α- and β-Lipomycin: Total syntheses by sequential stille couplings and assignment of the absolute configuration of all stereogenic centers

Article abstract of DOI:10.1002/anie.201402255

40 years ago spectroscopy, derivatization, and degradation revealed the structures of α-lipomycin and its aglycon β-lipomycin except for the configurations of their side-chain stereocenters. We synthesized all relevant β-lipomycin candidates: the (12R,13S

Full text of DOI:10.1002/anie.201402255

Angewandte
Chemie
DOI: 10.1002/anie.201402255
Natural Product Synthesis
a- and b-Lipomycin: Total Syntheses by Sequential Stille Couplings
and Assignment of the Absolute Configuration of All Stereogenic
Centers **
Max L. Hofferberth and Reinhard Brꢀckner*
Dedicated to Professor Axel Zeeck on the occasion of his 75th birthday
Abstract: 40 years ago spectroscopy, derivatization, and
degradation revealed the structures of a-lipomycin and its
aglycon b-lipomycin except for the configurations of their side-
chain stereocenters. We synthesized all relevant b-lipomycin
candidates: the (12R,13S) isomer has the same specific rota-
tional value as the natural product. By the same criterion the
(12R,13S)-configured d-digitoxide is identical to a-lipomycin.
We double-checked our assignments by degrading a- and b-
lipomycin to the diesters 33 and 34 and proving their 3D
structures synthetically.
database.[6] Of these, 59 have the enol structure (Z)-2a; this
means that they are derived from saturated acyl substituents.
Twelve naturally occurring 3-acyltetramic acids are dienol
tautomers (Z)-2b of 3-enoyl tetramic acids. Fourteen natu-
rally occurring 3-dienoyl tetramic acids are known as well
which correspond to trienols (Z)-2c. 3-Polyenoyl tetramic
acids from nature are known, too. Their number, at present
33,[6] has been growing steadily.[7] They are vinylogues,
bis(vinylogues), and tris(vinylogues) of the trienols (Z)-2c.
Naturally occurring tetramic acids, the configurations of
which were recently (claimed to be) fully elucidated by total
synthesis, are penicillenol A2,[8] penicillenol C1,[9] epicoccari-
ne A,[10] and virginenone[11] (Figure 2). The structurally most
complex 3-acyltetramic acids isolated from natural sources to
date include aflastatin A,[12] which comprises 28 acyclic
stereocenters, and aurantoside A,[13] which is an N-glycoside
of a trisaccharide.[14]
T
etramic acids[1] are weakly acidic heterocycles 1 (Figure 1;
Rx = H: pKa = 6.4[2]). 3-Acyltetramic acids (2), which consist
mostly of enols (Z)-2 and to a lesser extent their stereoiso-
mers (E)-2,[3,4] are more acidic than acetic acid (3-R’ = 3-
acetyl, Rx = 5-sBu: pKa = 3.35[5]). There are at least 118
naturally occurring 3-acyltetramic acids 2 in the REAXYS
Figure 2. Naturally occurring tetramic acids, the configurations of
which were recently fully elucidated by total synthesis.
Figure 1. Tetramic acids (1), 3-acyl tetramic acids (2; preferred tauto-
mers shown), and their subclasses 2a–c (major isomer shown).
The 3-polyenoyltetramic acid natural product a-lipomycin
(3) and its aglycon b-lipomycin (8) were isolated from
Streptomyces aureofaciens by Zeeck et al. 40 years ago
(Scheme 1).[15] Neither 3 nor 8 was crystalline. As a conse-
quence, structure elucidation was based upon mass spectros-
[*] Dr. M. L. Hofferberth, Prof. Dr. R. Brꢀckner
Institut fꢀr Organische Chemie, Albert-Ludwigs-Universitꢁt Freiburg
Albertstrasse 21, 79104 Freiburg (Germany)
1
E-mail: reinhard.brueckner@ocbc.uni-freiburg.de
Homepage: http://www.brueckner.uni-freiburg.de
metry and IR, UV/Vis, and 100 MHz H NMR spectroscopy
in conjunction with chemical degradation studies.[15–17] Only
the configurations of the stereocenters in the polyenoyl chain
could not be determined. In 2006 the gene cluster that
encodes its biosynthesis was characterized.[18] A recent
comparison of the amino acid sequence of the ketoreductase,
which this gene cluster encodes and which establishes the
mentioned stereocenters, with 78 other ketoreductase
sequences read from 38 other polyketide synthase gene
clusters revealed clues for predicting the configurations of the
[**] We are grateful to Prof. Dr. Andreas Bechthold and to Dipl.-Biol. E.
Welle (both Institut fꢀr Pharmazeutische Wissenschaften, Albert-
Ludwigs-Universitꢁt Freiburg) for an authentic sample of a-lip-
omycin and for assisting in growing and extracting Streptomyces
aureofaciens (Tꢀ 117) to obtain a- and b-lipomycin. Financial
support for this study by the DFG (IRTG 1038) is gratefully
acknowledged.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201402255.
Angew. Chem. Int. Ed. 2014, 53, 1 – 8
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
stereocenters in the corresponding gene products in general
and in b-lipomycin (8) in particular.[19,20]
Lacey–Dieckmann cyclocondensation routes to tetramic
acids.[26] Since (S,E)-5 reacted as envisaged it represents
a worthwhile addition to the synthetic arsenal.
