Toward the synthesis of cobyric acid. Enantioselective syntheses of completely differentiated ring D synthons

Article abstract of DOI:10.1021/ol800985m

(Chemical Equation Presented) Alkyne acids 11 were prepared in an enantioselective fashion from allylic ester derivatives 18 or 20 by Ireland-Claisen rearrangement, followed by Si-assisted elimination of HBr. The title compounds are attractive ring D synt

Full text of DOI:10.1021/ol800985m

ORGANIC
LETTERS
2008
Vol. 10, No. 13
2837-2840
Toward the Synthesis of Cobyric Acid.
Enantioselective Syntheses of Completely
Differentiated Ring D Synthons
Hui Wang, Carlos Tassa, and Peter A. Jacobi*
Department of Chemistry, Dartmouth College, HanoVer, New Hampshire 03755
p.jacobi@dartmouth.edu
Received April 29, 2008
ABSTRACT
Alkyne acids 11 were prepared in an enantioselective fashion from allylic ester derivatives 18 or 20 by Ireland-Claisen rearrangement, followed
by Si-assisted elimination of HBr. The title compounds are attractive ring D synthons for an ongoing synthesis of cobyric acid.
For several years, we have been exploring an iterative synthesis
of semicorrins of type 5 and higher homologues, in which the
Scheme 1. Iteritive Synthesis of Semicorrins and Tripyrrolines
pyrroline and lactam rings are derived from suitably function-
alized alkyne acids (Scheme 1).1 Alkyne acids 1 are first
converted to imidoyl chlorides or triflates 3 by a four-step
sequence consisting of (1) Pd(II)-catalyzed cyclization; (2)
aminolysis of the resultant ene lactone to give lactam 2; (3)
enamide protection (KCN); and (4) lactam activation employing
either CCl4/PPh3 (X ) Cl) or Tf2O/imidazole (X ) OTf).
Imidoyl derivatives 3 are then transformed to semicorrins 5 by
Pd(0)-mediated coupling-cyclization with alkyne acids 4, fol-
lowed by aminolysis (R ) H, Me). In an analogous fashion,
repetition of this cycle with 5 affords tri- and tetrapyrroline
derivatives.1a An attractive feature of this approach is that the
coupling of 3 and 4 is relatively insensitive to steric factors, in
contrast to more traditional methodology employing thio-
Wittig2a or sulfide contraction protocols.2b
photochemical ring closure to produce corrins (Scheme 2). Es-
chenmoser et al. pioneered this route to corrins in their elegant
synthesis of cobyric acid (7) and then vitamin B12.3,4 We are
engaged in extending our alkyne acid methodology to the
synthesis of 7, which requires a ready source of homochiral
A-D ring synthons 8-11. Linking of these fragments as
described above would then afford a fully functionalized
seco-corrin precursor to 7, corresponding in structural
features to 6.
Semicorrins 5 are themselves versatile building blocks for a
variety of pyrrole-derived natural products. For example, condensa-
tion of 5 with a similarly derived C,D-ring dipyrrin provides direct
access to seco-corrins 6,1 which are properly functionalized for
(1) (a) Jacobi, P. A.; Liu, H. J. Org. Chem. 2000, 65, 7676, and
references therein. (b) Jacobi, P. A.; Tassa, C. Org. Lett. 2003, 5, 4879.
(2) (a) Bishop, J. E.; O’Connell, J. F.; Rapoport, H. J. Org. Chem. 1991,
56, 5079. (b) Eschenmoser, A. Pure Appl. Chem. Suppl. 1971, 2, 69.
Previously, we described an enantioselective synthesis of the
C-ring synthon 10, a key step of which was an Ireland-Claisen
rearrangement of the homochiral allylic ester (S)-12 (Scheme 3).1b,5
10.1021/ol800985m CCC: $40.75
Published on Web 06/07/2008
2008 American Chemical Society

