Toward the total synthesis of hygrocin B and divergolide C: Construction of the naphthoquinone-azepinone core

Article abstract of DOI:10.1021/ol5003847

A highly regioselective Diels-Alder approach toward the bioactive natural products hygrocin B and divergolide C is presented. The route uses an unusual benzoquinone-azepinone dienophile prepared in 8 steps from ethyl 8-methoxy-1-naphthoate, by a route which includes, as key steps, a Birch alkylation and a Beckmann rearrangement of a tetralone oxime, both of which are demonstrated on multigram scale. The naphthoquinone-azepinone core is suitably functionalized for addition of the ansa-chain, found in the natural products.

Full text of DOI:10.1021/ol5003847

Letter
pubs.acs.org/OrgLett
Toward the Total Synthesis of Hygrocin B and Divergolide C:
Construction of the Naphthoquinone−Azepinone Core
Christopher C. Nawrat, Russell R. A. Kitson, and Christopher J. Moody*
School of Chemistry, University of Nottingham, University Park, Nottingham NG7 2RD, U.K.
S
* Supporting Information
ABSTRACT: A highly regioselective Diels−Alder approach
toward the bioactive natural products hygrocin B and
divergolide C is presented. The route uses an unusual
benzoquinone−azepinone dienophile prepared in 8 steps
from ethyl 8-methoxy-1-naphthoate, by a route which includes,
as key steps, a Birch alkylation and a Beckmann rearrangement
of a tetralone oxime, both of which are demonstrated on
multigram scale. The naphthoquinone−azepinone core is
suitably functionalized for addition of the ansa-chain, found in the natural products.
he ansamycin antibiotics are a group of structurally
diverse, biologically active, and historically important
although hygrocin A is reported as quite unstable as a result of
T
this functionality, the degradation products obtained are those
of cyclization of the ansa chain onto the quinone carbonyl
rather than the quinone olefin.5 Even if the desired azepinone
formation could be carried out, it is unclear whether the newly
formed quaternary center would have the required config-
uration or how the stereochemical outcome of this reaction
might be controlled. It should also be noted that, in the case of
divergolide C 2, the putative precursor 5 to the azepinone-
forming cyclization has not yet been isolated, although, as for
hygrocin A, products of the ‘unwanted’ cyclization onto the
quinone carbonyls such as divergolide D 6 do occur in Nature.
This suggests that the mode of cyclization required to form the
azepinone found in divergolide C may not be particularly
favorable and designing a synthesis that relies on carrying out
this step at a late stage may be unwise.
In view of these considerations, we decided to eschew the
‘macrocycle-first’ approach recently reported by Rasapalli9−11
and Trauner12 and, instead, consider two potential synthetic
routes: the first employing an ambitious intramolecular Diels−
Alder approach from the pendant diene-containing benzoqui-
none-azepinone 7, and the second aiming to prepare the
tricyclic naphthoquinone−azepinone core 8, common to both
divergolide C 2 and hygrocin B 3, to which suitable ansa chains
could later be attached, either by modifying a pendant ester,
nitrile, or allyl group or by introducing a halogen or a formyl
group ortho to the naphthol (Scheme 1) (cf. Rasapalli).9−11 By
focusing initially on the second approach, it was anticipated that
the core 8, bearing a point of attachment for the ansa chain,
could be prepared by a Diels−Alder reaction of a suitable diene
such as 10−13 with a benzoquinone−azepinone 9. This
dienophile could itself be prepared from a suitable tetralone
using classic aromatic chemistry. This strategy will allow the
natural products that have been known for over 50 years.1 This
class of compounds has provided humanity with a number of
important drugs that are used worldwide; for example rifabutin
1 and other semisynthetic derivatives of the rifamycins are
current frontline treatments for tuberculosis (Figure 1).2,3
