On the Origin of Dolabriferol: Total Synthesis via Its Putative Contiguous Precursor

Article abstract of DOI:10.1021/acs.orglett.6b01798

The putative contiguous polypropionate precursor of dolabriferol was synthesized using as the key step a rationally designed enantiomer-selective aldol coupling (i.e., with kinetic resolution) of a racemic C1-C8 ketone fragment with an enantiopure C9-C15

Full text of DOI:10.1021/acs.orglett.6b01798

Letter
pubs.acs.org/OrgLett
On the Origin of Dolabriferol: Total Synthesis via Its Putative
Contiguous Precursor
Athanasios Karagiannis, Naveen Diddi, and Dale E. Ward*
Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, SK S7N 5C9, Canada
S
* Supporting Information
ABSTRACT: The putative contiguous polypropionate pre-
cursor of dolabriferol was synthesized using as the key step a
rationally designed enantiomer-selective aldol coupling (i.e.,
with kinetic resolution) of a racemic C1−C8 ketone fragment
with an enantiopure C9−C15 aldehyde fragment. When
exposed to alumina, the precursor was cleanly transformed
into dolabriferol via a regioselective retro-Claisen fragmenta-
tion, providing the first experimental evidence for the proposed origin of dolabriferol and demonstrating that it is a plausible
isolation artifact.
olabriferol (1), isolated from the anaspidean mollusk
Dolabrifera dolabrifera,1 belongs to a small family of
epi-1 (−12.3 kJ/mol), the 13→9 (9S)-hemiacetal of (8R)-2
(0.0 kJ/mol), and 1 (2.3 kJ/mol) from which it was concluded
that the Gavagnin pathway was feasible. However, the question
of whether 1 is a biosynthetic product or a plausible isolation
artifact cannot be addressed by this approach. Moreover, if
retro-Claisen fragmentations are irreversible, the pathway from
2 to 1 will be under kinetic control and therefore depend on
the relative concentration and retro-Claisen reactivity of 3 (the
11→7 hemiacetal of 2) compared to those of the regioisomeric
13→9 and 5→9 hemiacetals of 2 and other tautomers or
intermediates capable of competing irreversible reactions (e.g.,
retro-aldol, elimination). If hemiacetal equilibration is much
faster than retro-Claisen fragmentation, product distribution
will be governed by reactivity (Curtin−Hammett principle);
otherwise, the relative concentrations of the hemiacetals will
increasingly dominate product formation. For example, in
contrast to the alumina-mediated transformation of 7 into 8
noted above, analogous treatment of 9, a complex mixture of
ring−chain and keto−enol tautomers containing both 11→7
and 5→9 hemiacetals gave a 1:1.5 mixture of 8 and 10
suggesting that retro-Claisen fragmentation was faster than
hemiacetal equilibration under these conditions.8 In this
context, an additional complication in the proposed trans-
formation of 2 into 1 arises from Goodman’s calculations9 that
suggest the formation of the 13→9 hemiacetal regioisomer is
very strongly favored (>99.8% at rt) for both (8R)-2 and (8S)-
2.10 In light of the above ambiguities, we set out to prepare the
putative contiguous precursor 2 and to evaluate its propensity
to undergo the proposed rearrangement to dolabriferol (1). In
this paper, we report a concise total synthesis of 2 and its
remarkably facile conversion into 1.
D
marine natural products known as noncontiguous polypropi-
onates. In contrast to most polypropionates found in nature,
members of this group have structures composed of a
polypropionate carboxylic acid fragment joined to a poly-
propionate alcohol fragment via an ester linkage and hence
possess noncontiguous carbon skeletons.2 These structures are
generally believed to result from retro-Claisen fragmentations
of hemiacetals formed by cyclization of 5-hydroxy-1,3-dione
motifs embedded in biosynthetically derived polypropionate
chains.3 For example, Gavagnin et al. proposed a hypothetical
biosynthesis of 1 involving retro-Claisen rearrangement of
hemiacetal 3, a ring−chain tautomer of the putative contiguous
polypropionate precursor 2 (Scheme 1).1 Although such
processes might be enzyme-mediated, the natural product
status of noncontiguous polypropionates is uncertain, and they
are often viewed as plausible isolation artifacts.3 Evidence for
such hypotheses is largely circumstantial and includes the
failure to detect the expected noncontiguous polypropionates
after a careful extraction procedure2d and the observation2a,d,4−8
of several related retro-Claisen transformations in vitro.
