Total synthesis of thuggacin B

Article abstract of DOI:10.1002/anie.200803271

Final proof of the complete structures of thuggacins A-C was gained by a highly convergent and flexible total synthesis. Notable features include a substrate-controlled, titanium-mediated aldol reaction, a cross-metathesis approach for converting terminal alkenes into the corresponding vinyl iodides, and the cross-coupling of a complex vinyl iodide and terminal alkyne in the Sonogashira reaction. (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200803271

Communications
DOI: 10.1002/anie.200803271
Natural Product Synthesis
Total Synthesis of Thuggacin B**
Martin Bock, Richard Dehn, and Andreas Kirschning*
Dedicated to Ernst Schaumann on the occasion of his 65th birthday
After AIDS, tuberculosis (TB) is the infectious disease with
the highest mortality worldwide.[1] TB and HIVepidemics fuel
one another in co-infected people, and at least 11 million
adults are infected with both pathogens.[2] The efficacy of first-
line anti-TB drug regiments is often reduced owing to drug
resistance.[3] Additionally, the ability of Mycobacterium tuber-
culosis to persist in latent infections necessitates the develop-
ment of alternative antibiotics, preferably with novel modes
of action.[1]
Thuggacin A (1) features a 17-membered a,b-unsaturated
macrolactone with a thiazole ring, a diene (11E,13Z), and an
n-hexyl side chain at C2. A side chain at C16 bearing three
hydroxy groups and a diene unit complements the structure.
Thuggacin B (2) shares these structural features except for the
ring size: the lactone is closed at O17 instead of at O16 in
thuggacin A. Thuggacin C (3) is macrocyclized at O18.
Importantly, the thuggacins slowly equilibrate in methanol
within five days at room temperature by transacylation (1/2/
3 = 1:2.1:2.7).[4] The rate of interconversion can be slowed by
acidifying the methanolic solution or by switching to an
aprotic solvent.
To further evaluate the biological properties of the
thuggacins we planned a total synthesis which should also
pave the way for accessing analogues. Therefore, we devel-
oped a highly convergent approach that could be used later to
prepare simplified macrocycles.
Recently, Jansen et al.[4] and we[5] disclosed the complete
structures of the polyketide natural products thuggacin A (1),
B (2), and C (3) which had been isolated from the
myxobacterium Sorangium cellulosum (Figure 1). These
compounds show strong antibiotic activity against various
organisms including Mycobacterium tuberculosis by targeting
the bacterial respiratory chain.[6]
Our retrosynthetic analysis (Scheme 1) of the thuggacins
1–3 relies on a macrolactonization and a Pd-mediated cross-
coupling reaction to form the bond linking C12 and C13, such
that major fragments 4 and 5 should serve as key building
blocks in this project. Precursor 4 was further disconnected to
carboxylic acid 6, which can be obtained from acrolein and
cysteine-derived amine 7. The vinyl iodide moiety at C11/C12
Figure 1. Thuggacins A, B, and C.
[*] M. Bock, R. Dehn, Prof. Dr. A. Kirschning
Institut für Organische Chemie und
Biomolekulares Wirkstoffzentrum (BMWZ)
Leibniz Universität Hannover
Schneiderberg 1B, 30167 Hannover (Germany)
Fax: (+49)511-762-3011
E-mail: andreas.kirschning@oci.uni-hannover.de
[**] This research work was funded bythe Stiftung Stipendien-Fonds des
Verbandes der Chemischen Industrie (graduate scholarship for
R.D.). We are indebted to H. Irschik (HZI, Braunschweig) and R.
Jansen (HZI, Braunschweig) for authentic samples of the thugga-
cins. We thank E. Hofer for expert NMR support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200803271.
Scheme 1. Retrosynthetic analysis of thuggacin B (2). TBS=tert-butyl-
dimethylsilyl, Bz=benzoyl, Tr=triphenylmethyl.
