Non-canonical regioisomerizations and a 'Diels-Alderase' are likely essential in the biosynthesis of Spiculoic acid A

Article abstract of DOI:10.1016/j.bmcl.2012.06.052

The regiochemistry of dehydration and cyclization steps of the linear biosynthetic precursor of the polyketide natural product Spiculoic acid A (1) were examined. Herein we describe the synthesis of polyene-containing aldehyde 21, a counterpart to the metabolite's putative polyketide intermediate and demonstrate its inability to undergo facile IMDA chemistry. These results suggest the involvement of a non-canonical regioisomerization in the biosynthesis of 1, and that the IMDA reaction is likely enzyme-catalyzed.

Full text of DOI:10.1016/j.bmcl.2012.06.052

Bioorganic & Medicinal Chemistry Letters 22 (2012) 5253–5256
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Non-canonical regioisomerizations and a ‘Diels–Alderase’ are likely essential
in the biosynthesis of Spiculoic acid A
Atahualpa Pinto a,b,c, Christopher N. Boddy b,c,
⇑
a Department of Chemistry, Syracuse University, Syracuse, NY 13244, USA
b Department of Chemistry, Centre for Catalysis Research and Innovation, University of Ottawa, Ottawa, Canada ON K1N 6N5
c Department of Biology, University of Ottawa, Ottawa, Canada ON K1N 6N5
a r t i c l e i n f o
a b s t r a c t
Article history:
The regiochemistry of dehydration and cyclization steps of the linear biosynthetic precursor of the
polyketide natural product Spiculoic acid A (1) were examined. Herein we describe the synthesis of
polyene-containing aldehyde 21, a counterpart to the metabolite’s putative polyketide intermediate
and demonstrate its inability to undergo facile IMDA chemistry. These results suggest the involvement
of a non-canonical regioisomerization in the biosynthesis of 1, and that the IMDA reaction is likely
enzyme-catalyzed.
Received 14 May 2012
Revised 14 June 2012
Accepted 15 June 2012
Available online 23 June 2012
Keywords:
Spiculoic acid A
Ó 2012 Elsevier Ltd. All rights reserved.
Polyketide biosynthesis
Double bond regioisomerization
Diels–Alderase
IMDA reaction
Spiculoic acid A (1, Fig. 1) is a polyketide natural product iso-
lated from the tissue extracts of the Caribbean sponge Plakortis
angulospiculatus.1 1 and the structurally related zyggomphic acids2
are characterized by a highly unusual trans-hydrindane framework
with six stereocenters, two of which are adjacent, strained quater-
nary carbons. Spiculoic acid A possesses modest antiproliferative
a
,b-unsaturated olefins from b-hydroxy thioester polyketide inter-
mediates. The proposed biosynthesis for spiculoic acid involves re-
peated b,
-unsaturations, presumably catalyzed by DH⁄ domains,
c
which lack the key residues required for elimination but can cata-
lyze isomerization.8 A consequence of this biosynthetic proposal is
a large HOMOdiene–LUMOdienophile energy gap of the cyclization pre-
cursor, precluding physiologically relevant cyclizations in the ab-
sence of a catalyst.
