Concise synthesis and revision of the proposed biogenesis of helicascolides

Article abstract of DOI:10.1016/j.tetlet.2017.10.021

A concise synthesis of helicascolides A, B and C was achieved in three to five steps from commercially available materials. The key transformations of the synthesis include an Evans-Metternich anti-aldol reaction of the known β-keto imide 10 and strategic

Full text of DOI:10.1016/j.tetlet.2017.10.021

Tetrahedron Letters 58 (2017) 4459–4464
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Concise synthesis and revision of the proposed biogenesis
of helicascolides
Kuan Zheng a,b, Changmin Xie a,b, Ran Hong a,
⇑
a CAS Key Laboratory of Synthetic Chemistry of Natural Substances, Shanghai Institute of Organic Chemistry (CAS), Shanghai 200032, China
b University of Chinese Academy of Sciences, 19A Yuquan Road, Beijing 100049, China
a r t i c l e i n f o
a b s t r a c t
Article history:
A concise synthesis of helicascolides A, B and C was achieved in three to five steps from commercially
available materials. The key transformations of the synthesis include an Evans-Metternich anti-aldol
reaction of the known b-keto imide 10 and strategic base-mediated one-pot cyclization/alkylation.
Based on the new chemical evidence of ring-opening reaction of b-keto ester under biocompatible basic
conditions, Krohn’s proposal for the biosynthetic relationship between helicascolide A (1) and a naturally
co-existing acyclic dienone 4 was suggested in a reverse manner.
Received 19 September 2017
Accepted 9 October 2017
Available online 13 October 2017
Keywords:
Helicascolides A–C
d-Lactones
Ó 2017 Elsevier Ltd. All rights reserved.
Biogenesis
Biomimetic synthesis
Boronate
Marine microorganisms have been intensely exploited for their
novel bioactive secondary metabolites from diverse biosynthetic
origins.1 Helicascolides A (1) and B (2) (Fig. 1) were first isolated
by Gloer and co-workers from the marine fungus Helicascus kana-
loanus (ATCC 18591) in 1989.2 Recently, Tarman and co-workers
reported the characterization of structurally related helicascolide
C (3) from an Indonesian marine algicolous strain Gracilaria sp.
SGR-1.3 Interestingly, only helicascolide C (3) shows a notable abil-
ity to inhibit the growth of Cladosporium cucumerinum, a patho-
genic plant fungus that affects cucumbers. In 2006, Krohn and
co-workers4 isolated the biologically active linear metabolite
(4E,6E)-2,4,6-trimethylocta-4,6-dien-3-one (4) along with helicas-
colide A (1) from the endophytic fungus Nodulisporium sp. From
a structural perspective, helicascolides (1–3) belong to a small
family of partially dehydrated tetraketides that are characterized
by a highly congested d-lactone ring system with gem-dimethyl
substitution at the C2 position.
With the ongoing interest in the development of bioinspired
synthetic strategies for the efficient syntheses of bioactive polyke-
tides,5 the molecular architecture of helicascolides captured our
interest, since the specific scaffold is found in various natural prod-
ucts or natural product-like synthetic intermediates (5–7 in
Fig. 1).6,7 The first total synthesis of 1 and 3, disclosed by Krishna
and co-workers,8 featured a one-pot acid-catalyzed acetonide
removal and subsequent lactonization. The Yadav group9 achieved
a concise total synthesis of 1–3 by employing a similar ring-form-
ing strategy and utilizing iterative aldolization to access the pre-
cursor in a more efficient manner.10 In 2016, Breit and co-
workers11 also accomplished the total syntheses of 1–3 by
intramolecular Rh-catalyzed diastereoselective additions of chiral
x
-allenyl carboxylic acids. Herein, we present a short synthesis
of all the representatives of the helicascolide family, which were
assembled with the longest linear sequences ranging from three
to five steps.