Nonetheless, only two examples of conversion of a contiguous
polypropionate “natural product” into an isomeric non-
contiguous “natural product” are known. The hemiacetal 5
and isomeric ester 6 were isolated2a from Siphonaria australis,
and transformation of the former into the latter by retro-
Claisen rearrangement was accomplished under relatively mild
conditions (DBU/benzene/rt2a,5b or chromatography over
alumina5a). Similarly, treatment of siphonarin B (7)4 with
alumina in refluxing ethanol gave baconipyrone C (8),2b both
previously isolated from S. zelandica.8
Several groups have reported synthetic studies on dolabrifer-
ol (1), and all approaches envisaged an esterification or retro-
Using a reaction prediction program, Goodman et al.
conducted a computational study of Gavagnin’s proposed
pathway from 2 to 1 via 3 and 4 assuming thermodynamic
control.9 The relevant structures of lowest relative energy
(RHF/3-21G/water basis set) among the >200 found were 12-
Received: June 20, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b01798
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Our retrosynthesis of 2 is presented in Scheme 2. Previous
Scheme 1. Proposed Origin of Dolabriferol and Related in
Vitro Retro-Claisen Fragmentations of Contiguous
Polypropionates
synthetic studies6,8,15 have amply demonstrated the sensitivity
Scheme 2. Retrosynthetic Analysis of Putative Precursor 2
of 5,11-dihydroxy 3,7,9-triones toward both acidic and basic
reaction conditions and underscore the need to reveal the
corresponding functionality in 2 under mild conditions.
Accordingly, we selected 11 as the intermediate objective
with protecting groups (Pg) carefully selected to permit its
transformation into 2. Disconnection of 11 at the C8−C9 bond
via an oxidation/directed−aldol transform generates enolate 12
and aldehyde 13, and suitably protected synthetic equivalents of
each should be available by straightforward transformations of
known compounds 1416 and (−)-15,17 respectively. Although
both new stereocenters generated in the anticipated aldol
coupling are absent in 2, a highly stereoselective reaction was
desirable. Among the four possible diastereomers, 6,8-syn-8,9-
syn-9,10-syn-18 was particularly attractive as previous work has
shown that aldol reactions of (Z)-enolates 16 [MLn = B(c-
Hex)2 or Ti(OiPr)3; Pg = CH2OMe but not SiEt3] with
iPrCHO give 6,8-syn-8,9-syn adducts with excellent diaster-
eoselectivity15 and aldol additions to tetrahydrothiopyran-3-
carbaldehydes related to 17 are highly Felkin selective18 even
with (Z)-enolates.19,20 Moreover, we have demonstrated
synthetically useful levels of kinetic resolution (i.e., s > 10) in
related aldol reactions rationally designed to have the three
stereocontrol elements (i.e., enolate and aldehyde diastereoface
selectivities and relative topicity) highly biased.21 All previous
examples have involved couplings of enantiopure (or achiral)
ketones with racemic aldehydes; however, because ( )-14 was
in hand, aldol coupling of the derived ( )-16 with enantiopure
17 provided an opportunity to test this design with the reactant
roles reversed.