activity against MCF-7 breast cancer cells (IC50 = 8 l
g/mL).1
The unusual hydrindane framework was proposed to be of poly-
ketide origin, arising from an intramolecular Diels–Alder-like reac-
tion.1 A type I polyketide synthase (PKS)3 was proposed to
condense ethyl and methylmalonyl-CoA building blocks with a
phenylacetyl-CoA starter unit4 to generate a linear polyketide
chain. While formation of [4.4.0]-bicyclo frameworks from linear
polyketides is well documented,5 the construction of the [4.3.0]-
bicyclo framework, as is seen in the hydrindane skeleton, requires
non-standard regiochemistry for the olefins in the polyketide cycli-
zation precursor.6 For example the [4.3.0]-bicyclo framework of
spinosyn A can only be accessed after the SpnM-catalyzed dehy-
dration generates a non-standard unsaturation pattern.7
The non-standard olefin regiochemistry required for Spiculoic
acid A cyclization was originally proposed by Andersen and co-
workers to be accessed through non-standard dehydratase (DH⁄)
domains (Fig. 2-IIa), as seen in rhizoxin D biosynthesis.8 In the vast
majority of examples polyketide synthase DH domains introduce
Consistent with this large HOMOdiene–LUMOdienophile energy gap,
synthetic studies based on this biosynthetic hypothesis have re-
quired elevated temperatures to effect an intramolecular Diels–Al-
der (IMDA) cyclization.9,10 Kirkham et al. showed that under the
reaction conditions used to install the molecule’s dienophile
(100 °C), their advanced intermediate underwent IMDA cycliza-
tion.9 Matsumura et al. successfully isolated an IMDA precursor
and showed it was slow to react at elevated temperatures
(70 °C).10 A truncated analog of 1 synthesized by Perkins11 is the
sole example of the IMDA reaction occurring at mild and physio-
logically relevant temperatures. This reactivity is attributed to
the pre-organization of the sterically crowded intermediate to-
wards a favorable reactive IMDA conformation.
To address both the unusual dehydration and large HOMOdiene
–
LUMOdienophile energy gap, we propose an alternative biosynthetic
hypothesis (Fig. 2-I and IIb). In our proposal, dehydration occurs
in a standard fashion generating a linear intermediate unable to
undergo an IMDA reaction. However due to the conjugated trie-
none system, enol or enolate formation should be facile, leading
⇑
Corresponding author. Tel.: +1 613 562 5800x8970; fax: +1 613 562 5170.
E-mail address: cboddy@uottawa.ca (C.N. Boddy).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.06.052

5254
A. Pinto, C. N. Boddy / Bioorg. Med. Chem. Lett. 22 (2012) 5253–5256
Ph
spiculoic acid structure was a thermodynamic minimum. These re-
sults indicate that enol or enolate activated IMDA reaction can
potentially access the desired diastereomer of 1 via a thermody-
namically controlled epimerization. Encouraged by these results,
we embarked on the synthesis of aldehyde 21, a biomimetic IMDA
precursor. We selected the aldehyde oxidation state because it
electronically resembles the PKS-bound linear thioester intermedi-
ate. The limited back-donation of the thioester’s sulfur lone pairs
into the carbonyl renders its LUMO energy similar to the aldehyde
oxidation state as compared to an ester, acid, or carboxylate of 1.
Our synthetic strategy relied on a late stage coupling of two
equivalently complex fragments (Fig. 4). Originally this was envi-
sioned via a Horner–Wadsworth–Emmons olefination, however
the phosphonate 10, proved to be unstable under all reaction con-
ditions employed, consistent with Heathcock’s results during the
synthesis of halichlorine.12 Ultimately we relied on a Lewis acid-
mediated aldol condensation–dehydration, followed by a Swern
oxidation to access the IMDA precursor 21.
The synthesis of 8 (Scheme 1) was rapidly achieved relying on a
series of successive Wittig olefinations13 from phenylacetaldehyde.
The ethyl ketone 9 was generated by exhaustive reduction of
diethyl ethylmalonate (16) followed by desymmetrization of the
diol using a biphasic monosilylation strategy. Oxidation followed
by Wittig olefination gave ester intermediate 19, which was con-
verted to the ethyl ketone via the Weinreb amide.