Hence, the isolation of new members of this class of
compounds continues to draw interest from the chemical
community. Recently, the ansamycin divergolide C 2 was
isolated along with a number of its congeners from the
bacterium Streptomyces sp. HKI0576,4 exhibiting a highly
unusual naphthoquinone−azepinone core as well as moderate
antibacterial activity. The only previously reported example of
this ring system is found in the closely related naphthoquinone
ansamycin hygrocin B 3, isolated in 2005 along with hygrocin A
4 from Streptomyces hygroscopicus,5 an organism that has proved
to be a goldmine of powerful bioactive compounds such as
milbemycin, geldanamycin, and rapamycin. Indeed, the
hygrocins were first isolated by chemists at Wyeth as impurities
in a batch of rapamycin. However, despite this auspicious
provenance, they display none of the characteristic biological
activity of their more famous coisolates, exhibiting only very
weak antibiotic activity, a fact attributed to their remarkably
short 13 carbon aliphatic chains, which place them among the
smallest known ansamycins. Additionally, the hygrocins C−G
were recently isolated from the bacterium gdmAI-disrupted
Streptomyces sp. LZ35.6
Considering the structures of these compounds, hygrocin A 4
appears to be an immediate biosynthetic precursor to hygrocin
B 3,7 requiring just enolization and cyclization of the
macrocyclic ester. Indeed, the biosynthesis of the hygrocins
has been the subject of a recent article by Shen and co-
workers.8 However, despite this relationship, the synthesis of 3
via 4 is a daunting prospect, due to the challenge of forming the
macrocycle of A, which contains a trisubstituted Z-olefin
adjacent to a highly active methylene group. Furthermore,
Received: February 5, 2014
Published: March 25, 2014
© 2014 American Chemical Society
1896
dx.doi.org/10.1021/ol5003847 | Org. Lett. 2014, 16, 1896−1899

Organic Letters
Letter
Scheme 1. Initial Retrosynthetic Analysis
found to be inert toward standard Beckmann conditions, so was
activated by tosylation, and a range of conditions were screened
for the key rearrangement. Unfortunately, no desired product
benzazepinone was obtained under Brønsted (PPA, H2SO4,
HCl, TFA) or Lewis (AlCl3,18 CuSO4, BF3·OEt219) acid
activation of 18, even at high temperatures. In view of this
failure, and the very limited literature precedent for Beckmann
rearrangement of α,β-unsaturated oximes, we decided to
prepare the corresponding saturated tosyl oxime 22 in order
to compare its reactivity to that of 18. As removal of the double
bond at the oxime stage was anticipated to be problematic,
hydrogenation of enone 16 was investigated instead. However,
to our surprise, reduction of enone 16 using palladium-on-
carbon under an atmosphere of hydrogen at room temperature
did not give tetralone 19 but instead yielded tetralin 20. This
unexpected reaction appeared to be quite rapid, and complete
conversion into the tetralin occurred in just 2 h. Further
investigation revealed that the desired product could not be
obtained by simply shortening the reaction time, as this only
produced mixtures of both products. Fortunately it was found
that hydrogenation over a platinum catalyst, in the form of
platinum dioxide or platinum-on-carbon, prevented over-
reduction and gave the desired tetralone in quantitative yield.
Oxime formation and tosylation gave Beckmann precursors 21
and 22, whose rearrangement was now investigated. Although
oxime 21 and tosyl oxime 22 proved quite unreactive toward
Brønsted acid activation, it was found that rearrangement to the
desired lactam 23 could be effected in good yield upon
treatment of 22 with 4 equiv of aluminum chloride at 0 °C,18
even on gram scale.
Figure 1. Selected naphthoquinone ansamycins, including hygrocins
and divergolides.
challenging quaternary center present in these natural products
to be set early on in the synthesis, and the divergent nature of
the route should allow access to both natural products from a
common late-stage intermediate.