Our first goal was the synthesis of b-ketothioester 12
(Scheme 2). Isomerically pure trans-bromoacrylic acid (10)
was activated as the chloride or mixed anhydride. Surpris-
ingly, lithiated tBuSAc reacted beyond the desired mono-
We have now established the configurations of all of the
stereocenters of a-lipomycin (3) and b-lipomycin (8) by total
synthesis. Retrosynthetically we dissected the conceivable
candidate structures—four diastereomers per target mole-
cule—into three building blocks each (Scheme 1). The
distannane (E,E,E)-6 was meant to provide the all-E-con-
figured triene core of 3 and 8 by stepwise couplings of the two
C(sp2)–Sn moieties; this distannane had been introduced and
a related biscoupling strategy successfully exploited in earlier
work from our group.[21] The coupling “on the left” should
engage the bromoalkene building block (S,E)-5, the coupling
“on the right” one of the four iodoalkene diastereomers (E)-7
or one of their glycosides (E)-4; the latter were unknown yet
looked accessible from additions of the former to the bis(tert-
butyldimethylsilylether) 9[22] of d-digitoxal.
The iodoalkenes (E)-7 of Scheme 1 had not been not
reported previously in the literature. Syntheses of the benzyl
and tert-butyldimethylsilyl ether of the racemic syn-alcohol
[(4,5ul,E)-7] were the closest precedence.[23] There was an
enantioselective synthesis of the same unprotected homopro-
pargyl alcohol, which we used as a precursor of the
iodoalkene isomer (4R,5R,E)-7.[24] However, as will be
shown in Scheme 3 we accessed this alcohol differently. The
bromoalkene (S,E)-5 of Scheme 1 contains an enolized b-
ketoamide and an ester group. These entities should make it
possible to establish the 3-acyltetramic acid motif by an
intramolecular acylation known as the Lacey–Dieckmann
condensation.[25] Building block (S,E)-5 is polarity-reversed
relative to all previously used substrates for cross-coupling/
Scheme 2. Preparation of building block (S,E)-5 by aminolysis of the b-
ketothioester 12: a) T3P, Me(MeO)NH, NMP, RT, 25 min; 86%.
b) Solution from nBuLi (in hexane) and HN[Si(Me3)]2 in THF; addition
of tBuSAc, THF, ꢀ788C, 1 h; MgBr2·OEt2, 1 h; addition of 11
(1.0 equiv), 2 h; 83%. c) b-Ketothioester 12, AgO2CCF3 (1.1 equiv), MS
4 ꢂ, THF, 08C, 1 h; 83% over the 2 steps. d) Me3Si(CH2)2OH, DMAP,
CH2Cl2/DMF (10:1), 08C, dropwise addition of DCC in CH2Cl2, !RT,
15 h; 82%. e) MeI, Ag2O, DMF, RT, 18 h; 93%. f) Pd (10% on C), H2
(ca. 1.3 bar), EtOAc, 30 min; the crude product was carried on without
purification. DCC=dicyclohexylcarbodiimide, MS=molecular sieves,
T3P=propylphosphonic anhydride, TMSE=2-(trimethylsilyl)ethyl.
(thioester) stage (12) giving some bis(thioester) 18. As
a bypass trans-bromoacrylic acid (10) was converted into
the Weinreb amide[27] 11 although we were unaware of the
acylation of thioester enolates by Weinreb amides prior to our
work. Reaction with DCC, cat. DMAP, and Me-
(MeO)NH·HCl[28] provided 11 in up to 65% yield in a mixture
with the analogous chloride (17, 3%). Bromoacrylic acid 10,
propylphosphonic anhydride (T3P), N-methylmorpholine,[29]
and Me(MeO)NH gave 86% Weinreb amide 11 selectively.
Compound 11 and lithiated tBuSAc afforded not only the
desired acylation product 12 (40% yield) but also a 68:32
mixture (19% yield) of the 1,4-addition/b-elimination prod-
ucts 19 and iso-19. Gratifyingly, transmetalation of lithio-
tBuSAc with 2.2 equiv of MgBr2·OEt2 before addition of the
Weinreb amide favored selective acylation. The b-keto-
thioester 12 resulted without by-products according to the
300 MHz 1H NMR spectrum. It was isolated in 83% yield by
flash chromatograpy over a pad of silica gel.[30]
The bis[(trimethylsilyl)ethyl] ester 16 of l-N-methylglu-
tamic acid, not previously described, was prepared next
(Scheme 2). It was designed to undergo a deprotection under
mild conditions with Bu4NF at the polyenoyltetramic acid
stage.[31] Accordingly, N-(benzyloxycarbonyl)glutamic acid
(13) was esterified in 82% yield with 2-(trimethylsilyl)ethanol
and DCC.[32] Treatment with MeI and Ag2O[33] in DMF—as
described more generally by Olsen[34] and applied to a diester
of l-N-(benzyloxycarbonyl)aspartic acid by de Meijere
Scheme 1. The stereochemically incompletely characterized[15–17] poly-
enoyltetramic acids a-lipomycin (3) and b-lipomycin (8). Tracing back
the isomeric candidates to the building blocks (S,E)-5 (novel), (E,E,E)-
6,[21] (E)-4 or (E)-7, and to the disilylether 9[22] of d-digitoxal. TBS=tert-
butyldimethylsilyl, TMSE=2-(trimethylsilyl)ethyl.
2
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 8
These are not the final page numbers!

Angewandte
Chemie
syn-configured ethers (3R,4R)-24 and (3S,4S)-24, respectively.
Each of the four dibromoalkenes was subjected separately to
Br!Li exchange, which induced a Fritsch–Buttenberg–Wie-
chell rearrangement.[42] The resulting lithioalkynes were C-
methylated in situ.[43] Hydrozirconation, iodolysis, and
removal of the respective protecting group afforded the
anti-configured iodoalkenes (4R,5S,E)- and (4S,5R,E)-7 via
steps d–f and their syn diastereomers (4R,5R,E)- and
(4S,5S,E)-7 via steps l–n.
