Synthesis of the Aziridinomitosene Skeleton by Intramolecular Michael Addition: α-Lithioaziridines and Nonaromatic Substrates

Article abstract of DOI:10.1021/jo030224a

The bicyclic pyrrole ketone 16 has been prepared by using an oxaza-Claisen rearrangement, followed by nitrogen deprotection. Coupling with the stannylaziridine mesylate 15a or nosylate 15b affords 17. Conversion to 40 provides a substrate for generation of an α-lithioaziridine 41 by tin lithium exchange. An intramolecular Michael addition pathway for 41 has been demonstrated by the isolation of 46 in 20% yield under conditions where the intermediate enolate 43 is trapped by selenenylation, but competing proton transfer gives 42. The synthetic potential of the process is limited by stability problems at the stage of the enolate 43 or the protonated product 44.

Full text of DOI:10.1021/jo030224a

Syn th esis of th e Azir id in om itosen e Sk eleton by In tr a m olecu la r
Mich a el Ad d ition : r-Lith ioa zir id in es a n d Non a r om a tic Su bstr a tes
Edwin Vedejs,*,‡ J eremy D. Little,‡ and Lisa M. Seaney†
Chemistry Department, University of Wisconsin, Madison, Wisconsin 53706, and
Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109
edved@umich.edu
Received J uly 14, 2003
The bicyclic pyrrole ketone 16 has been prepared by using an oxaza-Claisen rearrangement, followed
by nitrogen deprotection. Coupling with the stannylaziridine mesylate 15a or nosylate 15b affords
17. Conversion to 40 provides a substrate for generation of an R-lithioaziridine 41 by tin lithium
exchange. An intramolecular Michael addition pathway for 41 has been demonstrated by the
isolation of 46 in 20% yield under conditions where the intermediate enolate 43 is trapped by
selenenylation, but competing proton transfer gives 42. The synthetic potential of the process is
limited by stability problems at the stage of the enolate 43 or the protonated product 44.
There has been considerable interest in the “FR”
compounds 1 and 2 due to their potent antitumor
activity.1,2 However, clinical trials of a related structure
(FK973) encountered vascular leak syndrome (VLS).3
Subsequent efforts to develop a drug candidate lacking
this side effect have resulted in the discovery of FK317,
a semisynthetic substance that retains high potency
without inducing VLS.4 This observation has revived
interest in the “FR” series, and has refocused attention
on the mode of action of these unusual molecules. A
multistage activation mechanism has been proposed for
the “FR” compounds, featuring reductive N-O bond
cleavage followed by cyclization and aromatization to the
labile tetracyclic intermediates 4 and 5.5 These structures
bear a close resemblance to the leucoaziridinomitosenes
that play an important role as intermediates in the
reductive activation of mitomycins, and in DNA cross-
linking events that are responsible for their antitumor
activity.6 Strong evidence has been advanced to support
a similar mode of action by 4 and 5,7 and it is likely that
analogous structures such as 6 and 7 may be involved in
the activation sequence from FK317 (3).4 So far, none of
the proposed structures 4-7 have been observed directly
or generated by total synthesis, although progress toward
the tetracyclic skeleton has been reported.2
We have initiated a program designed to prepare 6 and
7 for the eventual study of their role in the activation
cascade from FK317. Two closely related approaches have
been investigated in parallel studies, based on the
proposition that intramolecular nucleophilic addition
starting from R-tributylstannylaziridines 8 or 10 may
allow synthetic access to the tetracyclic structures 9 or
11. The key cyclization step in the aromatic series (8 to
9) has been achieved, as described in a preliminary report
from our laboratory, and detailed in the accompanying
paper.8 Here, we report our earlier efforts to develop an
approach based on nonaromatic precursors related to 10.
The synthesis exploits prior methodology developed in
our group for the generation of R-lithioaziridines.9 While
our work was under way,10 Ziegler et al. described a
strategically similar approach to tetracyclic aziridino-
mitosanes such as 14, based on the cyclization of aziri-
dinyl radicals generated from the intermediates 12 and
13.11 The sequence is similar for its timing of bond-
forming events leading to the key tetracycle 14. However,
product 14 does not have the indole double bond that is
essential for activation of potential leaving groups (CH2-
OC(O)NH2; aziridine C-N) and for DNA cross-linking.
