Stereocontrolled total synthesis of (-)-kaitocephalin

Stereocontrolled Total Synthesis of (-)-Kaitocephalin

Rishi G. Vaswani^†and A. Richard Chamberlin*^,†,‡

Department of Chemistry and the Department of Pharmaceutical Sciences, UniVersity of Californias

IrVine, IrVine, California 92697-2025

richard.chamberlin@uci.edu

ReceiVed October 29, 2007

This paper describes the successful implementation of a stereocontrolled strategy for the total chemical

synthesis of the pyrrolidine-based alkaloid (-)-kaitocephalin. This scalable synthetic route profits from

the strategic utilization of substrate-controlled manipulations for the iterative installation of the requisite

stereogenic centers. The key transformations include a diastereoselective modified Claisen condensation,

a chemo- and diastereoselective reduction of a â-keto ester, and the substrate-directed hydrogenation of

a dehydroamino ester derivative. During the course of our investigations, an interesting stereoconvergent

cyclization reaction was discovered for the efficient assembly of the kaitocephalin 2,2,5-trisubstituted

pyrrolidine core.

Introduction

L-Glutamic acid (L-Glu) is recognized as the principle

excitatory neurotransmitter in the mammalian central nervous

system (CNS) and is involved in numerous important physi-

ological mechanisms, such as fast interneuronal signaling,

sensory perception, memory, learning, and synaptic plasticity.¹

Conversely, the overarching physiological significance of L-Glu

in normal CNS function is counterbalanced by a number of

pathological processes.²Thus, in addition to the central role of

ionotropic glutamate receptor (iGluR) activation in normal

neuronal function, excessive stimulation of the iGluRs by either

L-Glu itself or by any of a large number of iGluR agonists can

trigger a cascade of detrimental cellular events that ultimately

lead to neuronal injury or death. This abnormal response to

iGluR activation, known as excitotoxicity, is postulated to play

a role in a variety of neurodegenerative disorders including

epilepsy,^3a-cParkinson’s disease,^3a,cand Alzheimer’s disease.^3a,c

Because injury to neurons in many such neurological disorders

may be caused, at least in part, by overstimulation of the

iGluRs,³agents capable of attenuating receptor activity have

FIGURE 1. Structure of (-)-kaitocephalin.

attracted considerable interest as potential neuroprotectants. The

possibility that selective antagonists of the iGluRs may ame-

liorate the destructive effects of glutamate excitotoxicity has

motivated extensive research efforts by both academic and

pharmaceutical institutions to identify novel iGluR antagonists

capable of therapeutic intervention for certain neuropathologies.^3c,4

Our own interest in this area was stimulated by the discovery

of the first naturally occurring glutamate receptor antagonist,

kaitocephalin (1) (Figure 1). In 1997, Shin-ya and co-workers

isolated this novel pyrrolidine-based alkaloid from the filamen-

tous fungus Eupenicillium shearii PF1191.^5,6In initial biological

screening, kaitocephalin suppressed kainic acid toxicity with a

reported EC₅₀value of 0.68 µM at 500 µM concentration of

* Corresponding author. Phone: (949) 824-7089. Fax: (949) 824-7089.

^†Department of Chemistry.

^‡Department of Pharmaceutical Sciences.

(1) Cavalheiro, E. A.; Olney, J. W. Proc. Natl. Acad. Sci. U.S.A. 2001,

98, 5947-5948.

(2) Gill, A.; Madden, D. R. Glutamate Receptor Ion Channels: Structural

Insights Into Molecular Mechanisms; Royal Chemical Society: Cambridge,

UK, 2006.

(4) Moloney, M. G. Nat. Prod. Rep. 2002, 19, 597-616.

(5) (a) Isolation: Shin-ya, K.; Kim, J.-S.; Furihata, K.; Hayakawa, Y.;

Seto, H. Tetrahedron Lett. 1997, 38, 7079-7082. (b) Initial stereochemical

assignment: Kobayashi, H.; Shin-ya, K.; Furihata, K.; Hayakawa, Y.; Seto,

H. Tetrahedron Lett. 2001, 42, 4021-4023.

(3) (a) Doble, A. Pharmacol. Ther. 1999, 81, 163-221. (b) Loscher,

W. Prog. Neurbiol. 1998, 54, 721-741. (c) Brauner-Osborne, H.; Egebjerg,

J.; Nielsen, E. O.; Madsen, U.; Krogsgaard-Larsen, P. J. Med. Chem. 2000,

43, 2609-2645.

Published on Web 01/29/2008

J. Org. Chem. 2008, 73, 1661-1681

1661

Vaswani and Chamberlin

SCHEME 1. Retrosynthetic Analysis of Kaitocephalin (1)

the agonist kainic acid.^5aAnalogously, when 1 was tested for

antagonist activity against R-amino-3-hydroxy-5-methyl-4-isox-

azole propionate (AMPA), an EC₅₀value of 0.60 µM was

reported at 500 µM/50 µM concentration of AMPA/

cyclothiazide.^5aThus, as the first naturally occurring iGluR

antagonist, 1 is a potential lead compound for the development

of therapeutic agents to treat neurological diseases caused by

glutamate excitotoxicity. In that regard, a better understanding

of the origin of kaitocephalin’s antagonist activity at the iGluRs

would require not only ample quantities of kaitocephalin for

further physiological investigations but also the biological

evaluation of various kaitocephalin analogues to develop

structure-activity relationships (SAR) and elucidate the struc-

tural features that contribute to iGluR antagonism. These needs

coupled with limited availability from natural and synthetic

sources motivated us to develop a stereocontrolled total synthesis

of (-)-kaitocephalin that also provides scaffolds to be employed

later for analog library synthesis.

elaborate bidirectionally on an enantiomercally pure substituted

pyrrolidine scaffold. Retrosynthetic simplification of 1 first

exposes saturated amino ester 2. In the forward synthetic sense,

the construction of 1 would be accomplished by hydrolytic

cleavage of the N,O-acetal of 2, followed by chemoselective

oxidation of the resultant C1 primary alcohol, and finally

deprotection of the protecting groups (Scheme 1). In turn, we

planned to install the C9 stereocenter of amino ester 2 from

aldehyde 4 by a sequential Horner-Wadsworth-Emmons

olefination and hydrogenation of the resultant dehydroamino

ester (E)-3 or (Z)-3. Aldehyde 4 could be prepared from the

oxidative cleavage of the olefinated pyrrolidine 5, which

contains the requisite C2, C3, C4, and C7 stereocenters. Further

disconnection of pyrrolidine 5 suggests that the C4-tetrasubsti-

tuted center and the C3 secondary alcohol center could be

accessed from an aldol reaction between a metal enolate derived

from pyrrolidine ester 6 and an appropriate chiral aldehyde.

Alternatively, a stepwise protocol involving the C-acylation of

the enolate derived from 6 followed by a stereoselective

reduction of the resultant â-keto ester could also allow entry to

pyrrolidine 5. The 2,5-disubstituted pyrrolidine ester (6) would

most obviously be prepared from the amino acid chiral pool

(e.g., from pyroglutamate or a related structure).

Aside from the general synthetic challenges posed by the

dense array of stereocenters, we viewed two specific issues as

the most challenging requirements in developing an efficient

approach: (1) the olefination of pyrrolidine aldehyde 4 and

diastereoselective reduction of the resultant dehydroamino ester

(E)-3 or (Z)-3 to install the C9 stereocenter and (2) the

stereochemical outcome of the proposed aldol reaction between

Garner’s aldehyde and the 2,5-disubstituted pyrrolidine ester

6, in which the possible occurrence of complicated double

stereodifferentiation could undermine our prediction that the

desired diastereoselectivity would result from steric influence

of the vinyl substituent.

Synthesis Plan

Both the unusual structure and unique biological profile of

kaitocephalin make this natural product a compelling target for

further study. Since the elucidation of its structure,⁶the

Kitahara,⁷Ohfune,⁸and Ma⁹research groups have independently

reported total syntheses of 1 that illustrate a multiplicity of

noteworthy synthetic challenges in a relatively simple natural

product, including the formation of the substituted pyrrolidine

core, the installation of the tetrasubstituted C4 stereocenter, the

stereoselective introduction of the C3 secondary alcohol ste-

reocenter, and the efficient incorporation of the C2 and C9

amino acid moieties.

Our goal was to implement a concise, scalable, and stereo-

controlled total synthesis of 1 via a general strategy that would

(6) The initial absolute stereochemical assignment of kaitocephalin was

based on chemical derivatization and NMR studies, which led to the

misassignment of the C2 center as C2-(S) (see ref 5b). This misassignment

was later corrected through the first total synthesis of kaitocephalin by

Kitahara and co-workers and established as C2-(R) (see ref 7).

(7) (a) Watanabe, H.; Okue, M.; Kobayashi, H.; Kitahara, T. Tetrahedron

Lett. 2002, 43, 861-864. (b) Okue, M.; Kobayashi, H.; Shin-ya, K.; Furihata,

K.; Hayakawa, Y.; Seto, H.; Watanabe, H.; Kitahara, T. Tetrahedron Lett.

2002, 43, 857-860.

(8) Kawasaki, M.; Shinada, T.; Hamada, M.; Ohfune, Y. Org. Lett. 2005,

7, 4165-4167.

(9) Ma, D.; Yang, J. J. Am. Chem. Soc 2001, 123, 9706-9707. The

synthesis of the C2-(S) isomer of 1 was reported.

Results and Discussion

Synthesis of 2,5-Disubstituted Pyrrolidine Ester. The

synthesis began with the exploration of methods for the

construction of the trans-2,5-disubstituted pyrrolidine ring 6.¹⁰

The addition of organocopper reagents to the chiral N-

acyliminium ion derived from (R)-pyroglutamic acid (7) report-

edly provides an efficient means of synthesizing pure trans-

2,5-disubstituted pyrrolidines,^10,11a-fan approach that in this case

1662 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

SCHEME 2. Preparation of Aminal 10

TABLE 1. N-Acyliminium Approach to the Formation of Pyrrolidine 6

6

11

entry

X

conditions

t-BuLi, THF -78 ˚C to 0 ˚C

yield (%), dr (trans/cis)^a

1

2

3

4

5

Br

I

45 (>20:1)

20

18

Li(0) (5% Na content) Et₂O, -10 ˚C to 0 ˚C

t-BuLi, THF -78 ˚C to 0 ˚C

45 (>20:1)

89 (8:1)

Br

Mg(0), THF rt to 60 ˚C

commercially available 2-methylpropenylmagnesium bromide

69-78 (15:1-20:1)

50-86 (15:1-20:1)

5-10

0-30

1

^aTrans/cis ratios were determined by H NMR spectroscopic analysis of crude reaction mixtures.

required formation of aminal 10. Accordingly, esterification of

(R)-pyroglutamic acid (7) using catalytic SOCl₂in MeOH

provided lactam ester 8 in 88% yield (Scheme 2).¹²Protection

of the lactam nitrogen of 8 was carried out using Boc₂O and

DMAP to furnish imide 9 in 94% yield after crystallization.¹³

Chemoselective reduction of 9 was accomplished by careful

addition of lithium triethylborohydride in THF at -78 °C.^11a,14

The resultant hemiaminal was immediately subjected to catalytic

p-TsOH in MeOH to afford N,O-acetal 10 as a 1:1 mixture of

diastereomers in 95% yield.^11aAminal 10 was prepared on a

25 g scale in 79% yield over four steps (starting from 7) and

could be stored indefinitely at -20 °C with negligible hydroly-

sis.

The successful preparation of aminal 10 led next to investiga-

tions into the addition of organocopper reagents to the corre-

sponding chiral N-acyliminium ion (Table 1). Treatment of

1-bromo-2-methylpropene with either t-BuLi or Li(0) metal,

followed by transmetalation with CuBr‚DMS and addition of

BF₃‚OEt₂and 10, provided the trans-2,5-disubstituted pyrrolidine

6 in 45% yield as a >20:1 diastereomeric mixture (entry 1-2).

In addition to the isolation of the desired pyrrolidine product 6,

diene-incorporated trans-2,5-disubstituted pyrrolidine product

11 was also produced in 20% yield.^10,15-17In an effort to

improve on the yield of this organocuprate, vinylation coupled

with the unanticipated formation of diene 11 prompted us to

investigate alternative metal-halogen exchange conditions. We

rationalized that employing a vinyl iodide for the lithium-

halogen exchange should ultimately prevent the formation of

11.¹⁸Subjection of iodo-2-methylpropene¹⁹to the lithium-

halogen exchange conditions followed by copper-mediated

transmetalation and finally addition of BF₃‚OEt₂and 10 afforded

6 in a trans/cis diastereomeric ratio of 8:1 in 89% yield (entry

3). As expected under these conditions, the formation of the

diene-incorporated pyrrolidine 11 was suppressed; however, the

erosion in diastereoselection from 20:1 to 8:1 was considered

unsatisfactory for our synthesis, thus necessitating further

examination of this key transformation.