Retrosynthetically, we envisioned that structurally simpler heli-
cascolide C (3), for which the most oxidized carbon is C3, could
serve as the primary target and a branching intermediate to access
the other congeners of the family via late-stage stereoselective
reduction of the C3-ketone (Scheme 1). By taking advantage of this
distinctly different tactic, methylation at C2 of cyclic b-keto ester
enolate 8 should be feasible. This synthetic strategy is largely sim-
ulating a possible biosynthetic pathway that involves chain elon-
gation and methylation initiated by SAM (S-adenosyl
methionine). Anionic intermediate 8 should be accessible from
the cyclization of aldolate anion 9, in which the entire tetraketide
backbone could be derived from the reaction of known Evans’
dipropionate synthon 1012 with tiglic aldehyde. Integral to this
approach was the dramatic increase in the kinetic acidity of the
C2 position of 8, which was a result of the enolization favoring a
cyclized product conformation in which better orbital overlap
between
r .
-C2-H and p⁄-C3-O could be realized13
Chiral b-keto imide 10 was readily prepared in a two-step pro-
tocol (Scheme 2) that included an Evans syn-aldol reaction
⇑
Corresponding author.
E-mail address: rhong@sioc.ac.cn (R. Hong).
https://doi.org/10.1016/j.tetlet.2017.10.021
0040-4039/Ó 2017 Elsevier Ltd. All rights reserved.

4460
K. Zheng et al. / Tetrahedron Letters 58 (2017) 4459–4464
Fig. 1. Helicascolides and structurally related compounds.
Scheme 1. Retrosynthetic analysis of the helicascolides. [H] = reduction; R = substituent; LG = leaving group; LLS = the longest linear steps; PG = protecting group.
Scheme 2. Synthesis of compound 13.
between commercially available N-acyloxazolidinone 11 and pro-
panal followed by Dess-Martin oxidation of corresponding alcohol
12. Addition of the (E)-boron enolate pre-formed from 10 to tiglic
aldehyde provided two anti-aldol adducts in 90% combined yield
with a diastereomeric ratio (d.r.) of 13:1.14 A second purification
using flash column chromatography further enriched the d.r. up
to 30:1.
With several grams of the requisite acyclic tetraketide 13 in
hand, we were in a position to test the feasibility of the crucial
cyclization/alkylation cascade (conversion of 9 to 3 shown in
Scheme 1). Initially, the cyclization reaction was carried out under
conventional conditions15 for analogous substrates bearing a C3-
hydroxyl group; however, extensive decomposition was observed
(entries 1 and 2 in Table 1). Next, we found that treatment of 13

K. Zheng et al. / Tetrahedron Letters 58 (2017) 4459–4464
4461
Table 1
Optimization for ‘‘one-pot” cyclization/alkylation of 13.a
Entry
Conditions
Temperature
Reaction time
Yield (%)b
3
14
15
1c
2c
3c
4
5
6
7
8
9
10
11
LiOOH (2.0 equiv), THF/H2O
DBU (1.0 equiv), DCM
tBuOK, toluene
0 °C
0 °C
rt
2 h
2 h
–
–
–
0
0
0
90
88
90
92
67
tracec
tracec
80
78
72
–
–
–
0
0
0
4
4
2
0
0
1 min
10 min
10 min
10 min
10 min
10 min
10 min
10 min
10 min
tBuOK, then MeI, toluene
tBuOK, then MeI, toluene
tBuOK, then MeI, benzene
tBuOK, then MeI, DMSO
tBuOK, then MeI, DMF
tBuOK, then MeI, DMF
tBuOK, then MeI, THF
MeI, then tBuOK, THF
rt
55 °C
55 °C
55 °C
55 °C
rt
75
traced
traced
traced
traced
11
rt
rt
[Note]: (a) t-BuOK (1.05 equiv) and MeI (2.0 equiv) were utilized unless otherwise noted. (b) Isolated yield. (c) For entries 1–3, the reaction mixture was analyzed by 1H NMR
(400 MHz, CDCl3) after a simple work-up. (d) Less than 2% yield.
with 1.05 equiv of t-BuOK in toluene at room temperature resulted
in rapid formation of b-keto lactone 14, which was isolated as an
unstable colorless oil (entry 3). To our surprise, introduction of
the methyl group by MeI under the same conditions did not pro-
vide the desired alkylation product helicascolide C (3) in an appre-
ciable yield, even at elevated temperature (entries 4 and 5).
