New syntheses of E7389 C14-C35 and halichondrin C14-C38 building blocks: Reductive cyclization and oxy-Michael cyclization approaches

Article abstract of DOI:10.1021/ja9058487

Cr-mediated coupling reactions are usually achieved with a slight excess of a given nucleophile. To develop a cost-effective use of this process, two different approaches have been studied. The first approach depends on two consecutive catalytic asymmetri

Full text of DOI:10.1021/ja9058487

Published on Web 10/06/2009
New Syntheses of E7389 C14-C35 and Halichondrin
C14-C38 Building Blocks: Reductive Cyclization and
Oxy-Michael Cyclization Approaches
Cheng-Guo Dong, James A. Henderson, Yosuke Kaburagi, Takeo Sasaki,
Dae-Shik Kim, Joseph T. Kim, Daisuke Urabe, Haibing Guo, and Yoshito Kishi*
Department of Chemistry and Chemical Biology, HarVard UniVersity, 12 Oxford Street,
Cambridge, Massachusetts 02138
Received July 20, 2009; E-mail: kishi@chemistry.harvard.edu
Abstract: Cr-mediated coupling reactions are usually achieved with a slight excess of a given
nucleophile. To develop a cost-effective use of this process, two different approaches have been studied.
The first approach depends on two consecutive catalytic asymmetric Cr-mediated couplings, with use
of coupling partners purposely being of unbalanced molecular size and complexity. The second approach
rests on the success in identifying the nucleophile, which allows us to achieve the coupling satisfactorily
with a 1:1 molar ratio of the coupling partners. The C23-O bond is stereospecifically constructed via
reductive cyclization of the oxonium ion, or oxy-Michael cyclization. Both syntheses have a high overall
efficiency: E7389 C14-C35 and halichondrin C14-C38 building blocks have been synthesized from
the corresponding C27-C35 and C27-C38 aldehydes, respectively, in high overall yields with an
excellent stereoselectivity. Because of operational simplicity, the synthesis outlined herein appears to
be well suited for scaling.
Scheme 1. Synthesis of E7389 C14-C35 Building Block via
Catalytic Asymmetric Ni/Cr-Mediated Coupling of 1a with 2aa
1. Introduction
In the preceding paper,1 we discussed two convergent
approaches to synthesize E7389 C14-C35 and halichondrin
C14-C38 building blocks. For example, a catalytic asymmetric
Ni/Cr-mediated reaction is used to couple the two building
blocks, e.g., 1a + 2a f 3a (Scheme 1). Two aspects of this
coupling are noteworthy. First, the catalyst, prepared in situ from
CrCl2 and (S)-sulfonamide A, allows us to achieve the coupling
with excellent yield and stereoselectivity (yield ) 91%; dr )
>55:1). Second, this coupling gives a high level of convergency,
as the target molecule is dissected into two coupling partners
with roughly equal molecular size and structural complexity.
However, it is achieved best with use of 1.5 equiv of the vinyl
iodide for this Ni/Cr-mediated coupling. For this reason, it is
not necessarily most cost-effective on the basis of 2a, unless
the synthetic cost of 2a compensates the 1.5:1 molar ratio of
the two coupling partners.2 In this paper, we will present two
approaches to address this issue.
a Catalytic asymmetric Ni/Cr-mediated coupling in the presence of the
Cr catalyst derived from (S)-sulfonamide A, followed by AgBF4-promoted
cyclization. Abbreviation: TBDPS ) t-Bu(Ph)2Si-.
2. Results and Discussions
1.5 equiv, of a given nucleophile.3 Therefore, for cost-effective
use of Cr-mediated couplings, one needs to pay attention not
only to the relative molecular size and structural complexity of
the coupling partners but also to their molar ratio. One obvious
solution is to dissect a target molecule accordingly. Scheme 2
outlines such an approach. Taking these factors together into
account, the target molecule 5a is dissected into vinyl iodide 4
2.1. First Approach: Unbalancing the Molecular Size and
Structural Complexity of Coupling Partners. 2.1.1. C23-O
Bond-Formation via SN2 Cyclization. Cr-mediated coupling
reactions are usually carried out with a slight excess, typically
(1) Kim, D.-S.; Dong, C.-G.; Kim, J. T.; Guo, H.; Huang, J.; Tiseni, P. S.;
Kishi, Y. J. Am. Chem. Soc. 2009, 131, http://dx.doi.org/10.1021/
ja9058475.