After exhaustive screening, it was found that the optimal

conditions for the diastereoselective synthesis of pyrrolidine 6

utilized the 2-methylpropenlymagnesium bromide, derived from

1-bromo-2-methylpropene (entry 4). Transmetalation of the

Grignard reagent with CuBr‚DMS and sequential addition of

(15) The formation of 11 was speculated to arise as a result of competing

R-lithiation of 1-bromo-2-methylpropene, leading to the formation of a

reactive carbene intermediate III (see below). Addition of another 2-

methylpropenyl lithium species to the carbene intermediate would produce

the lithiated diene nucleophile. Transmetalation with CuBr‚DMS followed

by sequential treatment with BF₃‚OEt₂and aminal 10 would result in the

formation of side product 11.

(10) (a) Tong, Y.; Fobian, Y. M.; Wu, M.; Boyd, N. D.; Moeller, K. D.

J. Org. Chem. 2000, 65, 2484-2493. (b) Fobian, Y. M.; Moeller, K. D. In

Methods in Molecular Medicine: Peptidomimetic Protocols; Kazmierski,

W. M., Ed.; Humana Press, Inc.: Totowa, NJ, 1988; p 259.

(11) (a) Collado, I.; Ezquerra, J.; Pedregal, C. J. Org. Chem. 1995, 60,

5011-5015. (b) Ludwig, C.; Wistrand, L. G. Acta. Chem. Scand. 1994,

48, 367-371. (c) Thaning, M.; Wistrand, L. G. Acta. Chem. Scand. 1992,

46, 194-199. (d) Wistrand, L. G.; Skrinjar, M. Tetrahedron 1991, 47, 573-

582. (e) Skrinjar, M.; Wistrand, L. G. Tetrahedron Lett. 1990, 31, 1775-

1778. (f) Ludwig, C.; Wistrand, L. G. Acta. Chem. Scand. 1990, 44, 707-

710.

(12) Doyle, M. P.; Winchester, W. R.; Protopopova, M. N.; Kazala, A.

P.; Westrum, L. J. Org. Synth. 1996, 73, 13-19.

(13) Jain, R. Org. Prep. Proced. Int. 2001, 33, 405-409.

(14) (a) Brown, H. C.; Krishnamurthy, S. J. Am. Chem. Soc. 1973, 95,

1669-1671. (b) Polniaszek, R. P.; Belmont, S. E.; Alvarez, R. J. Org. Chem.

1990, 55, 215-223. (c) Fisher, M. J.; Overman, L. E. J. Org. Chem. 1990,

55, 1447-1459. (d) Pedregal, C.; Ezquerra, J.; Escribano, A.; Carreno, M.

C.; Garcia Ruano, J. L. Tetrahedron Lett. 1994, 35, 2053-2056.

J. Org. Chem, Vol. 73, No. 5, 2008 1663

Vaswani and Chamberlin

SCHEME 3. Retrosynthetic Analysis for the Cyclization Reaction of anti-12

SCHEME 4. Chemoselective Ring Opening of Imide 9

BF₃‚OEt₂and N,O-acetal 10 successfully produced the desired

pyrrolidine 6 in 69-78% with diastereomeric ratios ranging

from 15:1 to 20:1. The formation of side product 11 was

minimized under these conditions to 5-10% yield.²⁰

Stereoconvergent Cyclization Reaction for the Synthesis

of 6. Although the cuprate procedure gave reasonable yields,

there were several limitations that prompted us to continue

searching for a more satisfactory process. First, the formation

of the dienyl pyrrolidine 11 could not be completely suppressed,

and the separation of desired pyrrolidine 6 from 11 by silica

gel chromatography required multiple chromatographic runs to

obtain pure samples of 6. Second, several operational difficulties

were encountered during the generation of the Grignard reagent

on large scale: 2-methylpropenylmagnesium bromide had a

propensity to precipitate from the solution, and the resultant

heterogeneous mixture could not be transferred efficiently via

cannula. Consequently, the Grignard mixture required relatively

high dilution in THF before a homogeneous solution could be

attained, resulting in disproportionately large reaction volumes.

Last, excessive quantities of CuBr‚DMS, Grignard reagent, and

BF₃‚OEt₂(at least 3 equiv of each reagent) were required to

obtain diastereoselectivities of >20:1. The copious amounts of

CuBr‚DMS produced considerable quantities of copper waste

and limited the workable scale of this transformation. Ultimately,

these constraints were considered to be sufficiently unsatisfac-

tory that an alternative method for the synthesis of the

pyrrolidine core was pursued.

In an alternative route that ultimately provided not only a

workable solution but also an interesting surprise, the pyrrolidine

6 was envisioned to arise from allylic alcohol anti-12 via an

S_N2 cyclization of the nitrogen atom onto an appropriately

activated allylic alcohol (Scheme 3). The cyclization substrate,

anti-12, would derive from imide 9 by sequential ring opening

and stereoselective reduction of enone 13. The possibly delicate

nature of the allylic substitution in anti-12 would require that

alcohol activation and ring formation occur under mild condi-

tions in order to preserve the stereochemical integrity of the

hydroxyl moiety, which we assumed would be relayed ste-

reospecifically during the cyclization event to give the desired

trans diastereomer.

This cyclization approach to give the pyrrolidine 6 com-

menced with the chemoselective ring opening of imide 9

employing 2-methylpropenylmagnesium bromide in the presence

of TMEDA to furnish enone 13 in 50-78% yield (Scheme 4).²¹

The remainder of the mass consisted of dienyl-enone 14 and

the recovered starting material (9). Next, the stereoselective 1,2-

reduction was explored to prepare the requisite allylic alcohol

anti-12 (Table 2). Initial investigations examined the diaste-

reoselective outcome of the reduction of enone 13 using achiral

reagents. Thus, treatment of enone 13 under Luche conditions,²²

NaBH₄/CeCl₃‚7H₂O (entry 1), produced an equimolar mixture

of allylic alcohols anti-12 and syn-12, respectively. The

anticipated lack of stereoselectivity observed in the substrate-

(16) Analogous dienyl lithium formation (IV) during lithium-halogen

exchange of 1-bromo-2-methylpropene has previously been observed: Piers,

E.; Oballa, R. M. J. Org. Chem. 1996, 61, 8439-8447.

(17) For other references on R-lithiation or lithium carbenoid formation

during Li-X exchange, see: (a) Novikov, Y. Y.; Sampson, P. J. Org. Chem.

2005, 70, 10247-10259. (b) Novikov, Y. Y.; Sampson, P. Org. Lett. 2003,

5, 2263-2266. (c) Gernot Boche, Marsch, M.; Mu¨ller, A.; Harms, K. Angew.

Chem., Int. Ed. 1993, 32, 1032-1033. (d) Harada, T.; Nozaki, Y.; Yamaura,

Y.; Oku, A. J. Am. Chem. Soc. 1985, 107, 2189-2190.

(18) (a) Bailey, W. F.; Patricia, J. J. J. Organomet. Chem. 1988, 352,

1-46. (b) Curtin, D. Y.; Richardson, W. H. J. Am. Chem. Soc. 1959, 81,

4719-1728. (c) Curtin, D. Y.; Flynn, E. W. J. Am. Chem. Soc. 1959, 81,

4714-4719.

(19) 1-Iodo-2-methylpropene was synthesized from 1-bromo-2-methyl-

propene utilizing Buchwald’s “Aryl-Finkelstein” methodology: (a) Buch-

wald, S. L.; Klapars, A.; Kwong, F. Y.; Streiter, E.; Zanon, J. (Massachusetts

Institute of Technology) US Pat. Appl. US6888032B2, 2005. (b) Buchwald,

S. L.; Klapars, A.; Kwong, F. Y.; Streiter, E.; Zanon, J. (Massachusetts

Institute of Technology) PCT Int. Appl. WO2004013094 A2, 2004, p 128.

(c) Klapars, A.; Buchwald, S. L. J. Am. Chem. Soc. 2002, 124, 14844-

14845.

(20) Although 2-methylpropenylmagnesium bromide is commercially

available as a 0.5 M solution in THF, the significance of the efficient

generation of the Grignard reagent from the corresponding vinyl bromide

is underscored by the inconsistent results observed when commercially

available Grignard reagent was used in the organocuprate-mediated alk-

enylation. With commercially available 2-methylpropenylmagnesium bro-

mide, depending on the batch number, pyrrolidine 9 was furnished in yields

varying from 50-86%, and the dienyl pyrrolidine 12 was produced in

0-30% yield, albeit with little effect on diastereoselectivity (see Table 1,

entry 5). Thus, optimized conditions for this transformation required the

generation of the Grignard to deliver pyrrolidine 9 in 69-78% yield with

minimal formation of the dienyl pyrrolidine 12.

(22) (a) Gemal, A. L.; Luche, J. L. J. Am. Chem. Soc. 1981, 103, 5454-

5459. (b) Luche, J. L. J. Am. Chem. Soc. 1978, 100, 2226-2227. (c) Luche,

J. L.; Rodriguez-Hahn, L.; Crabbe, P. J. Chem. Soc., Chem. Commun. 1978,

601-602.

(21) Rudolph, A. C.; Machauer, R.; Martin, S. F. Tetrahedron Lett. 2004,

45, 4895-4898.

1664 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

TABLE 2. Stereoselective Reduction of Enone 13

entry

conditions

dr (anti:syn)^aof 12

yield (%)

1

2

3

NaBH₄, CeCl₃‚7H₂O, MeOH, -78 °C

BH₃‚THF (1.00 equiv) (R)-Me-CBS (1.20 equiv) THF, -78 °C to rt

BH₃‚THF (1.00 equiv) (S)-Me-CBS (1.20 equiv) THF, -78 °C to rt

1:1

15:1

1:33

88

72

41

^aAnti/syn ratios were determined by normal-phase analytical HPLC of crude reaction mixtures.

TABLE 3. Cyclization Reaction of Allylic Alcohol anti-12 with Ph₃P, I₂, and Imidazole

dr

(trans/cis)^a

of 6

(anti/syn)

of 12

yield (%)

of 6

yield (%)

entry

conditions

of 15

1

2

3

4

5

6

Ph₃P (1.10 equiv), I₂(1.12), imid (2.11 equiv), THF -78 °C to rt, 16-24 h

Ph₃P (1.24 equiv), I₂(1.23), imid (2.63 equiv), THF -78 °C to rt, 24 h

Ph₃P (2.50 equiv), I₂(2.57), imid (5.10 equiv), THF -78 °C to rt, 16 h

Ph₃P (2.00 equiv), I₂(2.00), imid (5.00 equiv), THF -78 °C to rt, 24 h

Ph₃P (1.30 equiv), I₂(1.00), imid (1.69 equiv), THF 0 °C to rt, 16-24 h

Ph₃P (2.00 equiv), I₂(2.00), imid (4.00 equiv), THF 0 °C to rt, 24 h

14:1

4:1

1:1

1:33

1:5

50

94

74

50

93

22:1

26:1

24:1

14:1

10:1

12:1

3-4

2-5

^aTrans/cis ratios were determined by normal-phase analytical HPLC of crude reaction mixtures.

directed reduction of enone 14 was corrected by the use of chiral

reagents to enforce the desired diastereofacial preference in the

ketone reduction, and ultimately, use of the Corey-Bakshi-

Shibata (CBS) reagent provided superior diastereoselection for

the reduction of 13.²³It was determined, however, that optimal

levels of diastereoselectivity required the use of stoichiometric

quantities of CBS reagent; for example, the reduction of enone

13 necessitated the use of 120 mol % of (R)-Me-CBS, and

BH₃‚THF furnished allylic alcohol anti-12 in 72% yield (dr anti/

syn ) 15:1) (entry 2).²⁴The reduction of enone 13 with the

enantiomeric (S)-Me-CBS reagent afforded the expected ste-

reocomplementary product syn-12 (entry 3), confirming that lack

of any significant stereochemical influence of the relatively

remote extant stereogenic center under these conditions.

Having successfully established conditions for the diastereo-

selective formation of allylic alcohol anti-12 (see Table 2), the

key cyclization reaction to form trans-6 was addressed (Table

3). After considerable experimentation, we observed that the

cyclization of allylic alcohol anti-12 proceeded smoothly in the

presence of Ph₃P, I₂, and imidazole to afford the desired

pyrrolidine trans-6 in 50% yield (entry 1). HPLC analysis of

the pyrrolidine product mixture, however, revealed an unex-

pected enhancementssmall but reproduciblesin the diastere-

omeric ratio; that is, exposure of a 14:1 (anti-/syn-) diastereo-

meric mixture of allylic alcohol 12 to Ph₃P, I₂, and imidazole

produced pyrrolidine trans-6 as a 22:1 diastereomeric mixture

(entry 1) instead of the anticipated 14:1 ratio of products.