At a later stage of our work we learnt that synthesizing
naturally configured a-lipomycin (3) most likely required
including the anti-configured iodoalkene enantiomer
(4R,5S,E)-7 as a b-digitoxide in our route. With this goal in
mind we synthesized the digitoxide b-(4R,5S,E)-4 from the
digitoxyl donor d-28 (Scheme 4), which had been synthesized
once (in 9 steps).[22] We made d-28 differently, though, namely
by silylating the underlying diol 9, for which a 7-step synthesis
Scheme 3. Preparation of the four stereoisomers of iodoalkene (E)-7 as
pure enantiomers: a) Aldehyde 20, l-proline, DMF, 48C; addition of
aldehyde 22 in DMF over 30 h; 48C, 15 h; 41%. b) CBr4, PPh3, CH2Cl2,
08C, 10 min; addition of product from step (a), ꢀ788C, 1 h; 73%,
99% ee (GC). c) CH3OCH2Cl, iPr2NEt, CH2Cl2, 508C sealed tube, 4 h;
84%. d) nBuLi (in hexane), THF, ꢀ788C, 1 h; MeI, !RT, 1 h; 87%.
e) [Cp2ZrHCl], THF, ꢀ108C; !408C, 45 min; addition of solution of I2
in THF, 08C, 5 min; 75%. f) HClconc, MeOH, 608C, 30 min; 87%.
g) Bu2BOTf, NEt3, CH2Cl2, 08C, 15 min; ꢀ788C; aldehyde 22, 1 h; !
08C; addition of H2O2 (30% in H2O), MeOH, phosphate buffer
(pH 7), 1.5 h; 91% (Ref. [39]: 99%), d.r. >95:5 (1H NMR, 400 MHz,
CDCl3). h) Me(MeO)NH·HCl, AlMe3, CH2Cl2, ꢀ158C!RT, 3 h; 88%
(Ref. [39]: 88%). i) TBSOTf, 2,6-lutidine, CH2Cl2, RT, 20 min; 97%
(Ref. [39]: 99%). j) (iBu)2AlH, THF, ꢀ788C, 1.5 h; the crude product
was processed further without purification. k) PPh3, CBr4, CH2Cl2, 08C,
30 min; 80% over the 2 steps. l) Same as (d) but 90%. m) [Cp2ZrHCl],
MS 4 ꢂ, THF, ꢀ108C!RT, 1.5 h; I2, 08C, 5 min; 79%. n) BF3·OEt2,
CH2Cl2, 08C!RT, 2 h; 95%. MOM=methoxymethyl, MS=molecular
sieves, TBSOTf=tBuMe2SiO3SCF3.
Scheme 4. Preparation of d-digitoxal (d-28) by literature proce-
dures,[44–47] by a silylation (!9) in analogy to Ref. [47], and by the b-
selective digitoxylation of iodoalkene (4R,5S)-7: a) pTsCl, pyridine,
558C, 4 d; 95% (Ref. [44]: 78%). b) NaOMe, CH2Cl2/MeOH (6:1),
08C, 2 d; RT, 1 d; 86% (Ref. [45]: 85%). c) NBS, AIBN, benzene,
reflux, 15 min; again AIBN, reflux, 15 min; 84% [Ref. [46] published no
yield for a related procedure employing (PhCO2)2 instead of AIBN and
30 min reaction time]. d) NaOMe, MeOH, RT, 20 min; 92% (77%
over the 2 steps; Ref. [46] “overnight”: 80% over the 2 steps).
e) NaBH4, NiCl2·6H2O, H2O/EtOH (3:1), reflux, 50 min; addition of
more NaBH4 in H2O, reflux, 40 min; 84% (Ref. [46]: 95%).
f) LiI·(H2O)1.5ꢀ3.0, HOAc, CHCl3, RT, 2 h; 408C, 1.5 h; 61% (Ref. [47]:
52%). g) nBuLi (in hexane), THF, ꢀ158C, 30 min; 408C, 1 h; 80% of
the crude product (Ref. [47] with MeLi instead of nBuLi: 67%).
h) TBSOTf, imidazole, DMAP, DMF, 558C, 2 h; 63%. i) PPh3·HBr,
508C, 3 d; 80% (4R,5S,E)-4. j) Bu4N+Fꢀ, THF, RT, 5 h; 73%. AIBN=
azobis(isobutyronitrile), DMAP=4-(N,N-dimethylamino)pyridine,
NBS=N-bromosuccinimide, pTs=para-toluenesulfonyl.
et al.[35]—furnished the N-methylated diester 15 in 93% yield
and reliably with 97–98% ee. The benzyloxycarbonyl group
was removed by hydrogenolysis within 30 min. The resulting
aminodiester 16 lactamized at room temperature within
2 days.[36] In order to circumvent this, 16 had to be engaged
without delay in the aminolysis of b-ketothioester 12. We used
AgO2CCF3 as a promoter as described for related reactants[37]
and the b-ketoamide (S,E)-5 was prepared in 83% yield over
the two steps.
The four stereoisomeric iodoalkenes (E)-7 were obtained
by known diastereo- and enantioselective aldol additions to
isobutanal (Scheme 3). l- and d-Proline-catalyzed crossed
additions of propionaldehyde to isobutanal provided the anti-
aldols with (2R,3S) and (2S,3R) configuration, respectively
(step a[38]), albeit in lower yields than previously reported.