9134
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 9134 –9137

Angewandte
Chemie
could be installed by a cross-metathesis strategy. Major
fragment 5 was planned to be prepared by the substrate-
controlled aldol reaction of aldehyde 9[7] and ketone 8, which
in turn should be generated from benzoic acid 3-oxo-propyl
ester 19.[8]
Our synthesis commenced with carboxylic acid 6, which
was prepared by a standard sequence that involved the known
Nagao aldol reaction[9] with acrolein followed by an Evans
aldol reaction[10,11] with the intermediate aldehyde
(Scheme 2). After protection of the alcohol, the auxiliary
Scheme 4. Preparation of vinyl iodide 4. a) TBTU, HOBt, DIEA, CH2Cl2,
RT, quant.; b) Hg(OAc)2, EtOH/ethyl acetate, NaBH4, RT, 82%; c) 17,
Dean–Stark conditions, benzene, 95%; d) NiO2, CH2Cl2, RT, 74%;
=
e) TBAF, THF, RT, 74%; f) (EtO)3SiCH CH2, cat. Grubbs II 18, CH2Cl2,
reflux, 61%; g) 2,2-DMP, cat. PPTS, RT, 2 h, 72%; h) MeOH, KHF2, RT,
12 h, then I2, RT, 4 h, 94%. TBTU=2-(1H-benzotriazole-1-yl)-1,1,3,3-
tetramethyluronium tetrafluoroborate, HOBt=hydroxybenzotriazole,
TBAF=tetra-n-butylammonium fluoride, 2,2-DMP=2,2-dimethoxypro-
pane, CSA=camphorsulfonic acid.
Scheme 2. Synthesis of 6. a) TiCl4, DIEA, CH2Cl2, ꢀ408C; then acrolein,
ꢀ788C, quant. d.r 5:1; b) TBSOTf, 2,6-lutidine, CH2Cl2, ꢀ788C, 93%;
c) DIBAL, toluene, ꢀ788C, 94%; d) Bu2BOTf, DIEA, 12, CH2Cl2, 08C;
then ꢀ788C, quant.; e) TBSOTf, 2,6-lutidine, CH2Cl2, ꢀ788C, 96%;
f) LiOH, H2O2, THF/H2O, 08C, 74%; g) 40% HF in H2O, MeCN, RT,
59%. DIEA=diisopropylethyl amine; Tf=trifluoromethanesulfonyl;
DIBAL=diisobutylaluminum hydride; Bn=benzyl.
Aromatization was best achieved with freshly prepared NiO2
as the oxidant.[16]
Conversion of the terminal olefinic double bond into the
E-configured vinyl iodide was achieved via vinylsilane 16,
which was prepared by cross-metathesis with vinyl triethyoxy-
silane in the presence of the Grubbs II complex 18.[17]
Removal of the two TBS protecting groups was necessary to
allow cross-metathesis to occur in satisfactory yields. Most
likely the size of the protecting groups hampered addition of
the catalyst to the olefinic double bond. Prior to formation of
the vinyl iodide moiety the diol had to be protected as
acetonide. Based on the Kumada protocol, synthesis of the
intermediate fluoro silicate was achieved with KHF2 and
subsequent addition of molecular iodine yielded vinyl iodide
4.[18]
was removed under basic conditions to liberate the carboxylic
acid 6. Since the analytical data of this compound does not
entirely correspond to reported values,[11] we unequivocally
confirmed the relative stereochemistry of acid 6 after
formation of lactone 11 and characterization by NMR
spectroscopy.
The a,b-unsaturated ethyl ester 7 was prepared from l-
cysteine via the known aldehyde 13[12] (Scheme 3). Olefina-
The synthesis of the northern fragment started with the
Wittig olefination of aldehyde 19[8] with ylide 20[19]
(Scheme 5). Treatment of the resulting a,b-unsaturated
ketone with ADmix-a[20] followed by TBSOTf/2,6-lutidine
gave ketone 8. The aldol reaction between ketone 8 and
aldehyde 9[7] served as a key reaction in the construction of
the northern fragment. Two aspects make this step a
challenge. Firstly, dienal 9 proved to be rather unreactive
towards nucleophilic attack because of the extended conju-
gation. Secondly, there is no precedence in the literature for
the stereochemical outcome of this type of aldol reaction. In
fact, no systematic studies exist on the influence of 1,4 versus
1,5 induction in aldol reactions of a,b-bis-siloxy ketones like
8. Initial experiments with the boron enolate afforded the
aldol product in only moderate yield and selectivity.[21] After
switching from boron to titanium (TiCl4)[22] and substantially
optimizing the reaction time and temperature, we could
significantly improve the outcome of the reaction (Table 1).
To our delight, the desired hydroxyketone 21 was formed in
88% yield with good selectivity for the desired diastereoiso-
mer under the optimized conditions.