Fragments 8 and 9 were coupled (Scheme 2) by a Lewis acid-
mediated aldol reaction yielding a complex mixture of diastereo-
mers. A mesylation/elimination strategy followed by deprotection,
provided stable hydroxyketone tetraene 20 as a mixture of diaste-
reomers. The regioisomeric configuration of the olefins was con-
firmed through 2D NMR analysis. Isomerization of the triene
system occurred at this stage, likely due to the strongly destabiliz-
ing effect of the in-plane 1,3-ethyl groups. Swern oxidation of 20
furnished tetraene aldehyde 21. Migration of the olefin into conju-
gation with the newly formed aldehyde generated an intermediate
with the olefination pattern observed in the advanced aldehyde
Ph
H
H
H
H
O
O
HO
HO
O
O
1, spiculoic acid A
2, zyggomphic acid
Figure 1. Spiculoic acid A (1) and zyggomphic acid (2), representative examples of
the hydrindane framework.
to the correct olefin orientation for [4.3.0]-bicycle formation. In
addition, we speculate that a mild, spontaneous IMDA cyclization
may take place prior to the polyketide’s release from the PKS,
and be facilitated by the favorable electronics of the extended enol
or enolate intermediate. The increased electron density of the
diene provided by the extended conjugation of the enol moiety will
decrease the HOMOdiene–LUMOdienophile energy gap of the system.
Guided by our biosynthetic rationale, we describe the synthesis
of a novel biomimetic acyclic polyketide intermediate of 1 and
demonstrate that the isomerization of the olefins in the linear
polyketide is non-facile and IMDA chemistry does not occur. These
results suggest that a discrete catalytic domain, such as a DH⁄ do-
main, is essential to isomerize the olefins into the correct regio-
chemistry for an IMDA reaction and that the IMDA reaction is
likely enzyme-catalyzed.
Given that our proposed IMDA pathway requires enolization at
C6, the stereochemistry of this carbon must be set either via a
kinetically controlled protonation or via equilibration to a thermo-
dynamic minimum. To gain insight into the energetic landscape of
the products, we employed gas-phase ab initio density functional
theory (DFT) computational methods at the B3LYP/6-31G(d,p) level
of theory. Computational results indicated the desired configura-
tion of the C6 stereocenter was energetically favored (Fig. 3). Be-
cause the C4 proton is also sensitive to enolization, we calculated
the relative energies of the epimers at this center. Again the native
I.
II.
a)
Loading
Module 2
Module 4
Release
Module 1
DH* KR
Module 1
DH* KR
Module 1
KR
Module 3
KR
Module 5
KR
DH*
KR
DH
DH
DH
DH
KS AT
ACP
KS AT
ACP
AT ACP KS AT
ACP KS AT
ACP KS AT
ACP KS AT ACP KS AT
ACP TE
H2O
SH OH
Ppant
S
SH OH
Ppant
S
S
S
S
S
S
S
O
O
O
O
O
O
O
O
O
O
3
OH
7
9
b)
Module 1
DH KR
Module 1
DH KR
DH
KS AT
ACP
KS AT
ACP
H2O
SH OH
Ppant
S
SH OH
Ppant
S
H
O
O
9
7
3
9
7
O
OH
OH
3
H
HO
R
O
O
R = S-ACP or
1
R = OH
Figure 2. (I) Proposed spiculoic acid A (1) linear chain and PKS extender unit framework. Either while covalently bound to the PKS or following thioesterase (TE) hydrolysis,
we hypothesize enol/enolate formation leads to a facile intramolecular Diels–Alder cycloaddition. (II) Contrasting biosynthetic hypotheses: (a) Abnormal PKS-bound
dehydration/regioisomerization mediated by a non-canonical dehydratase domain (DH⁄) proposed by Andersen et al. (b) Normal dehydrated product catalyzed by canonical
DH domains.

A. Pinto, C. N. Boddy / Bioorg. Med. Chem. Lett. 22 (2012) 5253–5256
5255
O
O
O
H
a
OH
OR
O
H
9
20
H
4
HO
O
R = TBDPS
O
O
b
H
16.1
kcal/mol
8
H
O
HO
O
5
O
H
H
12.4
HO
O
kcal/mol
3
O
21
H
9.2
kcal/mol
6
4
Scheme 2. Reagents and conditions: Key (a) 8 TiCl4, (i-Pr)2NEt; MsCl, Et3N; TBAF,
6% over three steps. (b) (COCl)2, DMSO, Et3N, 100%.