Additions of the boron enolate of Evansꢀ propionyl oxazo-
lidinone 23 and of its enantiomer to isobutanal delivered the
expected syn-hydroxyimides with (2R,3R) and (2S,3S) con-
figuration, respectively (step g[39]). Without protection of the
OH group[40] each of the anti-aldols was C1-elongated by the
Wittig reagent formed from CBr4 and PPh3 to give the
corresponding gem-dibromoalkene (step b[41]). The latter was
MOM-protected, which furnished the anti-configured ethers
(3R,4S)-21 and (3S,4R)-21, respectively. The Evans-type syn-
aldols were processed further through known transformations
(Weinreb amide formation; tert-butyldimethylsilylation of the
OH group; DIBAH reduction; steps h–j[39]) and gem-dibro-
momethylenation as before (step k[41]). This furnished the
existed. First, the commercially available methyl a-d-gluco-
side 25 was tosylated (step a[44]). Epoxide formation in the
presence of base (step b[46]) and defunctionalization of C-6
ensued (steps c–e[47]). The resulting epoxide 26 was ring-
opened with lithium iodide with a 2:1 bias (Ref.[47]: 4:1) for
the Fꢁrst–Plattner product 27 (step f). The latter was sepa-
rated from the minor ring-opening product iso-27 by flash
chromatography on silica gel.[30] Though it is an iodohydrin,
when compound 27 was treated with BuLi it did not re-form
epoxide 26 but underwent I!Li exchange. Thereupon
elimination of LiOMe afforded d-digitoxal (d-28; step g[47]).
The latter was bis(tert-butyldimethylsilylated)[48] giving glycal
9 (step h, 63% yield).
Angew. Chem. Int. Ed. 2014, 53, 1 – 8
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
Glycal 28 and 1 mol% PPh3·HBr had been shown to effect
the b-selective (97:3) digitoxylation of a sterically demanding
secondary alcohol at room temperature in 88% yield.[22] An
analogous digitoxylation of our secondary alcohol (“iodo-
alkene”) (4R,5S,E)-7 proceeded much more slowly. Complete
conversion required more catalyst (11 mol% PPh3·HBr),
higher temperature (508C), and an extended reaction time
(3 d). Gratifyingly, a single digitoxide was formed (80%
yield). Its anomeric center was b-configured as evidenced by
the magnitudes of the vicinal coupling constants in the
1H NMR spectrum (400.1 MHz, CDCl3).[49] Deprotection
with Bu4N+Fꢀ afforded the desired building block
(4R,5S,E)-4 in 73% yield.
Having prepared the six building blocks identified by the
retrosynthetic analysis of Scheme 1 for the assembly of four
diastereomers of b-lipomycin (8), including the one from
nature, we proceeded to the Stille couplings (Scheme 5). We
obtained the best overall yields when any iodoalkene
stereoisomer
7 was first coupled with the distannane
(E,E,E)-6[21] (step a; 49–55% yield). Secondly, each of the
resulting monostannanes 29 was Stille-coupled with the
bromovinylated b-ketoamide (S,E)-5 (step c; 53–85%
yield). The resulting pentaenes 31 contained the full carbon
backbone of b-lipomycin (8).[50] Each pentaene 31 underwent
a Lacey–Dieckmann cyclization[25] in the presence of tetra-
methylguanidine; tetrabutylammonium fluoride was then
added to remove the TMSE group (one-pot transformation;
step d, 56–89% yield). This provided the desired set of four
diastereomers with the constitution of b-lipomycin (8). None
of their 1H or 13C NMR spectra in CDCl3 revealed a sizeable
amount of isomers and the spectra of (12R,13S)-8 showed no
Scheme 5. Five stepwise Stille biscouplings of the bisstannane
(E,E,E)-6,[21] each setting the stage for a terminating Lacey–Dieckmann
condensation[25] (31!8; 32!3)/desilylation sequence: a) 6,
Bu4N+Ph2PO2ꢀ, AsPh3, CuI, [Pd(dba)2], molecular sieves (4 ꢂ), NMP,
RT, 0.25–2 h; 49–55%.[+] b) Same as (a) but using b-(4R,5S,E)-4,
NMP/THF (10:1), 30 min; 60%. c) (S,E)-5, AsPh3, [Pd(dba)2], NMP,
1.25–3 h; 53–85%.[++] d) N,N,N’,N’-Tetramethylguanidine, THF, RT,
2–3 h; addition of Bu4N+Fꢀ, 40–508C, 1–3 h; 56–89%.
=
isomers at all. In each isomer of 8 the disubstituted C C
bonds were E-configured; this followed from the correspond-
3
ing JCH
values. In compound (12R,13S)-8 the trisubsti-
=
CH
e) N,N,N’,N’-Tetramethylguanidine, THF, RT, 30 min; addition of
Bu4N+Fꢀ, 508C, 1 h; mixture of 49% (12R,13S)-3 and 24% (12R,13S)-
8; after preparative HPLC 25% (12R,13S)-3. BHT=Di-tert-butylated
para-hydroxytoluene, dba=dibenzylidenacetone, NMP=N-methylpyr-
rolidone. [+] Isꢀomer (9S,10S)-29 was prepared by replacing
=
tuted C C bond was also E-configured, as shown by a NOE
between 12-H and 10-CH3.
Methanol solutions of (12R,13S)-, (12S,13R)-, (12R,13R)-,
and (12S,13S)-8 were uniformly levorotatory. The specific
20
rotation of (12R,13S)-8 (½aꢁD ¼ꢀ180) matched the value
Bu4N+Ph2PO2 and the molecular sieves by BHT (2 mol%). [++] For
preparing isomers (14S,15S)- and (14R,15S)-31 we employed
reported for natural b-lipomycin (8; ꢀ176[15]) closely whereas
the specific rotations of the other diastereomers (ꢀ45, ꢀ136,
and ꢀ120, respectively) differered substantially (Table 1).
