Direct Entry to 4,10-Didesmethyl (9S)-Dihydroerythronolide A via Catalytic Allene Osmylation

Article abstract of DOI:10.1021/acs.orglett.6b01151

Desmethyl erythronolides have emerged as macrolide targets that may prove effective against resistant bacteria. A five-step sequence to 4,10-didesmethyl (9S)-dihydroerythronolide A (1) from known cyclic bis[allene] 13 is reported. Key structural and mechanistic aspects of the synthesis are discussed along with catalytic allene osmylation. An improved route to 13 is also described.

Full text of DOI:10.1021/acs.orglett.6b01151

Letter
pubs.acs.org/OrgLett
Direct Entry to 4,10-Didesmethyl (9S)‑Dihydroerythronolide A via
Catalytic Allene Osmylation
Libing Yu,† Huan Wang,† Novruz G. Akhmedov,‡ Christopher Sowa,† Kai Liu,† Hiyun Kim,†
and Lawrence Williams*,†
†Department of Chemistry and Chemical Biology, Rutgers, The State University of New Jersey, Piscataway, New Jersey 08854,
United States
‡Department of Chemistry, West Virginia University, 406 Clark Hall, Prospect Street, Morgantown, West Virginia 26506, United
States
S
* Supporting Information
ABSTRACT: Desmethyl erythronolides have emerged as macrolide targets
that may prove effective against resistant bacteria. A five-step sequence to 4,10-
didesmethyl (9S)-dihydroerythronolide A (1) from known cyclic bis[allene] 13
is reported. Key structural and mechanistic aspects of the synthesis are
discussed along with catalytic allene osmylation. An improved route to 13 is
also described.
athogenic bacteria resistant to standard and last-resort
therapies have prompted an intense search for novel
deliberately incorporates sites of unsaturation into the scaffold
for late-stage oxidation and diversification. Although our
strategy deviates from some notions of synthetic ideality,6 it
offers a much-improved cumulative step economy. In this
report we demonstrate the superiority of this strategy with a
concise synthesis of 1, the 4,10-didesmethyl variant of (9S)-
dihydroerythronolide A: compound 1 is prepared in five steps
from 13, our advanced intermediate, and in 16 linear steps
overall; 13 is prepared by an improved and convenient route; a
new procedure for catalytic osmium tetroxide oxidation of
allenes is also described.
P
antibiotics.1 The complex macrolide erythromycin A (Figure 1)
Schemes 1 and 2 summarize the synthesis of 1. In our first-
generation synthesis5b certain intermediates proved either
reluctant to react or sensitive to storage. Accordingly,
intermediates 3−8 were modified, and the sequence was
improved for operational convenience and scale and to ensure
that key compounds could be readily stockpiled. As before, we
lactonized prior to allene elaboration, since macrolide
cyclization efficiency is notoriously sensitive to the functionality
and stereochemistry of the seco acid. We eliminated the use of
dimethoxy acetaldehyde as an aldol partner and adopted the
use of acrolien, to which the chiral anion of 2 added smoothly,
as reported.7 The resultant alcohol was readily masked as the
benzyl ether (3 → 4).8 Hydride reduction provided primary
alcohol 5, which was converted to the corresponding tosylate
and then displaced with acetylide (→ 7). Aldehyde 6 was
prepared from the antipode of 4 (ent-4) using Lemieux−
Johnson conditions.9,10 Zinc bromide mediated addition of
alkyne 7 to aldehyde 6, as before,5b afforded alkyne 8 with
concomitant lactonization.11 Oxidative cleavage of the terminal
Figure 1. Erythromycin A, (9S)-dihydroerythronolide A, 1.
and its analogues have been used widely as first-line
antibacterial agents.2 However, resistance mechanisms in
growing populations compromise their effectiveness. For
example, the ribosomal A2058G mutation appears to render
this binding site incompatible with the methyl substitution
pattern of the macrolide.3 In principle, semisynthesis, total
synthesis, and bioengineering methods would enable removal
of certain combinations of methyl groups, but the task is not
trivial. The complexity of erythromycin A limits direct access to
these and related targets. The pressing importance of this
problem prompted Andrade et al., in a tour de force, to evaluate
desmethyl combinations through total syntheses.4 These highly
refined syntheses required more than 25 linear steps each. We
have demonstrated the benefits of an alternative strategy that
uses an advanced, structurally complex macrocyclic bis[allene]
to generate diverse 14-membered polyketides. In one to three
steps we have been able to gain entry to diverse erythromycin
A-related targets from this intermediate.5 This approach
Received: April 20, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b01151
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Scheme 1. Synthesis of 13
Scheme 2. Synthesis of 1 and Cyclization of 20
Scheme 3. Proposed Catalytic Cycle
alkene revealed aldehyde 9,12 which was joined with 10 and
advanced to 13 as reported.