(2) Two major modifications have been made on the synthesis of 1a,
reported in ref 2c in the preceding paper. In our estimation, the cost
for the synthesis of 1a now appears to be very close to, even lower
than, that for 2a.
(3) Guo, H.; Dong, C.-G.; Kim, D.-S.; Urabe, D.; Wang, J.; Kim, J. T.;
Liu, X.; Sasaki, T.; Kishi, Y. J. Am. Chem. Soc. 2009, 131, http://
dx.doi.org/10.1021/ja905843e and references cited therein.
9
15642 J. AM. CHEM. SOC. 2009, 131, 15642–15646
10.1021/ja9058487 CCC: $40.75  2009 American Chemical Society

New Syntheses of E7389 and Halichondrin Building Blocks
A R T I C L E S
Scheme 2. Synthesis of E7389 C20-C35 Building Block 5aa
Scheme 4. Attempted Synthesis of E7389 C20-C35 Building
Block 3b via the Oxonium Route
triethylsilane (10 equiv) at -78 °C, the crude allylic alcohols
gave the cyclized products with the C20 silyl protecting group
intact.
When this reaction mixture was warmed to 0 °C before the
workup, the C20 silyl protecting group was cleanly removed
in situ, to furnish directly the deprotected product 8b in 95%
yield from the Ni/Cr-mediated coupling product. NMR and TLC
analyses demonstrated that the reductive cyclization did not yield
the undesired C23-diastereomer at detectable levels.6
a Reagents and conditions: (a) catalytic asymmetric Ni/Cr-mediated
coupling in the presence of the Cr catalyst derived from (S)-sulfonamide
A. (b) KHMDS/THF/-15 °C. Abbreviation: KHMDS ) KN(SiMe3)2.
Scheme 3. Synthesis of E7389 C20-C35 Building Block 8ba
Related to the synthesis outlined in Scheme 3, we should
comment on our attempt to extend the oxonium ion approach
to the vinyl iodide 9 (Scheme 4).7 The catalytic asymmetric
Ni/Cr-mediated coupling worked well, but the following reduc-
tive cyclization gave a mixture of two products. On comparison
with the authentic sample, one of the products was found to be
the desired product. The spectroscopic data (MS and NMR)
suggested the other product to be 10, which presumably arose
through participation of the C19-exocyclic olefin to the oxonium
ion II. In spite of extensive efforts (varying reducing agents,
Lewis acids, and solvents), we were unable to prevent formation
of this byproduct.
The primary alcohol 8b was oxidized to the corresponding
aldehyde 11b and then subjected to catalytic asymmetric Ni/
Cr-mediated coupling, followed by base-induced cyclization,
to furnish stereochemically homogeneous E7389 C14-C35
building block 3b (Scheme 5). This transformation deserves
several comments. First, considering the unbalanced molec-
ular size and structural complexity, the coupling was carried
out with an excess (1.5 equiv) of the vinyl iodide to give the
products in 85% yield based on the aldehyde.8 However, it
is noteworthy that the coupling efficiency was not noticeably
affected by the use of 1.2 equiv of vinyl iodide. Second, as
discussed in the preceding paper, asymmetric induction (dr
) ca. 32:1) was best achieved with the Cr catalyst derived
from (R)-sulfonamide B. Interestingly, this Cr catalyst gave
a higher asymmetric induction for a structurally complex
aldehyde such as 11b than for structurally simple aldehydes
previously reported.1 Third, the base-induced cyclization was
effected with KH in toluene containing 18-crown-6 at -78
°C f -20 °C with no detectable level of olefin byproduct.
a Reagents and conditions: (a) Catalytic asymmetric Ni/Cr-mediated
coupling in the presence of the Cr catalyst derived from (S)-sulfonamide
A. (b) Et3SiH/TMSOTf/CH2Cl2/-78 °C f 0 °C. Abbreviation: Bz )
PhCO-; TMSOTf ) (Me)3SiOSO2CF3.
and aldehyde 1a. The catalytic asymmetric Ni/Cr-mediated
coupling of 1a with 4 (1.5 equiv) furnished the desired coupling
product in ca. 80% yield with excellent stereoselectivity (dr )
ca. 50:1).