Unless contaminants contributed to these data one could
conclude that the side-chain of b-lipomycin (8) is (12R,13S)-
configured. This analysis was corroborated by completing the
synthesis of diastereomer (12R,13S)-3 of a-lipomycin
(Scheme 5). This compound possessed the same side-chain
configuration as that deduced for natural b-lipomycin,
(12R,13S)-8, and was reached from the digitoxylated iodo-
alkene building block b-(4R,5S,E)-4 by the analogous trans-
formations: 1) Stille coupling with distannane (E,E,E)-6[21]
(step b; 60% yield); 2) Stille coupling of the resulting
monostannane (9R,10S)-30 with the bromovinylated b-keto-
amide (S,E)-5 (step c; 67% yield); 3) Lacey–Dieckmann
cyclization[25]/removal of the TMSE group (step e). A 67:33
mixture of a-lipomycin [(12R,13S)-3] and b-lipomycin
[(12R,13S)-8] resulted at first. Separation by preparative
HPLC afforded (12R,13S)-3 in 25% yield. The 1H NMR
(500 MHz, CDCl3) spectrum of this compound showed very
Bu4N+Ph2PO2 (1.5 equiv).
ꢀ
Table 1: Specific rotations of synthetic b-lipomycin [(12R,13S)-8], its
diastereomers (12S,13R)-, (12R,13R)-, and (12S,13S)-8, natural b-lipo-
mycin (8),[15] synthetic a-lipomycin [(12R,13S)-3], and natural a-lipo-
mycin (3).[15]
20
Tetramic acid
½aꢁD in MeOH
Natural lipomycins:
20
½aꢁD in MeOH[15]
(12R,13S)-8
(12S,13R)-8
(12R,13R)-8
(12S,13S)-8
(12R,13S)-3
ꢀ180 (c=0.080)
ꢀ45 (c=0.031)
ꢀ136 (c=0.035)
ꢀ120 (c=0.045)
ꢀ237 (c=0.10)
ꢀ176 (c=0.09)
ꢀ229 (c=0.10)
4
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 8
These are not the final page numbers!

Angewandte
Chemie
broad signals. Broadened 1H and weak 13C NMR resonances,
which we observed for (12R,13S)-3, too, are known in
acyltetramic acids.[3,50] The phenomenon was explained by
the fact that these compounds sequester Mg2+, Ca2+, and Fe2+
contaminants of silica gel.[51] Accordingly, we considered the
NMR spectra of (12R,13S)-3 to be normal and did not resort
to any countermeasures (cf. Ref. [3]). Synthetic (12R,13S)-3
We subjected the above-mentioned 7.8:0.9 mixture of b-
lipomycin (8) and a presumed isomer with a (Z)-configured
=
disubstituted C C bond to a Lemieux–Johnson cleavage of
=
(most of) the C C bonds. We reduced the expected carbonyl
compounds by NaBH4 and isolated a mixture wherein we
=
=
identified one C C-containing 1,5-diol and one C C-free 1,3-
diol by GC–MS analysis. This mixture was esterified with
bis(trifluoroacet)imide[52] giving the corresponding bis(tri-
fluoroacetates) 33 and 34 (Table 2). Comprehensive sets of all
20
exhibited ½aꢁD ¼ꢀ237 in methanol solution (Table 1). This was
20
only 3.5% off the value ½aꢁD ¼ꢀ229 for natural a-lipomycin
(3[15]).
As likely as it appeared that we had identified configura-
tionally and achieved to synthesize both a- (3) and b-
Table 2: Proving the 3D structure of the side-chains of a-lipomycin (3)
and b-lipomycin (8) by GC comparisons of their trifluoroacetylated
degradation products 33from 3 via 8 and 34from 3 via 8 with two comprehensive
sets of synthetic reference compounds.
20
lipomycin (8), we were hesitant to base our analysis on ½aꢁD
values alone. The non-existence of high-field NMR data for
the natural products[15–17] thwarted 1H and 13C NMR compar-
isons as identity proofs. We therefore sought independent
evidence for the lipomycin side-chain configurations by
isolating/preparing b-lipomycin (8) from the natural source
(Streptomyces aureofaciens Tꢁ 177; Scheme 6). We extracted
Degradation GC retention time Synthetic
GC reten-
product of
lipomycin
of degradation
bis(trifluoroacetate) tion time[a]
for reference
product[a]
33from 3 via 8
45.5 min
(4R,5S)-33
(4S,5R)-33
(4R,5R)-33
(4S,5S)-33
45.7 min
43.4 min
60.1 min
59.7 min
Scheme 6. Chemical degradation of a-lipomycin (3) via b-lipomycin
(8); yields were not determined. a) Aq. HCl (10%)/H2O/MeCN
(4:32:64 v/v/v), 5 h; preparative HPLC [XBridge, MeOH/H2O (45:55
v/v) + NH3 (10 mm), 16 mLminꢀ1, lUV/vis =300 nm, tr,8 =8.0 min].
b) NaIO4 (13.2 equiv), K2OsO4·2H2O (30 mol%), EtOH/H2O (1:1 v/v),
RT, 1.5 h; NaBH4 (120 equiv), 1 h. c) (CF3CO)2NMe (50 equiv), Et2O,
RT, 2 h. The resulting 33/34 mixture (ca. 1:1) was analyzed by GC
(Table 2).
Degradation GC retention time Synthetic
GC reten-
product of
lipomycin
of degradation
bis(trifluoroacetate) tion time[b]
for reference
product[b]
34from 3 via 8
38.9 min
(2R,3S)-34
(2S,3R)-34
(2R,3R)-34
(2S,3S)-34
38.9 min
38.0 min
37.1 min
36.1 min
[a] GC conditions: Astec Chiraldex G-TA, column length 30 m, column
diameter 0.25 mm, coating thickness 0.12 mm, Tinjector =2008C,
4.4 L of the corresponding culture broth with an equal volume
of ethyl acetate, evaporated the solvent, purified the residue
by HPLC, and dissolved the isolate in acetonitrile/H2O (2:1).