Scheme 3. Preparation of amine 7. a) 14, CHCl3, reflux, 94%;
b) TMSOTf, 2,6-lutidine, CH2Cl2, 08C, 79%. Cys=cysteine, Boc=tert-
butyloxycarbonyl, TMS=trimethylsilyl.
tion with ylide 14[13] had to be optimized with respect to the
desired E selectivity and complete conversion of aldehyde 13.
We found that a high excess of ylide 14 and low concentration
of aldehyde are required to prevent decomposition of
aldehyde 13. The Boc group was then efficiently removed
using TMSOTf/2,6-lutidine.[12]
Formation of the amide bond between fragments 6 and 7
was achieved using TBTU as the coupling reagent
(Scheme 4). Removal of the trityl protecting group liberated
the thiol group which was cyclized to give the thiazoline using
Mo complex 17, which was disclosed by Sakakura et al. very
recently.[14] Other Lewis acids including the related [MoO2-
(acac)2][15] did not effect ring closure in sufficient yields.
Angew. Chem. Int. Ed. 2008, 47, 9134 –9137
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
9135

Communications
Scheme 6. Completion of the total synthesis of thuggacin B (2).
a) [PdCl2(PPh3)2], CuI, Et3N, MeCN, ꢀ208C!RT, 74%; b) TBAF, THF,
99%; c) KOH (3m), 90%; d) Lindlar catalyst, py, 56%; e) MS 4 Š,
MNBA, DMAP, toluene, 708C; 54%; f) CSA, MeOH; 38% 2, and
<7% 1. DMAP=N,N-dimethylaminopyridine), MNBA=2-methyl-6-
nitrobenzoic anhydride;
Scheme 5. Preparation of alkyne 5. a) CHCl3, 408C, 70%; b) AD-mix a,
occurred selectively with the 17-hydroxy group. This unex-
pected selectivity for lactonization in favor of the thuggacin B
derivative is most likely a result of structural and conforma-
tional changes caused by the two acetonides. It should be
noted that we had isolated the same compound as a by-
product after Mosher ester formation during our studies on
the structure elucidation of the thuggacins.[5]
water/tBuOH, 08C, 78%; >96% ee (recryst.); c) TBSOTf, 2,6-lutidine,
CH2Cl2, ꢀ788C!RT, 79%; d) aldehyde 9 (see Scheme 1), TiCl4, DIEA,
CH2Cl2, ꢀ788C!ꢀ408C, (88%; d.r.>20:1); e) Et2BOMe, NaBH4,
THF/MeOH, ꢀ308C!ꢀ158C, 89%, (syn/anti=8:1); f) 2,2-DMP, CSA,
DMF, RT, 92%; g) K2CO3, MeOH,RT, 95%; h) cat. TPAP, NMO, MS
3 Š, CH2Cl2, RT, 90%; i) Bestmann’s reagent, K2CO3, MeOH, 08C,
89%. TPAP=tetrapropylammonium perruthenate, NMO=N-methyl-
morpholine-N-oxide; Bestmann’s reagent=CH3(CO)C(N2)P(O)-
(OCH3)2.
Removal of the acetonides could be accomplished by
repeated action of CSA in methanol. After a reaction time of
30 minutes thuggacin B (2) was isolated and the residue was
again treated with CSA (three times). This procedure
provided thuggacin B (2) in 38% yield and thuggacin A (1)
in < 7% yield. Since the presence of acids is known to hamper
acyl migration of the thuggacins, thuggacin B (2) was isolated
as major product after acetal cleavage. Extending the reaction
time without removing the freshly formed thuggacin B led to
decomposition (presumably elimination of water at C7/C8 as
well as methanolysis). The spectroscopic data (1H NMR,
13C NMR, HRMS, CD) of the synthetic thuggacin B (2) are
identical to those of an authentic sample of 2. Moreover, we
observed slow rearrangement of thuggacin B into the thug-
gacins A (1) and C (3) in the NMR tube.[4,29] The intercon-
version was judged from analysis of diagnostic signals (3-H
and 5-H), thus, confirming that the three thuggacins inter-
convert by a transacylation mechanism.