Upon treatment with a base, the resultant allylic alcohols gave
the desired cyclized product in ca. 70% yield. However, as
observed before,1 the base-induced cyclization was accompanied
with an olefin byproduct (ca. 10%). In order to overcome this
drawback, we studied an oxonium ion route.
2.1.2. C23-O Bond Formation via Reductive Cyclization.
Once again, the catalytic asymmetric coupling of aldehyde 1b
with vinyl iodide 64 (1.5 equiv) furnished the expected, desired
coupling product 7b in excellent yield (92%) and stereoselec-
tivity (dr ) ca. 50:1) (Scheme 3).
(5) (a) Lewis, M. D.; Cha, J. K.; Kishi, Y. J. Am. Chem. Soc. 1982, 104,
4976. For recent examples, see: (b) Ayala, L.; Lucero, C. G.; Romero,
J. A. C.; Tabacco, S. A.; Woerpel, K. A. J. Am. Chem. Soc. 2003,
125, 15521. (c) Henderson, J. A.; Jackson, K. L.; Phillips, A. J. Org.
Lett. 2007, 9, 5299.
Based on previous work from this laboratory,5 we anticipated
the desired C23-diastereomers to be preferentially formed on
reductive cyclization of the resultant allylic alcohols through
an oxonium ion such as I. Experimentally, on treatment with
trimethylsilyl triflate (2 equiv) in methylene chloride containing
(6) The ethylene glycol ketal corresponding to 6 gave virtually the same
results as summarized in Schemes 3 and 5.
(7) Xie, C.; Nowark, P.; Kishi, Y. Org. Lett. 2002, 4, 4427. For an
alternative synthesis of this alcohol, see the Supporting Information
in one of our preceding papers.
(8) The catalytic asymmetric Ni/Cr-mediated coupling with the vinyl
iodide bearing C17 R-2,4,6-trimethylbenzenesulfonate (2 equiv) gave
results similar to 12.
(4) For the preparation of vinyl iodide 6, see the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 131, NO. 43, 2009 15643

A R T I C L E S
Dong et al.
Scheme 5. Synthesis of E7389 C14-C35 Building Block 3ba
series, the C30 protecting group (PMB ) p-methoxybenzyl)
was concomitantly removed under the conditions given in
Scheme 3. Therefore, a stepwise protocol was used to obtain
the C20 primary alcohol, i.e., reductive cyclization at -78 °C,
followed by TBDPS deprotection. The overall yield in the
halichondrin series was 60%.10,11
In summary, E7389 C14-C35 building block 3b and
halichondrin C14-C38 building block 15b have been
synthesized from 1b and 14b in five and six steps, respec-
tively, in 60-65% overall yields with excellent stereoselec-
tivities. This synthesis route relies on (1) two consecutive
catalytic asymmetric Cr-mediated couplings, utilizing cou-
pling partners with unbalanced molecular size and complexity
for consideration of cost-effectiveness, and (2) stereoselective
reductive cyclization to form the C23-O bond. We should
also note that none of the steps outlined in Schemes 3 and 5
presents any apparent problem on scaling. In our view, this
synthesis offers several appealing features, including high
overall yield, high stereoselectivity, operational simplicity,
and overall cost-effectiveness.
2.2. Second Approach: Convergent Synthesis with
Approximately Equal Size and Complexity of Coupling
Partners in a ca. 1:1 Molar Ratio. In order to take advantage
of the highest degree of convergency, it is ideal to find a
coupling reaction that can be efficiently achieved with a 1:1
mixture of coupling partners with approximately equal molecular
size and structural complexity. The case shown in Scheme 1
used coupling partners with approximately equal molecular size
and structural complexity. However, to achieve this coupling
with a high efficiency, an excess (1.5 equiv) of the nucleophile
was required.
a Reagents and conditions: (a) Catalytic asymmetric Ni/Cr-mediated
coupling in the presence of the Cr catalyst derived from (R)-sulfonamide
B. (b) KH/18-crown-6/toluene/-78 °C f -20 °C. Abbreviation: Pv )
(Me)3CCO-.
