyene fragment
2 and macrolactonization.
Deprotection of the primary C12-benzyl group under reduc-
tive conditions (LiDBB, -78 °C)43,44 followed by Dess-Martin
oxidation41 provided the epimerization-prone aldehyde 38 in
90% yield (Scheme 10). Although Horner-Wadsworth-Emmons
reaction with phosphonate 2 (Scheme 3) was the next obvious
step, deprotection of the PMB ether would be troublesome in
the presence of the oxidizable polyene moiety. Hence, the
oxidative deprotection of the C30-methoxybenzyl ether was
immediately carried out under buffered conditions (DDQ,45,46
pH 7 phosphate buffer, CH2Cl2) to deliver hydroxy aldehyde
a Key: (a) 2,4,6-trichlorobenzoyl chloride, NEt3, then DMAP, toluene,
6 h, 110 °C.
39 in 77% yield. The olefination step was then carried out by
prior to the addition of DMAP to avoid complete decomposi-
tion of the substrate, presumably initiated by acid-catalyzed
(Et3NHCl) destruction of the polyene. Final deprotection and
product isolation were eventually accomplished with PPTS in
MeOH at ambient temperature (<1 h) to deliver synthetic
roxaticin 1 in 63% yield after purification by reverse phase
HPLC. The synthetic material proved to be identical in all
aspects to the natural product.
While this synthesis appeared to proceed with few major
obstacles, we encountered significant difficulties when an earlier
version of this synthesis was executed with acetonide-based 1,3-
diol protection. In a comparison of the lactonization reactions,
deprotonation of phosphonate 2 with freshly prepared LiHMDS,
generating the deep purple lithium anion, which when added
to aldehyde 39 afforded the conjugated ethyl ester 40 in 83%
yield. Ester hydrolysis was straightforward under mild conditions
(LiOH/H2O/MeOH), and the unpurified acid was immediately
subjected to Yamaguchi macrolactonization conditions (2,4,6-
trichlorobenzoyl chloride, Et3N, THF, then DMAP, toluene, rt)
to provide the macrolactone 41 in 66% isolated yield for the
two steps.47 It was imperative to isolate the mixed anhydride
(43) Freeman, P. K.; Hutchinson, L. L. J. Org. Chem. 1980, 45, 1924-1930.
(44) Ireland, R. E.; Smith, M. G. J. Am. Chem. Soc. 1988, 110, 854-860.
(45) Horita, K.; Yoshioka, T.; Tanaka, T.; Oikawa, Y.; Yonemitsu, O.
Tetrahedron 1986, 42, 3021-3028.
(46) Oikawa, Y.; Yoshioka, T.; Yonemitsu, O. Tetrahedron Lett. 1982, 23, 885-
888.
(47) Inanaga, J. H. K.; Saeki, H.; Katsuki, T.; Yamaguchi, M. Bull. Chem. Soc.
Jpn. 1979, 52, 1989-1993.
9
10904 J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003

Antifungal Macrolide Antibiotic (+)-Roxaticin
A R T I C L E S
while the macrocyclization of the cyclopentylidine-protected
hydroxyacid proceeded at room temperature in toluene (40f41),
the equivalent acetonide-protected hydroxyacid 42 (Scheme 11)
required refluxing toluene to achieve formation of the desired
macrocycle 43. More significantly, acetonide deprotection
(43froxaticin) proceeded in low yield providing less than 10%
of the natural product. The greater lability48 of the cyclopen-
tylidene ketals proved to be crucial to the isolation of this acid-
sensitive natural product in reasonable yield, and this protecting
group is recommended to those who undertake the synthesis of
related structures.
macrolide synthesis, and this reaction has already been used in
our syntheses of altohyrtin C49 and phorboxazole B,50 as well
as by others in their approaches to polyol containing natural
products.51
Acknowledgment. Support has been provided by the National
Institutes of Health (GM-33328-18) and the National Science
Foundation. Fellowship support from the NSF for B.T.C. is also
acknowledged.
Supporting Information Available: Experimental details and
analytical data for all new compounds and comparison data for
natural and synthetic roxaticin (PDF) (27 pages). This material
is available free of charge via the Internet at http://pubs.acs.org.
Conclusion
The objective in this study has been to develop and integrate
stereoselective aldol addition reactions (eqs 1-3) into an
efficient synthesis of roxaticin. Stereochemical control in these
reactions has been derived from the use of either chiral aux-
iliaries or chiral catalysts. The boron enolate-mediated aldol
fragment coupling reaction employed to join the C20-C30
aldehyde and C12-C19 methyl ketone fragments (Scheme 10)
should be an important fragment coupling process in polyene
JA027638Q
(49) Evans, D. A.; Trotter, B. W.; Coleman, P. J.; Coˆte´, B.; Dias, L. C.;
Rajapakse, H.; Tyler, A. N. Tetrahedron 1999, 55, 8671-8726.
(50) Evans, D. A.; Fitch D. M.; Smith, T. E.; Cee, V. J. J. Am. Chem. Soc.
2000, 122, 10033-10046.
(51) (a) Bhattacharjee, A.; Soltani, O.; De Brabander, J. K. Org. Lett. 2002, 4,
481-484. (b) Kozmin, S. A. Org. Lett. 2001, 3, 755-758. (c) Paterson, I.;
Chen, D. Y. K.; Coster, M. J.; Acena, J. L.; Bach, J.; Gibson, K. R.; Keown,
L. E.; Oballa, R. M.; Trieselmann, T.; Wallace, D. J.; Hodgson, A. P.;
Norcross, R. D. Angew. Chem., Int. Ed. 2001, 40, 4055-4060. (d) Paterson,
I.; Collett, L. A. Tetrahedron Lett. 2001, 42, 1187-1191. (e) Trieselmann,
T.; Hoffman, R. W. Org. Lett. 2000, 2, 1209-1212.
(48) Hampton, A.; Fratantoni, J. C.; Carroll, P. M.; Wang, S.-C. J. Am. Chem.
Soc. 1965, 87, 5481-5487.
9
J. AM. CHEM. SOC. VOL. 125, NO. 36, 2003 10905