Claisen fragmentation to construct the key ester linkage in a
hydroxy-protected intermediate followed by deprotec-
tion.6,7,11−14 Vogel et al.13 used the esterification approach to
complete the first total synthesis of 1 establishing its absolute
configuration. The choice of an Alloc (−C(O)OCH2CH
CH2) protecting group for the C-7′ OH group proved crucial
for success and underscores the sensitivity of 1. Toste et al.,14
also using the esterification strategy, prepared several 7′-O-
protected derivatives of 4 (SEM = CH2OC2H4SiMe3; TES =
SiEt3; TBS = SiMe2-t-Bu; BOM = CH2OCH2Ph) via
esterification but were unable to remove the protecting group
without decomposition in the final step. Lister and Perkins,6 the
first to use the retro-Claisen approach, obtained the 7′-O-TBS
derivative of ent-4 by fragmentation of ent-13-O-TBS-2 in the
presence of DBU; however, advancing to ent-1 was thwarted as
the TBS protecting group could not be removed. Following a
closely related route, Currie and Goodman7 successfully
obtained 1 by retro-Claisen fragmentation of a triply protected
derivative of 2 (functionalities at C-3, C-5, and C-13 blocked)
followed by removal of the protecting groups (from 7′-O-PMB-
4 in the final step). Although these synthetic studies clearly
demonstrated the feasibility of obtaining the dolabriferol
skeleton via a retro-Claisen fragmentation, the origin of 1 was
not addressed because the presence of protecting groups on 2
serves to block various competing reactions along Gavagnin’s
proposed pathway (Scheme 1).
The synthesis commenced with the preparation of (−)-15
(77%, 92% ee) on a decagram scale by L-proline catalyzed aldol
reaction of 19 with iPrCHO using a modification of the
published protocols17 to achieve higher yield with respect to
the more valuable 19 (Scheme 3).22 The triethylsilyl ether
derivative of (−)-15 was prepared under standard conditions
B
DOI: 10.1021/acs.orglett.6b01798
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(+)-18a are assigned based on the known diastereoselectivities
of aldol reactions of 16 [Pg = CH2OMe; MLn = Ti(O-i-Pr)3)]16
and of additions to aldehydes related to 17;18 the 8,9-syn
Scheme 3. Synthesis of Dolabriferol (1)
relative configuration in 18a is supported by NMR (3JH8−H9
=
3.5 Hz; δC CH3C-8 = 10.2; δC C-9 = 69.5).16 The minor aldol
adduct presumably results from the mismatched reaction of
(+)-17a with (+)-24 but was not isolated in pure form or
characterized.
Following FeCl3-mediated removal of the ethylene ketal
protecting group and IBX oxidation of the alcohol in (+)-18a,
desulfurization of the resulting trione with Raney Ni occurred
with concomitant hydrogenolysis of the BOM group to afford
25 (68%) that was a 1:1:0.1 mixture of two keto and enol
tautomeric forms, respectively, in CDCl3 solution (Scheme
3).25 The benzylidene acetal in 25 was surprisingly resistant to
hydrogenolysis.26 After considerable experimentation, stirring a
suspension of excess freshly prepared Pd black and 25 in
iPrOH under a H2 atmosphere (4 bar) afforded 2 in good yield.
Among the plethora of possible keto−enol and ring−chain
tautomeric forms, in C6D6 solution 2 was predominantly a 1.5:1
mixture of hemiacetals 2a [(9S) 13→9; epimeric at C-8]
consistent with Goodman’s calculations.9,10 Treatment of 2a
with DBU in C6D6 slowly produced a mixture of products
including a 3:6:1 mixture of 1 and two unidentified
diastereomers,27 respectively (38%), and 2628 (ca. 44%),
among others. Presumably 1 (and its diastereomers) results
from retro-Claisen fragmentation of the C7−C8 bond in the
11→7 hemiacetal 3 (and its diastereomers)27 and 26 arises
from fragmentation of C8−C9 bond in the 13→9 hemiacetal
2a (or perhaps by fragmentation of the same bond in the 5→9
hemiacetal of 2 followed by lactonization). To improve the
retro-Claisen regioselectivity and reduce base-mediated epime-
rization, we subjected 2a to alumina.5a,8 Although 2a was stable
to neutral alumina in hot EtOH, brief exposure to neutral
alumina in benzene cleanly produced dolabriferol (1) in
excellent yield. This remarkably selective retro-Claisen process
presumably results from a much higher reactivity of hemiacetal
3 (compared to its regioisomers) under conditions (alumina/
benzene) where hemiacetal equilibration is rapid. The facile
conversion of 2 into 1 clearly establishes the latter as a plausible
artifact of isolation.