Because of this difference, 14 is a relatively stable
molecule compared to 11 and related structures where
the allylic aziridine C-N bond is activated for heterolysis
by the pyrrole subunit.
‡ University of Michigan.
† University of Wisconsin.
(1) Kiyoto, S.; Shibata, T.; Yamashita, M.; Komori, T.; Okuhara, M.;
Terano, H.; Kohsaka, M.; Aoki, H.; Imanaka, H. J . Antibiot. 1987, 40,
594. Kiyoto, S.; Shibata, T.; Yamashita, M.; Komori, T.; Okuhara, M.;
Terano, H.; Kohsaka, M.; Aoki, H.; Imanaka, H. J . Antibiot. 1989, 42,
145.
(2) Review: Danishefsky, S. J .; Schkeryantz, J . M. Synlett 1995,
475. A summary and listing of key references is provided in the
following article.
(3) Pazdur, R.; Ho, D. H.; Daugherty, K.; Bradner, W. T.; Krakoff,
I. H.; Raber, M. N. Invest. New Drugs 1991, 9, 337.
(4) Naoe, Y.; Inami, M.; Matsumoto, S.; Nishigaki, F.; Tsujimoto,
S.; Kawamura, I.; Miyayasu, K.; Manda, T.; Shimomura, K. Cancer
Chemother. Pharmacol. 1998, 42, 31.
(5) Fukuyama, T.; Goto, S. Tetrahedron Lett. 1989, 30, 6491.
(6) DNA alkylation reviews: (a) Rajski, S. R.; Williams, R. M. Chem.
Rev. 1998, 98, 2723. (b) Tomasz, M.; Palom, Y. Pharmacol. Ther. 1997,
76, 73. (c) Tomasz, M. Chem. Biol. 1995, 2, 575. For a review of
synthetic approaches see ref 2.
(7) Huang, H.; Pratum, T. K.; Hopkins, P. B. J . Am. Chem. Soc.
1994, 116, 2703. Paz, M. M.; Hopkins, P. B. J . Am. Chem. Soc. 1997,
119, 5999. J udd, T. C.; Williams, R. M. Org. Lett. 2002, 4, 3711.
(8) Companion paper : Vedejs, E.; Little, J . D. J . Org. Chem. 2003,
68, 1794.
(9) Vedejs, E.; Moss, W. O. J . Am. Chem. Soc. 1993, 115, 1607.
(10) Seaney, L. M. Ph.D. Dissertation, University of Wisconsin, 1995.
(11) (a) Ziegler, F. E.; Belema, M. J . Org. Chem. 1994, 59, 7962. (b)
Ziegler, F. E.; Belema, M. J . Org. Chem. 1997, 62, 1083. (c) Ziegler, F.
E.; Berlin, M. Y. Tetrahedron Lett. 1998, 39, 2455.
10.1021/jo030224a CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/15/2004
1788
J . Org. Chem. 2004, 69, 1788-1793

Synthesis of the Aziridinomitosene Skeleton
SCHEME 1
SCHEME 2
more convenient for monitoring the reaction, but they
were complicated by competing hydrolysis of the ester
upon prolonged exposure to the water released during
cyclization. Conversion was minor on the 3 h time scale
of the reaction, but facile ester hydrolysis proved to be
more problematic in the next stage, during attempts to
oxidize the sulfide 25 to the sulfone 26. Although the
sulfide was cleanly oxidized with oxone in an EtOH/H2O
mixture, hydrolysis of the ester to the acid occurred under
these conditions. Fortunately, buffering the oxone with
a pH 5 NaOAc/HOAc buffer14 resulted in rapid (15 min)
oxidation to the sulfone 26. Deprotection of the benze-
nesulfonylethyl functionality15 was then accomplished
with excess t-BuOK in THF/CH2Cl2 (-45 °C to room
temperature). A 76% yield of the deprotected 16 was
obtained.
The synthesis of the stannylaziridine coupling partner
was based on an earlier report from our laboratory
describing analogous structures,9 as shown in Scheme 3.
The primary alcohol group of N-tritylserine methyl ester
(27)16 was protected with (tert-butyldimethylsiloxy)m-
ethyl chloride17 and the resulting ester 28 was reduced
with LAH to yield alcohol 29. Oxidation of 29 under
Swern conditions provided aldehyde 30 in 86% yield and
incorporation of the tributylstannyl functionality was
accomplished via addition of Bu3SnLi to 30 to afford the
amino alcohol 31. The stereochemistry of this addition
follows the earlier precedent,9 as evidenced by conversion
of crude 31 to the cis-aziridine 32 under Mitsunobu
conditions (70% yield for the two steps). Attempted
Our initial goal was to learn whether derivatives of
11 (X, Y ) carbonyl equivalent) might be accessible by
intramolecular Michael addition of an R-metalloaziridine.