A five-step sequence from 13 to 1 is shown in Scheme 2a
along with catalytic osmylation studies in Scheme 3 and Table
1. Owing to the need to effect divergent stereochemical
outcomes for the hydroxyls at C5 and C11 (vide infra), we used
a two-stage oxidation procedure. Treatment of 13 with osmium
tetroxide gave hydroxy ketone 14 in excellent yield,
regioselectivity, and stereoselectivity. Reduction with sodium
borohydride afforded diol 15 (dr = 6:1). Catalytic osmylation
of 15 was also highly efficient and selective. Reduction of the
resultant ketone with zinc borohydride provided tetraol 16 (dr
= 8:1), and hydrogenolysis cleanly furnished didesmethyl target
1.
Lactonization, osmylation, and reduction of the C5 and C11
keto functionalities were examined in detail.13 The landmark
total synthesis of erythromycin14 showed that successful
macrolactonization of this target can be problematic. Cyclic
motifs spanning C3−C5 and C9−C11 as well as the S
configuration of the C9 combine to enable lactonization
(Figure 1). These design principles have been followed in
virtually all syntheses of the erythronolides.15 Although we
opted to prepare the 9S configuration of 12, we could not be
certain that this feature and the allenic moieties spanning C4−
C6 and C10−C12 would facilitate cyclization. In the event, 12
readily cyclized (Scheme 1). Still, we doubted that the
configuration at C9 was essential for the success of this
strategy and were anxious to test this hypothesis. Compound
20, the C9 epimer of 12, was prepared from 13 (Scheme 2b).
The action of 2,3-dichloro-5,6-dicyano-1,4-benzoquinone
(DDQ) on 13 selectively cleaved the C9 benzyl ether in 60%
yield.16 Mitsunobu conditions17 were successfully applied to
afford, subsequent to methanolysis, 18. After the matter of
benzyl group reinstallation, lactone 19 was opened to give 20,
the 9R variant of 12, and was then smoothly closed again (20
→ 21). Although the somewhat stringent conditions required
to open 19 also promoted epimerization at C2 (dr = 2:1),18 the
mixture was taken on and efficiently cyclized in 70% overall
yield (dr = 2:1).19 Indeed, this strategy is compatible with
either configuration at C9.
Efficient catalytic allene oxidation with osmium tetroxide
transforms each allenic moiety of substrate 13 to the
corresponding hydroxy ketone with no evidence of regioiso-
meric or stereoisomeric products (Table 1). Excess osmium
tetroxide (2.5 equiv) gave bicycle 22 (Table 1, entry 1).
Inclusion of a tertiary amine greatly accelerated the reaction
and gave 23 as the major product (Table 1, entry 2). DABCO
B
DOI: 10.1021/acs.orglett.6b01151
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Table 1. Osmylation of Cyclic Bis[allene] 13
Scheme 4. Two-Step Synthesis of Macrolides from 13
yield (%)
a
entry
OsO4 (equiv) additive (equiv) time (h) 14
22
23
1
2
3
4
5
6
2.5
2.5
0.3
0.1
0.1
0.02
none
4
0
0
0
0
0
0
46
20
20
15
15
15
0
44
60
60
77
75
osmium tetroxide that made direct access to these targets
possible.
DABCO (4)
DABCO (4)
DABCO (4)
DABCO (1)
DABCO (0.2)
20 min
40 min
2
ASSOCIATED CONTENT
■
2
S
* Supporting Information
20
a
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.6b01151.
Reaction conditions: 0.02−0.04 mmol of 13 in 1:1 t-BuOH/H2O
(0.03 M), rt; under catalytic conditions (Table 1, entries 3−6): N-
methylmorpholine-N-oxide (2 equiv) was used as terminal oxidant.
Yields were determined after FCC purification. DABCO = 1,4-
diazabicyclo[2.2.2]octane.
Procedures and additional data (PDF)
AUTHOR INFORMATION
Corresponding Author
■
outperformed pyridine and quinuclidine in terms of both yield
and rate enhancement.20 Catalysis was achieved using N-
methylmorpholine-N-oxide (NMO) as terminal oxidant (Table
1, entries 3−6). In Table 1, entry 6 demonstrates that these
conditions support low catalyst loading (2 mol %) for overnight
reactions with virtually no change in yield.