In summary, the first total synthesis of 2, the putative
contiguous precursor of dolabriferol (1), was achieved via a
rationally designed enantioselective aldol coupling of ( )-24
with (+)-17a as the key step.29 A unique feature of this
“fragment coupling with kinetic resolution” strategy is that the
stereocenters in the racemic fragment are set under substrate-
control, thereby allowing syntheses of either enantiomer or
racemic product from the identical reactants via the same
convergent route by simply changing the catalyst used in the
initial preparation of 15. Although 2 existed predominantly in
the undesired hemiacetal form 2a, exposure to neutral alumina
resulted in a highly regioselective retro-Claisen fragmentation
via hemiacetal 3 to efficiently afford 1. These results cannot rule
out an enzyme-mediated retro-Claisen process in vivo;
however, they demonstrate for the first time that 2 is a
competent biosynthetic end product capable of in vitro
transformation into dolabriferol (1) under very mild con-
ditions.
and treated with t-BuLi followed by BnO2CCN to afford β-
ketoester 20 in 82% yield over two steps. Hydroxy-directed
reduction23 of 20 gave a separable 1:1 mixture of 1,3-syn diols
21 and 3-epi-21. During attempted benzylation of 3-epi-21 by
reaction with KH and BnBr, a rapid isomerization to 21,
presumably via a retroaldol−aldol mechanism, was noted. After
extensive optimization, subjecting the crude 1:1 mixture of diols
to t-BuOK (2 equiv) and CuI (0.3 equiv) in Et2O at 0 °C for 5
min produced 21 (dr >19:1; ca. 80% yield over two steps) on
decagram scale. A three-step sequence involving formation of
the benzylidene acetal, reduction of the ester, and oxidation of
the resulting primary alcohol transformed 21 into aldehyde
(+)-17a in 80% yield.
The reported procedure for the synthesis of ( )-14 gave a
separable 1:1 mixture of diastereomers.16 In an improved route,
( )-14 was obtained by MgBr2·Et2O-mediated Mukaiyama
aldol reaction of 2316 with ( )-2221c (dr = 4:1) followed by
Raney Ni desulfurization and was subsequently converted to
the BOM derivative ( )-24 under standard conditions
(Scheme 3). Conversion of ( )-24 into its Ti(IV) (Z)-enolate
by transmetalation16,24 of the corresponding Li enolate
followed by addition of (+)-17a (0.33 equiv) gave a 7:1
mixture of aldol adducts from which (+)-18a (79%) and (+)-24
(70%, ca. 20% optical purity) were isolated. The recovery of 24
enriched in the dextrorotatory enantiomer confirms that
(+)-18a results from preferential reaction of (+)-17a with
(−)-24, as expected. The C-8 and C-9 configurations in
C
DOI: 10.1021/acs.orglett.6b01798
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(13) (a) Laclef, S.; Turks, M.; Vogel, P. Angew. Chem., Int. Ed. 2010,
ASSOCIATED CONTENT
■
́
49, 8525−8527. (b) Laclef, S. Ph.D. Thesis, Ecole Polytechnique
S
* Supporting Information
Federale de Lausanne, Switzerland, 2010.
́
́
(14) Gesinski, M. R.; Brenzovich, W. E., Jr.; Staben, S. T.; Srinilta, D.
J.; Toste, F. D. Tetrahedron Lett. 2015, 56, 3643−3646.
(15) (a) Paterson, I.; Chen, D. Y.-K.; Franklin, A. S. Org. Lett. 2002,
4, 391−394. (b) Beye, G. E.; Goodman, J. M.; Ward, D. E. Org. Lett.
2009, 11, 1373−1376.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.6b01798.