Eventually, we plan to evaluate the synthetic potential
of related compounds for the preparation of the proposed
FK317 metabolites 6 or 7. The strategy is similar to
Ziegler’s in that 17 would be assembled by N-alkylation,
using an aziridinyl-containing alkylating agent 15 and
a bicyclic pyrrole derivative 16. Earlier and later steps
would be quite different, however, due to the planned
cyclization sequence based on tin-lithium exchange, and
the reliance on nonaromatic precursors.
The synthetic effort began with an adaptation of the
oxaza-Claisen method12 for synthesis of a bicyclic pyrrole,
as outlined in Scheme 2. The main difference in this route
compared to the precedent was in the use of a potentially
removable nitrogen substituent, the N-benzenesulfonyl-
ethyl group. Thus, the known oxime 1813 was reduced
with NaCNBH3 to afford hydroxylamine 19 and conden-
sation of the crude 19 with 1,3-cyclohexanedione (20)
afforded the vinylogous hydroxamic acid 21. Treatment
of 21 with ethyl propiolate and i-Pr2NEt at room tem-
perature then produced aminal 24 via Michael addition
of 21, oxaza-Claisen rearrangement of 22, and ring
closure of the aldehyde intermediate 23.
Conversion of the crude aminal 24 to 25 could be
achieved by heating in toluene with azeotropic removal
of water, or by room temperature treatment with TsOH
(90% isolated overall from 21). The TsOH conditions were
(14) Trost, B. M.; Curran, D. P. Tetrahedron Lett. 1981, 22, 1287.
(15) Gonzalez, C.; Greenhouse, R.; Tallabs, R.; Muchowski, J . M.
Can. J . Chem. 1983, 61, 1697.
(12) Reis, L. V.; Lobo, A. M.; Prabhakar, S. Tetrahedron Lett. 1994,
35, 2747.
(13) Kanemasa, S.; Norisue, Y.; Suga, H.; Tsuge, O. Bull. Chem. Soc.
J pn. 1988, 61, 3973.
(16) Baldwin, J . E.; Spivey, A. C.; Schofield, C. J .; Sweeney, J . B.
Tetrahedron 1993, 49, 6309.
(17) Benneche, T.; Gundersen, L.; Undheim, K. Acta Chem. Scand.
1988, B42, 384.
J . Org. Chem, Vol. 69, No. 6, 2004 1789

Vedejs et al.
SCHEME 3
SCHEME 4
In view of the long reaction time required in the
coupling reaction between 16 and 15a , the use of more
reactive leaving groups in place of the mesylate was
investigated. Treatment of aziridinol 33 with 4-nitroben-
zenesulfonyl chloride (NosCl) in the presence of NEt3/
DMAP afforded the crude nosylate 15b (Scheme 3). The
nosylate was not stable to flash chromatography and was
therefore used directly in the coupling step. Deprotona-
tion of 16 with NaN(TMS)2 in THF/DMPU (2:1) at -78
°C as before, followed by addition of crude 15b in THF
and warming to 60 °C for 24 h resulted in consumption
of 15b. After flash chromatography, 17 was isolated in
56% yield, somewhat lower than using the mesylate 15a .
However, the reaction was more convenient due to the
shorter time scale, and was preferred for preparative
experiments. To further improve the rate for the coupling
process, the aziridinyltriflate 15c was considered, as in
the analogous reaction reported by Ziegler et al. with use
of an N-alkoxycarbonyl protected aziridinyl fragment.11c
Unfortunately, attempts to prepare 15c resulted in
decomposition, and an experiment where 15c was gener-
ated in situ and coupling was attempted as before gave
none of the coupling product 17.
purification of 31 by column chromatography was not
successful due to facile conversion to 3-(tert-butyldi-
methylsiloxy)methoxypropanal on silica gel. For this
reason, 31 was used directly in the Mitsunobu cyclization
(2 equiv of both DEAD and PPh3) and purification was
performed at the stage of the aziridine 32. Deprotection
of 32 occurred readily upon treatment with TBAF at room
temperature, affording aziridinol 33 (91%), and subse-
quent reaction with methanesulfonyl chloride/triethyl-
amine gave the mesylate 15a in 91% yield.