Although we might have rationalized this small difference as

due to a minor fluctuation in reaction conditions and moved

on, we wondered if it might instead be signaling something more

significant. We therefore checked whether this enhancement in

diastereoselectivity occurred with other reactant ratios and found

that when a 4:1 (anti-/syn-) diastereomeric mixture of 12 was

subjected to identical reaction conditions (entry 2) the reaction

produced 6 with a similar enhancement in diastereoselectivity

to give a 26:1 mixture of trans- and cis-diastereomers in 50%

combined yield. Under optimized conditions employing 2.5

equiv of I₂, 2.5 equiv of Ph₃P, and 5 equiv of imidazole, the

cyclization of a 4:1 (anti-/syn-) mixture of allylic alcohols 12

afforded 6 as a 24:1 (trans-/cis-) mixture of diastereomers in

94% combined yield (entry 3).²⁵

With our interest now stimulated, mixtures of starting allylic

alcohol favoring the diastereomeric (“wrong”) substrate (syn-

12) were tested under the same reaction conditions (Table 3,

entries 5 and 6). Interestingly, we found that treatment of a 1:5

(anti-/syn-) diastereomeric mixture of 12 with Ph₃P, I₂, and

imidazole still provided predominantly trans-6 as a 12:1 (trans-/

cis-) mixture of diastereomers (93% combined yield). Clearly,

the cyclization of allylic alcohol syn-12 was proceeding

selectively but with net retention of stereochemistry; that is,

allylic alcohol syn-12 cyclized to furnish predominantly the

pyrrolidine trans-6, the same product as produced in the

(25) It should be noted that the potential for the imidazole-mediated

epimerization at the C2 R-center of cis-6 to produce the ent-trans-6 was

not observed. Regardless of the diastereomeric ratio of the starting allylic

alcohol 12, the cyclization event produced trans-6 with no significant

degradation in the optical purity. The optical rotation measured for trans-

6, produce during the stereoconvergent cyclization strategy, was similar in

sign and magnitude to that produced in the organocuprate reaction.

(23) For an excellent review on the CBS reagents, see: Corey, E. J.;

Helal, C. J. Angew. Chem., Int. Ed. 1998, 37, 1986-2012.

(24) The stereochemical outcome of the reduction of 13 using (R)-Me-

CBS reagent was inferred based on literature precedent: Li, A.; Yue, G.;

Li, Y.; Pan, X.; Yang, T.-K. Tetrahedron: Asymmetry 2003, 14, 75-78.

J. Org. Chem, Vol. 73, No. 5, 2008 1665

Vaswani and Chamberlin

SCHEME 5. Proposed Mechanistic Rationale for Stereoconvergent Cyclization of 12

cyclization of the diastereomeric allylic alcohol precursor (anti-

12). Thus, both the allylic alcohol diastereomers cyclize to

produce pyrrolidine product ratios enhanced in the trans isomer;

the cyclization of anti-12 proceeds with inVersion of configu-

ration with respect to the alcohol stereocenter, while cyclization

of syn-12 proceeds with retention of configuration.

alcohol 16a. Ring closure by attack of the weakly nucleophilic

carbamate nitrogen would be facilitated by carbocation-like

transition states 17 or 18 resulting from loss of the triphenyl-

phosphine oxide leaving group from 16a, with free rotation in

the resultant carbocation giving rise to several possible transition

state geometries. It is also possible that the iodide from the

reagent displaces the triphenylphosphine oxide from 16a with

inversion (and even subsequently another iodide, resulting in

multiple inversions) and that it is the resultant allylic iodide

16b, or the carbocation derived from it, that undergoes cycliza-

tion, again via the transition states shown. Either way, the

stereochemistry of the starting alcohol is lost (by carbocation

formation or multiple inversions), and thus, the stereochemical

outcome of the reaction would not be determined by the

stereochemistry of the alcohol starting material as originally

supposed, but instead by the relative energies of the respective

cyclic transition states (17 and 18), which in turn are governed

by well-established stereoelectronic and conformational fac-

tors.²⁷Specifically, in this case, 17 would be energetically

preferred over 18 because of the indicated steric strain in the

latter, and thus, trans-6 is the preferred product formed from

either anti- or syn-12 under these reaction conditions.

On the basis of these results, it was now obvious that despite

the demonstrated lack of a directing effect by the protected

amino group in a kinetically controlled reduction to the precursor

alcohol, we had stumbled upon a more subtle mechanism of

substrate-controlled diastereoselectivity in the subsequent cy-

clization that renders the “chiral reduction” unnecessary.²⁶Thus,

treatment of a 1:1 diastereomeric mixture of allylic alcohols 12

with 2 equiv of Ph₃P, I₂, and 5 equiv of imidazole in THF

furnished the desired 2,5-disubstituted pyrrolidine (trans-6) as

a 14:1 (trans-/cis-) mixture of diastereomers in 74% yield (entry

4). This simple procedure obviated the need for stoichiometric

quantities of a chiral reducing agent to establish the stereo-

chemical configuration of the allylic alcohol and simply relies

on substrate control for the formation of trans-6, which in turn

greatly expedites synthetic access to the pyrrolidine core of

kaitocephalin.

Establishing the Tetrasubstituted C4 Stereocenter. With

a viable synthesis of the requisite pyrrolidine core (trans-6) in

hand, attention next focused on constructing the tetrasubstituted

C4 stereocenter of kaitocephalin. A lithium enolate aldol reaction

between the trans-6 and (S)-Garner’s aldehyde²⁸(19) was

envisioned for the simultaneous introduction of the C4 and C3

stereocenters of kaitocephalin. To that end, trans-6 was

Although an extensive mechanistic analysis has not been

performed on this interesting stereoselective cyclization reaction,

the intriguing and useful result may be simply rationalized by

ring closure in a conformation that minimizes allylic strain in

one of two alternative transition states (Scheme 5). The

cyclization is initiated by activation of both anti- and syn-12

with iodotriphenylphosphonium iodide to give the activated

(26) In a related system, utilizing MsCl and Et₃N for the cyclization of

a 1:1 mixture of allylic alcohols reported similar enhancement for the

formation of a trans-pyrrolidine isomer, albeit with much lower diastere-

omeric ratios (2:1): (a) Gosselin, F.; Lubell, W. D. J. Org. Chem. 2000,

65, 2163-2171. (b) Ohta, T.; Hosoi, A.; Nozoe, S. Tetrahedron Lett. 1988,

29, 329-332. (c) Ohta, T.; Hosoi, A.; Kimura, T.; Nozoe, S. Chem. Lett.

1987, 2091-2094.

(27) (a) The stereoelectronic factor that contributes to 17 and 18 include

the parallel alignment of the σ_C-Xbond of the leaving group with the

π-system of the double bond. The conformational factor that contributes to

17 and 18 involves minimizing allylic strain (A^1,3strain) between the methyl

substituent, occupying the (Z)-position on the double bond, and the allylic

center, where the H-atom is placed in plane. (b) For an excellent review on

allylic strain, see: Hoffmann, R. W. Chem. ReV. 1989, 89, 1841-1860.

1666 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

SCHEME 6. Aldol Reaction of Pyrrolidine trans-6 with Garner’s Aldehyde

SCHEME 7. Unsuccessful C-Acylation of trans-6 with

Serine-Derived Electrophiles

treated with LDA at -78 °C, followed by the addition of (S)-

1

Garner’s aldehyde (19) (Scheme 6). H NMR analysis of the

unpurified reaction mixture revealed a complex mixture of four

aldol adducts in a ratio of 2:2:1:1 based on comparison of the

integration values for the vinyl proton; in addition, unreacted

aldehyde 19 and a mixture of 2,5-disubstituted pyrrolidines,

trans-6 and C2-epi-6, were also detected (Scheme 6).²⁹

The stereochemical outcome of the aldol reaction between

pyrrolidine ester trans-6 and (S)-Garner’s aldehyde (19) clearly

suggested that the desired aldol reaction was a case of

mismatched double stereodifferentiation. To confirm this as-

sertion, an aldol reaction between trans-6 and (R)-Garner’s

aldehyde (ent-19) was performed; indeed, this reaction provided

one diastereomer of the aldol product 21 in an unoptimized yield

of 40%, indicating a matched double stereodifferentiation

(Scheme 6). The unfortunate mismatched stereochemical out-

come of the key aldol reaction between trans-6 and aldehyde

19 necessitated a revision of our strategy for the formation of

the C3 and C4 stereocenters.

C-Acylation Chemistry with N-Acylimidazole. An alterna-

tive approach to the kaitocephalin intermediate 20a simply

would generate the C4 and C3 centers sequentially: first,

C-acylation of the ester enolate derived from trans-6 with a

serine-derived electrophilic acylating agent to form the C4

stereocenter and, second, a diastereoselective reduction of the

resultant â-keto ester 23 to install the C3 secondary alcohol.

To implement this plan, various serine-derived electrophiles for

the C-acylation of trans-6 were surveyed (Scheme 7). Unfor-

tunately, addition of Weinreb amide 24, methyl ester 25a, or

pentafluorophenyl ester 25b to the lithium enolate of pyrrolidine

ester trans-6 failed to afford any of the desired C-acylation

product (23).

Uncertain at this point whether the lack of success was due

to kinetic or thermodynamic problems, we sought other leaving

groups in the modified Claisen condensation for this key bond

construction. The utility of N-acylimidazoles for the C-acylation

of enolates, especially in the context of natural product synthesis,

has been widely established.³⁰To that end, serine-derived

N-acylimidazole 26 was synthesized from L-serine on scales

greater than 20 g and was stable indefinitely when stored under

an atmosphere of Ar at -20 °C (Scheme 8).

Following the successful preparation of N-acylimidazole 26,

the C-acylation of trans-6 was conducted by the addition of 26

to a solution of the lithium enolate derived from pyrrolidine

ester trans-6.^{31 1}H NMR and normal-phase HPLC analysis of

the crude reaction mixture indicated that the â-keto ester 30

(28) (S)-Garner’s aldehyde was synthesized according to the procedure

reported by Dondoni: Dondoni, A.; Perrone, D. Org. Synth. 2000, 77, 64-

77. For other syntheses of (S)-Garner’s aldehyde, see: Garner, P.; Park, J.

M. Org. Synth. 1992, 70, 18-28. Garner, P.; Park, J. M. J. Org. Chem.

1987, 52, 2361-2364.

(29) Chiral HPLC analysis of the unreacted (S)-Garner’s aldehyde starting

material indicated that 19 was formed in high enantiomeric purity. The

formation of aldol adduct 21, therefore, suggested that some epimerization

of the chiral aldehyde occurred under the basic reaction conditions. This

conjecture was confirmed after chiral HPLC analysis of the Garner’s

aldehyde recovered from the aldol reaction, which revealed significant

racemization.

(30) For some examples on the use of N-acylimidazoles as C-acylation

agents in the context of natural product synthesis see: (a) Ing, A.; Hasegawa,

Y.; Murabayashi, A. Tetrahedron Lett. 1998, 39, 3509-3512. (b) Hoffmann,

R. V.; Tao, J. J. Org. Chem. 1997, 62, 2292-2297. (c) Bergmeier, S. C.;

Cobas, A. A.; Rapoport, H. J. Org. Chem. 1993, 58, 2369-2376. (d) Turner,

J. A.; Jacks, W. S. J. Org. Chem. 1989, 54, 4229-4231. (e) Brooks, D.

W.; Lu, L. D.-L.; Masamune, S. Angew. Chem., Int. Ed. 1979, 18, 72-74.

J. Org. Chem, Vol. 73, No. 5, 2008 1667

Vaswani and Chamberlin

SCHEME 8. Synthesis of N-Acylimidazole 26

was generated as a single detectable diastereomer in 40-50%

yield (Scheme 9). Significantly, the C-acylation of trans-6

exclusively produced the desired C4-(R)-stereocenter; the ster-

eochemistry at the C4 center was determined by NOESY

correlation studies and was unambiguously established by a

single-crystal X-ray structure of a derivatized late-stage inter-

mediate (vide infra).

to provide 30 in an additional 40% yield, thereby increasing

the overall yield of the Claisen product 30 to ∼80% after one

recycle. Overall, the diastereoselective Claisen reaction between

pyrrolidine trans-6 and N-acylimidazole 26 reproducibly fur-

nished â-keto ester 30 as a single diastereomer in multigram

quantities that could be advanced in the synthetic route to

kaitocephalin.