*E-mail: lwilliams@chem.rutgers.edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We gratefully acknowledge financial support from the NIH
(GM078145).
■
Allene oxidation with osmium tetroxide proceeds by way of
the osmium(VI) enolate21 (e.g., 13 → 24). Scheme 3 outlines a
mechanistic framework for the conversion of 13 to 23. The
osmium reagent approaches the most nucleophilic double bond
(C5−C6) of 13 from the most accessible face to generate the
E-enolate 24, which hydrolyzes to give 14. Similarly, 25 forms
from approach of the oxidant to the most accessible face of the
more substituted double bond (C11−C12) of the remaining
allene to give the E-enolate. In our earlier work,5b
stoichiometric osmylation of the C10−C12 allene was
confounded by an unwanted elimination/addition sequence
that gave 22 instead of the desired 23. We posit that the
persistent problem faced in previous allene osmylation
reactions was the slow hydrolysis of the osmate ester and
that DABCO, and other additives, accelerate this hydrolysis. In
the present case, the ligand releases the macrolide from the
osmium before this unwanted transformation can dominate. On
the basis of these data, we conclude that the formation of 25 is
effected by way of the osmium enolate or a close variant.
Catalytic allene oxidation enabled a two-step maneuver from
allenic substrate 13 to macrolide analogues 27 and 28 (Scheme
4). Zinc borohydride reduction of 23 yielded 5β,11β-tetraol 27,
whereas sodium borohydride reduction yielded 5α,11α-tetraol
28. These conditions furnished the products in excellent overall
yield and complement the synthesis of 5α,11β-target 1.22
We have reported a series of allene oxidation/elaboration
methods in both simplified and complex settings.23 Catalytic
methods to effect allene oxidation are underexplored and hold
considerable promise for concise, practical, and broadly useful
routes to high-value molecular architectures.24 The present
disclosure describes an improved route to 13, its subsequent
conversionin five or fewer stepsto erythronolides 1, 27,
and 28, and catalytic oxidation of the allenic moieties with
REFERENCES
■
(1) (a) Fishbach, M. A.; Walsh, C. T. Science 2009, 325, 1089.
(b) Walsh, C. T. Nat. Rev. Microbiol. 2003, 1, 65. (c) Wright, G. D.
Chem. Commun. 2011, 47, 4055. (d) Wright, G. D. Nat. Rev. Microbiol.
2007, 5, 175−186.
(2) (a) Katz, L.; Ashley, G. W. Chem. Rev. 2005, 105, 499. (b) Pal, S.
Tetrahedron 2006, 62, 3171.
(3) (a) Bulkley, D.; Innis, C. A.; Blaha, G.; Steitz, T. A. Proc. Natl.
Acad. Sci. U. S. A. 2010, 107, 17158. (b) Dunkle, J. A.; Xiong, L.;
Mankin, A. S.; Cate, J. H. D. Proc. Natl. Acad. Sci. U. S. A. 2010, 107,
17152. (c) Tu, D.; Blaha, G.; Moore, P. B.; Steitz, T. A. Cell 2005, 121,
257. (d) Hansen, J. L.; Ippolito, J. A.; Ban, N.; Nissen, P.; Moore, P.
B.; Steitz, T. A. Mol. Cell 2002, 10, 117. (e) Schlunzen, F.; Zarivach,
R.; Harms, J.; Bashan, A.; Tocilj, A.; Albrecht, R.; Yonath, A.;
Franceschl, F. Nature 2001, 413, 814.
(4) (a) Velvadapu, V.; Paul, T.; Wagh, B.; Klepacki, D.; Guvench, O.;
MacKerell, A., Jr; Andrade, R. B. ACS Med. Chem. Lett. 2011, 2, 68.
(b) Velvadapu, V.; Paul, T.; Wagh, B.; Glassford, I.; DeBrosse, C.;
Andrade, R. B. J. Org. Chem. 2011, 76, 7516. (c) Velvadapu, V.; Lee,
M.; Paul, T.; DeBrosse, C.; Klepacki, D.; Small, M. C.; MacKerell, A.
D., Jr; Andrade, R. B. ACS Med. Chem. Lett. 2012, 3, 211. (d) Wagh,
B.; Paul, T.; Glassford, I.; DeBrosse, C.; Klepacki, D.; Small, M. C.;
MacKerell, A. D., Jr; Andrade, R. B. ACS Med. Chem. Lett. 2012, 3,
1013. (e) Glassford, I.; Lee, M.; Wagh, B.; Velvadapu, V.; Paul, T.;
Sandelin, G.; Debrosse, C.; Klepacki, D.; Small, M.; Mackerell, A.;
Andrade, R. B. ACS Med. Chem. Lett. 2014, 5, 1021. (f) Andrade, R. B.
