Enantioselective syntheses of ring-C precursors of vit. B12. Reagent control

Article abstract of DOI:10.1021/ol0275116

(Figure presented) Enelactones of the general structure S-(-)-I were prepared in three steps from alcohol 21 and acids 22 (ee ≈ 85%). Lactones S-(-)-I are versatile precursors to enelactams II of the type found in Vitamin B₁₂.

Full text of DOI:10.1021/ol0275116

ORGANIC
LETTERS
2003
Vol. 5, No. 5
701-704
Enantioselective Syntheses of Ring-C
Precursors of Vit. B12. Reagent Control
Peter A. Jacobi* and Yongkai Li
Burke Chemical Laboratory, Dartmouth College, HanoVer, New Hampshire 03755
peter.jacobi@dartmouth.edu
Received December 19, 2002
ABSTRACT
Enelactones of the general structure S-(−)-I were prepared in three steps from alcohol 21 and acids 22 (ee ≈ 85%). Lactones S-(−)-I are
versatile precursors to enelactams II of the type found in Vitamin B12.
In a recent series of papers, we described a general synthesis
of semicorrins of type 4, in which the A- and B-rings were
derived from suitably functionalized alkyne acids (Figure 1).1
Acids 1 were first converted to imidoyl chlorides 2 by a four-
step sequence involving (1) Pd(II)-catalyzed cyclization, (2)
aminolysis of the resultant enelactone followed by cyclo-
dehydration, (3) enamide protection (KCN), and (4) chlori-
nation using CCl4/PPh3. Imidoyl chlorides 2 were then
transformed to semicorrins 4 by Pd(0)-mediated coupling/
cyclization with alkyne acids 3 followed by aminolysis. A
significant advantage to this route is that the coupling of 2
and 3 is relatively insensitive to steric factors, in contrast to
more traditional methodology employing thio-Wittig2 or
sulfide contraction protocols.3 Therefore, meso-substituents
R can be incorporated directly into the semicorrin ring.
Semicorrins 4 are important building blocks for a variety
of linear and macrocyclic tetrapyrroles. For example, repeti-
tion of the sequence of enamide activation and Pd(0)-
mediated coupling-cyclization affords tripyrrolines and
higher analogues.1a Alternatively, condensation of 4 with a
similarly derived C,D-ring dipyrrin provides direct access
to seco-corrins 5,1c which are properly functionalized for
photochemical ring closure to produce corrins (Figure 1).
Eschenmoser pioneered this route to corrins in his extra-
ordinary synthesis of Vitamin B12.4 We are investigating
using the alkyne acid methodology for the synthesis of
Cobyric Acid (10), a known precursor to Vitamin B12 (Figure
2). Our initial objective was to develop enantioselective
syntheses of alkyne acids 6-9 or closely related synthons.
(1) (a) Jacobi, P. A.; Liu, H. J. Am. Chem. Soc. 1999, 121, 1958. (b)
Jacobi, P. A.; Liu, H. J. Org. Chem. 1999, 64, 1778. (c) Jacobi, P. A.; Liu,
H. Organic Lett. 1999, 1, 341. (d) Jacobi, P. A.; Liu, H. J. Org. Chem.
2000, 65, 7676.
(2) Bishop, J. E.; O’Connell, J. F.; Rapoport, H. J. Org. Chem. 1991,
56, 5079.
(3) (a) Eschenmoser, A. Pure Appl. Chem. Suppl. 1971, 2, 69. (b) Arnott,
D. M.; Harrison, P. J.; Henderson, G. B.; Sheng, Z.-C.; Leeper, F. J.;
Battersby, A. R. J. Chem. Soc., Perkin Trans. 1 1989, 265.
(4) (a) Eschenmoser, A.; Wintner, C. E. Science 1977, 196, 1410, and
references therein.
Figure 1. Iterative synthesis of tetrapyrrole derivatives.
10.1021/ol0275116 CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/07/2003