Although the C-acylation of trans-6 proceeded in modest

yield, this reaction produced no major side products; the

remainder of the mass consisted of unreacted starting material

and its C2-epimer, trans-6 and C2-epi-6, respectively, as an

inseparable mixture in 42% yield, and the calculated total mass

recovery for this transformation was 85-92%. Whether the

problem is a thermodynamic issue or one of competing

O-acylation followed by hydrolysis during workup, this yield

could not be improved; however, the starting pyrrolidine trans-6

and the epimerized C2-epi-6 (presumably produced by nonse-

lective protonation of the corresponding enolate) were easily

recovered and re-subjected to the C-acylation reaction conditions

Generation of the C3 Stereocenter: Reduction Studies and

Stereochemical Determination. With the Claisen product

successfully prepared and the tetrasubstituted C4 stereocenter

secured, we directed our attention to the diastereoselective

reduction of the â-keto ester 30 to produce the desired â-hydroxy

ester 31a and establish the C3 secondary alcohol of kaitocephalin

(Scheme 10). An investigation of various borane reducing agents

reducing agents (e.g., NaBH₄, Li(s-Bu)₃BH, Li(Et)₃BH, BH₃‚THF,

BH₃‚DMS, BH₃‚PBu₃) identified BH₃‚NH₃as a highly chemose-

lective reducing agent: the reduction of â-keto ester 30 with

excess BH₃‚NH₃(7 equiv) in a solvent mixture of THF/MeOH

at ambient temperature afforded a chromatographically insepa-

rable mixture of â-hydroxy ester 31a and 31b in >90%

combined yield, albeit in modest diastereoselectivity (dr 31a/

31b ) 2:1). Attempts to improve the diastereoselectivity of this

transformation by performing the reaction at lower temperatures

(0 °C) proved unproductive, as the BH₃‚NH₃reduction of 30

at ambient temperature required unacceptably long reaction

times ranging from 4 to 7 days at room temperature.

(31) Our initial investigation on the C-acylation of trans-6 involved the

use of serine-derived N,O-acetonide N-acylimidazole (see below). Although

this reaction also afforded the desired â-keto ester as a single diastereomer

in a modest 40-50% yield, the isolation of analytically pure quantities of

the desired â-keto ester by silica gel chromatography, however, proved

difficult since the unreacted pyrrolidine starting material trans-6 and its

C2-epimer (C2-epi-6) coeluted with the desired C-acylated product. The

desired product was easily separated from trans-6 and C2-epi-6 by reversed-

phase HPLC; however, HPLC purification on preparatory scales of greater

than 100 mg was prohibitive, which compromised synthetic access to useful

quantities of this key kaitocephalin intermediate. The purification problem

was circumvented by replacing the N,O-acetonide with a N,O-cyclo-

hexylidine acetal (26).

While the BH₃‚NH₃reduction of â-keto ester 30 was highly

chemoselective, the 2:1 diastereomeric ratio observed for this

reaction was unsatisfactory and mandated further screening of

reducing agents. After some experimentation, it was found that

the addition of 1 equiv of DIBALH to a solution of 30 in toluene

at -78 °C resulted in the formation of â-hydroxy aldehyde 32

as a single diastereomer. Although this reduction proceeded with

excellent diastereoselectivity at the C3 secondary alcohol, it was

plagued by poor chemoselectivity and low conversion that could

not be circumvented: the complete consumption of â-keto ester

30 could only be achieved using 3.7 equiv of DIBALH, but

this excess also produced â-hydroxy aldehyde 32 (20%), diol

33 (53%), and N-Me diol 34 (26%) arising from competitive

reduction of the ester and N-Boc groups (Scheme 11).

Although chemoselectivity in the reduction of â-keto ester

30 with DIBALH in toluene was poor, it is important to note

that the diastereoselection in the production of the C3 secondary

alcohol in each isolated product (32-34) was excellent. We

speculated that having performed the reduction in a nonpolar

solvent such as toluene might have facilitated the competitive

1668 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

SCHEME 9. Successful C-Acylation of the Pyrrolidine Ester trans-6 with N-Acylimidazole 26

SCHEME 10. Chemoselective Reduction of 30 with BH₃‚NH₃

reduction of the methyl ester (CO₂R f CHO f CH₂OH) and

the Boc carbamate (Boc f Me) by favoring coordination to

multiple Lewis basic sites in 30 (e.g., O atom of CO₂R, RCOR′,

and Boc) and promoting activation of each carbonyl toward

reduction. We therefore rationalized that performing the reduc-

tion in the presence of Lewis basic solvent, such as THF, might

attenuate this promiscuous coordination and minimize the

formation of over-reduced products.

To investigate this possibility, â-keto ester 30 was treated

with DIBALH in THF at -78 °C, followed by gradual warming

to ambient temperature over 24 h, giving the desired â-hydroxy

ester 31a as a single diastereomer in 86-93% yield (Scheme

11).³²Thus, the simple expedient of employing THF as the

reaction solvent strikingly increased the chemoselectivity of the

transformation while maintaining the desired level of diaste-

reoselection. Neither aldehyde 32 nor N-Me diol 34 were

detected by H NMR analysis of crude reaction mixture; the

1

only other product isolated from this reaction was N-Boc diol

33, again as a single diastereomer, in 10% yield, which could

also be easily converted, if desired, into ester 31a by oxidation

of the primary alcohol to the carboxylic acid^33,34and subsequent

esterification with trimethylsilyldiazomethane.

The configurations of the reduction products 31a and 31b

were determined individually by initial derivatization to the

6-membered O,O-benzylidene acetal (Scheme 12), followed by

¹H NMR evaluation of the coupling constants between the

vicinal protons located on the C3 and C2 centers.³⁵To that end,

an inseparable 2.5:1 mixture of 31a and 31b was treated with

80% aqueous AcOH to effect the hydrolysis of the N,O-

cyclohexylidene acetal to yield a chromatographically separable

mixture of amino diols 35a (65%) and 35b (26%) (Scheme 12a).

Exposure of diol 35a to benzaldehyde dimethyl acetal in the

presence of catalytic (S)-(+)-CSA and 4 Å molecular sieves

resulted in the formation of the six-membered O,O-benzylidene

acetal 36a and the five-membered N,O-benzylidene acetal 37

in a ratio of 2:1, respectively (Scheme 12b). Alternatively,

addition of catalytic p-TsOH to a solution of amino diol 35a

and benzaldehyde dimethyl acetal in CH₂Cl₂at room temper-

ature furnished the thermodynamic six-membered O,O-ben-

zylidene acetal 36a as a single diastereomer in 77% yield.

Treatment of amino diol 35b under analogous conditions

furnished benzylidene acetal 36b in 48% yield (Scheme 12c).

Due to the presence of carbamate rotamer peaks at room

(32) (a) The diastereoselective reduction of â-keto ester 30 was rational-

ized on the basis of a Felkin-Anh transition-state model shown below, in

which the re-face of the carbonyl is sterically accessible to a bulky reducing

agent, such as DIBALH. As a result, addition of the hydride produced the

desired C3-(S)-stereochemistry (see below). (b) Anh, N. T. Top. Curr. Chem.

1980, 88, 145.

(33) For the use TEMPO and bis-acetoxyiodobenzene in the chemose-

lective oxidation of primary alcohols to aldehydes, see: De Mico, A.;

Margarita, R.; Parlanti, L.; Vescovi, A.; Piancatelli, G. J. Org. Chem. 1997,

62, 6974-6977.

(34) For the use of sodium chlorite in the oxidation of aldehydes to acids,

see: (a) Bal, B. S.; Childers, W. E.; Pinnick, H. W. Tetrahedron 1981, 37,

2091-2096. (b) Lindgren, B. O.; Hilsson, T. Acta Chem. Scand. 1973, 27,

888-890.

(35) (a) Herold, P. HelV. Chim. Acta 1988, 71, 354-362. (b) Garner, P.

Tetrahedron Lett. 1984, 25, 5855-5858.

J. Org. Chem, Vol. 73, No. 5, 2008 1669

Vaswani and Chamberlin

SCHEME 11. Reduction of 30 with DIBALH: Solvent Effects^a

^aKey: (a) TEMPO (10 mol %), BAIB, CH₂Cl₂, rt (78%); (b) NaClO₂, 2-Me-butene NaH₂PO₄, t-BuOH/H₂O (4:1), then TMSCHN₂, MeOH/EtOAc

(90%).

SCHEME 12. Conversion of 31 to Benzylidene Acetals

temperature, ¹H NMR spectroscopic analysis of the benzylidene

acetals 36a and 36b was performed above the coalescence

temperature of 375 K. As depicted in Figure 2, a coupling

constant of 9.6 Hz was calculated between the C3 methine

proton and the C2 methine proton, H_aand H_b, respectively,

clearly representing a trans-diaxial relationship between these

protons (nearly 180° alignment based on a vicinal Karplus

correlation).³⁶The trans-diaxial relationship (between H_aand

H_b) observed in acetal 36a ultimately indicated that the

cyclohexylidene acetal 31a has the desired C3-(S) stereochem-

1670 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

FIGURE 2. Coupling constant analyses of benzylidene acetals 36a and 36b.

TABLE 4. Olefination Reaction of D-Prolinal Under Various

Conditions

ical configuration. In contrast to the large coupling constant

observed for acetal 36a, the C3 methine proton and the C2

methine proton of acetal 36b, H_aand H_b, respectively, exhibited

J ) 1.5 Hz (Figure 2); the small coupling constant observed

for 36b is representative of an axial-equatorial relationship

between H_aand H_b, respectively. Based on this analysis, it was

deduced that the cyclohexylidene acetal 31b was the undesired

C3-(R) diastereomer. Additionally, formation of the five-

membered N,O-benzylidene acetal 37, followed by crystalliza-

tion allowed us to procure a single-crystal X-ray structure that

served to establish the absolute configuration of the C3

stereocenter, the tetrasubstituted C4 (formed via the C-acylation),

and C7 stereocenter (generated by the organocuprate or stereo-

convergent cyclization reaction).³⁷

yield

(%)

39 (Z/E)

entry

conditions

ratio^a

1

2

3

4

NaH, THF, -78 °C to rt

50

72

73

1:1

3:1

KHMDS/18-C-6, THF, -78 °C to rt, 16 h

DBU, LiCl, CH₂Cl₂, 0 °C to rt, 24 h

DBU, CH₂Cl₂, 0 °C to rt, 24 h

>30:1

^aZ/E ratios determined by ¹H NMR spectroscopic analysis of crude

Generation of the C9 Stereocenter. Having developed an

efficient and stereoselective approach for the construction of

the C2, C3, C4, and C7 stereocenters of kaitocephalin, formation

the remaining C9 R-amino acid stereocenter could now be

addressed. The C9 stereocenter was envisioned to arise from

the reduction of dehydroamino esters (Z)-4 or (E)-4 to generate

an R-amino ester with the desired (S)-stereochemistry (Scheme

1). As we could not predict a priori which olefin isomer would

undergo hydrogenation to give the desired C9-(S)-diastereomer,

we elected to perform model studies of this dehydroamino ester

hydrogenation reaction on a simplified system before advancing

precious kaitocephalin intermediates through the synthetic

sequence.

Model System: Synthesis of (Z)-Dehydroamino Ester. The

model studies on the formation and hydrogenation of pyrrolidine

dehydroamino esters were initiated with the olefination of

D-prolinal³⁸using N-Bz-protected glycine phosphonate 38.^39,40

A variety of bases were screened to examine the influence of

the reaction conditions on the (E)-/(Z)-dehydroamino ester

product ratio (Table 4). The use of metal hydrides (NaH) or

metal amides (KHMDS) in THF resulted in the formation of

mixtures of dehydroamino esters, (Z)-39 and (E)-39, in ratios

of 1:1-3:1, respectively (entries 1 and 2). Olefination utilizing

Roush-Masamune conditions⁴¹(LiCl/DBU in CH₂Cl₂) pro-

duced a 3:1 mixture of dehydroamino esters (Z)-39 and (E)-39,

respectively, in 73% combined yield (entry 3). The exclusion

of LiCl, however, while retaining the use of DBU as a base,

reaction mixtures

furnished exclusively the (Z)-dehydroamino ester 39 with a (Z)/

(E) ratio of 30:1 in 73% yield. Conditions to exclusively generate

the stereochemically complementary dehydroamino ester (E)-

39 could not be identified. Additionally, a single-crystal X-ray

structure of (Z)-39 was obtained to unambiguously establish the

(Z)-configuration.³⁷

Model System: Hydrogenation Studies of (Z)-Dehy-

droamino Ester. Following the successful synthesis of the (Z)-

dehydroamino ester (Z)-39, the key hydrogenation reaction was

examined. Catalytic hydrogenation of (Z)-39 over 10% Pd/C

in i-PrOH at 15 psi of H₂furnished 40a as a 12:1 mixture of

diastereomers in 92% yield (Scheme 13).⁴²The stereochemistry

of the newly generated center of 40a was elucidated by NOESY

correlation studies performed on the derived rigid bicyclic lactam

41a and its C2-epimer 41b (see below).^43,44In contrast to the

diastereomer 41b (see below), which displayed a positive NOE

correlation between the R-H and methine C-H, lactam 41a

lacked the relevant NOE correlation between the R-H and

methine C-H; the lack of the requisite NOE suggested that

the newly generated stereocenter of 40a was the undesired (R)-

stereochemistry.