Direct DIBAL reduction of ester 28 to the aldehyde 30
was also possible, although the product was contami-
nated with ca. 9% unreacted 28. The mixture was
sufficiently pure for conversion to 31, 32, and 33 in a
sequence where only the final product was purified by
chromatography. This approach was more convenient on
preparative scale, and provided the aziridine alcohol 33
in 45% overall yield from 28.
The coupling of bicyclic ketone 16 with mesylate 15a
proved to be challenging (Scheme 4). After considerable
optimization, it was found that deprotonation of 16 with
sodium hexamethyldisilazide (NaHMDS) in THF/DMPU
(2:1) at room temperature followed by addition of 15a ,
concentration of the resulting solution to 0.6 M, and
heating at 60 °C for 3 days afforded the desired coupling
product 17 in 68% yield.
Sufficient 17 was in hand for a preliminary test of the
proposed tin lithium exchange; internal Michael addition
sequence (Scheme 4). We were interested in knowing
whether metal exchange might be faster than addition
to the presumably deactivated vinylogous amide carbonyl
group. However, treatment of stannylaziridine 17 with
methyllithium at -65 °C followed by quenching with
ethanol gave the destannylated tertiary alcohol 34 (71%),
but no products that could be attributed to the formation
of tetracyclic structures related to 11. According to
trapping experiments, tin-lithium exchange had indeed
occurred. Thus, treatment of 17 with methyllithium as
before, followed by addition of Me2S2, gave 35 as well as
34, together with complex decomposition products. This
result indicates that the C-lithioaziridine is formed and
survives long enough to undergo intermolecular sulfen-
ylation, but for some reason does not cyclize. One possible
interpretation of the data is that the undesired 1,2-
carbonyl addition interferes with cyclization. Attention
was therefore turned to preparation of suitably protected
derivatives of 17.
Temporary protection of the carbonyl group by conver-
sion to a silyl enol ether was investigated first (Scheme
1790 J . Org. Chem., Vol. 69, No. 6, 2004

Synthesis of the Aziridinomitosene Skeleton
SCHEME 5
5). Addition of LiN(TMS)2 (LiHMDS) to a cooled solution
of 17 and tert-butyldimethylchlorosilane (TBSCl) in THF
resulted in 87% conversion to silyl enol ether 36 based
mixture obscured the outcome and further conclusions
were not possible.
Encouraged by the finding that at least the selective
deuterium incorporation had occurred without destruc-
tion of the stannylaziridine, we explored a more conven-
tional protecting group strategy for the problematic
ketone carbonyl group. Treatment of 17 with excess
NaBH4 at room temperature in ethanol afforded alcohol
38 in 85% yield as a 1:1 mixture of diastereomers
(Scheme 5). The alcohol 38 was then protected as the
benzyl ether by treatment with excess sodium hydride
in DMF, followed by addition of benzyl bromide to give
39 (66% isolated).
As a prelude to cyclization studies, the feasibility of
replacing the pyrrole ring proton with deuterium was
investigated. It was hoped that this precaution would
help protect against the complexity of product mixtures
that had been observed in previous cyclization attempts.
Treatment of 39 with phenyllithium at -78 °C in THF
followed by EtOD quench resulted in 40 containing 90%
deuterium according to 1H NMR integration. In contrast
to the enol silane experiments, the deuterated 40 could
be purified by chromatography, but complex minor
byproducts were also formed, resulting in significant
material loss after purification (66% of 40 isolated).
However, sufficient material was obtained to explore
cyclization conditions with use of the deuterated sub-
strate 40 (Scheme 6).
1
on assay by H NMR integration. However, 36 proved to
be exceptionally sensitive to hydrolysis and could not be
purified nor fully characterized. The characteristic dou-
blet of doublets at δ 4.8 ppm (CDCl3), along with the tert-
butyldimethyl resonances, left little doubt that the silyl
enol ether had been formed, so tin-lithium exchange
with the crude 36 was attempted under conditions that
had induced the destannylation of 33.