Experimental procedures, characterization data, and
1
copies of H and 13C NMR spectra for all reported
(16) Ward, D. E.; Kundu, D.; Biniaz, M.; Jana, S. J. Org. Chem. 2014,
79, 6868−6894.
compounds; comparison of NMR data for natural and
synthetic 1; structure assignments for 7-epi-1 (tentative)
and 26 (PDF)
(17) (a) Ward, D. E.; Jheengut, V. Tetrahedron Lett. 2004, 45, 8347−
8350. (b) Pihko, P. M.; Laurikainen, K. M.; Usano, A.; Nyberg, A. I.;
Kaavi, J. A. Tetrahedron 2006, 62, 317−328.
(18) Ward, D. E.; Sales, M.; Man, C. C.; Shen, J.; Sasmal, P. K.; Guo,
C. J. Org. Chem. 2002, 67, 1618−1629.
(19) Roush, W. R. J. Org. Chem. 1991, 56, 4151−4157.
(20) Ward, D. E. In Modern Methods in Stereoselective Aldol Reactions;
Mahrwald, R., Ed.; Wiley−VCH: Weinheim, 2013; Chapter 6, pp
377−429.
(21) (a) Ward, D. E.; Jheengut, V.; Akinnusi, O. T. Org. Lett. 2005, 7,
1181−1184. (b) Ward, D. E.; Becerril-Jimenez, F.; Zahedi, M. M. J.
Org. Chem. 2009, 74, 4447−4454. (c) Ward, D. E.; Jheengut, V.; Beye,
G. E.; Gillis, H. M.; Karagiannis, A.; Becerril-Jimenez, F. Synlett 2011,
2011, 508−512. (d) Becerril-Jimenez, F.; Ward, D. E. Org. Lett. 2012,
́
14, 1648−1651. (e) Ward, D. E.; Kazemeini, A. J. Org. Chem. 2012, 77,
10789−10803.
(22) On a per mole basis, iPrCHO is ca. 3 orders of magnitude less
expensive than 19 (ca. $4K/mol). The latter is readily available in two
simple steps (>70% yield; 50 g scale) from dimethyl 3,3′-
thiodipropanoate (ca. $25/mol): Ward, D. E.; Rasheed, M. A.;
Gillis, H. M.; Beye, G. E.; Jheengut, V.; Achonduh, G. T. Synthesis
2007, 2007, 1584−1586.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: dale.ward@usask.ca.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Pramod Jadhav and Sushital Jana (University of
Saskatchewan) for preliminary experiments on the synthesis of
21 and the improved route to ( )-14, respectively. Financial
support from the Natural Sciences and Engineering Research
Council (Canada) (RGPIN 1536-2013) and the University of
Saskatchewan is gratefully acknowledged.
REFERENCES
■
(23) Chen, K. M.; Hardtmann, G. E.; Prasad, K.; Repic, O.; Shapiro,
M. J. Tetrahedron Lett. 1987, 28, 155−158.
(24) (a) Reetz, M. T.; Peter, R. Tetrahedron Lett. 1981, 22, 4691−
4694. (b) Siegel, C.; Thornton, E. R. J. Am. Chem. Soc. 1989, 111,
5722−5728.
(1) Ciavatta, M. L.; Gavagnin, M.; Puliti, R.; Cimino, G.; Martinez,
E.; Ortea, J.; Mattia, C. A. Tetrahedron 1996, 52, 12831−12838.
(2) (a) Ester 6 (unnamed): Hochlowski, J. E.; Faulkner, D. J. J. Org.
Chem. 1984, 49, 3838−3840. (b) Baconipyrones A−D: Manker, D.
C.; Faulkner, D. J.; Stout, T. J.; Clardy, J. J. Org. Chem. 1989, 54,
5371−5374. (c) Membrenones A and B: Ciavatta, M. L.; Trivellone,
E.; Villani, G.; Cimino, G. Tetrahedron Lett. 1993, 34, 6791−6794. (d)
Sisserone A: Brecknell, D. J.; Collett, L. A.; Davies-Coleman, M. T.;
Garson, M. J.; Jones, D. D. Tetrahedron 2000, 56, 2497−2502. (e)
Micromelones A and B: Napolitano, J. G.; Souto, M. L.; Fernandez, J.