We next examined the hydrogenation of the dehydroamino

ester (Z)-39 using chiral catalysis (Table 5) in an attempt to

(42) For other examples of highly diastereoselective substrate directed

hydrogenation of dehydroamino acid derivatives, see: (a) Avenoza, A.;

Cativiela, C.; Peregrina, J. M.; Zurbano, M. M. Tetrahedron: Asymmetry

1997, 8, 863-871. (b) Avenoza, A.; Cativiela, C.; Peregrina, J. M.; Zurbano,

M. M. Tetrahedron: Asymmetry 1996, 7, 1555-1558. (c) Schmidt, U.;

Kumpf, S.; Neumann, K. J. Chem. Soc., Chem. Commun. 1994, 1915-

1916.

(36) (a) Gutowsky, H. S.; Karplus, M.; Grant, D. M. J. Chem. Phys.

1959, 31, 1278-1289. (b) Karplus, M. J. Chem. Phys. 1959, 30, 11-15.

(37) For X-ray crystallographic data, refer to section III of the Supporting

Information.

(38) Diaba, F.; Ricou, E.; Bonjoch, J. Tetrahedron: Asymmetry 2006,

17, 1437-1443.

(43) For the use of TMSOTf in the cleavage of Boc carbamates, see:

(a) Sakaitani, M.; Ohfune, Y. J. Am. Chem. Soc. 1990, 112, 1150-1158.

(b) Sakaitani, M.; Ohfune, Y. J. Org. Chem. 1990, 55, 870-876. (c)

Sakaitani, M.; Ohfune, Y. Tetrahedron Lett. 1985, 26, 5543-5546.

(44) . It should be noted that addition of 2 equiv of TMSOTf was crucial

as the use of only 1 equiv of TMSOTf resulted in the isolation of only

starting material, presumably due to initial reaction of TMSOTf with the

benzamide to generate a TMS imidate.

(39) Schmidt, U.; Lieberknecht, A.; Wild, J. Synthesis 1984, 53-60.

(40) Coleman, R. S.; Carpenter, A. J. J. Org. Chem. 1993, 58, 4452-

4461.

(41) Blanchette, M. A.; Choy, W.; Davis, J. T.; Essenfeld, A. P.;

Masamune, S.; Roush, W. R.; Sakai, T. Tetrahedron Lett. 1984, 25, 2183-

2186.

J. Org. Chem, Vol. 73, No. 5, 2008 1671

Vaswani and Chamberlin

SCHEME 13. Hydrogenation of (Z)-39 and Stereochemical Elucidation

TABLE 5. Hydrogenation of (Z)-39 with Homogeneous Chiral

Catalyst

The predominating diastereofacial preference for the forma-

tion of the (R)-stereoisomer during the hydrogenation of (Z)-

39 was rationalized by a reactive conformation that minimizes

allylic (A^1,3) strain in (Z)-39 (Figure 3a), resulting in the

coalignment of the allylic methine (C-H) with the amine

N-H.^27bIn addition, this model places the pyrrolidine σ_C-NBoc

bond parallel to the π-system of the olefin and effectively allows

the Boc group to shield the re-face of the olefin. This

conformational and stereoelectronic arrangement is mirrored in

the solid state, as evidenced by the X-ray crystal structure of

(Z)-39.³⁷As a consequence, hydrogenation under heterogeneous

conditions predominately delivers H₂to the more accessible face

of the olefin (si-face) via a Felkin-type transition state.

Model System: Synthesis and Hydrogenation of Struc-

tural Congener of “(E)-Dehydroamino Ester”. Based on the

model proposed in Figure 3a, the diastereoselective formation

of the requisite (S)-stereochemistry would necessitate the

addition of H₂to the re-face of the olefin (Z)-39; however, this

face is hindered by the Boc group. By simple extension of this

model, the hydrogenation of the corresponding trans-isomer (E)-

39, on the now-more-accessible re-face, would result in the

formation of the desired (S)-stereoisomer (Figure 3b). Synthetic

access to (E)-39 proved to be elusive since the Horner-

Wadsworth-Emmons olefination of D-prolinal failed to deliver

the (E)-isomer under all conditions tested (Table 4). As a result,

the alternative strategy employing a functional equivalent of

(E)-39 was explored.

The key element of the ultimately successful approach

circumvented the problematic olefination step by substituting

(E)-39 with a synthetically accessible and chemically equivalent

surrogate (Figure 4). Of particular interest in this regard was

the (Z)-â-brominated dehydroamino ester ((Z)-Br-42), which is

a synthetic equivalent of (E)-39 given the propensity of vinyl

halides to undergo dehalogenation. We therefore selected (Z)-

â-brominated dehydroamino ester ((Z)-Br-42) as a target to

verify the stereochemical outcome of the hydrogenation.

The synthesis of the requisite (Z)-â-brominated dehydroamino

ester ((Z)-Br-42) commenced with treatment of (Z)-39 with NBS

in CH₂Cl₂using a protocol adapted from Coleman and co-

workers (Scheme 14).^40,45Subsequent addition of DABCO

immediately promoted tautomerization of the R-bromo imine(s)

(43) to yield mixtures of (Z)- and (E)-â-Br-dehydroamino acid

isomers (Z)-Br-42 and (E)-Br-42, respectively, as judged by ¹H

NMR spectroscopic analysis. Allowing the reaction mixture to

stir over a period of 24 h at room temperature promoted the

^aRatios determined by ¹H NMR spectroscopic analysis of crude reaction

mixtures.

overwhelm the apparent innate preference for reduction from

the “wrong” face noted for achiral catalysts. The hydrogenation

of (Z)-39 using Rh-(2S,5S)-DuPHOS catalyst furnished 40b in

diastereomeric ratio of 2:1 in favor of the desired (S)-

stereochemistry (entry 1). The hydrogenation of (Z)-39 using

Rh-(2S,5S)-DuPHOS catalyst was presumably a consequence

of a mismatched stereodifferentiation, because utilizing the

enantiomeric catalyst, Rh-(2R, 5R)-DuPHOS, resulted in matched

stereodifferentiation between the substrate and the catalyst to

furnish 40a, containing the undesired (R)-stereochemistry in a

15:1 ratio (entry 2).

FIGURE 4. (Z)-Br-42 a structural congener of dehydroamino ester

(E)-39.

FIGURE 3. Proposed models for the reduction of (Z)-39 and (E)-39.

1672 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

SCHEME 14. Bromination/Isomerization of (Z)-42 Using NBS and DABCO

SCHEME 15. Hydrogenation of (Z)-Br-42 and Stereochemical Elucidation

complete conversion of (E)-Br-42 into the desired (Z)-â-Br-

dehydroamino ester ((Z)-Br-42), which was isolated in 74%

yield as a single detectable isomer. The stereochemistry of the

olefin was inferred on the basis of the results of the subsequent

hydrogenation and bicyclic lactam formation (see Scheme 15).

While the major product results from thermodynamic equilibra-

tion of the (E)-isomer, presumably via conjugate addition/

elimination of the base, as shown, it is not obvious why the

(Z)-isomer is the more stable of the two despite the fact that

the product ratio >10:1 clearly suggests a difference in energy

of several kcal/mol favoring the (Z)-isomer (Scheme 14).

prevent catalyst poisoning.⁴⁶Careful analysis of the reaction

using ESI, TLC, and reverse phase LC/MS revealed that

dehalogenation of the C-Br bond is complete within 10-15

min after H₂is introduced, while hydrogenation of the olefin

occurs over a period of 6 h. Furthermore, we see no indication

of isomerization of the alkene bond during this process, which

would result in subsequent alkene hydrogenation from the

opposite face as discussed above.

Comparison of the ¹H NMR spectra in the R-H region of the

newly generated amino ester (4.50-4.60 ppm) of 40b (tenta-

tively assigned as the (S)-stereochemistry) showed clear dif-

ferences from that observed for 40a. To elucidate the stereo-

chemistry of the hydrogenated product, 40b was transformed

into the bicyclic lactam 41b (Scheme 15). Unlike 41a, the

undesired stereoisomer, the NOESY spectrum of 41b revealed

a strong NOE correlation between the R-H atom and the methine

Following the successful assembly of (Z)-Br-42, the key

hydrogenation was investigated (Scheme 15), which requires

not only alkene hydrogenation but also stereoselective removal

of the Br atom. We expected that both of these transformation

of (Z)-Br-42 could be accomplished in a one-pot procedure using

a heterogeneous Pd/C catalyst, since vinyl halides are readily

dehalogenated with Pd/C. Treatment of (Z)-Br-42 with 10%

Pd/C and K₂CO₃at 15 psi of H₂thus uneventfully furnished

40b in 95% yield as a 15:1 mixture of diastereomers, although

the use of K₂CO₃as an acid (HBr) scavenger was essential to

(45) For related studies on the synthesis of â-halogenation of dehy-

droamino esters, see: (a)Yamada, M.; Nakao, K.; Fukui, T.; Nunami, K.

Tetrahedron 1996, 52, 5751-5764. (b) Nunami, K.; Yamada, M.; Fukui,

T.; Matsumoto, K. J. Org. Chem. 1994, 59, 7635-7642. (c) Armstrong, R.

W.; Tellew, J. E.; Moran, E. J. Org. Chem. 1992, 57, 2208-2211.

(46) Rylander, P. N. Catalytic Hydrogenations OVer Platinum Metals;

Academic Press: New York 1967.

J. Org. Chem, Vol. 73, No. 5, 2008 1673

Vaswani and Chamberlin

SCHEME 16. Proposed Conversion of (E)-44 and (Z)-44 to Amino Ester 46a

SCHEME 17. Synthesis of 3,5-Dichloro-4-methoxybenzoic Acid (51) and Phosphonate 49

C-H, providing strong evidence that the desired (S)-stereo-

chemistry was produced during the dehalogenation/hydrogena-

tion of (Z)-Br-42 (Scheme 15). In addition, recrystallization of

bicyclic lactam 41b from EtOAc/hexanes produced crystals that

were suitable for X-ray analysis, thereby unambiguously

establishing the (S)-configuration at the newly generated ste-

reocenter.³⁷

The overwhelming diastereofacial preference for the (S)-

stereochemistry during the hydrogenation of (Z)-Br-42 is

consistent with the hydrogenation model that was proposed

(Scheme 15), in which minimization of allylic (A^1,3) strain in

(Z)-Br-42 results in the co-alignment of the allylic methine (C-

H) with the ester (CO₂Me).^27bIn addition, this model places

the pyrrolidine σ_C-NBocbond parallel to the π-system of the

enamide olefin and effectively allows the Boc group to shield

the si-face of the olefin. As a consequence, the hydrogenation

of (Z)-Br-42 under heterogeneous conditions predominately

delivers H₂to the most accessible re-face of the olefin via a

Felkin-type transition state, generating the (S)-stereochemistry

required for kaitocephalin.

tected 3,5-dichloro-4-hydroxy-benzamide group, followed by

hydrogenation to complete the kaitocephalin carbon skeleton

(Scheme 16). The key element in this strategy was to introduce

simultaneously the C9-stereocenter and the 3,5-dichloro-4-

hydroxybenzamide group of kaitocephalin, thereby minimizing

protecting group manipulations of late stage intermediates. The

main challenge in this plan is to achieve the chemo- and

diastereoselective hydrogenation of the dehydroamino ester

(either (E)-44 or (Z)-Br-45) without the concurrent dehaloge-

nation of the aryl chloride groups to yield amino ester 46a.

The synthesis of the dehydroamino esters commenced with

the preparation of the 3,5-dichloro-4-methoxybenzamide phos-

phonate 49, from phosphonate 50³⁹and acid 51 (Scheme 17).

The olefination partner, aldehyde 52, was prepared in >91%

yield by the oxidative cleavage of alkene 31a with ozone

followed by a reductive workup (Scheme 18). Using the

optimized conditions for the Horner-Wadsworth-Emmons

olefination established in the prolinal model system (see entry

5, Table 4), the olefination of aldehyde 52 using phosphonate

49 in the presence of DBU furnished a 2:1 mixture of (Z)- and

(E)-dehydroamino esters, (Z)-44 and (E)-44, respectively, in

91% combined yield.