We added HCl for deglycosylating a-lipomycin (3) to obtain
b-lipomycin (8). Extractive isolation with tBuOMe after
saturation with solid NaCl and final purification by HPLC
delivered 7.8 mg of b-lipomycin (8) admixed with 0.9 mg of
a presumed (Z)-isomer. The 400 MHz 1H NMR spectrum
(CDCl3) of this sample of b-lipomycin (8) from nature
resembled the 1H NMR spectra of (12R,13S)-8 and
(12S,13R)-8 (i.e., the ul isomers) in all respects but differed
from the spectra of (12R,13R)-8 and 12S,13S)-8 (i.e., the lk
isomers) in two regards: 1) the three CH-CH3 groups gave
rise to two doublets in natural 8 and in ul-8 while they caused
Tcolumn =858C (isocratic), flame ionization detector. [b] GC conditions:
FS-Cyclodex beta-I/P, column length 50 m, column diameter 0.32 mm,
coating thickness not published on manufacturer’s website (http://www.
cs-chromatographie.de) and accordingly unknown, Tinjector =2508C,
Tcolumn =808C (isocratic), flame ionization detector.
stereoisomers of both bis(trifluoroacetates) were synthesized
for comparison.[53] GC analyses of each bis(trifluoroacetate)
from the degradation of lipomycin and of the pertinent
reference compounds from synthesis allowed the identifica-
tion of bis(trifluoroacetate) 33from 3 via 8 as compound (4R,5S)-
33 and bis(trifluoroacetate) 34from 3 via 8 as compound (2R,3S)-
34 by their pairwise matching retention times (Table 2). The
two identities proved that the sterocenters in the side-chain of
a-lipomycin (3) and b-lipomycin (8) were configured 12R and
13S. This result coincided with what we had concluded from
our synthetic endeavour.
=
three doublets in lk-8; 2) the CMe CH resonance of natural 8
and ul-8 was at d = 5.60 ppm whereas it appeared at d =
5.48 ppm in lk-8. These coincidences and discrepancies
proved unambiguously that the side-chains of a- (3) and b-
lipomycin (8) were anti-configured in accordance with our
earlier conclusions.
Angew. Chem. Int. Ed. 2014, 53, 1 – 8
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
[8] T. Sengoku, Y. Nagae, Y. Ujihara, M. Takahashi, H. Yoda, J. Org.
Chem. 2012, 77, 4391 – 4401.
[9] K. Kempf, A. Raja, F. Sasse, R. Schobert, J. Org. Chem. 2013, 78,
2455 – 2461.
[10] Y. Ujihara, K. Nakayama, T. Sengoku, M. Takahashi, H. Yoda,
Org. Lett. 2012, 14, 5142 – 5145.
[11] A. Yajima, C. Ida, K. Taniguchi, S. Murata, R. Katsuta, T.
Nukada, Tetrahedron Lett. 2013, 54, 2497 – 2501.
[12] H. Ikeda, N. Matsumori, M. Ono, A. Suzuki, A. Isogai, H.
Nagasawa, S. Sakuda, J. Org. Chem. 2000, 65, 438 – 444.
[13] S. Matsunaga, N. Fusetani, Y. Kato, J. Am. Chem. Soc. 1991, 113,
9690 – 9692.
[14] Bicyclic tetramic acids represent another field of recent interest.
See: a) C. A. Holloway, C. J. Matthews, Y.-C. Jeong, M. G.
Moloney, C. F. Roberts, M. Yaqoob, Chem. Biol. Drug Des. 2011,
78, 229 – 235; b) Y.-C. Jeong, M. Anwar, T. M. Nguyen, B. S. W.
Tan, C. L. L. Chai, M. G. Moloney, Org. Biomol. Chem. 2011, 9,
6663 – 6669; c) M. Anwar, M. G. Moloney, Chem. Biol. Drug
Des. 2013, 80, 645 – 649.
[15] B. Kunze, K. Schabacher, H. Zꢂhner, A. Zeeck, Arch. Mikrobiol.
1972, 86, 147 – 174.
[16] K. Schabacher, A. Zeeck, Tetrahedron Lett. 1973, 14, 2691 –
2694.
[17] A. Zeeck, Justus Liebigs Ann. Chem. 1975, 2079 – 2088.
[18] A. Bihlmaier, E. Welle, C. Hofmann, K. Welzel, A. Vente, E.
Breitling, M. Mꢁller, S. Glaser, A. Bechthold, Antimicrob.
Agents Chemother. 2006, 50, 2113 – 2121.
[19] A. Kitsche, M. Kalesse, ChemBioChem 2013, 14, 851 – 861.
[20] This insight set the stage for an independent second total
synthesis of 8 (O. Hartmann, M. Kalesse, Angew. Chem. 2014,
DOI: 10.1002/ange.201402259; Angew. Chem. Int. Ed. 2014,
DOI: 10.1002/anie.201402259).
[21] a) R. Brꢁckner, K. Siegel, A. Sorg in Strategies and Tactics in
Organic Synthesis, Vol. 5 (Ed.: M. Harmata), Elsevier, Amster-
dam, 2004, pp. 437 – 473; b) A. Sorg, R. Brꢁckner, Angew. Chem.
2004, 116, 4623 – 4626; Angew. Chem. Int. Ed. 2004, 43, 4523 –
4526; c) A. Sorg, K. Siegel, R. Brꢁckner, Chem. Eur. J. 2005, 11,
1610 – 1624; d) J. Burghart, R. Brꢁckner, Eur. J. Org. Chem.
2011, 150 – 165.
Our synthesis of b-lipomycin [(12R,13S-8)] required 16
steps, comprised 9 steps in the longest linear sequence, and
gave an overall yield of 6%. The following Communication
describes an independent different synthesis of b-lipomycin
(12R,13S-8).[20] It was realized after a comprehensive analysis
of the polyketide synthase genome,[19] by which the config-
urations of C-12 and C-13 were predicted correctly. Whatever
concerns one might have held against “assignments derived
from synthesis and analysis”, there are none against the
identical “assignments from chemical degradation”, which we
provided.