H
HO
1
O
0
kcal/mol
Figure 3. Relative energies of C4 and C6 diastereomers (1, 3–5) of spiculoic acid A.
O
O
R
R
15
: R = OEt
8: R = H
IMDA
Dehydration
Figure 5. Regioisomerization of intermediates 8 and 15.
H
H
O
O
7
3
6
4
2
O
reaction strategy.17 Upon heating at 80 °C for five days 21 re-
mained unreactive as monitored by 1H NMR.
The stability of 21 was surprising. The absence of a complex
mixture of olefin regioisomers in solution suggests either a high
barrier for enolization or a strong thermodynamic preference for
the observed isomer. A structure optimization of 21 showed a
lowest energy conformation where the C10 proton was co-planar
H
R
7
O
6
: R = H
Pinnick
Aldol
1: R = OH
O
O
O
11
7
4
2
P
6
EtO
OR
OR
+
O
EtO
to the C8–C9 and C11–C12
p systems, minimizing the proton’s
10
8
9
R = TBDPS
acidity. In addition to this conformational effect, the enol struc-
tures were very high in energy. For example the Z enol was
+10.8 kcal/mol higher in energy than the keto tautomer, presum-
ably due to the 1,3-strain of having multiple trisubstituted olefins
in-plane. Combined, these observations suggest that non-enzyme-
catalyzed isomerization to the postulated conjugated polyenol
Diels–Alder precursor is disfavored.
While non-enzymatic isomerization of 21 is disfavored, non-en-
zyme catalyzed isomerization did occur readily at earlier steps of
our synthesis. Under multiple reaction conditions, with substrates
such as 8 and 15, we observed regioisomer formation consistent
with elimination of 1,3-dialkyl strain via olefin migration (Fig. 5).
Presumably these regioisomers are accessed via protonation of
the extended enolate under basic conditions or extended enol un-
R = TBDPS
Figure 4. Biomimetic retrosynthetic analysis of spiculoic acid A.
a
O
a
O
O
R
R
O
Ph3P
O
OEt
11
12
13
15
: R = OEt
: R = OEt
b,c
b,c
14: R = H
8: R = H
O
O
R
d
f ,a
EtO
OEt
RO
OH
TBDPSO
17: R = H
18
19: R = OEt
der acidic conditions at the
a or c positions.
16
e
g,h
9
: R = TBDPS
: R = Et
It is interesting to note that the regiochemical outcome ob-
served for 21 matches precisely with the olefination pattern ob-
served for putative biosynthetic shunt products recently isolated
from zyggomphic acid-containing sponge samples.2b It is possible
that these shunt products are formed when this regiochemical sink
Scheme 1. Reagents and conditions: Key (a) 12, PhMe, reflux, 18 h; (b) DIBAL-H,
CH2Cl2, 0 °C, 1 h; (c) IBX, DMSO, 1.5 h; (d) LAH, THF, 36 h, 82%; (e) TBDPSCl, 1:4
MeCN–Hexanes, Et3N, 72 h, 56%; (f) (COCl)2, DMSO, Et3N, 5 h, 64%; (g) i-PrMgCl,
HClÁHN(OMe)Me, THF, À20 °C, 2 h, 61%; (h) EtMgBr, THF, 0 °C, 2 h, 62%.
is accessed via unwanted
these compounds.
e-protonation during the biosynthesis of
intermediate from the synthesis of khafrefungin,14 and consistent
with minimization of in-plane 1,3-dialkyl strain. This provided us
with the key compound needed to investigate our proposed IMDA
strategy.