Generation of C9-(S)-Stereocenter of Kaitocephalin. For-

mation of N-3,5-Dichloro-4-Methoxybenzamide Protected

Dehydroamino Esters. Guided by the outcome of the model

studies, the initial approach for installing the C9-(S)-stereocenter

of kaitocephalin required the formation of the (E)-dehydroamino

ester ((E)-44)sor its synthetic equivalent the (Z)-Br-dehy-

droamino ester ((Z)-Br-45)scontaining the appropriately pro-

The (Z)- and (E)-dehydroamino esters, (Z)-44 and (E)-44,

respectively, were separable by silica gel chromatography, and

each olefin isomer was carried through the reaction sequence

developed in the model system: the (E)-dehydroamino ester

((E)-44) was immediately treated under hydrogenation condi-

1674 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

SCHEME 18. Synthesis of Dehydroamino Acids (Z)-44 and (E)-44

SCHEME 19. Attempted Conversion of (Z)-44 and (E)-44 to 46a

tions, while the (Z)-dehydroamino ester ((Z)-44) was to be

converted into the (Z)-â-Br-dehydroamino ester ((Z)-Br-45) prior

to hydrogenation. Analogous to the model (see Scheme 16),

these parallel transformations would allow the convergence of

(Z)-44 and (E)-44 to a single desired product, 46a. The

hydrogenation of (E)-44 using 10% Pd/C in i-PrOH, however,

failed to yield any of the desired product 46a, resulting instead

in considerable decomposition of the starting material (Scheme

19). Attempts to advance (Z)-44 through the bromination and

isomerization reaction sequence using NBS and DABCO also

met with limited success. Heating the reaction mixture in an

external bath (65 °C) furnished (Z)-Br-45 in only 10% yield

with considerable recovery of starting material (Z)-44. The use

of a more reactive brominating reagent, bromonium di-sym-

collidine perchlorate (Br⁺[collidine]₂ClO₄^-),⁴⁷only afforded an

intractable mixture that did not contain any detectable quantities

of the desired product. Because of these difficulties, a slightly

more circuitous route was developed to avoid subjecting the

electron-deficient aryl ring to electrophilic halogenation condi-

tions.

Formation of N-Teoc-Protected Dehydroamino Esters. The

failure of the (Z)-dehydroamino ester bromination reaction was

attributed to possible deactivation of the olefin as a result of

inductive effects of the electron-withdrawing 3,5-dichloro-4-

methoxyaryl amide. It was hypothesized that replacing the 3,5-

dichloro-4-methoxyaryl amide with a less electron-withdrawing

group might significantly improve the reactivity of the olefin

toward the brominating reagent, not to mention removing the

aryl ring from harm’s way during the bromination step.

Consequently, the amino group was protected as a trimethyl-

silylethoxy (Teoc) carbamate, rather than being directly acylated

with the natural aryl acyl group. The general plan, outlined in

Scheme 16, envisioned the formation of 46a, containing the

(47) (a) Lemieux, R. U.; Levine, S. Can. J. Chem. 1964, 42, 1473-

1480. (b) Lemieux, R. U.; Morgan, A. R. Can. J. Chem. 1965, 43, 2190-

2197.

J. Org. Chem, Vol. 73, No. 5, 2008 1675

Vaswani and Chamberlin

SCHEME 20. Synthesis of N-Teoc-Protected Glycine

Phosphonate 56

of oxazolidinone side-product 61 (10%). The geometry of (Z)-

Br-48 was inferred based on the results obtained from the

subsequent dehalogenation and hydrogenation reaction (see

Scheme 23).

Subsequent dehalogenation and hydrogenation of (Z)-Br-48

was accomplished using 20 mol % of 10% Pd/C and K₂CO₃in

i-PrOH under 15 psi of H₂to afford 62a as a >13:1 mixture of

C9 diastereomers in 98% yield (Scheme 23a). Careful monitor-

ing of the reaction by ESI-MS and TLC provided evidence for

the dehalogenation of the vinyl-Br (C-Br f C-H) within 10-

15 min, followed by slow hydrogenation of the olefin to yield

62a. The catalytic hydrogenation of the dehydroamino ester (E)-

47 using 20 mol % of 10% Pd/C in i-PrOH at one atmosphere

of H₂afforded 62a in 95% yield as a single diastereomer

(Scheme 23), thereby successfully converging both dehy-

droamino esters to the desired intermediate. The diastereose-

lectivity observed for the hydrogenation of (Z)-Br-48 and (E)-

47 can be rationalized via a model analogous to that proposed

for the diastereofacial hydrogenation of the proline model system

(see Figure 3).

desired C9-(S)-stereochemistry, from dehydroamino esters (E)-

47 and (Z)-47 (Via (Z)-Br-48).

The formation of the requisite N-Teoc protected dehy-

droamino ester necessitated the synthesis of the N-Teoc

protected phosphonate 56 (Scheme 20). Initial attempts to

prepare 56 by hydrogenolysis of the N-Cbz-protected phospho-

nate 50 followed by treatment of the resultant amine 55 with

trimethylsilylethoxy chloroformate (Teoc-Cl)⁴⁸or trimethyl-

silylethoxy carbonylimidazole afforded 56 in low yields (5-

29%). An alternative method for the construction of phosphonate

56 was devised starting from the glycine methyl ester hydro-

chloride salt (57). N-Acylation of 57 was accomplished by

treatment with freshly prepared Teoc-Cl and K₂CO₃in a biphasic

mixture of THF and H₂O to furnish 58 in 97% yield (Scheme

20). Radical bromination of 58 employing NBS in CCl₄

furnished bromide 59 after 41/2 h at ambient temperature.⁴⁹The

unstable bromide 59 was immediately treated with P(OMe)₃in

THF to produce phosphonate 56 in 50-89% by an Arbuzov

reaction.

The initial experiments for the installation of the C9-

stereocenter of kaitocephalin were performed on separated

samples of (E)-47 and (Z)-47. For practical reasons, however,

we sought to minimize the separate processing of isolated

samples of (E)-47 and (Z)-47 through the reaction sequence.

Ultimately, it was determined that the diastereomeric mixture

of (E)-47 and (Z)-47 produced from the Horner-Wadsworth-

Emmons olefination reaction could be directly and efficiently

converted into pure (Z)-Br-48 (Scheme 24) without separation.

To that end, a 1:1 diastereomeric mixture of (E)-47 and (Z)-47,

obtained from the olefination of aldehyde 52, was subjected to

Br⁺[collidine]₂PF₆^-, followed by the addition of DABCO to

provide the (Z)-Br-48 in 90% yield as a chromatographically

inseparable >13:1 mixture of (Z)-Br and (E)-Br isomers.⁵¹

Catalytic hydrogenation of (Z)-Br-48 over Pd/C in the presence

of K₂CO₃furnished 62a as a single detectable diastereomer in

98% yield.

Optimal conditions for the Horner-Wadsworth-Emmons

olefination of aldehyde 52 and N-Teoc phosphonate 56 em-

ployed LDA in THF, and warming the reaction mixture from

-78 °C to ambient temperatures afforded a mixture of dehy-

droamino esters, (E)-47 and (Z)-47 (dr (E)/(Z) ) 1:1) respec-

tively, in quantitative yield (Scheme 21). The (E)-dehydroamino

ester and (Z)-dehydroamino ester were separable by silica gel

chromatography and each isomer was characterized by diag-

nostic ¹H NMR chemical shifts and NOESY correlation studies.

¹H NMR spectroscopic analysis of (Z)-47 revealed a downfield

chemical shift for the vinyl C-H when compared to (E)-47.

The downfield chemical shift of the vinyl C-H of (Z)-47 is

indicative of (Z)-olefin geometry. NOESY correlation studies

disclosed characteristic NOEs between the N-H and allylic

C-H for (Z)-47, whereas this NOE was absent in (E)-47.

Following the preparation of (E)-47 and (Z)-47, the instal-

lation of the C9-stereocenter of kaitocephalin was addressed

successfully. First, the conversion of the dehydroamino ester

(Z)-47 to product 46a, containing the requisite C9-(S) isomer,

demanded the successful bromination-isomerization, hydro-

genation, Teoc removal, and finally amide bond formation.

Accordingly, the (Z)-47 was treated with bromonium di-sym-

End Game: Completion of the Synthesis of Kaitocephalin

Following the successful preparation of 62a, which contains

the complete carbon backbone and all of stereogenic centers of

the natural product, we were poised to complete the total

synthesis of 1. At that point, all that remained was the

introduction of the 3,5-dichloro-4-hydroxyaryl amide, formation

of the C1 carboxylic acid, and removal of the protecting groups.

Introduction of the 4-methoxy-3,5-dichloroaryl amide neces-

sitated the removal of the Teoc protecting group, which proved

to be a delicate task. Exposure of 62a to commercially available

TBAF (1.0 M in THF) or solid TBAF‚3H₂O cleanly removed

the Teoc carbamate; the resultant amino ester, however, was

generated with varying degrees of epimerization at the C9 center,

which presumably occurred as a consequence of the basic

reaction conditions. In an attempt to attenuate the basicity of

the F^-anion in the reaction mixture, the TBAF deprotection of

62a was buffered with AcOH, which unfortunately suppressed

the deprotection process completely.

collidine hexafluorophosphorus⁵⁰and Br⁺[collidine]₂PF₆(to

-

generate R-bromo imine 60), followed by the addition of

DABCO to yield the (Z)-Br-48 as a 13:1 mixture of (Z)-/(E)-

isomers in 86% yield (Scheme 22). Temperature control during

this transformation was essential to minimize the production

(48) Shute, R. E.; Rich, D. H. Synthesis 1987, 346-348.

(49) Nakamura, Y.; Matsubara, R.; Kiyohara, H.; Kobayashi, S. Org.

Lett. 2003, 5, 2481-2484.

(50) Homsi, F.; Robin, S.; Rousseau, G. Org. Synth. 2000, 77, 206-

211.

(51) The (E)-dehydroamino ester ((E)-47) reacted with Br⁺[collidine]₂PF₆^-

within 5-6 min to yield the R-Br imine, while the conversion of the (Z)-

dehydroamino ester ((Z)-47) to the corresponding R-Br imine occurred over

60-90 min.

1676 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

SCHEME 21. Olefination of 52 with N-Teoc Phosphonate 56

SCHEME 22. Bromination/Isomerization of Dehydroamino Ester (Z)-47

We then conjectured that sparingly soluble sources of F^-

anion,⁵²such as CsF⁵³or KF,⁵⁴might sufficiently attenuate the

basicity of the reaction mixture to allow for deprotection of the

Teoc group without compromising the stereochemical integrity

of the C9-stereocenter. Gratifyingly, addition of a solution of

62a in DMSO to freshly dried CsF cleanly deprotected the Teoc

carbamate after 13 h to afford amino ester 63 in near quantitative

yield, with no detectable epimerization at the C9-stereocenter

(Scheme 25). Subsequent amide coupling with 4-methoxy-3,5-

dichlorobenzoic acid 51 was effected with the aid of EDCI‚HCl

to yield 46a in 94% yield over two steps.

oxidation of diol 64 was then achieved with catalytic TEMPO

and NaOCl₂to provide carboxylic acid 65 in 88% yield from

acetal 46a.⁵⁵The susceptibility of the C2-N in 65 toward lactam

formation (with the C18 ester) precluded the use of harsh Lewis

acids, mineral acids, or strong base for the deprotection of the

methyl ether and methyl esters. Ultimately, it was determined

that treatment of 65 with LiI in EtOAc (60-80 °C) over a period

of 1 week effectively promoted the cleavage of the methyl ether

and esters,⁵⁶which was followed by the in situ deprotection of

the Boc carbamates using trifluoroacetic acid (TFA) to deliver

crude kaitocephalin (1) as a dark brown solid.

Completion of the synthesis proceeded uneventfully with the

hydrolysis of the cyclohexylidene N,O-acetal 46a, which was

accomplished using aqueous 80% AcOH at ambient tempera-

tures to furnish amino diol 64 (Scheme 25). Chemoselective

Purification of the crude product by reversed-phase HPLC

under acidic conditions (10:90 CH₃CN/H₂O with 0.1% TFA)

followed by lyophilization afforded kaitocephalin (1) as the bis-

(55) (a) Palian, M. M.; Polt, R. J. Org. Chem. 2001, 66, 7178-7183.

(b) Zhao, M.; Li, J.; Mano, E.; Song, Z.; Tschaen, D. M.; Grabowski, E. J.

J.; Reider, P. J. J. Org. Chem. 1999, 64, 2564-2566.

(56) Fisher, J. W.; Trinkle, K. L. Tetrahedron Lett. 1994, 35, 2505-

2508.

(52) Clark, J. H. Chem. ReV. 1980, 80, 429-452.

(53) Crowley, B. M.; Mori, Y.; McComas, C. C.; Tang, D.; Boger, D.

L. J. Am. Chem. Soc. 2004, 126, 4310-4317.

(54) Roush, W. R.; Blizzard, T. A. J. Org. Chem. 1984, 49, 4332-4339.

J. Org. Chem, Vol. 73, No. 5, 2008 1677

Vaswani and Chamberlin

SCHEME 23. Hydrogenation of Dehydroamino Esters (Z)-Br-48 and (E)-47

SCHEME 24. Efficient Conversion of Mixtures of (E)-47 and (Z)-47 to 62a

TFA salt in 20% yield (>72% yield per protecting group

removed, five in all). The lactam 67 was also isolated during

HPLC purification, which presumably arises as a result of acid

(TFA)-catalyzed cyclization of the C2 amine onto the C18 ester.