The synthesis of dienophile 9 began from ethyl 8-hydroxy-1-
naphthoate 14 (Scheme 2),13 which was readily prepared from
commercial naphthalic anhydride or naphtholactam using
known chemistry (see Supporting Information for details).14,15
This was protected as its methyl ether and subjected to Birch
reductive alkylation by treatment with lithium in liquid
ammonia−THF−t-BuOH, followed by quenching with iodo-
methane to give ester 15. This reaction proceeded in excellent
yield, even on decagram scale and was found to be tolerant of
various O-protecting groups including TES and TIPS ethers as
well as several different esters (methyl, ethyl, and tert-butyl).
Furthermore, as asymmetric Birch alkylations are well
precedented,16 this step could be adapted to provide a single
enantiomer. Allylic oxidation was carried out using Corey’s
procedure,17 which proved superior to chromium based
methods on a large scale, even with very low catalyst loadings
(1 mol % Pd), and the resulting enone 16 was converted into
the corresponding oxime 17. Surprisingly, this second step
required quite forcing conditions involving the use of a large
excess of hydroxylamine hydrochloride in anhydrous pyridine at
80 °C in order to achieve good conversion. Oxime 17 was
Initially we had planned to demethylate benzazepinone 23
prior to oxidation to the quinone, as oxidation of phenols is far
easier than that of anisoles and can be achieved under mild and
́
specific conditions using salcomine or Fremy’s salt. However,
1897
dx.doi.org/10.1021/ol5003847 | Org. Lett. 2014, 16, 1896−1899

Organic Letters
Letter
Scheme 2. Synthesis of a Suitable Benzazepinone Precursor
this deprotection proved challenging; although demethylation
occurred rapidly, the product tended to undergo lactonization
to 24 under the reaction conditions (Scheme 3). Attempted
ammonium nitrate in aqueous acetonitrile, which delivered the
desired quinone 26 in moderate yield. Unfortunately, the yields
decreased on scale-up beyond 100 mg, although owing to the
efficiency of the rest of the synthetic route enough material
could be obtained to allow examination of the pivotal Diels−
Alder reaction.
At this stage oxidation of benzazepinone 23 to reintroduce
the α,β-unsaturation was investigated. Although such a
transformation is commonplace in the γ-lactam series,25 it is
almost unknown in benzazepinones and was crucial to the
viability of the proposed route.26 Fortunately, after screening a
number of oxidants known to effect this transformation, it was
found that simply heating 23 with 2 equiv of benzeneselenic
anhydride in chlorobenzene gave a reasonable yield of the
desired compound in a single step (Scheme 3).25 Benzaze-
pinone 27 could also be oxidized successfully to the
corresponding benzoquinone−azepinone 28.
Scheme 3. Oxidation of Lactam 23 to the Quinone 26 and
Reintroduction of the α,β-Unsaturation with Subsequent
Oxidation to the Benzoquinone−Azepinone 28
With dienophiles 26 and 28 in hand, their reactivity in the
key Diels−Alder step was now investigated (Scheme 4).
Disappointingly, 26 did not undergo a reaction with the ester
containing diene 1024 and did not appear to tolerate the
relatively high reaction temperatures (130−150 °C) typically
Scheme 4. Diels−Alder Reactions of Quinone 26; Synthesis
of the Naphthoquinone−Azepinone Core
opening of lactone 24 to obtain the desired phenol 25 was not
successful, generally resulting in decomposition and mixtures of
products. Nucleophilic removal of the methyl group was
unsuccessful, and thus we were forced investigate the possibility
of oxidizing 23 directly to the quinone. The use of Dess-Martin
periodinane for this transformation as recently described by
Rasapalli was unsuccessful,10 as was the use of IBX as described
by Nicolaou and co-workers for the oxidation of a range of
acetanilides.20−23 Our previously reported conditions employ-
ing bis(trifluoroacetoxy)iodobenzene and di(acetoxy)iodoben-
zene did result in formation of some product (15−20%),24 but
the best conditions involved the use of excess cerium(IV)
1898
dx.doi.org/10.1021/ol5003847 | Org. Lett. 2014, 16, 1896−1899

Organic Letters
Letter
(2) Brogden, R. N.; Fitton, A. Drugs 1994, 47, 983−1009.
required for this partner. Next, we prepared and tested
Trauner’s nitrile-containing diene 13,27 which has been
shown to be somewhat more reactive than 10 and quite
amenable to Lewis acid catalyzed Diels−Alder reactions, but
this was also unproductive. Additionally, studies on fully
unsaturated benzoquinone−azepinone 28 were unsuccessful.