A number of reaction conditions were attempted to convert 21
into the bicyclic core of spiculoic acid via an IMDA reaction, how-
ever none were successful. For example, we employed MacMillan’s
catalyst15 under acidic conditions to effect proton transfer, enoliza-
tion and product formation. After five days and gradual increase in
temperature, 21 slowly decomposed. In an attempt to increase the
dienophile’s polarization and reactivity, we employed MeAlCl2.16
Upon treatment of 21 with the Lewis acid at À78 °C and its slow
heating to room temperature, unreacted starting material was
This study thus demonstrates that non-enzyme catalyzed isom-
erization of the olefins generated by standard dehydratase chemis-
try in a linear spiculoic acid precursor is non-facile and IMDA
chemistry does not readily occur. These results suggest that a dis-
crete catalytic domain, such as a DH⁄ domain or potentially a post-
PKS elimination as seen in spinosyn A biosynthesis, is essential to
isomerize the olefins into the correct regiochemistry for an IMDA
cycloaddition.6 Finally our work supports the hypothesis that the
IMDA reaction is enzyme-catalyzed.
Acknowledgments
L. Barriault (University of Ottawa), N. Totah (Syracuse Univer-
sity) and T. Korter (Syracuse University) are thanked for insightful
recovered. We investigated
a water-accelerated Diels–Alder

5256
A. Pinto, C. N. Boddy / Bioorg. Med. Chem. Lett. 22 (2012) 5253–5256
4. Moore, B. S.; Hertweck, C. Nat. Prod. Rep. 2002, 19, 70.
5. (a) Kelly, W. L. Org. Biomol. Chem. 2008, 6, 4483; (b) Campbell, C. D.; Vederas, J.
C. Biopolymers 2012, 93, 755.
discussion. This work was funded by NSERC and an NSF Predoc-
toral Fellowship to A.P..
6. Kim, H. J.; Ruszczycky, R. W.; Choi, S.-H.; Liu, Y -n.; Liu, H.-W. Nature 2011, 473,
109.
7. Gulder, T. A. M.; Freeman, M. F.; Piel, J. Top. Curr. Chem. 2012, 1.
8. Kusebauch, B.; Busch, B.; Scherlach, K.; Roth, M.; Hertweck, C. Angew. Chem., Int.
Ed. 2010, 49, 1460.
9. Kirkham, J. E. D.; Lee, V.; Baldwin, J. E. Chem. Commun. 2006, 2863.
10. Matsumura, D.; Takarabe, T.; Toda, T.; Hayamizu, T.; Sawamura, K.; Takao, K.-I.;
Tadano, K.-I. Tetrahedron 2011, 67, 673.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.
06.052.
11. Crossman, J. S.; Perkins, M. V. Tetrahedron 2008, 64, 4852.
12. Christie, H. S.; Heathcock, C. H. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 12079.
13. Baldwin, J. E.; Moloney, M. G.; Parsons, A. F. Tetrahedron 1992, 48, 9373.
14. Shirokawa, S.-i.; Shinoyama, M.; Ooi, I.; Hosokawa, S.; Nakazaki, S.; Kobayashi,
S. Org. Lett. 2007, 9, 849.
15. Northrup, A. B.; MacMillan, D. W. C. J. Am. Chem. Soc. 2002, 124, 2458.
16. Roush, W. R.; Barda, D. A. J. Am. Chem. Soc. 1997, 119, 7402.
17. Otto, S.; Engberts, J. B. F. N. Pure Appl. Chem. 2000, 72, 1365.
References and notes
1. Huang, X.-H.; van Soest, R.; Roberge, M.; Andersen, R. J. Org. Lett. 2004, 6, 75.
2. (a) Berrué, F.; Thomas, O. P.; Rogelio Fernández, R.; Amade, P. J. Nat. Prod. 2005,
68, 547; (b) Ankisetty, S.; Gochfeld, D. J.; Diaz, M. C.; Khan, S. I.; Slattery, M. J.
Nat. Prod. 2010, 73, 1494.
3. Fischbach, M. A.; Walsh, C. T. Chem. Rev. 2006, 106, 3468.