High-resolution mass spectrometry analysis of purified 1

correlated with the reported mass of kaitocephalin, and ¹H NMR,

¹³C NMR, and optical rotation [observed [R]²⁵_D-35.4 (c 0.13,

H₂O) [lit.^5a[R]²¹-31.0 (c 0.70, H₂O), lit.^7a[R]²⁵-26.6 (c

Conclusion

The studies reported in this paper culminated in the total

synthesis of (-)-kaitocephalin in 11 steps (eight isolated and

purified intermediates) and 6% overall yield starting from the

known intermediate trans-6. The overall efficiency profited from

the use of substrate-controlled manipulations to establish the

requisite C3, C4, C7, and C9 stereocenters. Two strategies were

developed for the diastereoselective formation of the trans-2,5-

disubstitited pyrrolidine core (trans-6) of kaitocephalin contain-

ing the C7 stereocenter. The first approach utilized a standard

diastereoselective organocuprate addition to a chiral N-acylimin-

ium ion; the second and ultimately superior strategy employed

a novel stereoconvergent cyclization reaction for the preparation

of the trans-2,5-disubstituted pyrrolidine core. The C4 and C3

D

0.19, H₂O), lit.⁸[R]²⁴-29.0 (c 0.30, H₂O)]] were consistent

D

1

with the natural compound, although the H NMR coincided

perfectly with the published spectrum of kaitocephalin only after

conversion of the bis-TFA salt of 1 to its mono-diethylammo-

nium salt with a 2 mM aqueous solution of pH 7.0 Et₂NH/CO₂

buffer.^57,58

(57) The published proton NMR spectrum for the kaitocephalin clearly

shows signals assignable to the diethylammonium counterion; however, they

were not included in the tabulated ¹H NMR data included in the

Experimental Section.

(58) We thank Dr. Shin-ya for sending a sample of kaitocephalin that

we requested for HPLC comparison; unfortunately, HPLC and HRMS

analysis of the material that was received revealed only decomposition

products.

1678 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

SCHEME 25. Completion of the Synthesis of Kaitocephalin

quenched by the addition of AcOH in THF (50.0 mL, 10% AcOH

in THF). The mixture was placed in a 0 °C bath and allowed to

stir for 15 min, and then saturated NaHCO₃(50.0 mL) was added.

After 20 h, the blue reaction mixture was extracted,diluted with

H₂O (50.0 mL), and extracted with EtOAc (4 × 150 mL). The

combined organic layers were washed with 1.0 N NaHSO₄(100

mL) and brine (500 mL), dried over MgSO₄, and concentrated in

vacuo to afford a green foam. ¹H NMR analysis of unpurified

material revealed that â-keto ester 30 was formed as a single

diastereomer. The foam was purified by flash chromatography

(gradient elution 5:95-10:90-20:80-100:0 EtOAc/hexanes) to

yield â-keto ester 30 as a white foam (4.46 g, 49%, single

diastereomer): ¹H NMR (500 MHz, DMSO-d₆, 375 K) δ 5.35 (app

dsep, J ) 8.9, 1.3, 1H), 4.92-4.90 (m, 1H), 4.63 (td, J ) 8.8, 1.8,

1H), 4.01-3.98 (m, 2H), 3.71 (s, 3H), 2.55 (ddd, J ) 12.2, 6.1,

2.7, 1H), 2.30-2.08 (td, J ) 12.0, 6.5, 1H and m, 3H), 1.86-1.82

(m, 1H), 1.72 (d, J ) 1.1, 3H), 1.67 (d, J ) 1.2, 3H), 1.62-1.54

(m, 4H), 1.54-1.50 (m, 1H), 1.50-1.45 (m, 2H), 1.42 (s, 9H),

1.38 (s, 9H), 1.18-1.07 (m, 1H); ¹³C NMR (125 MHz, DMSO-d₆,

only major rotamer reported at 298 K) δ 204.6, 170.8, 154.8, 151.8,

131.5, 126.4, 95.7, 80.4, 80.3, 77.3, 66.3, 60.7, 57.9, 52.9, 35.7,

stereocenters of kaitocephalin were introduced sequentially using

a diastereoselective modified Claisen condensation and a chemo-

and diastereoselective DIBALH reduction of the â-keto ester

30. The stereochemistry of the C9-(S)-amino ester was estab-

lished with excellent control by the substrate directed hydro-

genation of a dehydroamino ester derivative to complete the

carbon skeleton of kaitocephalin and introduce the C9-(S)

stereocenter as a single diastereomer. This route allows for the

preparation of several key intermediates in gram-scale quantities,

which in turn is provides relatively large quantities of kaito-

cephalin for complete biological evaluation.

Experimental Section

â-Keto Ester 30. Under a constant atmosphere of N₂, ester

trans-6 (4.70 g, 16.6 mmol) was diluted with THF (31.0 mL) and

cooled to -78 °C. Freshly prepared LDA (24.0 mL, 24.0 mmol,

1.00 M in THF) was added dropwise over 10 min. The reaction

mixture was allowed to stir for 3 h at -78 °C, after which time a

solution of 26 (8.44 g, 25.2 mmol) in THF (20.0 mL) was added

dropwise over 15 min. After 90 min at -78 °C, the reaction was

J. Org. Chem, Vol. 73, No. 5, 2008 1679

Vaswani and Chamberlin

33.4, 33.2, 31.5, 28.5, 28.3, 25.7, 25.2, 23.7, 23.5, 18.1; IR (KBr)

2976, 2933, 2868, 1754, 1720, 1694, 1454, 1392, 1366, 1257, 1159

cm^-1; HRMS (ESI/methanol) m/z calcd for C₂₉H₄₆N₂O₈(M + Na)⁺

47 was detected. Solid DABCO (1.50 g, 13.4 mmol) was added in

one portion, and the reaction mixture was allowed to gradually

warm to room temperature over 3 h. After being stirred for an

additional 16 h, the yellow suspension was concentrated in vacuo,

diluted with EtOAc (100 mL), and washed with saturated aqueous

Na₂S₂O₃(2 × 50 mL). The aqueous layer was extracted with EtOAc

(3 × 100 mL). The combined organic extracts were washed with

1.0 N NaHSO₄(aq) (3 × 50 mL), saturated NaHCO₃(100 mL),

and brine (100 mL). The organic extract was dried over MgSO₄

and concentrated to give a white foam. Normal-phase HPLC

analysis (2:98 i-PrOH/hexanes, 1.0 mL/min, 254 nm) of the

unpurified material revealed a ratio of (Z)-Br-48/(E)-Br-48 ) 13.3:

1.00. The foam was purified by flash chromatography (gradient

elution 20:80-60:40 EtOAc/hexanes) to yield an inseparable

mixture of (Z)-Br-48:(E)-Br-48 as a white foam (2.00 g, 90%, dr

(Z)-Br-48/(E)-Br-48 ) 13.0:1.0).

573.3152, found 573.3137; [R]²⁴+66.6 (c 0.64, CHCl₃).

D

â-Hydroxy Ester 31a. Under a constant atmosphere of Ar,

â-keto ester 30 (3.98 g, 7.23 mmol) was diluted with THF (102

mL) and cooled to -78 °C. To this cooled solution was added

DIBALH (1.0 M in toluene, 10.0 mL, 10.0 mmol, 1.38 equiv). The

reaction mixture was allowed to warm to ambient temperature

overnight. The reaction was cooled (0 °C) and quenched with

saturated aqueous sodium-potassium tartarate (10 mL). The

resultant biphasic suspension was warmed to room temperature and

stirred vigorously for 30 min, and aqueous 1.0 N NaHSO₄(30 mL)

was added to remove emulsions. The organic phase was separated,

and the aqueous phase was extracted with EtOAc (4 × 50 mL).

The combined organic extracts were washed with brine (50 mL),

dried over MgSO₄, and concentrated in vacuo to provide a pink

foam. Normal-phase HPLC analysis (2:98 i-PrOH/hexanes, 1.0 mL/

min, 220 nm) of the unpurified material revealed dr (31a/31b))

30.0:1.00. The foam was purified by flash chromatography (gradient

elution 10:90-40:60 EtOAc/hexanes) to yield â-hydroxy ester 31a

as a white foam (3.45 g, 86%, dr 31a/31b ) > 30:1).31a: ¹H NMR

(500 MHz, DMSO-d₆, 375 K) δ 5.32 (app dsep, J ) 8.9, 1.4, 1H),

5.25 (br s, 1H), 4.57 (td, J ) 8.4, 3.3, 1H), 4.43 (br s, 1H), 4.05

(ddd, J ) 6.9, 3.5, 1.8, 1H), 4.01 (dd, J ) 8.8, 3.5, 1H), 3.84 (dd,

J ) 8.8, 6.9, 1H), 3.70 (s, 3H), 2.41-2.36 (m, 1H), 2.25-2.05 (m,

4H), 1.81-1.78 (m, 1H), 1.69 (d, J ) 1.3, 3H), 1.63 (d, J ) 1.2,

3H), 1.61-1.55 (m, 3H), 1.53-1.47 (m, 4H and s, 9H), 1.40 (s,

9H), 1.15-1.07 (m, 1H); ¹³C NMR (125 MHz, CDCl₃, only major

rotamer reported at 298 K) δ 173.2, 157.2, 152.8, 131.1, 126.1,

94.6, 81.1, 80.1, 75.3, 73.9, 62.2, 59.9, 58.1, 52.8, 34.5, 34.4, 32.7,

30.9, 28.6, 28.4, 25.6, 25.2, 23.7, 23.5, 17.9; IR (KBr), 3358, 2977,

(Z)-Br-48: ¹H NMR (500 MHz, DMSO-d₆, 385 K) δ 8.16 (br s,

1H), 5.08 (app t, J ) 7.5, 1H), 4.71 (d, J ) 2.9, 1H), 4.16-4.12

(m, 3H), 4.07 (ddd, J ) 6.4, 3.2, 3.2, 1H), 3.79 (dd, J ) 8.8, 6.6,

1H), 3.73 (s, 3H), 3.66 (s, 3H), 2.35 (ddd, J ) 12.8, 7.9, 4.9, 1H),

2.26-2.15 (m, 3H), 2.09 (td, J ) 13.3, 4.3, 1H), 1.82-1.80 (m,

1H), 1.78-1.71 (m, 1H), 1.61-1.58 (m, 3H), 1.54-1.50 (m, 3H),

1.46 (s, 9H), 1.39 (s, 9H), 1.16-1.08 (m, 1H), 0.99-0.96 (m, 2H),

0.06 (s, 9H); ¹³C NMR (125 MHz, CDCl₃, only major rotamer

reported at 298 K) δ 171.5, 163.0, 155.4, 153.5, 152.8, 127.3, 126.5,

94.8, 82.1, 80.3, 75.5, 75.0, 64.8, 62.7, 62.1, 57.8, 52.9, 52.5, 35.0,

32.1, 29.6, 28.6, 28.6, 28.3, 25.1, 23.6, 23.5, 17.7, -1.5; IR (thin

film) 3414, 3319, 3006, 2952, 2865, 1732, 1693, 1632, 1479, 1393,

1367, 1320, 1250, 1172 cm^-1; HRMS (ESI/methanol) m/z calcd

for C₃₅H₅₈BrN₃O₁₂Si (M + Na)⁺842.2871, found 842.2856; [R]²³

+5.2 (c 0.20, CHCl₃).

D

Amino Ester 62a. Under a constant atmosphere of N₂(g), 10%

Pd/C (0.344 g, 0.325 mmol, 20 mol %) was added to a suspension

of (Z)-Br-48 (1.34 g, 1.63 mmol, dr (Z)-Br-48/(E)-Br-48 ) 13.0:

1.0) and K₂CO₃(0.378 g, 2.73 mmol) in i-PrOH (67 mL). Three

cycles of evacuation (30 s) followed by purging with N₂(g) was

conducted. To the partially degassed suspension was introduced

H₂(g) (1atm) via a H₂(g) filled balloon, followed by three cycles

of evacuation (30 s) and purging with H₂(g). After 6 h, excess H₂

(g) was removed by three cycles of evacuation (30 s) and purging

with N₂(g). The suspension was filtered over a pad of Celite (1.5

in. thick), the pad was washed with CH₂Cl₂(3 × 75 mL), and the

2932, 2863, 1750, 1693, 1455, 1392, 1366, 1253, 1160 cm^-1

;

HRMS (ESI/methanol) m/z calcd for C₂₉H₄₈N₂O₈(M + Na)⁺

575.3308, found 575.3300; [R]²⁴-60.1 (c 0.78, CHCl₃).