Reaction of 36 with methyllithium in THF at -65 °C
resulted in extensive degradation to give a complex
mixture that could not be separated. In the course of
optimization attempts, it was observed that treatment
of crude 36 with phenyllithium at -78 °C was less
destructive than was the analogous methyllithium ex-
periment. Although tetracyclic products were not de-
tected and the NMR signals due to the tributylstannyl
group were still present, there were indications that the
phenyllithium reagent had been consumed. The reason
became clear when the reaction was quenched with d1-
ethanol (EtOD). This gave the deuterated enol silane 37,
as determined from the disappearance of the signal at δ
6.99 ppm in the NMR spectrum.
The formation of 37 demonstrates that lithiation of the
pyrrole ring is faster than tin-lithium exchange in the
phenyllithium; DOEt quenching experiment. It also
suggests that ring lithiation R to the pyrrole nitrogen may
have been a complicating factor in the methyllithium
experiments, and may have prevented the intramolecular
Micheal addition. By this time, we had completed pre-
liminary experiments in the aromatic series (8 to 9) and
had confirmed that ring lithiation is the major pathway
in that series if methyllithium is used for the tin-lithium
exchange.8 Furthermore, we had observed that the
undesired pyrrole lithiation can be minimized by taking
advantage of the kinetic deuterium isotope effect during
the metalation step.8,18 In the first attempt to probe this
possibility in the aliphatic series, the deuterated silyl enol
ether 37 was reacted with methyllithium at -65 °C.
However, once again a complex product mixture was
obtained from which little could be learned. The silyl enol
ether did not survive the reaction conditions, judging
from the disappearance of the δ 4.9 ppm signal in the
1H NMR spectrum, but the complexity of the product
Treatment of 40 with excess MeLi in THF/Et2O at -65
°C (15 min) followed by quenching with ethanol gave the
uncyclized product 42 (67%), together with yet another
complex array of minor, unknown byproducts. Analysis
by HRMS and 1H NMR revealed that 42 still contains
(18) (a) Knaus, G.; Meyers, A. I. J . Org. Chem. 1974, 39, 1192.
Warmus, J . S.; Rodkin, M. A.; Barkley, R.; Meyers, A. I. J . Chem. Soc.,
Chem. Commun. 1993, 1357. Kopach, M. E.; Meyers, A. I. J . Org.
Chem. 1996, 61, 6764. (b) Clive, D. L. J .; Khodabocus, A.; Cantin, M.;
Tao, Y. J . Chem. Soc., Chem. Commun. 1991, 1755. Clive, D. L. J .;
Cantin, M.; Khodabocus, A.; Kong, X.; Tao, Y. Tetrahedron 1993, 49,
7917. Clive, D. L. J .; Tao, Y.; Khodabocus, A.; Wu, Y.-J .; Angoh, A. G.;
Bennett, S. M.; Boddy, C. N.; Bordeleau, L.; Kellner, D.; Kleiner, G.;
Middleton, D. S.; Nichols, C. J .; Richardson, S. R.; Vernon, P. G. J .
Am. Chem. Soc. 1994, 116, 11275. (c) Hoppe, D.; Paetow, M.; Hintze,
F. Angew. Chem., Int. Ed. Engl. 1993, 32, 394. Hoppe, D.; Kaiser, B.
Angew. Chem., Int. Ed. Engl. 1995, 34, 323. (d) Laplaza, C. E.; Davis,
W. M.; Cummins, C. C. Organometallics 1995, 14, 577. (e) Ahmed, A.;
Clayden, J .; Rowley, M. Tetrahedron Lett. 1998, 39, 6103. (f) Mathieu,
J .; Gros, P.; Fort, Y. Chem. Commun. 2000, 951. (g) Dudley, G. B.;
Danishefsky, S. J .; Sukenick, G. Terahedron Lett. 2002, 43, 5605.
J . Org. Chem, Vol. 69, No. 6, 2004 1791

Vedejs et al.
SCHEME 6
the original deuterium label, indicating that no signifi-
cant pyrrole ring lithiation had occurred prior to forma-
tion of 42. The desired intermediate 41 must therefore
have been generated, suggesting that perhaps the cy-
clization to 43 is unexpectedly slow. However, it was now
clear that all of the methyllithium cyclization attempts
afford exceptionally complex product mixtures, an ob-
servation that could be taken to imply a stability problem
for the key intermediate 43 or for the desired product
44. We therefore considered ways to convert 43 directly
into the presumably stable, aromatic pyrrole 46.