Altamycin[54] and oleficin[55] are 3-polyenoyl tetramic
acids, which possess the constitution of a lower and a higher
vinylogue, respectively, of a-lipomycin (3). This similarity
leaves one wondering whether their unexplored side-chain
stereocenters have the same configurations as those of a-
lipomycin (3), in other words, whether altamycin is bisnor-a-
lipomycin and whether oleficin is bishomo-a-lipomycin. Our
conditions for degrading a-lipomycin (3) to the bis(trifluoro-
acetates) 33 and 34 and our pool of reference compounds[53]
for assessing their configurations should reveal the 3D
structures of altamycin and oleficin readily. Moreover the
retrosynthetic disconnection on which we based our total
synthesis of a-lipomycin (3) should be applicable to the
syntheses of altamycin, oleficin, and other 3-(polyenoyl)tetr-
amic acids as well.
Received: February 17, 2014
Published online: && &&, &&&&
Keywords: configuration determination · cross-coupling ·
.
natural products · polyenes · tetramic acids
[22] Bis(tert-butyldimethylsilyl)-protected d-digitoxal (9) as a glyco-
sylating agent: a) F. E. McDonald, K. Subba Reddy, Y. Diaz, J.
Am. Chem. Soc. 2000, 122, 4304 – 4309; b) F. E. McDonald, K.
Subba Reddy, Angew. Chem. Int. Ed. 2001, 40, 3653 – 3655;
Angew. Chem. 2001, 113, 3765 – 3767.
[23] M. J. C. Buckle, I. Fleming, S. Gil, K. L. C. Pang, Org. Biomol.
Chem. 2004, 2, 749 – 769.
[24] S. C. Archibald, D. J. Barden, J. F. Y. Bazin, I. Fleming, C. F.
Foster, A. K. Mandal, A. K. Mandal, D. Parker, K. Takaki, A. C.
Ware, A. R. B. Williams, A. B. Zwicky, Org. Biomol. Chem.
2004, 2, 1051 – 1064.
[1] Reviews: a) H.-G. Henning, A. Gelbin, Adv. Heterocycl. Chem.
1993, 57, 139 – 185; b) B. J. L. Royles, Chem. Rev. 1995, 95, 1981 –
2001; c) E. L. Ghisalberti in Studies in Natural Product Chemis-
try, Vol. 28 (Ed.: Atta-ur-Rahman), Elsevier, Dordrecht, 2003,
pp. 109 – 163; d) R. Schobert, A. Schlenk, Bioorg. Med. Chem.
2008, 16, 4203 – 4221; e) Y.-C. Jeong, M. Anwar, Z. Bikadi, E.
Hazai, M. G. Moloney, Chem. Sci. 2013, 4, 1008 – 1015.
[2] L. Jurd, J. Heterocycl. Chem. 1996, 33, 1227 – 1232.
[3] Y.-C. Jeong, M. G. Moloney, J. Org. Chem. 2011, 76, 1342 – 1354.
[4] V. V. Gromak, V. G. Avakyan, O. F. Lakhvich, J. Appl. Spectrosc.
2000, 67, 205 – 215.
[25] R. N. Lacey, J. Chem. Soc. 1954, 850 – 854.
[5] C. E. Stickings, Biochem. J. 1959, 72, 332 – 340.
[26] Earlier substrates of cross-coupling/Lacey – Dieckmann cyclo-
condensation[25] routes to tetramic acids were the stannylated b-
ketoamide esters 35 (D. J. Dixon, S. V. Ley, T. Gracza, P.
Szolcsanyi, J. Chem. Soc. Perkin Trans. 1 1999, 839 – 841) and 36
(D. A. Longbottom, A. J. Morrison, D. J. Dixon, S. V. Ley,
Angew. Chem. Int. Ed. 2002, 41, 2786 – 2790; Angew. Chem.
2002, 114, 2910 – 2914; D. A. Longbottom, A. J. Morrison, D. J.
Dixon, S. V. Ley, Tetrahedron 2003, 59, 6955 – 6966). In contrast
to our b-ketoamide ester 5, 35 and 36 act as nucleophiles in their
cross-couplings.
[6] January 29, 2013: The search items were tautomers 1 (125 hits)
and their enols (90 hits); altogether they comprised 177 different
hits.
[7] For example: F. Vinale, G. Flematti, K. Sivasithamparam, M.
Lorito, R. Marra, B. W. Skelton, E. L. Ghisalberti, J. Nat. Prod.
2009, 72, 2032 – 2035; b) J. C. Carlson, S. Li, D. A. Burr, D. H.
Sherman, J. Nat. Prod. 2009, 72, 2076 – 2079; c) S. Cao, J. A. V.
Blodgett, J. Clardy, Org. Lett. 2010, 12, 4652 – 4654; d) M. Isaka,
P. Chinthanom, S. Supothina, P. Tobwor, N. L. Hywel-Jones, J.
Nat. Prod. 2010, 73, 2057 – 2060; e) Z. Yu, S. Vodanovic-
Jankovic, N. Ledeboer, S.-X. Huang, S. R. Rajski, M. Kron, B.
Shen, Org. Lett. 2011, 13, 2034 – 2037; f) R. Kumar, R. Sub-
ramani, K.-D. Feussner, W. Aalbersberg, Mar. Drugs 2012, 10,
200 – 208; g) M. E. Rateb, Z. Yu, Y. Yan, D. Yang, T. Huang, S.
Vodanovic-Jankovic, M. A. Kron, B. Shen, J. Antibiot. 2014, 67,
127 – 132.
6
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 8
These are not the final page numbers!