In summary, we have achieved the first total synthesis of
the thuggacins A–C (1–3) by a highly stereoselective and
modular route which should be amenable to the preparation
of analogues. Thuggacin B (2) was synthesized in 23 linear
steps (longest linear sequence) from acrolein in 0.6% overall
yield. Notable features include a substrate-controlled, tita-
nium-mediated aldol reaction, a cross-metathesis approach
for converting terminal alkenes into the corresponding E-
vinyl iodides, and the cross-coupling reaction of a complex
vinyl iodide and terminal alkyne by the Sonogashira reaction.
The synthesis gives final proof of the reported structures and
paves the way for detailed studies on structure–activity
relationships.
Table 1: Conditions for the aldol reaction of 8 and 9.
Lewis acid Conditions (enolate formation/reaction) Yield (d.r.)
Bu2BOTf[a]
1 h, ꢀ788C/16 h, ꢀ788C!08C
51% (4:1)
65% (>20:1)
88% (>20:1)
[b]
TiCl4
3.5 h, ꢀ788C/4.5 h ꢀ788C!ꢀ458C
[b]
TiCl4
3.5 h, ꢀ788C/16 h, ꢀ788C!ꢀ358C
[a] 1.2 equiv Bu2BOTf, 1.3 equiv DIEA, 4 equiv 9. [b] 1.2 equiv TiCl4,
1.4 equiv DIEA, 4 equiv 9.
This complex substrate-controlled titanium aldol reaction
presumably proceeds via transition state TS-I.[22] Facial
selectivity is governed by electrostatic repulsion between
the enolate and the a-OTBS group. The correct 1,4-stereo-
induction was confirmed by analysis of the Mosher esters
prepared from hydroxyketone 21. The formation of the syn-
aldol product was confirmed after preparation of the 1,3-syn-
diol using NaBH4 in the presence of Et2BOMe[23] and
formation of the acetonide. The acetonide showed the
expected coupling constants (J) and 13C NMR chemical
shifts[24] of the acetonide carbon atoms (Scheme 5). After
saponification and oxidation alkyne 5 was generated by action
of Bestmannꢀs reagent.[25]
The complete carbon backbone of the thuggacins was
assembled by a Sonogashira cross-coupling reaction in
excellent yield; the reaction conditions described by Wipf
et al. in their synthesis of disorazole C were applied
(Scheme 6).[26] After desilylation and ester hydrolysis, the
enyne was reduced to the E,Z-diene by Lindlar hydrogena-
tion. The hydrogenation proceeded with moderate selectivity
for monoreduction of the alkyne.[27] The inseparable mixture
was subjected to the macrocyclization, which was carried out
according to the protocol of Shiina et al.[28] Ring closure
Received: July 5, 2008
Revised: August 11, 2008
Published online: October 16, 2008
9136
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 9134 –9137

Angewandte
Chemie
[18]K. Tamao, M. Akita, K. Maeda, M. Kumada, J. Org. Chem. 1987,
Keywords: aldol reaction · antibiotics · Sonogashira reaction ·
total synthesis · tuberculosis
.
52, 1100 – 1106.
[19]W. Sucrow, B. Schubert, W. Richter, M. Slopianka, Chem. Ber.
1971, 104, 3689 – 3703.
[20]a) W. Amberg, Y. L. Bennani, G. A. Crispino, J. Hartung, K. S.
Jeong, H. L. Kwong, K. Morikawa, Z. M. Wang, D. Xu, X. L.
Zhang, K. B. Sharpless, J. Org. Chem. 1992, 57, 2768 – 2771;
b) P. J. Walsh, K. B. Sharpless, Synlett 1993, 605 – 606.
[21]I. Paterson, Pure Appl. Chem. 1992, 64, 1821 – 1830.
[22]S. Figueras, R. Martín, P. Romea, F. Urpí, J. Vilarrasa,
Tetrahedron Lett. 1997, 38, 1637 – 1640.
[1]D. McKinney, Nat. Med. 2000, 6, 1330 – 1333.
[2]E. L. Corbett, C. J. Watt, N. Walker, D. Maher B. G. Williams,
M. C. Raviglione, C. Dye, Arch. Intern. Med. 2003, 163, 1009 –
1021.
[3]R. J. OꢀBrian, P. P. Nunn, Am. J. Respir. Crit. Care Med. 2001,
163, 1055 – 1058.
[4]H. Steinmetz, H. Irschik, B. Kunze, H. Reichenbach, G. Höfle, R.