Although an initial reaction occurred with diene 12, no desired
product was observed. Fortunately, quinone 26 did undergo a
Diels−Alder reaction with diene 12 to give the naphthoqui-
none−azepinone 29 in reasonable yield, with the allyl group
serving as a latent aldehyde, as has already been demonstrated
in the context of aminonaphthoquinone synthesis.24 Alter-
natively, allylic/benzylic oxidation would give an enone suitable
for attaching the ansa-chain by the organocatalytic asymmetric
conjugate addition methodology developed by Gellman and co-
workers,28 allowing access to both configurations of the
stereocenter at position 14, as yet undetermined in the
hygrocins A and B (Figure 1). The simpler diene 1129 was
also found to be a suitable reaction partner, allowing tricyclic
quinone 30 to be prepared in moderate yield. This was slightly
improved by employing a telescoped procedure, avoiding
purification of the somewhat sensitive quinone 26, giving the
Diels−Alder adduct 30 in 33% yield over 2 steps. In both
Diels−Alder reactions, there was no sign of the alternative
regioisomer, and the regiochemistry of this product was
confirmed by HMBC NMR spectroscopy experiments and is
in accordance with those previously observed for the reactions
of aminonaphthoquinones.22,24,30 Further functionalization of
30 was also shown to be possible: bromination of the aromatic
ring9−11 gave the bromoaminonaphthoquinone−azepinone 31
in excellent yield. This allows for lithium−halogen exchange
and trapping of the aryllithium with a suitable electrophile.
In conclusion, a Diels−Alder based strategy to reach the
naphthoquinone−azepinone core of divergolide C and
hygrocin B has been realized. Central to this was the
development of a highly scalable route to key benzazepinone
building block 23, which should be easily adapted to allow an
asymmetric synthesis of these natural products through the use
of a chiral auxiliary in the Birch reductive alkylation step.
Further development of this transformation, optimization of the
Diels−Alder chemistry, and investigations into the potential of
an intramolecular Diels−Alder approach are currently in
progress.
(3) O’Brien, R. J.; Lyle, M. A.; Snider, D. E. Rev. Infect. Dis. 1987, 9,
519−530.
(4) Ding, L.; Maier, A.; Fiebig, H.-H.; Gorls, H.; Lin, W.-H.; Peschel,
̈
G.; Hertweck, C. Angew. Chem., Int. Ed. 2011, 50, 1630−1634.
(5) Cai, P.; Kong, F.; Ruppen, M. E.; Glasier, G.; Carter, G. T. J. Nat.
Prod. 2005, 68, 1736−1742.
(6) Lu, C.; Li, Y.; Deng, J.; Li, S.; Shen, Y.; Wang, H.; Shen, Y. J. Nat.
Prod. 2014, 76, 2175−2179.
(7) Kang, Q.; Shen, Y.; Bai, L. Nat. Prod. Rep. 2012, 29, 243−263.
(8) Li, S.; Wang, H.; Li, Y.; Deng, J.; Lu, C.; Shen, Y.; Shen, Y.
ChemBioChem 2014, 15, 94−102.
(9) Rasapalli, S.; Jarugumilli, G.; Yarrapothu, G. R.; Golen, J. A.;
Rheingold, A. L. Org. Lett. 2013, 15, 1736−1739.