Interestingly, when more polar aprotic solvents, such as DMF and
DMSO, were used, C2-alkylation exclusively occurred, and UV-
inactive material 3 was generated in excellent yield (entries 7
and 8). Further optimizations proved that the above-described
one-pot protocol proceeded reliably and reproducibly in 90% yield
on a multi-gram scale (entry 10).16 The NMR data as well as speci-
fic rotation of synthetic helicascolide C (3) were in good agreement
with those reported in the literature.3 When MeI was introduced
prior to t-BuOK, the desired natural product can also be obtained,
albeit in a lower yield (entry 11). To further expedite the synthetic
route, we telescoped the anti-aldol reaction and the above-
described cyclization/methylation (Scheme 3). The resulting aldol
reaction was quenched with a solution of methanol in THF. The
solution was then titrated with t-BuOK to smoothly provide a solu-
tion of the enolate form of lactone 14, which was then subjected to
methylation to give 3 by sequential addition of DMF and MeI. This
unoptimized protocol directly delivered helicascolide C (3) in one
flask and in a synthetically acceptable yield from known building
block 10, provided an expedited protocol for purification, and fur-
ther reduced the overall step count.
We then focused our attention on the late-stage reduction of
the C3-keto group (Scheme 4). Unfortunately, initial screening of
various reducing agents failed to directly convert 3 to 1 or 2 with-
out reducing the ester functionality. In some cases, such as treat-
ment with L-selectride at low temperature, an epimeric mixture
of b-keto- d-lactols 16 was isolated in 88% yield. Nevertheless, glo-
bal reduction of 3 with 5 equiv of NaBH4 smoothly furnished 17 in
nearly quantitative yield albeit as an inseparable mixture. Careful
analysis of the NMR spectra of the mixture confirmed the com-
pounds are epimeric at the anomeric carbon (C1), and the configu-
ration of C3 was determined to be (R) from the diagnostic NOE
correlations of each anomer. These structures were confirmed by
successful elaboration of 17 to helicascolide B (2) via MnO2
oxidation.
To override the intrinsic 1,3-diaxial strain of axial hydride deliv-
ery, a hindered reagent like LiBHEt3 was presumed to deliver an
axial alcohol through equatorial attack of the ketone.17 At this
point, the crude 1H NMR spectrum of the organic layer taken
immediately after work-up indicated that a new pair of anomeric
(3S)-lactols 18 were formed as the major products in a moderate
ratio of 2.7:1. Surprisingly, during the conventional work-up pro-
cedure, less polar component 19 was isolated, and after extensive
Scheme 3. Expedite ‘‘one-pot” aldolization/cyclization/alkylation of the known b-keto imide 10.

4462
K. Zheng et al. / Tetrahedron Letters 58 (2017) 4459–4464
Scheme 4. Divergent synthesis of helicascolide A (1) and B (2) via late-stage reduction.
Scheme 5. Biosynthetic relationships between helicascolide C (3) and 4.

K. Zheng et al. / Tetrahedron Letters 58 (2017) 4459–4464
4463
NMR analysis, 19 was determined to be a cyclic ethylboronate
derivative of 18.18,19 Boronate 19 could be isolated in 51% yield
from the chromatographically inseparable lactols 17 and 18, thus
providing an opportunity for the establishment of the correct
stereochemistry at C3 for helicascolide A (1). As anticipated, heat-
ing 19 on neat silica gel yielded (3S)-lactol 18 in 86% yield as a mix-
ture of anomeric isomers, which were further subjected to MnO2
oxidation to give helicascolide A (1) in 85% yield. Gratifyingly,
direct oxidation of cyclic ethylboronate 19 under the same reaction
conditions also provided 1 in excellent yield (89%). The NMR spec-
tra and optical rotation of the synthetic samples and natural heli-
cascolides A (1) and B (2) were identical.2
Research Program (XDB20000000), and the Key Research Program
of Frontier Sciences (QYZDY-SSW-SLH026) of the Chinese Academy
of Sciences are highly appreciated.