Scheme 6. Synthesis of Halichondrin C14-C38 Building Block
15ba
As previously noted, the Ni/Cr-mediated couplings involve
at least four discrete steps.3 Based on the X-ray structure for
two Cr-sulfonamide complexes,12 we speculated that a group
or atom with a high Cr affinity present in the nucleophile could
occupy the coordination site for the electrophile. With this
speculation, we proposed a possible approach to improve the
catalyst loading.3 At the same time, we noticed that this working
hypothesis might be applicable to improve the molar ratio of
coupling partners.
With this background, we searched for a nucleophile with
no capacity for coordination to the Cr metal center and found
that the C14-C26 building block 16, previously reported from
this laboratory,13 exhibits the ideal property for our purpose.
Even with a 1:1 molar ratio of the coupling partners,14 the
catalytic asymmetric Ni/Cr-mediated coupling in the presence
of the Cr catalyst derived from (S)-sulfonamide A proceeded
smoothly to furnish the expected, desired product 17 in ca. 80%
yield in both the E7389 and halichondrin series (Scheme 7).
For preparative purposes, we routinely used a 1:1.2 molar ratio
and obtained the coupling product in a better than 85% yield.
a Reagents and conditions: see Schemes 3 and 5. Abbreviation: TBAF
) tetra-n-butylammonium fluoride; PMB ) p-methoxybenzyl.
The minor C20 diastereomer was readily removed by flash
chromatography on silica gel.
The synthesis in Scheme 5 was illustrated with the vinyl
iodide 12 with the TBDPS protecting group at C14. It is worth
noting that (1) the transformation of 11b + 12 f 13b f 3b
was equally effective for two additional vinyl iodides bearing
OPv and 2,2-dimethylpropylene acetal at C14 and (2) the
C34 and C35 benzoate protecting groups were selectively
removed without affecting the C14 protecting group in all
the series.9 Thus, the protecting group in 3b is readily
adjustable as required.
(10) As expected, a change of the C30-protecting group from PMB to
MeOCH2CO allowed for the reductive cyclization and C20-deprotec-
tion in one pot in the halichondrin series as well.
(11) The ethylene glycol ketal corresponding to 6 gave virtually the same
results as summarized in Schemes 3 and 5.
(12) Wan, Z.-K.; Choi, H.-w.; Kang, F.-A.; Nakajima, K.; Demeke, D.;
Kishi, Y. Org. Lett. 2002, 4, 4431.
(13) See ref 2b in the preceding paper.
Halichondrin C14-C38 building block 15b was synthesized
via the same synthetic sequence, with one minor modification
(Scheme 6). On the reductive cyclization in the halichondrin
(14) The following yields were obtained with varied ratios of the coupling
partners: 1a:16 ) 1:1 (85% yield), 1:1.2 (91%), 1:1.5 (90%). Similar
results were observed for the coupling with the Cr catalyst derived
from (S)-sulfonamide with R1 ) i-Pr, R2 ) Me, and R3 ) OMe: 1:1
(85%), 1:1.2 (90%), and 1:1.8 (>90%). However, the stereoselectivity
with this catalyst was ca. 15:1.
(9) For details, see the Supporting Information.
9
15644 J. AM. CHEM. SOC. VOL. 131, NO. 43, 2009

New Syntheses of E7389 and Halichondrin Building Blocks
A R T I C L E S
Scheme 7. Oxy-Michael Route to E7389 C14-C35 Building Block
Scheme 8. Synthesis of Halichondrin C14-C38 Building Block
15ca
3aa
a Reagents and conditions: (a) Catalytic asymmetric Ni/Cr-mediated
coupling in the presence of the Cr catalyst derived from (S)-sulfonamide A
(86% yield). (b) t-BuOK/CH2Cl2/0 °C (88% yield). (c) LiI/pyridine/110
°C (90% corrected yield). (d) 2,4,6-Cl3PhCOCl/ polymer-supported Hu¨nig
base, followed by 1-hydroxypyridine-2-thione sodium salt/C6H6. (e) t-BuSH/
hν (82% overall yield for two steps).