(10) Rasapalli, S.; Jarugumilli, G.; Yarrapothu, G. R.; Golen, J. A.;
Rheingold, A. L. Tetrahedron Lett. 2013, 54, 2615−2618.
(11) Rasapalli, S.; Jarugumilli, G.; Yarrapothu, G. R.; Ijaz, H.; Golen,
J. A.; Williard, P. G. Tetrahedron Lett. 2014, 55, 821−825.
(12) Hager, A.; Kuttruff, C. A.; Hager, D.; Terwilliger, D. W.;
Trauner, D. Synlett 2013, 24, 1915−1920.
(13) Yang, Y.; Lin, Y.; Rao, Y. Org. Lett. 2012, 14, 2874−2877.
(14) Cammidge, A. N.; Oeztuerk, O. J. Org. Chem. 2002, 67, 7457−
7464.
(15) Packer, R. J.; Smith, D. C. C. J. Chem. Soc. C 1967, 2194−2201.
(16) Schultz, A. G. Chem. Commun. 1999, 1263−1271.
(17) Yu, J.-Q.; Corey, E. J. Org. Lett. 2002, 4, 2727−2730.
(18) Davies, R. J.; Xu, J. Vertex Pharmaceuticals Incorporated, USA.
WO2008021545A2, 2008.
(19) Anilkumar, R.; Chandrasekhar, S. Tetrahedron Lett. 2000, 41,
5427−5429.
(20) Nicolaou, K. C.; Baran, P. S.; Zhong, Y. L.; Barluenga, S.; Hunt,
K. W.; Kranich, R.; Vega, J. A. J. Am. Chem. Soc. 2002, 124, 2233−
2244.
(21) Nicolaou, K. C.; Baran, P. S.; Zhong, Y. L.; Sugita, K. J. Am.
Chem. Soc. 2002, 124, 2212−2220.
(22) Nicolaou, K. C.; Sugita, K.; Baran, P. S.; Zhong, Y.-L. Angew.
Chem., Int. Ed. 2001, 40, 207−210.
(23) Nicolaou, K. C.; Sugita, K.; Baran, P. S.; Zhong, Y. L. J. Am.
Chem. Soc. 2002, 124, 2221−2232.
(24) Nawrat, C. C.; Lewis, W.; Moody, C. J. J. Org. Chem. 2011, 76,
7872−7881.
(25) Rasmusson, G. H.; Reynolds, G. F.; Steinberg, N. G.; Walton,
E.; Patel, G. F.; Liang, T.; Cascieri, M. A.; Cheung, A. H.; Brooks, J. R.;
Berman, C. J. Med. Chem. 1986, 29, 2298−2315.
(26) Hino, K.; Nagai, Y.; Uno, H. Chem. Pharm. Bull. 1988, 36,
2386−2400.
(27) Kuttruff, C. A.; Geiger, S.; Cakmak, M.; Mayer, P.; Trauner, D.
Org. Lett. 2012, 14, 1070−1073.
(28) Peelen, T. J.; Chi, Y. G.; Gellman, S. H. J. Am. Chem. Soc. 2005,
127, 11598−11599.
(29) Anada, M.; Washio, T.; Shimada, N.; Kitagaki, S.; Nakajima, M.;
Shiro, M.; Hashimoto, S. Angew. Chem., Int. Ed. 2004, 43, 2665−2668.
(30) Nawrat, C. C.; Palmer, L. I.; Blake, A. J.; Moody, C. J. J. Org.
Chem. 2013, 78, 5587−5603.
ASSOCIATED CONTENT
■
S
* Supporting Information
Full experimental details and copies of H and 13C NMR
1
spectra for all compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: c.j.moody@nottingham.ac.uk.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We thank the EPSRC for funding.
■
REFERENCES
■
(1) Bryskier, A. Antimicrobial Agents: Antibacterials and Antifungals;
ASM Press: Washington DC, 2005.
1899
dx.doi.org/10.1021/ol5003847 | Org. Lett. 2014, 16, 1896−1899