A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.tetlet.2017.10.021.
References
1. (a) Selected reviews: Gerwick WH, Fenner AM. Microb Ecol. 2013;65:800–806;
(b) Kobayashi J. Chem Pharm Bull. 2016;64:1079–1083;
It is also noteworthy that, during the optimization of the
cyclization/alkylation cascade, methyl dienoate 15 was isolated
as a side product, albeit in very low yield (<5%). We were aware
of the structural similarity between this compound and an acid
intermediate in Krohn’s biogenesis proposal (Scheme 5 A).4 Corre-
sponding acid 20, arising from enzymatic carboxylation of natu-
rally occurring 4, was suggested to be biosynthetically involved
in the construction of the lactone ring of helicascolides through
an oxa-Michael addition. We considered that the reverse process
might be thermodynamically more favorable. To test our hypothe-
sis, helicascolide C (3) was treated with biocompatible conditions
like K2CO3 in methanol at ambient temperature. Interestingly,
the ring-opening process was rapid and completed within minutes.
The ring opening resulted in polar compound 20, which, based on
extensive NMR analysis, had a dienoic acid structure (Scheme 5
C).20,21 This observation also mechanistically accounts for the for-
mation of side product 15 via the O-methylation of an intermediate
carboxylate. However, with b-keto acid 20 in hand, all attempts to
realize the putative intramolecular oxa-Michael-type addition4
only resulted in decarboxylation to form a less polar, UV-active
material. The spectroscopic data were consistent with those
reported for isopropyl ketone 4.4 The conditions tested for the
oxa-Michael-type addition were as follows: (1) heating a salt-free
solution of 20 in methanol for 10 min, which led to the isolation of
4; (2) storage of a solution of 20 in d4-methanol at room temper-
ature for 3 days, which resulted in 50% conversion to the C2-
deuterated analogue 4a, as indicated by NMR analysis (Fig. S1);22
(3) treatment of a solution of 20 in THF with TBAF at room temper-
ature, which resulted in the rapid generation of isopropyl ketone 4,
and intermediate b-keto acid derivatives were not detected by
either TLC or NMR analyses.23 The facile one-pot elimination and
subsequent decarboxylation of 3 under mild conditions suggested
a more plausible biogenesis of 4.
(c) El-Hossary EM, Cheng C, Hamed MM, et al. Eur
2017;126:631–651;
(d) Lane AL, Moore BS. Nat Prod Rep. 2011;28:411–428.
2. Poch GK, Gloer JB. J Nat Prod. 1989;52:257–260.
3. Tarman K, Palm GJ, Porzel A, et al. Phytochem Lett. 2012;5:83–86.
4. Dai JQ, Krohn K, Flörke U, et al. Eur J Org Chem. 2006;3498–3506.
5. (a) Yang L, He G, Yin R, Zhu L, Wang X, Hong R. Angew Chem Int Ed.
2014;53:11600–11604;
(b) Yang L, Lin Z, Huang S-H, Hong R. Angew Chem Int Ed. 2016;55:6280–6284;
(c) He G, Wang Y, Lai C, Li W, Hong R. Sci China Chem. 2016;59:1197–1204;
(d) Zheng K, Shen D, Hong R. J Am Chem Soc. 2017;139:12939–12942. https://
doi.org/10.1021/jacs.7b08500.
6. (a) Kormann EC, Amaral Pde A, David M, Eifler-Lima VL, Cechinel Filho V,
Campos Buzzi F. Pharmacol Rep. 2012;64:1419–1426;
(b) Jayasuriya H, Guan Z, Dombrowski AW, et al.
2007;70:1364–1367.