Figure 3. Ester enolate Claisen route to syn-alkyne acids.
both of these approaches for controlling absolute stereo-
chemistry in alkyne acids of type 14. In this Letter we
describe our results using reagent control to synthesize ring-C
analogues of Vitamin B12.8
Our initial experiments were carried out with the model
system 17 to test the utility of the Ireland-Claisen re-
arrangement for preparing alkyne acids (Scheme 1). Racemic
Figure 2. Possible alkyne precursors for Cobyric Acid.
Alkyne acids 6-9 share a number of features in common
(Figure 2).5 Each has a C-3 quaternary center, and at least
one of these substituents is methyl. Also, in 6, 7, and 9, the
orientation of the acetate and propionate groups is syn. We
planned to establish these relationships employing a variant
of the Ireland-Claisen rearrangement, a powerful method
for synthesizing 1-pentenoic acid derivatives (Figure 3).6 In
principle, the alkyne oxidation level can be attained by
incorporating a leaving group “X” in allylic esters of type
11. Following 3,3-sigmatropic rearrangement to 13, elimina-
tion of HX would provide the desired alkyne 14.
Scheme 1
The stereochemical outcome depicted in 14 has excellent
precedent.6 Diastereoselectivity in this transformation is
controlled by both enolate and double-bond geometry, with
the stipulation that reaction occurs through the most stable
chair conformation. As indicated, the desired syn-selectivity
would be obtained from the (Z)-enolate-(Z)-alkene config-
uration of 12. Control of absolute stereochemistry is also
precedented and might be accomplished in one of two ways.
When R * H, C-3 is a chiral center that can be introduced
in enantioselective fashion or by alcohol resolution (substrate
control).6b Alternatively, with R ) H, facial selectivity might
be achieved using a chiral Lewis Acid (M*-Br; reagent
control). Corey et al. have reported promising results in this
area employing the boron reagent 15.7 We have investigated
17 was conveniently prepared by propionylation of the allylic
alcohol 16 (EtCOCl/pyr), itself derived in ∼90% overall yield
from mesityl oxide.9 We explored a number of procedures
for effecting the rearrangement of 17 to 19 under achiral
conditions. However, the best results were obtained employ-
ing the classic Ireland conditions.6a-c This involved silylation
(5) Acids 6 and 7 are identical except for the C-5 alkyne substituent (H
vs Me).
(6) (a) Ireland, R. E.; Mueller, R. H. J. Am. Chem. Soc. 1972, 94, 5897.
(b) Ireland, R. E.; Varney, M. D. J. Am. Chem. Soc. 1984, 106, 3668. (c)
Ireland, R. E.; Wipf, P.; Armstrong, J. D., III. J. Org. Chem. 1991, 56,
650. (d) Koch, G.; Janser, P.; Kottirsch, G.; Romero-Giron, E. Tetrahedron
Lett. 2002, 43, 4837, and references therein.
(7) Corey, E. J.; Lee, D. J. Am. Chem. Soc. 1991, 113, 4026.
(8) Use of substrate control will be described in a following paper.
(9) Mori, H.; Matsuo, T.; Yamashita, K.; Katsumura, S. Tetrahedron
Lett. 1999, 40, 6461.
702
Org. Lett., Vol. 5, No. 5, 2003

of 17 at -78 °C with 1.1 equiv each of LiHMDS/TBSCl,
using a solvent combination expected to favor (Z)-enolate
formation (THF/HMPA). No effort was made to isolate the
presumed intermediate 18, which was cleanly transformed
to the (Z)-bromoalkene 19 upon warming to room temper-
ature (75-85%).10 Finally, 19 afforded 45-50% yields of
the alkyne acid 20 upon treatment with NaH in DMF (not
optimized).
Scheme 3
We also tested the compatibility of the achiral Ireland-
Claisen rearrangement with sensitive functionality (Scheme
2). Allylic esters 23a-c were prepared by acylation of the
Scheme 2
Most likely, the nonreactive nature of 29, 31, and 32 stems
from a combination of factors. In the case of 29 an important
issue is competitive ester enolization, but with 31 and 32,
the principal effects are probably steric. Reagent 15 derives
much of its selectivity from its size and structural rigidity,7
both of which contribute to steric crowding. These interac-
tions are accentuated during 3,3-sigmatropic rearrangement
due to the formation of a quaternary center. Finally, the large
bromine atom imparts additional strain into what is already
a high-energy transition state, thereby inhibiting reaction.
This rationale is supported by experiments carried out with
the desbromo substrates 23a-d (Scheme 4). As with 29
commercially available alcohol 21 with carboxylic acid
derivatives 22a-c (X ) OH, Cl). As with the allylic ester
17 (cf. Scheme 1), 23b and 23c gave high yields of alkene
acids 25 and 26 using the Ireland protocol. However, ester
23a presented a special case, since competitive deprotonation
occurred at the R-position of the carbomethoxy group. As a
result we obtained only trace amounts of the desired alkene
24 under standard conditions. Interestingly, however, similar
substrates undergo clean rearrangement utilizing 2.2 equiv
of LiHMDS/TBSCl.11
Scheme 4
We next studied the reactivity of allylic esters 29, 31, and
32 with Lewis acids (Scheme 3). Ester 29 was prepared in
93% yield by condensation of acid chloride 22a with allylic
alcohol 28, itself derived by bromination of alkene 21.12 In
analogous fashion, esters 31 and 32 were obtained by DCC-
mediated coupling of 28 with the appropriate carboxylic acids
22e,f. Allylic ester 29 was then reacted with the Corey
reagent 15 in an attempt to effect 3,3-sigmatropic rearrange-
ment. Using the literature conditions, we obtained only trace
amounts of the desired product S-30 after several days at
temperatures from -20 to 0 °C.7 Similarly, we observed no
reaction employing the Oh reagent (-)-Ipc2BCl13 or with
achiral reagents such as Bu2BOTf.
above (cf. Scheme 3), allylic esters 23a,b failed to undergo
3,3-sigmatropic rearrangement, presumably due to competing
complexation of 15 with the carbomethoxy or nitrile groups.
In contrast, substrates 23c,d were transformed relatively
smoothly to the corresponding alkene acids S-(-)-26 and
S-(-)-27 with ee ≈ 85%.14 The isolated yields of these
materials depended strongly upon the concentration of 15
and reached a maximum of ∼50% utilizing a 3-fold excess
(10) Geometry of 19 was established by NOE studies, which showed a
strong interaction between the vinyl C-H and the geminal methyl groups
(curved arrow).
(11) Jacobi, P. A.; Tassa, C. Manuscript in preparation.
(12) Ito, M.; Koyama, T.; Ogura, K. Chem. Lett. 1986, 101.
(13) Oh, T.; Wrobel, Z.; Devine, P. N. Synlett. 1992, 81.
Org. Lett., Vol. 5, No. 5, 2003
703