to containing a second, quaternary chiral center (C3), requires
differentiation of the three carboxylate functionalities.
In principle, the desired stereo- and regiochemical features in
11 might be introduced employing either of the allylic ester
derivatives 18 or 20, themselves derived in enantioselective fashion
from enones 16 (Scheme 4). Thus, following path a, reduction of
Scheme 2. Alkyne Acid Precursors to Cobyric Acid (7)
Scheme 4. Potential Precursors to Ring D (11)
The tert-butyldimethylsilyl (TBS) group in (S)-12 served two roles:
(1) as a proton surrogate providing the chirality necessary for
inducing the C-2S configuration in 10, and (2) as an anchor for
Z-16 from the Si face would afford the homochiral allylic alcohol
17, which on coupling with methyl succinate (MeSuc) would give
18. Stereoselective Z-silylenolate formation (blue arrow), with in
situ I-C rearangement, would then produce a bromoalkene acid
having the 2R,3R-configuration found in 11 (analogous to 14 in
Scheme 3). Alternatively, the identical intermediate could be
derived by Re face reduction of E-16 (E-16 f 19), esterification
with MeSuc (19 f 20), and stereoselective E-silylenolate formation
(path b). Si-assisted elimination of HBr would then give 11 in four
steps starting with 16.
Scheme 3. Ireland-Claisen Route to Ring C (10)
We planned to synthesize 16 by regioselective cross-coupling
of an appropriate organocopper or zinc species with R,ꢀ-
dibromoenones 22, themselves derived by Br2 addition to the
corresponding propargyl ketones 21 (Scheme 5). There was
evidence that such additions might be stereoselective.6 However,
in the case of 21a (R ) Me),7a bromination gave an inseparable
∼1:1 E,Z-mixture of addition products 22a, while 21b (R )
TBS)7b afforded variable ratios of dibromides E- and Z-22b,
depending on the reaction conditions. In neither case could the
E- and Z-isomers be distinguished spectroscopically.
Our initial coupling experiments were carried out with 1:1
E,Z mixtures of dibromoenone 22a (R ) Me) to test the activity
of various combinations of organometallic species and catalysts
(Scheme 5). Both isomers proved to be relatively unreactive,
requiring much experimentation to develop suitable coupling
conditions. Even then, we consistently found that only one
isomer of this mixture underwent reaction. Among other
examples, the reagent combination of alkyl zinc iodide 23c (Y
) CH2CO2Me) and TMSCl/CuCN-2LiCl gave a 43% yield
of a single coupling product 16ac,8 together with a near-
quantitative return of one of the isomeric starting materials. This
stabilizing the desired chair conformation in 13. As indicated, the
combination of (S)-configuration in 12 and Z-enolate geometry in
13 afforded carboxylic acid 14 in 78% yield and with ee >95%.
Alkene 14 was then converted directly to the desired alkyne acid
10 by Si-assisted elimination of HBr,1b followed by TBS cleavage.
The structure of 10 was confirmed by its two-step conversion to
the cyclic enamide (S)-(-)-15, an intermediate in Eschenmoser’s
synthesis of 7.4 However, while the Ireland-Claisen (I-C)
methodology was readily adapted to the synthesis of 10, its
application to ring synthons of type 8, 9, and 11 provides a more
stringent test. This is especially the case with 11, which in addition
(3) We gratefully acknowledge many stimulating conversations with
Professor Eschenmoser on this topic.
(4) Eschenmoser, A.; Wintner, C. E. Science 1977, 196, 1410.
(5) Ireland, R. E.; Wipf, P.; Armstrong, J. D., III J. Org. Chem. 1991,
56, 650, and references therein.
(6) Myers, A. G.; Alauddin, M. M.; Fuhry, M. A. M.; Dragovich, P. S.;
Finney, N. S.; Harrington, P. M. Tetrahedron Lett. 1989, 30, 6997.
2838
Org. Lett., Vol. 10, No. 13, 2008