7. Chen Y, Li Y. United States patent US8030503 B2.
8. Krishna PR, Mallula VS, Arun Kumar PV. Tetrahedron Lett. 2012;53:4997–4999.
9. Yadav JS, Reddy AB, Shankar KS. Synthesis. 2013;45:1034–1038.
10. (a) For a general discussion on iterative approach in polyketide synthesis, see:
Zheng K, Xie C, Hong R. Front Chem. 2015;3:e32;
(b) Kumar P, Dwivedi N. Acc Chem Res. 2013;46:289–299;
(c) Horst BT, Feringa BL, Minnaard AJ. Chem Commun. 2010;46:2535–2547;
(d) Hanessian S, Giroux S, Mascitti V. Synthesis. 2006;38:1057–1076.
11. Haydl AM, Berthold D, Spreider PA, Breit B. Angew Chem Int Ed.
2016;55:5765–5769.
J Med Chem.
J
Nat Prod.
12. Evans DA, Clark JS, Metternich R, Novack VJ, Sheppard GS. J Am Chem Soc.
1990;112:866–868.
13. (a) For selected examples of configurationally stable chiral
a-substituted-1,3-
dicarbonyl compounds: Evans DA, Ennis MD, Le T, Mandel N, Mandel G. J Am
Chem Soc. 1984;106:1154–1156;
(b) Evans DA, Sjogren EB. Tetrahedron Lett. 1986;27:4961–4964;
(c) Soucy F, Grenier L, Behnke ML, et al. J Am Chem Soc. 1999;121:9967–9976;
(d) Calter MA, Liao W. J Am Chem Soc. 2002;124:13127–13129;
(e) Jeffery DW, Perkins MV, White JM. Org Lett. 2005;7:407–409;
(f) Lister T, Perkins MV. Org Lett. 2006;8:1827–1830.
14. Evans DA, Ng HP, Clark JS, Rieger DL. Tetrahedron. 1992;48:2127–2142.
15. (a) Castonguay R, He W, Chen AY, Khosla C, Cane DE.
2007;129:13758–13769;
J Am Chem Soc.
(b) Magauer T, Martin HJ, Mulzer J. Angew Chem Int Ed. 2009;48:6032–6036;
(c) Evans DA, Fitch DM, Smith TE, Cee VJ.
2000;122:10033–10046;
J
Am Chem Soc.
In summary, we have developed
a cyclization/alkylation
(d) Yuan Y, Men H, Lee C. J Am Chem Soc. 2004;126:14720–14721;
(e) Smith AB, Brandt BM. Org Lett. 2001;3:1685–1688;
(f) Beck BJ, Aldrich CC, Fecik RA, Reynolds KA, Sherman DH. J Am Chem Soc.
2003;125:12551–12557.
approach for the efficient total synthesis of helicascolide C (3) in
only two steps from chiral auxiliary-based b-keto imide 10. The
synthetic route can be further shortened by combining aldol reac-
tion and cyclization/methylation, which provided a platform for
other bioactive natural products bearing a critical d-lactone ring
system with gem-dimethyl substitution at C2. Based on the
practicability of the new route, other congeners of the family could
be divergently obtained with additional two steps. Moreover, a
chemical intermediacy of b-keto acid derived from eliminative
ring-opening of helicascolide C (3) implicates a revised biogenesis
of the decarboxylative derivative 4. Such biogenesis proposal may
be also possible to form similar natural products through decar-
boxylation of b-keto esters with or without cooperation of individ-
ual enzymes.
16. For a 4-g scale reaction performed in THF, the reaction mixture became vicious
due to the solubility issue, and the sluggish magnetic stirring led to
diminished yield.