(∼35% yield with 2.0 equiv 15). After this point, no further
improvement was realized with either additional 15 or longer
reaction periods. The reason for this behavior is unclear.
These transformations are quite clean with respect to
byproduct formation, affording >90% yields of S-(-)-26 and
S-(-)-27 based upon recovered 23c,d.
Scheme 6
With reasonable quantities of alkene acids S-(-)-26 and
S-(-)-27 in hand, we devoted considerable effort to oxidiz-
ing these materials to the corresponding alkynes. These
experiments were not fruitful. However, both S-(-)-26 and
S-(-)-27 were readily converted to the corresponding
enelactones 34 and 36. With S-(-)-27, this was initially
accomplished by a sequence involving iodolactonization to
afford S-33 (99%), followed by based-induced elimination
(Scheme 5). Unfortunately, however, dehydroiodination
Scheme 5
Finally, oxidation of S-(-)-39 with PDC/MeOH gave a 70%
overall yield of the lactone ester S-(-)-41.16b This ma-
terial is in the proper oxidation state for direct conversion
to ring-C analogues of Vitamin B12. Alternatively, oxidation
of S-(-)-39 with PDC/t-BuOH provided the tert-butyl ester
S-(-)-42 (31%, not optimized),17 with little or no loss in
optical activity. Lactone S-(-)-42 has previously been
described by Mulzer et al., who obtained a 90% yield of
enelactam S-(-)-43 upon aminolysis/cyclodehydration.18
occurred with complete racemization to give (()-34 in 57%
overall yield. Much more satisfactory results were obtained
employing the reagent system PdCl2/CuCl2/O2,15 which
afforded an 84% yield of S-(-)-34 directly (ee ) 86%).14
Finally, aminolysis of S-(-)-34 and cyclodehydration gave
a 40% yield of the enelactam S-(-)-35 (not optimized).1
In the case of alkene acid S-(-)-26, oxidative cyclization
provided the enelactone S-(-)-36 in 52% yield with ee )
84% (Scheme 6).14 However, most attempts at removing the
TBDPS protecting group gave the rearranged lactone 37.16a
This difficulty was circumvented by carrying out deprotection
of S-(-)-26 first (TBAF), which afforded a 90% yield of
the alcohol acid S-(-)-38. Cyclization then took place
normally to give the enelactone S-(-)-39 in 60% yield.
Acknowledgment. Financial support of this work by the
National Institutes of Health, NIGMS Grant GM38913, is
gratefully acknowledged. We are grateful to Mr. Carlos Tassa
of Dartmouth College for carrying out the experiments
described in Scheme 1.
Supporting Information Available: Experimental pro-
cedures and NMR spectra for all new compounds reported.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL0275116
(16) (a) Modest yields of S-(-)-39 were obtained using pyridine/HF;
cf.: Nicolaou, K. C.; Seitz, S. P.; Pavia, M. R.; Petasis, N. A. J. Org. Chem.
1979, 44, 4011. (b) O’Connor, B.; Just, G. Tetrahedron Lett. 1987, 28,
3235. Oxidation with PDC/CH2Cl2 gave a 78% yield of the aldehyde
S-(-)-40.
(14) Enantiomeric excess (ee) was determined at the enelactone stage
employing a Chiralpak AD column (cf. Supporting Information).
(15) Tanaka, M.; Urata, H.; Fuchikami, T. Tetrahedron Lett. 1986, 27,
3165.
(17) Corey, E. J.; Samuelsson, B. J. Org. Chem. 1984, 49, 4735.
(18) Mulzer, J.; Riether, D. Tetrahedron Lett. 1999, 40, 6197.
704
Org. Lett., Vol. 5, No. 5, 2003