Scheme 9. Synthesis of Vinyl Iodide 16a
a Reagents and conditions: (a) Catalytic asymmetric Ni/Cr-mediated
coupling in the presence of the Cr catalyst derived from (S)-sulfonamide
A. (b) t-BuOK/CH2Cl2/0 °C. (c) LiI/pyridine/110 °C. (d) 2,4,6-Cl3PhCOCl/
polymer-supported Hu¨nig base, followed by 1-hydroxypyridine-2-thione
sodium salt/C6H6. (e) t-BuSH/hν.
The vinyl iodide 16 used in this study was a 1.4:1 mixture
of E- and Z-unsaturated esters. Interestingly, E- and Z-isomers
were consumed at roughly the same rate to yield the coupling
products composed of a 1.4:1 mixture of E- and Z-unsaturated
esters. As might be expected, E- and Z-isomers of 16 gave
products with similar stereoselectivity; for example, in the
coupling with 1a, a stereoselectivity of ca. 55:1 was observed
for both E- and Z-isomers.15
a Reagents and conditions: (a) Catalytic asymmetric Ni/Cr-mediated
coupling in the presence of Cr catalyst prepared from (R)-sulfonamide C.
(b) 1. KH/18-crown-6/toluene/-78 °C f -20 °C; 2. AcCl/MeOH.
The allylic alcohols were then subjected to the base-induced
oxy-Michael reaction to furnish the cyclization products in 90%
yield with excellent stereoselectivity, e.g., 17 f 18. NMR and
TLC analyses at this stage, as well as after decarboxylation,
showed that the oxy-Michael reaction yielded no C23-diaste-
reomer at detectable levels.
of molecular oxygen; otherwise byproduct hydroxylated at C22
was formed in up to ca. 10% yield.19
With the same sequence of reactions, the halichondrin
C14-C38 building block 15c was synthesized (Scheme 8). The
overall yield (∼55%) in this series was slightly lower than that
in the E7389 series, primarily due to a partial cleavage of the
C38 silyl protecting group on the LiI-promoted transformation
of the methyl ester to the carboxylic acid.
Last, we should comment on the new synthesis of 16. Once
again, we relied on the catalytic asymmetric Ni/Cr-mediated
reaction to couple aldehyde 20 and vinyl iodide 12 (Scheme
9). Under the conditions discussed in the preceding paper,1
the Cr catalyst derived from (R)-sulfonamide C smoothly
effected the coupling to furnish the desired allylic alcohol
21 (dr ) 22:1). Interestingly, for coupling of aldehyde 20
with vinyl iodide 12, the sulfonamide C performed better
than the sulfonamide reported in the preceding paper. The
After hydrolysis of the C20 methyl ester to the corresponding
carboxylic acid, the oxy-Michael product was subjected to
Barton decarboxylation, e.g., 18 f 19 f 3a;16 this process was
best achieved with use of Yamaguchi’s protocol for activation
of carboxylic acids17 to furnish the E7389 C14-C35 building
block in 65% overall yield from aldehyde 1a.18 It is worth noting
that the decarboxylation should be carried out in the absence
(15) Although tedious, E- and Z-unsaturated esters 17 were chromato-
graphically separable. Using the separated E- and Z-unsaturated esters,
we confirmed the stereoselectivity for each diastereomer.
(16) (a) Barton, D. H. R.; Dowlatshahi, H. A.; Motherwell, W. B.; Villemin,
D. J. Chem. Soc., Chem. Commun. 1980, 732. (b) Barton, D. H. R.;
Crich, D.; Motherwell, W. B. J. Chem. Soc., Chem. Commun. 1983,
939.
(17) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M. Bull.
Chem. Soc. Jpn. 1979, 52, 1989.