Jansen, Chem. Eur. J. 2007, 13, 5822 – 5832.
ˇ
[23]K.-M. Chen, G. E. Hardtmann, K. Prasad, O. Repic , M. J.
Shapiro, Tetrahedron Lett. 1987, 28, 155 – 158.
[24]S. Rychnovsky, D. J. Skalitzky, Tetrahedron Lett. 1990, 31, 945 –
948. Diagnostic 13C NMR data (d) for 1,3-syn: Meax ꢁ 20 ppm,
Meeq ꢁ 30 ppm; for 1,3-anti: both Me ꢁ 25 ppm. Occasionally,
also the 13C d values for the quatenary C are analyzed: d <
100 ppm (for 1,3-syn) and d > 100 ppm (for 1,3-anti). Diagnostic
13C d values for 1,2-anti: Meb ꢁ 25 ppm, Meg ꢁ 28 ppm; for 1,2-
syn both Me ꢁ 27 ppm.
[5]M. Bock, R. Müller, K. Buntin, A. Kirschning, Angew. Chem.
2008, 120, 2341 – 2344; Angew. Chem. Int. Ed. 2008, 47, 2308 –
2311.
[6]H. Irschik, H. Reichenbach, G. Höfle, R. Jansen, J. Antibiot.
2007, 60, 733 – 738.
[7]J. E. Moses, S. Brückner, S. J. Eade, R. M. Adlington, J. E.
Baldwin, Org. Biomol. Chem. 2003, 1, 3670 – 3684.
[8]Shell Devel. Co. , US 2638479, 1949.
[9]a) Y. Nagao, Y. Hagiwara, T. Kumagai, M. Ochici, J. Org. Chem.
1986, 51, 2391 – 2393; b) D. Janssen, Angew. Chem. 2007, 119,
4985 – 4988; Angew. Chem. Int. Ed. 2007, 46, 4898 – 4901.
[10]D. A. Evans, T. L. Shih, J. Am. Chem. Soc. 1981, 103, 2127 – 2129.
[11]P. Curran, A. M. Brückner, C. O. Kangami, Org. Lett. 2005, 7,
379 – 382.
[12]T. Takahashi, A. Munakaba, Y. Memagiri, T. Doi, Org. Lett.
2006, 8, 531 – 534.
[13]Prepared following the procedure by H. J. Bestmann, H.
Hartung, Chem. Ber. 1966, 99, 1198 – 1207.
[14]A. Sakakura, R. Kondo, S. Umemura, K. Ishihara, Adv. Synth.
Catal. 2007, 349, 1641 – 1646.
[15]A. Sakakura, R. Kondo, K. Ishihara, Org. Lett. 2005, 7, 1971 –
1974.
[25]S. Müller, B. Leipold, G. J. Roth, H. J. Bestmann, Synlett 1996,
521 – 522.
[26]P. Wipf, T. H. Graham, J. Am. Chem. Soc. 2004, 126, 15346 –
15347.
[27]The reaction had to be terminated at 80% conversion furnishing
a mixture of starting material (20%), seco acid (56%), and
doubly hydrogenated alkyne (24%). Extended reaction times
led to full conversion of the starting alkyne into the doubly
hydrogenated product. Product ratios were determined by ESI-
MS analysis of the respective ions. NMR analysis of the mixture
revealed the diene to be 11E,13Z as expected.
[28]I. Shiina, M. Kubota, H. Oshiumi, M. Hashizume, J. Org. Chem.
2004, 69, 1822 – 1830.
[29]As acyl migration requires a protic solvent to occur, the rather
slow rearrangement in DMSO is probably a result of traces of
water in the solvent. This also explains the different ratios of the
thuggacins observed in DMSO and methanol (1/2/3 = 2.5:8.3:1)
(see the Supporting Information).
[16]a) K. Nakagawa, R. Konaka, T. Nakata, J. Org. Chem. 1962, 27,
1597 – 1601; b) D. L. Evans, D. K. Minster, U. Jordis, S. M.
Hecht, A. L. Mazzu, A. I. Meyers, J. Org. Chem. 1979, 44,
497 – 501.
[17]M. J. Comin, D. A. Parrish, J. R. Deschamps, V. E. Marquez,
Org. Lett. 2006, 8, 1417 – 1419.
Angew. Chem. Int. Ed. 2008, 47, 9134 –9137
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
9137