D

31b: ¹H NMR (500 MHz, DMSO-d₆, 375 K) δ 5.46 (app dsep,

J ) 8.8, 1.3, 1H), 4.56 (td, J ) 8.5, 3.1, 1H), 4.31 (dd, J ) 9.0,

5.5, 1H), 4.07 (d, J ) 5.3, 1H), 4.00 (dd, J ) 9.0, 4.8, 1H), 3.77

(dd, J ) 8.7, 5.0, 1H), 3.68 (s, 3H), 3.63 (d, J ) 8.7, 1H), 2.69

(ddd, J ) 13.2, 8.7, 4.2, 1H), 2.32-2.18 (m, 3H), 2.04 (ddd, J )

13.5, 9.2, 9.2, 1H), 1.70 (d, J ) 0.92, 3H), 1.64 (d, J ) 1.1, 3H),

1.63-1.56 (m, 3H), 1.54-1.47 (m, 4H), 1.44 (s, 9H), 1.36 (s, 9H),

1.17-1.10 (m, 1H); ¹³C NMR (125 MHz, CDCl₃, only major

rotamer reported at 298 K) δ 175.8, 154.8, 152.3, 132.2, 127.3,

94.6, 81.2, 80.0, 74.0, 72.1, 66.4, 59.3, 57.5, 52.5, 36.1, 35.6, 32.3,

31.2, 28.5, 28.3, 28.2, 25.9, 25.0, 23.3, 17.9; IR (KBr), 3358, 2977,

1

filtrate was concentrated in vacuo to give a white foam. H NMR

analysis of the unpurified material revealed one major diastereomer,

dr (62a/62b) ) >13.2:1.0). The foam was purified by flash

chromatography (gradient elution 10:90-20:80-40:60 EtOAc/

hexanes) yielded 62a as a white foam (1.18 g, 98%).

2932, 2863, 1750, 1693, 1455, 1392, 1366, 1253, 1160 cm^-1

;

HRMS (ESI/methanol) m/z calcd for C₂₉H₄₈N₂O₈(M + Na)⁺

62a: ¹H NMR (500 MHz, DMSO-d₆, 375 K) δ 6.94 (d, J ) 7.5,

1H), 5.26 (br s, 1H), 4.43 (dd, J ) 9.5, 1.4, 1H), 4.12-4.08 (m,

3H), 4.02-4.00 (m, 1H), 3.97 (dd, J ) 8.9, 3.4, 1H), 3.94-3.91

(m, 1H), 3.82 (dd, J ) 8.9, 7.0, 1H), 3.68 (s, 3H), 3.67 (s, 3H),

2.42-2.37 (app m, 1H), 2.29 (ddd, J ) 14.0, 5.9, 2.2, 1H), 2.24-

2.18 (m, 2H), 2.08 (td, J ) 13.4, 4.3, 1H), 2.00-1.92 (m, 1H),

1.81-1.76 (m, 2H), 1.67 (ddd, J ) 14.0, 9.9, 8.1, 1H), 1.62-1.56

(m, 3H), 1.52-1.49 (m, 3H and s, 9H), 1.47-1.42 (s, 9H), 1.15-

1.06 (m, 1H), 0.97-0.93 (m, 2H), 0.05 (s, 9H); ¹³C NMR (125

MHz, CDCl₃, only major rotamer reported at 298 K) δ 173.1, 172.6,

156.5, 156.2, 152.8, 94.6, 82.3, 80.2, 75.4, 73.4, 63.7, 61.8, 59.6,

57.9, 52.9, 52.8, 52.5, 36.1, 34.4, 33.8, 32.5, 28.8, 28.5, 28.4, 25.1,

23.6, 23.5, 17.7, -1.46; IR (thin film) 3332, 2953, 2943, 2863,

1749, 1720, 1696, 1681, 1390, 1367, 1250, 1161 cm^-1; HRMS

(ESI/methanol) m/z calcd for C₃₅H₆₁N₃O₁₂Si (M + Na)⁺766.3922,

575.3308, found 575.3305; [R]²⁵+12.7 (c 0.27, CHCl₃).

D

1

33: isolated as a white solid (0.302 g, 7.9%); H NMR (600

MHz, DMSO-d₆, 375 K) δ 5.28 (app dsep, J ) 8.7, 1.4, 1H), 4.75

(br s, 1H), 4.58 (td, J ) 8.2, 3.1, 1H), 4.28 (s, 1H), 4.06 (ddd, J )

7.1, 3.8, 1.3, 1H), 3.99 (dd, J ) 8.6, 3.7, 1H), 3.85-3.82 (dd, J )

8.6, 7.2, 1H and d, J ) 11.1, 1H), 3.75 (d, J ) 10.8, 1H), 2.26-

2.21 (m, 1H), 2.11-2.01 (m, 3H), 1.95-1.91 (m, 1H), 1.84-1.80

(m, 1H), 1.66 (d, J ) 1.2, 3H), 1.62 (d, J ) 1.1, 3H), 1.61-1.55

(m, 4H), 1.53-1.46 (m, 3H and s, 9H), 1.43-1.40 (m, 1H), 1.38

(s, 9H), 1.14-1.07 (m, 1H); ¹³C NMR (125 MHz, CDCl₃, only

major rotamer reported at 298 K) δ 156.6, 153.0, 130.8, 127.7,

94.6, 81.2, 80.5, 74.7, 70.5, 65.8, 65.1, 62.7, 59.4, 58.6, 35.7, 31.6,

30.2, 28.5, 28.4, 25.6, 25.0, 23.5, 23.3, 18.0; HRMS (ESI/methanol)

m/z calcd for C₂₈H₄₈N₂O₇(M + Na)⁺547.3359, found 547.3344.

â-Bromo Enamide 48. To a cooled (0 °C) solution of (E)-47

and (Z)-47 (2.00 g, 2.70 mmol, 1:1 mixture of (E)-47:(Z)-47) in

CHCl₃(35 mL) was added Br[collidine]₂PF₆(1.51 g, 3.23 mmol).

The resultant suspension was stirred vigorously at 0 °C. The

progress of the reaction was monitored by TLC (30:70 EtOAc/

hexanes). After 5 min, complete consumption of (E)-47 was

observed by TLC analysis, and after 1 h the disappearance of (Z)-

found 766.3942; [R]²³-21.1 (c 0.60, CHCl₃).

D

62b: ¹H NMR (500 MHz, DMSO-d₆, 375 K) δ 6.64 (d, J ) 7.6,

1H), 4.44 (d, J ) 7.6, 1H), 4.12-4.02 (m, 4H), 3.99-3.95 (m,

2H), 3.83 (dd, J ) 8.8, 7.0, 1H), 3.68 (s, 3H), 3.66 (s, 3H), 2.43-

2.39 (m, 1H), 2.24-2.17 (m, 2H), 2.15-2.05 (m, 2H), 2.02-1.94

(m, 1H), 1.82-1.77 (m, 2H), 1.70-1.65 (m, 1H), 1.61-1.55 (m,

3H), 1.54-1.49 (m, 2H), 1.47 (s, 9H), 1.47-1.44 (s, 9H and m,

1680 J. Org. Chem., Vol. 73, No. 5, 2008

Total Synthesis of (-)-Kaitocephalin

1H), 1.16-1.07 (m, 1H), 0.97-0.94 (m, 2H), 0.05 (s, 9H); ¹³C

NMR (125 MHz, CDCl₃, only major rotamer reported at 298 K) δ

173.1, 172.6, 156.6, 156.4, 152.8, 94.6, 82.3, 80.2, 75.4, 73.6, 63.7,

61.8, 58.6, 57.9, 53.0, 52.7, 52.0, 35.8, 34.4, 33.8, 32.6, 28.7, 28.6,

27.8, 25.1, 23.6, 23.4, 17.7, -1.44; IR (thin film) 3332, 2954, 2933,

Bis-TFA salt 1: ¹H NMR (500 MHz, D₂O) δ 7.81 (s, 2H), 4.60

(dd, J ) 8.4, 5.5, 1H), 4.54 (s, 1H), 4.30 (s, 1H), 3.89-3.82 (m,

1H), 2.59 (ddd, J ) 13.8, 6.8, 6.8, 1H), 2.42 (ddd, J ) 13.5, 6.4,

2.4, 1H), 2.28-2.20 (m, 2H), 2.15 (ddd, J ) 13.5, 11.2, 6.7, 1H),

1.78-1.70 (m, 1H); ¹³C NMR (125 MHz, D₂O, uncalibrated)⁵⁹

δ

2863, 1747, 1720, 1696, 1679, 1390, 1368, 1250, 1161 cm^-1

;

174.2, 170.6, 167.8, 151.5, 128.0, 125.9, 122.0, 76.6, 70.5, 58.6,

55.6, 52.4, 34.1, 31.9, 29.7; HRMS (ESI/methanol) m/z calcd for

C₁₈H₂₁Cl₂N₃O₉(M + Na)⁺516.0552, found 516.0557; [R]²⁵_D-35.4

(c 0.13, H₂O).

TFA salt 67: ¹H NMR (500 MHz, D₂O) δ 7.88 (s, 2H), 4.60 (d,

J ) 7.4, 1H), 4.53 (dd, J ) 9.1, 5.0, 1H), 3.94-3.89 (d, J ) 7.4,

1H and m, 1H), 2.63-2.57 (m, 2H), 2.39-2.33 (m, 1H), 2.27 (ddd,

J ) 14.8, 8.9, 6.2, 1H), 2.13 (ddd, J ) 13.7, 8.3, 3.3, 1H), 2.04-

1.96 (m, 1H); ¹³C NMR (125 MHz, D₂O, uncalibrated)⁵⁹δ 175.7,

174.9, 171.8, 167.7, 151.5, 127.9, 125.8, 122.0, 73.8, 72.5, 61.1,

59.9, 52.6, 34.0, 29.4, 27.2; HRMS (ESI/methanol) m/z calcd for

C₁₈H₁₉Cl₂N₃O₈(M + H)⁺476.0627, found 476.0613; [R]²⁵_D-7.4

(c 0.10, H₂O).

HRMS (ESI/methanol) m/z calcd for C₃₅H₆₁N₃O₁₂Si (M + Na)⁺

766.3922, found 766.3920; [R]²⁵-18.4 (c 0.17, CHCl₃).

D

Bis-TFA Salt 1. To a mixture of 65 (0.144 g, 0.195 mmol) in

anhydrous EtOAc (5 mL) was added LiI (0.276 g, 2.07 mmol) in

one portion. The resulting brown homogeneous reaction mixture

was heated (60 °C) in an oil bath. After 30 h, the formation of

solid precipitates was observed, and AcOH (0.10 mL) was added

to solubilize the precipitates. The consumption of starting material

65 was monitored by LC/MS (C18-reversed-phase HPLC, gradient

elution 2:98-95:5 CH₃CN/H₂O with 0.1% AcOH). Over the course

of 5 days, additional LiI (0.342 g, 2.56 mmol) and AcOH (0.50

mL) were introduced to the reaction mixture, and the external bath

temperature was increased (90 °C). After complete consumption

of 65 was judged by LC/MS analysis, the reaction was allowed to

cool to room temperature, concentrated in vacuo, and dried (0.50

mmHg) to afford a brown paste.

Acknowledgment. We thank the National Institute of

Neurological Disorders and Stroke (NS-27600) for generously

supporting this research. We thank Dr. John Greaves for mass

spectrometric data, Dr. Phil Dennison for NMR assistance, and

Dr. Joe Ziller for X-ray crystallographic analyses.

To a cooled (0 °C) solution of the resultant brown paste in

CH₂Cl₂(25 mL) was added TFA (10.0 mL, 130 mmol). After 1 h,

the reaction was concentrated in vacuo, and excess TFA was

azeotropically removed by co-distillation with hexanes (5 × 25 mL)

to afford a brown solid. The solid was partitioned between EtOAc

(10 mL) and H₂O (30 mL). The aqueous phase was separated, and

the brown organic phase was washed with H₂O (3 × 25 mL) and

saturated aqueous Na₂S₂O₃(5 mL). The combined aqueous phases

were lyophilized to afford an off-white solid. The resultant solid

was purified by preparatory reversed-phase HPLC (10:90 CH₃CN:

H₂O with 0.8% TFA, 210 nm), and the collected fractions were

lyophilized to furnish the bis-TFA salt of 1 as a white powder (0.028

g, 20%) and the TFA salt of 67 as a white powder (0.011 g, 9.6%).

Supporting Information Available: Experimental procedures,

compound characterization, X-ray crystallographic data, and spectral

data for selected compounds. This material is available free of

charge via the Internet at http://pubs.acs.org.

JO702329Z

(59) Due to observed broadness of some carbon signals, several ¹³C NMR

chemical shift values were obtained indirectly through ¹H-¹³C HMBC and

HMQC correlations.

J. Org. Chem, Vol. 73, No. 5, 2008 1681

Article Doi

Stereocontrolled total synthesis of (-)-kaitocephalin

DOI: 10.1021/jo702329z

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jo702329z

Full text of DOI:10.1021/jo702329z

Products guided by the article

R&D Labs maybe for 1010808-11-3

Relevant to this article

Hot Product