Angewandte
Chemie
[27] S. Nahm, S. M. Weinreb, Tetrahedron Lett. 1981, 22, 3815 – 3818.
[28] Method: R. H. Munday, R. M. Denton, J. C. Anderson, J. Org.
Chem. 2008, 73, 8033 – 8038.
[29] Procedure: http://www.euticals.com/attachments/082_t3p.pdf.
[30] W. C. Still, M. Kahn, A. Mitra, J. Org. Chem. 1978, 43, 2923 –
2925.
[43] Such dibromomethylenation / Br,Li exchange / rearrangement
sequences, possibly followed by a C-alkylation, were introduced
by E. J. Corey, P. L. Fuchs, Tetrahedron Lett. 1972, 13, 3769 –
3972.
[44] V. P. Miller, D.-y. Yang, T. M. Weigel, O. Han, H.-w. Liu, J. Org.
Chem. 1989, 54, 4175 – 4188.
[31] P. Sieber, Helv. Chim. Acta 1977, 60, 2711 – 2716.
[32] Method: A. J. Pearson, M. V. Chelliah, J. Org. Chem. 1998, 63,
3087 – 3098.
[33] Ag2O was prepared as described by D. E. Janssen, C. V. Wilson,
Org. Synth. 1956, 36, 46 – 47.
[34] R. K. Olsen, J. Org. Chem. 1970, 35, 1912 – 1915.
[35] V. N. Belov, M. Brands, S. Raddatz, J. Krꢁger, S. Nikolskaya, V.
Sokolov, A. de Meijere, Tetrahedron 2004, 60, 7579 – 7589.
[36] For details see the Supporting Information.
[37] S. V. Ley, S. C. Smith, P. R. Woodward, Tetrahedron 1992, 48,
1145 – 1174.
[38] a) A. B. Northrup, D. W. C. MacMillan, J. Am. Chem. Soc. 2002,
124, 6798 – 6799; b) P. M. Pihko, A. Erkkliꢂ, Tetrahedron Lett.
2003, 44, 7607 – 7609; c) A. Cꢃrdova, Tetrahedron Lett. 2004, 45,
3949 – 3952.
[45] P.-E. Sum, L. Weiler, Can. J. Chem. 1982, 60, 327 – 334.
[46] H. Paulsen, V. Sinnwell, Chem. Ber. 1978, 111, 879 – 889.
[47] J. Thiem, P. Ossowski, J. Schwentner, Chem. Ber. 1980, 113, 955 –
969.
[48] Procedure: J. Ma, Y. Zhao, S. Ng, J. Zhang, J. Zeng, A. Than, P.
Chen, X.-W. Liu, Chem. Eur. J. 2010, 16, 4533 – 4540.
[49] For details see the Supporting Information.
=
[50] We did not assess either the C C bond configurations or the
stereochemical homogeneity of the monostannanes 29 or the
pentaenes 31 because of signal overlap (29, 31) and the
interference of satellite peaks caused by 1H,117Sn, 1H,119Sn,
13C,117Sn, and 13C,119Sn couplings. Each isomer of 31 was
a mixture of the keto and the enol tautomers of both the (Z)-
and the (E)-configured amide “rotamers”. In contrast, the
isomers of compound 8 represented 9:1 mixtures of (Z)- and
(E)-configured enols.
[39] D. A. Evans, B. D. Allison, M. G. Yang, C. E. Masse, J. Am.
Chem. Soc. 2001, 123, 10840 – 10852.
[51] B. Barnickel, R. Schobert, J. Org. Chem. 2010, 75, 6716 – 6719.
[52] M. Donike, J. Chromatogr. A 1973, 78, 273 – 279.
[53] For syntheses see the Supporting Information.
[54] Y. D. Shenin, Antibiot. Med. Biotekhnol. 1986, 835 [Chem. Abstr.
1987, 106, 66971].
[55] a) G. Horvꢄth, J. Gyimesi, Z. Mꢅhesfalvi-Vajna, Tetrahedron
Lett. 1973, 14, 3643 – 3648; b) Structure revision: J. Gyimesi, Z.
Mꢅhesfalvi-Vajna, G. Horvꢄth, J. Antibiot. 1978, 31, 626 – 627.
[40] Dibromomethylenations of unprotected b-hydroxyaldehydes
with CBr4/PPh3: a) M. J. Robins, S. F. Wnuk, X. Yang, C.-S.
Yuan, R. T. Borchardt, J. Balzarini, E. De Clerq, J. Med. Chem.
1998, 41, 3857 – 3864; b) N. F. Langille, J. S. Panek, Org. Lett.
2004, 6, 3203 – 3206.
[41] Method: F. Ramirez, N. B. Desai, N. McKelvie, J. Am. Chem.
Soc. 1962, 84, 1745 – 1747.
[42] Method: G. Kçbrich, H. Trapp, K. Flory, W. Drischel, Chem. Ber.
1966, 99, 689 – 697.
Angew. Chem. Int. Ed. 2014, 53, 1 – 8
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7
These are not the final page numbers!

.
Angewandte
Communications
Communications
Natural Product Synthesis
M. L. Hofferberth,
R. Brꢁckner*
&&&& — &&&&
a- and b-Lipomycin: Total Syntheses by
Sequential Stille Couplings and
Assignment of the Absolute
Configuration of All Stereogenic Centers
Making doubly sure: 40 years ago the
structures of a-lipomycin and its aglycon
b-lipomycin were determined except for
the configurations of the side-chain ste-
reocenters. All of the relevant b-lipomycin
candidates have now been synthesized.
The optical rotation of the (12R,13S)
isomer matched that of the natural
product. The (12R,13S)-configured d-
digitoxide was synthesized too and iden-
tified as a-lipomycin.
8
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 8
These are not the final page numbers!