J.; Norte, M. J. Nat. Prod. 2008, 71, 281−284. (f) Dolabriferols B and
C: Jimenez-Romero, C.; Gonzalez, K.; Rodriguez, A. D. Tetrahedron
Lett. 2012, 53, 6641−6645.
(25) Perkins (ref 6) and Goodman (ref 7) reported 1,3-diones
related to 25 as single keto tautomers (but with opposite relative
configurations at C-8); however, under their retro-Claisen conditions
(DBU, benzene), epimerization at C-8 is likely. For selected examples
of kinetically stable chiral 1,3-diones that epimerize under mild
conditions, see refs 8 and 21d and: (a) Paterson, I.; Perkins, M. V.
Tetrahedron 1996, 52, 1811−1834. (b) De Brabander, J.; Oppolzer, W.
Tetrahedron 1997, 53, 9169−9202. (c) Calter, M. A.; Liao, W. S. J. Am.
Chem. Soc. 2002, 124, 13127−13129. (d) Ward, D. E.; Zahedi, M. M.
Org. Lett. 2012, 14, 6246−6249.
(3) Davies-Coleman, M. T.; Garson, M. J. Nat. Prod. Rep. 1998, 15,
477−493.
(26) For a related example, see: Muri, D.; Carreira, E. M. J. Org.
Chem. 2009, 74, 8695−8712.
(4) Hochlowski, J.; Coll, J.; Faulkner, D. J.; Clardy, J. J. Am. Chem.
Soc. 1984, 106, 6748−6750.
(27) The major diastereomer was tentatively assigned as 7-epi-1 by
comparison of its NMR data with those of 1 and 10. Because 7-epi-1
was unchanged when re-subjected to the reaction conditions,
epimerization presumably occurs prior to retro-Claisen fragmentation.
(28) Exner, C. J.; Laclef, S.; Poli, F.; Turks, M.; Vogel, P. J. Org.
Chem. 2011, 76, 840−845.
(5) (a) Sundram, U. N.; Albizati, K. F. Tetrahedron Lett. 1992, 33,
437−440. (b) Lister, T.; Perkins, M. V. Aust. J. Chem. 2004, 57, 787−
797.
(6) (a) Lister, T.; Perkins, M. V. Org. Lett. 2006, 8, 1827−1830.
(b) Lister, T. Ph.D. Thesis, Flinders University, Adelaide, Australia,
2006.
(7) Currie, R. H.; Goodman, J. M. Angew. Chem., Int. Ed. 2012, 51,
4695−4697.
(8) Beye, G. E.; Ward, D. E. J. Am. Chem. Soc. 2010, 132, 7210−7215.
(9) Socorro, I. M.; Taylor, K.; Goodman, J. M. Org. Lett. 2005, 7,
3541−3544.
(29) For the use of this strategy in the total synthesis of caloundrin B,
see ref 21d.
(10) The relative energies for all possible 6-membered ring
hemiacetal tautomers of 2 have been calculated (kJ/mol) (RHF/3-
21G/water basis set; ref 9). From (8R)-2: 13→9, 0.0; 11→7, 37.7; 5→
9, 25.5. From (8S)-2: 13→9, 6.7; 11→7, 23.0; 5→9, 27.1.
(11) (a) Chen
Asymmetry 2003, 14, 2567−2571. (b) Pelchat, N.; Caron, D.;
Chen
̂
evert, R.; Courchesne, G.; Caron, D. Tetrahedron:
̂
evert, R. J. Org. Chem. 2007, 72, 8484−8488.
(12) Dias, L. C.; de Sousa, M. A. Tetrahedron Lett. 2003, 44, 5625−
5628.
D
DOI: 10.1021/acs.orglett.6b01798
Org. Lett. XXXX, XXX, XXX−XXX