Synlett 2015, 26, 2199.
(5) (a) Ghosh, P.; Zhang, Y.; Emge, T. J.; Williams, L. J. Org. Lett.
2009, 11, 4402. (b) Liu, K.; Kim, H.; Ghosh, P.; Akhmedov, N. G.;
Williams, L. J. J. Am. Chem. Soc. 2011, 133, 14968.
(6) (a) Gaich, T.; Baran, P. S. J. Org. Chem. 2010, 75, 4657.
(b) Hendrickson, J. B. J. Am. Chem. Soc. 1971, 93, 6847.
C
DOI: 10.1021/acs.orglett.6b01151
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(7) (a) Evans, D. A.; Bartroli, J.; Shih, T. L. J. Am. Chem. Soc. 1981,
103, 2127. (b) Nicolaou, K. C.; Brenzovich, W.; Bulger, P.; Francis, T.
Org. Biomol. Chem. 2006, 4, 2119.
(8) The first-generation synthesis used dimethoxy acetaldehyde in
the aldol addition reaction and required 2 days to convert the alcohol
to the benzyl ether under more forcing conditions (3 equiv of Ag2O
and 2 equiv of BnBr).
(9) (a) Hua, Z.; Yu, W.; Jin, Z. Org. Lett. 2004, 6, 3217. (b) Because
the α-position of aldehyde 6 epimerizes during silica gel chromatog-
raphy, the crude material was used for the subsequent coupling
reaction without purification.
(10) The use of the alkene avoids the problematic dimethyl acetal of
the first-generation synthesis, which slowly hydrolyzed and decom-
posed upon storage.
(21) To our knowledge, this functionality has not been explored
outside our laboratory.
(22) The structure of new compounds were assigned unambiguously
by extensive NMR analysis; see the Supporting Information.
(23) (a) Joyasawal, S.; Lotesta, S. D.; Akhmedov, N. G.; Williams, L.
J. Org. Lett. 2010, 12, 988. (b) Ghosh, P.; Cusick, J. R.; Inghrim, J.;
Williams, L. J. Org. Lett. 2009, 11, 4672. (c) Zhang, Y.; Cusick, J. R.;
Ghosh, P.; Shangguan, N.; Katukojvala, S.; Inghrim, J.; Emge, T. J.;
Williams, L. J. J. Org. Chem. 2009, 74, 7707. (d) Wang, Z.; Shangguan,
N.; Cusick, J. R.; Williams, L. J. Synlett 2008, 2008, 213. (e) Lotesta, S.
D.; Kiren, S.; Sauers, R. R.; Williams, L. J. Angew. Chem., Int. Ed. 2007,
46, 7108. (f) Ghosh, P.; Lotesta, S. D.; Williams, L. J. J. Am. Chem. Soc.
2007, 129, 2438. (g) Lotesta, S. D.; Hou, Y.; Williams, L. J. Org. Lett.
2007, 9, 869. (h) Katukojvala, S.; Barlett, K. N.; Lotesta, S. D.;
Williams, L. J. J. Am. Chem. Soc. 2004, 126, 15348.
(11) Mead, K. T. Tetrahedron Lett. 1987, 28, 1019.
(24) Other than ref 5b, allene oxidation with osmium tetroxide has
been only tentatively examined: (a) Fleming, S. A.; Liu, R.; Redd, J. T.
Tetrahedron Lett. 2005, 46, 8095. (b) Fleming, S. A.; Carroll, S. M.;
Hirschi, J.; Liu, R.; Pace, J. L.; Redd, J. T. Tetrahedron Lett. 2004, 45,
3341. (c) Mereyala, H. B.; Gurrala, S. R.; Mohan, S. K. Tetrahedron
1999, 55, 11331. (d) Manabe, S. Chem. Commun. 1997, 737.
(12) This procedure was more expeditious than the earlier synthesis
(ref 5b), which generated the aldehyde through acetic acid/TFA/water
(4:1:1) cleavage of the corresponding acetal over the course of 14 h at
rt. Acetic acid and water were insufficient.
(13) Glycosylation, which is essential for erythronolide bioactivity,
will be discussed elsewhere.