(19) Using the nucleophile with an R,ꢀ-unsaturated nitrile group in 16,
the synthesis paralleling the one shown in Scheme 7 was carried out
to yield the oxy-Michael product in a high overall yield. However,
attempted reductive decyanation and DIBAL reduction/decarbonylation
did not give promising results.
(18) In the halichondrin series, the TBS group at C38 was partially
deprotected on treatment with LiI/py/120 °C. Thus, the reaction was
stopped at approximately 75% completion.
9
J. AM. CHEM. SOC. VOL. 131, NO. 43, 2009 15645

A R T I C L E S
Dong et al.
Scheme 10. Summary of the Current E7389 Synthesis from This
crude product was then subjected to the base-induced
cyclization, followed by acid treatment in methanol, to yield
22 in 80% overall yield from 12 as a 22:1 diastereomeric
mixture at C20. The undesired C20-diastereomer was readily
removed with medium-pressure column chromatography or
selective hydrolysis of the corresponding acetate by Amano
lipase PS-800 to afford stereochemically homogeneous 22.
Following the method previously reported,13 22 was then
converted to 16 in a good overall yield.
Laboratorya
3. Conclusion
In summary, Cr-mediated coupling reactions are usually
achieved best with an excess of a given nucleophile, calling
attention not only to the relative molecular size and structural
complexity of the coupling partners but also to their molar
ratio. To employ this reaction in a cost-effective manner,
two different approaches have been studied. The first
approach depends on two consecutive catalytic asymmetric
Cr-mediated couplings with use of coupling partners pur-
posely being of unbalanced molecular size and complexity.
The second approach rests on the success in identifying the
nucleophile, which allows us to achieve the coupling
satisfactorily with a 1:1 ratio of the coupling partners. Both
syntheses have a high overall efficiency: E7389 C14-C35
and halichondrin C14-C38 building blocks are obtained from
the corresponding C27-C35 and C27-C38 aldehydes,
respectively, in high overall yields with an excellent stereo-
selectivity. Because of operational simplicity, we feel that
the synthesis outlined in Schemes 3 and 5 is well suited for
cost-effective scaling. The progress reported here has had a
major impact on the overall synthetic efficiency of E7389 as
well as the right half of halichondrin B. Scheme 10
summarizes the current E7389 synthesis from this laboratory.
With the use of additional key transformations (highlighted),
the C14-C35 building block is converted into E7389.20-24
Acknowledgment. We are grateful to the National Institutes
of Health (CA 22215) and to the Eisai Research Institute for
generous financial support. D.U. gratefully acknowledges a post-
doctoral fellowship from Japan Society for the Promotion of Science
(JSPS).
Supporting Information Available: Experimental details,
including the syntheses outlined in Schemes 2, 3, 5, 6, 7, 8,
and 9, complete characterization of synthetic intermediates, and
1H and 13C NMR spectra. This material is available free of
charge via the Internet at http://pubs.acs.org.
a Catalytic asymmetric Ni/Cr- and Co/Cr-mediated couplings are used
for bond formation at the indicated sites. After coupling of 3b with C1-C13
building block under the Eisai protocol,20 the macrocyclization was effected
JA9058487
by use of the Cr catalyst prepared from 3,3′-dimethyl-2,2′-dipyridyl.21
A
convenient ion-exchange resin-based method is developed to deprotect TBS
groups.22 A device composed of two ion-exchange resin columns allows
us to construct the polycyclic ring system from the deprotected enone
quantitatively.23 In addition, a new condition is developed for ammonolysis
to introduce an amino group at C35.24 Abbreviation: nr ) not relevant.
(20) See ref 5 in the preceding paper.
(21) Namba, K.; Kishi, Y. J. Am. Chem. Soc. 2005, 127, 15382.
(22) Kaburagi, Y.; Kishi, Y. Org. Lett. 2007, 9, 723.
(23) Namba, K.; Jun, H.-S.; Kishi, Y. J. Am. Chem. Soc. 2004, 126, 7770.
(24) Kaburagi, Y.; Kishi, Y. Tetrahedron Lett. 2007, 48, 8967.
9
15646 J. AM. CHEM. SOC. VOL. 131, NO. 43, 2009