17. Wong SS, Paddon-Row MN. J Chem Soc, Chem Commun. 1990;456–458.
a
18. Data for cyclic ethylboronate 19: TLC (petroleum ether/ethyl acetate = 4:1 v/v,
KMnO4): Rf = 0.81; [
a]
D
25 = +2.7 (c = 0.69 in CHCl3); 11B NMR (128 MHz, CDCl3):
d = 31.7 ppm; 1H NMR (400 MHz, CDCl3): d = 5.52 (qq, J = 6.8, 0.8 Hz, 1H, HC7),
4.77 (d, J = 1.6 Hz, 1H, HC1), 3.60 (br t, J = 2.0 Hz, 1H, HC3), 3.53 (d, J = 11.2 Hz,
1H, HC5), 2.09 (m, 1H, HC4), 1.66 (s, 3H, H3C11), 1.62 (d, J = 6.4 Hz, 3H, H3C12),
1.12 (s, 3H, H3C8), 1.00 (s, 3H, H3C9), 0.96 (t, J = 7.2 Hz, 3H, H3C14), 0.80 (q, J =
7.2 Hz, 2H, H2C13), 0.78 (d, J = 6.8 Hz, 3H, H3C10) ppm; 13C NMR (125 MHz,
CDCl3): d = 132.4 (C6), 125.4 (C7H), 97.4 (C1H), 77.8 (C3H), 77.5 (C5H), 35.4 (C2),
31.8 (C4H), 23.1 (C9H3), 21.7 (C8H3), 14.4 (C10H3), 13.2 (C12H3), 11.2 (C11H3), 8.1
(C14H3) ppm; IR (thin film): vmax = 2962, 2934, 2878, 1739, 1461, 1404, 1378,
1328, 1266, 1217, 1188, 1106, 1072, 1048, 1030, 959, 935, 816, 758, 747 cmÀ1
;
HRMS-DART (m/z): calcd. for C14H26O130B[M+H]+: 252.2006, found: 252.2004.
19. (a) Reduction with L-selectride also provided analogous sec-butyl boronate
19a. See the Supporting information for details. For other examples of cyclic
boronate formation in the context of borohydride reductions: Garlaschelli L,
Mellerio G, Vidari G. Tetrahedron Lett. 1989;30:597–600;
Acknowledgments
Financial support from the Shanghai Science and Technology
Commission (15JC1400400), the National Natural Science Founda-
tion of China (21290184 and 21472223), the Strategic Priority
(b) Rodríguez-Hahn L, Cárdenas J, Esquivel B, Toscano RA, Rodríguez B.
Tetrahedron Lett. 1995;36:6391–6394;
(c) Renata H, Zhou Q, Baran PS. Science. 2013;339:59–63.

4464
K. Zheng et al. / Tetrahedron Letters 58 (2017) 4459–4464
20. Data for b-keto acid 20: TLC (petroleum ether/ethyl acetate = 4:1 v/v, KMnO4):
Rf = 0.01; 1H NMR (400 MHz, CD3OD): d = 7.03 (s, 1H, HC5), 5.70 (q, J = 7.2 Hz,
1H, HC7), 1.92 (s, 3H, H3C10), 1.84 (s, 3H, H3C11), 1.74 (d, J = 7.2 Hz, 3H, H3C12),
1.33 (s, 6H, H3C8, H3C9) ppm; 13C NMR (100 MHz, CD3OD): d = 206.8 (C3@O),
182.6 (C1@O), 144.3 (C5H), 134.6 (C6), 132.9 (C4), 131.4 (C7H), 56.3 (C2), 26.2
1048, 1034, 896, 845 cmÀ1. HRMS-ESI (m/z): calcd. for C12H18O3Na [M+Na]+:
233.1148, found: 233.1147.
21. For a recently reported b-keto acid with similar structure: Wu W, Wu Y, Liu B.
Tetrahedron. 2017;73:1265–1274.
22. See the Supporting Information for details.
(C8H3, C9H3), 16.3 (C11H3), 14.6 (C10H3), 14.1 (C12H3) ppm; IR (thin film): vmax
3394, 2991, 2965, 2928, 2864, 1643, 1617, 1575, 1396, 1348, 1310, 1243, 1158,
=
23. For C2-enolizable lactone 14, decarboxylation reactions using methanolic
K2CO3 or TBAF under similar conditions only resulted in recovery of starting
materials, and elimination products were not formed to any extent.