(14) Woodward, R. B.; Logusch, E.; Nambiar, K. P.; Sakan, K.; Ward,
D. E.; Au-Yeung, B.-W.; Balaram, P.; Browne, L. J.; Card, P. J.; Chen,
C. H.; Chenevert, R. B.; Fliri, A.; Frobel, K.; Gais, H.-J.; Garratt, D. G.;
Hayakawa, K.; Heggie, W.; Hesson, D. P.; Hoppe, D.; Hoppe, I.;
Hyatt, J. A.; Ikeda, D.; Jacobi, P. A.; Kim, K. S.; Kobuke, Y.; Kojima,
K.; Krowicki, K.; Lee, V. J.; Leutert, T.; Malchenko, S.; Martens, J.;
Matthews, R. S.; Ong, B. S.; Press, J. B.; Rajan Babu, T. V.; Rousseau,
G.; Sauter, H. M.; Suzuki, M.; Tatsuta, K.; Tolbert, L. M.; Truesdale,
E. A.; Uchida, I.; Ueda, Y.; Uyehara, T.; Vasella, A. T.; Vladuchick, W.
C.; Wade, P. A.; Williams, R. M.; Wong, H. N.-C. J. Am. Chem. Soc.
1981, 103, 3210 , 3213, 3215.
́
(e) David, K.; Ariente, C.; Greiner, A.; Gore, J.; Cazes, B. Tetrahedron
Lett. 1996, 37, 3335. (f) Wolff, S.; Agosta, W. C. Can. J. Chem. 1984,
62, 2429. (g) Biollaz, M.; Haefliger, W.; Velarde, E.; Crabbe, P.; Fried,
J. H. J. Chem. Soc. D 1971, 0, 1322.
(15) For cyclization with cyclic protecting groups spanning C3−C5
and C9−C11 as well as the S configuration at C9: (a) Stork, G.;
Rychnovsky, S. D. J. Am. Chem. Soc. 1987, 109, 1565. (b) Nakata, M.;
Arai, M.; Tomooka, K.; Ohsawa, N.; Kinoshita, M. Bull. Chem. Soc. Jpn.
1989, 62, 2618. (c) Hikota, M.; Tone, H.; Horita, K.; Yonemitsu, O. J.
Org. Chem. 1990, 55, 7. (d) Myles, D. C.; Danishefsky, S. J.; Schulte,
G. J. Org. Chem. 1990, 55, 1636. (e) Sviridov, A. F.; Borodkin, V. S.;
Ermolenko, M. S.; Yashunsky, D. V.; Kochetkov, N. K. Tetrahedron
1991, 47, 2317. (f) Sturmer, R.; Ritter, K.; Hoffmann, R. W. Angew.
̈
Chem., Int. Ed. Engl. 1993, 32, 101. (g) Evans, D. A.; Kim, A. S.;
Metternich, R.; Novack, V. J. J. Am. Chem. Soc. 1998, 120, 5921.
(h) Peng, Z. H.; Woerpel, K. J. Am. Chem. Soc. 2003, 125, 6018.
(i) Stang, E. M.; White, M. C. Nat. Chem. 2009, 1, 547. (j) Chandra,
B.; Fu, D.; Nelson, S. G. Angew. Chem., Int. Ed. 2010, 49, 2591. For
cyclization with a cyclic protecting group spanning C3−C5 and the S
configuration at C9: (k) Corey, E. J.; Hopkins, P. B.; Kim, S.; Yoo, S.
E.; Nambiar, K. P.; Falck, J. R. J. Am. Chem. Soc. 1979, 101, 7131. For
cyclization without cyclic protecting group and with the S
configuration at C9: (l) Paterson, I.; Rawson, D. J. Tetrahedron Lett.
1989, 30, 7463.
(16) This interesting selectivity may reflect the increased accessibility
of the C-9 benzyl compared to the C-3 benzyl.
(17) Mitsunobu, O.; Yamada, Y. Bull. Chem. Soc. Jpn. 1967, 40, 2380.
(18) The macrolactone 19 resisted various hydrolysis conditions
(e.g., LiOOH, 5 M KOH in H2O/MeOH). Nevertheless concentrated
sodium methoxide was found to open the macrolactone. The seco acid
20 was obtained as a mixture of inseparable C2 epimers (dr = 2:1).
This finding is reminiscent of the findings of Martin et al., where the
hydrolysis of a 14-membered lactone derived from erythromycin B
proved stubbornly resistant to cleavage under a wide variety of
reaction conditions. Martin, S. F.; Yamashita, M. J. Am. Chem. Soc.
1991, 113, 5478.
(19) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M.
Bull. Chem. Soc. Jpn. 1979, 52, 1989.
(20) See Supporting Information for more details, including effects of
HOAc and buffer.
D
DOI: 10.1021/acs.orglett.6b01151
Org. Lett. XXXX, XXX, XXX−XXX