silylenolate formation from (()-26. In preliminary studies, how-
ever, this transformation was only moderately selective using
standard literature conditions,5 and path b was temporarily put aside.
Instead, we opted to devote additional effort to realizing path a in
Scheme 4, although we had thus far been unsuccessful in
synthesizing the requisite starting materials Z-16.
Scheme 5. Preliminary Studies for Synthesizing Ring D
To evaluate path a, enone Z-16a was synthesized by an indirect
route, which provided ample material to test the crucial I-C
rearrangement (Scheme 6). This synthesis began with the readily
Scheme 6. Enantioselective Synthesis of D-Ring Synthon 11a
same reactivity pattern held for dibromoenones 22b (R ) TBS)
in that only one of these isomers underwent clean cross-coupling
with alkylzinc or cuprate reagents. Employing MeCNCuLi, for
example, the reactive isomer gave a ∼80% yield of the known
methylation product 16be (Y ) H).1b Also, the reagent
combination of alkylzinc iodide 23d (Y ) CH2CN) and
Pd(OAc)2/PBu3 gave a 79% yield of bromoenone 16bd,9 still
of unknown geometry.
The stereochemistry of 16bd could only be determined at
a later stage (Scheme 5). Thus, 16bd was first converted to
the I-C precursor (()-20b by NaBH4 reduction to (()-19b
followed by esterification with MeSuc. NOE studies on (()-
20b then strongly suggested the E-configuration as shown
(blue arrow). To confirm this assignment, (()-20b was
subjected to I-C rearrangement employing the more readily
controlled conditions for Z-silylenolate formation (LiHMDS/
TBSCl in THF/HMPA),5 which produced a ∼5:1 mixture
of alkene acids (()-24 (64%). The predominate isomer was
identified as (()-24-anti by its facile conversion to alkyne
acid (()-26 and then ene lactone (()-27, which was
amenable to detailed NOE analysis (blue arrows). Finally,
by the same sequence of reactions, the minor product from
I-C rearrangement of (()-20b was identified as the desired
syn-isomer corresponding to (()-24, which on elimination
gave alkyne acid (()-11-H in Scheme 2.
available ꢀ-bromoenone 28,10 which on Pd-catalyzed coupling with
iodozinc reagent 23d afforded 61-69% yields of the coresponding
E-enone 29. Enone 29 was then converted in 92% yield to a
mixture of R-bromoenones E- and Z-16a, by a one-pot sequence
consisting of bromination followed by dehydrobromination (DBU).
Not surprisingly, this procedure gave essentially 1:1 E,Z-mixtures
of 16a, which, however, were readily separated by chromatography.
Each isomer was then reduced separately to the corresponding
alcohols (()-17a and (()-19a employing NaBH4/CeCl3, at which
point their geometries were confirmed by NOE analysis (cf. blue
arrow in 17a). We were then pleased to find that enone Z-16a
underwent efficient enantioselective reduction with the reagent
system (S)-CBS-Me/BH3-DMS,11 giving a 76% yield of the chiral
alcohol (+)-17a (ee 93%). With a ready supply of (+)-17a now
in hand, the remaining steps leading to the target ring-D precursor
11-Me followed as planned. These involved acylation with MeSuc
to give ester (+)-18a (82%), followed by stereoselective Z-
silylenolate formation employing the same conditions as for ester
(()-20b in Scheme 5. We thus obtained a ∼65% yield of the
alkene acid (-)-30 with dr 10:1. Finally, HBr elimination using
TBAF/DMSO produced (-)-11-Me in 77% yield, completing a
six-step synthesis from enone 28. As with (()-26 in Scheme 5,
the structure of (-)-11-Me was corroborated by conversion to the
With the geometry of (()-24 now established, and by correla-
tion, that of enones E-16bd and E-22b, we set out to reverse the
stereochemical outcome of the Ireland-Claisen rearangement
leading from (()-20b to (()-24. This was to be accomplished
following path b in Scheme 4, requiring stereoselective E-
(7) (a) Hanack, M.; Hassdenteufel, J. R. Chem. Ber. 1982, 115, 764.
(b) Sakaguchi, K.; Fujita, M.; Suzukio, H.; Higashino, M.; Ohfune, Y.
Tetrahedron Lett. 2000, 41, 6589.
(8) (a) Corey, E. J.; Biaz, N. W. Tetrahedron Lett. 1985, 26, 6019.
(9) Mandai, T.; Matsumoto, T.; Tsuji, J.; Saito, S. Tetrahedron Lett.
1993, 34, 2513.
(10) Buono, G. Synthesis 1981, 872.
(11) Corey, E. J.; Helal, C. J. Angew. Chem., Int. Ed. 1998, 37, 1986,
and references therein.
Org. Lett., Vol. 10, No. 13, 2008
2839