The usual tin-lithium exchange of 40 was carried out,
followed by a quenching with PhSeCl in place of ethanol.
This gave the desired pyrrole 46 in 20% yield, isolated
as an inseparable 3:2 mixture of diastereomers, and
provided, at last, clear evidence that some of the lithiated
aziridine 41 survives long enough to cyclize to the lithium
enolate 43. Trapping with excess PhSeCl affords the
labile tetracycle 45, and spontaneous aromatization
occurs with expulsion of PhSeD to give 46. This result
demonstrates that slow cyclization by the lithiated aziri-
dine 41 is not the reason cyclized products were not
isolated in the prior experiment (simple protic quench of
43). More likely, the protonated product 44 is unstable,
perhaps because it contains a benzyloxy leaving group
adjacent to an electron-rich enamine double bond.
The structure of tetracycle 46 was assigned based on
1H NMR data and extensive homonuclear decoupling
studies, along with IR and HRMS evidence. A summary
of the key 1H NMR data for 46 (400 MHz, CDCl3) is
shown in Figure 1 along with data for the known
tetracycle 48.19 In general, chemical shifts for protons
Ha-Hd appear at somewhat lower field for 48, consistent
with the increased electron demand in the vinylogous
amide 48. In the two diastereomers of 46, the aziridine
F IGURE 1. Selected 1H NMR signals for 46 and 48.
proton Ha appears at δ 2.92 ppm (0.4 H) and δ 2.93 ppm
(0.6 H) with J values of 4.4 and 4.8 Hz, respectively. This
is in close agreement with the shifts and J values of Ha′
in 48 (Figure 1), confirming that 46 is a cis-fused
aziridine. The chemical shift of the second aziridine
proton, Hb, is shifted to ca. 1 ppm higher field in 46
compared to 48. The signal is obscured by a multiplet
due to one of the cyclohexane protons (δ 1.66-1.80 ppm),
but the assignment is clear from decoupling experiments.
(19) Vedejs, E.; Klapars, A.; Naidu, B. N.; Piotrowski, D. W.; Tucci,
F. C. J . Am. Chem. Soc. 2000, 122, 5401.
1792 J . Org. Chem., Vol. 69, No. 6, 2004

Synthesis of the Aziridinomitosene Skeleton
Differences in coupling constants between Hb and Hc or
Hd of 46 compared to the analogous signals in 48 indicate
differences in conformation between the two tetracyclic
structures, but the connectivity is clearly the same.
In addition to tetracycle 46, 20% of the protonated
product 42 was also isolated in the PhSeCl quench
experiment. The remainder of the reaction mixture was
even more complex than in the protic quench, and
material balance for products that could be purified was
considerably lower (40%). Some of the minor products
may be derived from the lithiated aziridine 41 via
reaction with PhSeCl. NMR signals that might indicate
the presence of the phenylselenenyl aziridine 47 were not
detected. Other explanations for the complexity of the
product mixture may be related to the reasons why 42
is formed in this experiment. Apparently, some of the
lithiated aziridine 41 is protonated by potentially acidic
protons derived from the substrate or its decomposition
products. The resulting anions may be formed next to
the pyrrole ring, and their selenenylation would afford
solvolytically labile phenylselenenyl derivatives, acti-
vated by the same donor effect of pyrrole nitrogen that
activates the aziridine C-N bond. This rationale is not
fully satisfying, but it is consistent with the observation
of numerous minor products from the cyclization experi-
ment.
The proposed cyclization by intramolecular Michael
addition of an R-lithio aziridine has been demonstrated,
but the synthetic potential of the process is limited by
stability problems at the stage of the enolate 43 or the
protonated product 44. Structure 44 was never observed,
despite the effort invested to detect it. The exact reasons
for these limitations remain unclear, but the principal
source of difficulty resides in the cylohexane ring, and
not in the aziridine. The conclusion follows from observa-
tions made in the aromatic series where cyclization
products were obtained with much better efficiency. This
work is described in the accompanying article,8 along
with a more detailed investigation of the kinetic deute-
rium isotope effect used to control pyrrole ring lithiation
during the tin lithium exchange; cyclization steps.
Ack n ow led gm en t. This work was supported by the
National Institutes of Health (CA17918).
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
1
cedures, characterization, and H NMR spectra. This material
is available free of charge via the Internet at http://pubs.acs.org.
J O030224A
J . Org. Chem, Vol. 69, No. 6, 2004 1793