corresponding ene lactone (+)-31 (not shown), in this case prepared
by CuI-catalyzed ring closure of bromoalkene acid (-)-30 to avoid
competitive 6-membered ring lactone formation.12a,b
ordering found in methyl ketone Z-16a. Working from this premise,
our initial reduction studies on E-16b were carried out with the
same (S)-CBS-Me/BH3-DMS combination employed in the reduc-
tion of Z-16a to R-(+)-17a, with the expectation that the S-
enantiomer (+)-19b would be favored. This turned out to be the
case, as verified by careful inspection of the corresponding Mosher
esters.14 However, the ee for this transformation was only 27%.
Following extensive screening, the combination of (S)-CBS-Bu
and catechol borane was found to give much better results,
affording (+)-19b in 62% yield and ee 91%.15 Alcohol (+)-19b
then gave a 90% yield of allylic ester (+)-20b on coupling with
MeSuc.
Following path a, we were routinely able to prepare alkyne acid
(-)-11-Me with high enantio- and diastereoselectivity. However,
a number of considerations led us to reexamine the viability of
path b in Scheme 4, in particular utilizing dibromides 22b (R )
TBS). One advantage was that isomers E,Z-22b were readily
separable, in contrast to E,Z-22a (R ) Me). Also, we had
significantly improved upon the ratio of Z-22b/E-22b obtained on
bromination (Scheme 7). As originally effected (Br2, CH2Cl2, -78
The remaining hurdle to be adressed was in developing reliable
conditions for effecting the E-silylenolate Ireland-Claisen rear-
rangement of (+)-20b,16 a transformation that had thus far been
problematic (vide supra). However, a solution was found based
on the recent studies of McIntosh et al.,17 who introduced the
reagent combination of KHMDS/TIPSOTf in Et2O (-78 °C).
While the TIPSOTf reagent proved too sterically hindered for use
with (+)-20b, the combination of excess KHMDS/TBSOTf in
Et2O (-78 °C f rt) routinely afforded ∼85% yields of rearrange-
ment products 2R,3R-32, incorporating an additional TBS group
R to the nitrile (mixture of epimers at C*).18 This was of little
consequence, however, since both silyl groups were cleanly
removed on treatment with TBAF in DMSO, giving a 90% yield
of ring-D synthon (-)-11-H with dr 11:1. As with alkyne acids
(()-26 and (-)-11-Me above, the structure of (-)-11-H was
corroborated by cyclization to the corresponding ene lactone (+)-
33 (not shown), which was subjected to NOE analysis.
Scheme 7. Completion of the Synthesis of (-)-11c
The described six-step route leading from ynone 21b to
alkyne acid (-)-11-H proceeds with excellent enantio- and
diastereoselectivity and provides efficient access to this
important ring-D synthon for cobyric acid (7). Moreover,
we believe that 21b will serve as a common precursor to
each of the remaining ring synthons 8-10.19
°C), bromination of 21b produced a ∼3:5 mixture of isomers E-and
Z-22b, the unreactive Z-isomer predominating.13 In contrast, on
bromination at rt, with catalytc NaI, the desired E-isomer was
favored by >2:1 (62%:27%). These conditions presumably pro-
mote thermodynamic control, since essentially the same product
ratio was established on subjecting pure Z-22b to equilibration with
BF3·Et2O (67% E-22b). By this means we were able to conve-
niently prepare multigram quantities of E-22b for elaboration to
alkyne acid (-)-11-H.
Pd-catalyzed cross-coupling of E-22b with alkylzinc reagent 23d
then gave a ∼80% yield of E-16b (Scheme 7). However, the
assymmetric reduction of E-16b proved to be a greater challenge
than with the related methyl derivative Z-16a, likely because of a
less definitive “size” difference in the substituents attached to the
ketone. Also, the bromoalkene in TBS-ketone E-16b is probably
the “smaller” of the two ketone substituents, a reversal of the
Acknowledgment. Financial support of this work by the
National Institutes of Health, NIGMS Grant No. GM38913,
is gratefully acknowledged.
Supporting Information Available: Experimental and
NMR spectra for all new compounds, NOE studies, and
X-ray crystal structure for (()-32 (major epimer). This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL800985M
(14) Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc. 1973, 95, 512, and
references therein.
(15) Corey, E. J.; Bakshi, R. Tetrahedron Lett. 1990, 31, 611.
(16) These studies were initially carried out with racemic (()-20b.
(17) McFarland, C.; Hutchison, J.; McIntosh, M. C. Org. Lett. 2005, 7,
3641, and references therein.
(12) Modeled on the conditions of Buchwald et al. for amidation of
vinyl bromides: (a) Jang, L.; Job, G. E.; Klapars, A.; Buchwald, S. L. Org.
Lett. 2003, 5, 3667. (b) In contrast to terminal alkyne acids of type 11-H,
internal alkyne acids of type 11-Me give mixtures of 5- and 6-membered
ring lactones on cyclization with PdCl2.
(13) The low reactivity of Z-22a and -b is surprising, since numerous
examples of selective ꢀ-coupling of cyclic R,ꢀ-dibromoenones are known.
See, for example: Pelphrey, P. M.; Orugunty, R. S.; Helmich, R. J.; Battiste,
M. A.; Wright, D. L. Eur. J. Org. Chem. 2005, 4926.
(18) The structure of the predominate (S)-relative epimer was confirmed
by X-ray analysis of a racemic sample (cf. ref 16). We thank the University
of Massachusetts Amherst X-ray Structural Characterization Facility (NSF
CHE 9974648) for providing diffractometer access and Mr. Travis Benanti
of that facility for data collection and refinement.
(19) A reviewer commented that the term “synthon” as originally defined
referred to an abstract entity arising from a retrosynthetic analysis (typically
a radical, cation or anion). Presently, though, this term is more commonly
associated with synthetic precursors or “equivalents”.
2840
Org. Lett., Vol. 10, No. 13, 2008