Rare example of nucleophilic substitution at vinylic carbon with inversion: Mechanism of methyleneaziridine formation by sodium amide induced ring closure revisited

Article abstract of DOI:10.1021/ja0482684

Intramolecular nucleophilic substitution of the C-Br bond of (E)- and (Z)-2-bromobut-2-enylamines by the pendant nitrogen atom leads to 2-ethyleneaziridines by way of stereochemical inversion at the vinylic carbon atom. The stereochemistry of the products is unambiguously established by X-ray crystallography performed on two derivatives. These cyclizations represent some of the first examples of substitution with inversion in unactivated vinylic substrates. In conjunction with additional deuterium-labeling experiments, the accepted mechanism for this reaction is shown to be flawed. Copyright

Full text of DOI:10.1021/ja0482684

Published on Web 05/13/2004
Rare Example of Nucleophilic Substitution at Vinylic Carbon with Inversion:
Mechanism of Methyleneaziridine Formation by Sodium Amide Induced Ring
Closure Revisited
Jason J. Shiers,† Michael Shipman,*,† Jerome F. Hayes,‡ and Alexandra M. Z. Slawin§
Department of Chemistry, UniVersity of Warwick, Gibbet Hill Road, CoVentry, CV4 7AL, U.K.,
GlaxoSmithKline, Old Powder Mills, Tonbridge, Kent, TN11 9AN, U.K., and School of Chemistry,
UniVersity of St. Andrews, Purdie Building, St. Andrews, Fife, KY16 9ST, U.K
Received March 26, 2004; E-mail: m.shipman@warwick.ac.uk
Scheme 1. Ring Closure to Ethyleneaziridines with
Stereochemical Inversion at the Vinylic Carbon Atom
Nucleophilic substitution reactions are very commonly encoun-
tered in mechanistic and synthetic chemistry. Bimolecular nucleo-
philic substitution (SN2) at a saturated carbon atom (sp3) proceeds
with inversion of configuration.1 Both stepwise and concerted
processes at vinylic carbon (sp2) normally afford substitution with
retention of configuration although when the carbanion intermediate
is long-lived, partial or complete stereoconvergence can result.2
Examples of nucleophilic substitutions at vinylic carbon proceeding
with inversion of configuration are rare. Typically, these involve
substrates bearing very good leaving groups such as vinyl triflates3
or vinyl iodonium salts.4 Inversion of configuration has not been
observed in unactivated systems such as vinyl chloride even though
high-level computational studies suggest that in-plane, σ-approach
of the nucleophile to the backside of the C-Cl bond is the preferred
pathway.5 Herein, we report examples of stereochemical inVersion
occurring during substitution at an unactiVated Vinylic carbon atom.
Specifically, we show that intramolecular substitution of the C-Br
bond of 2-bromobut-2-enylamines by the pendant nitrogen atom
leads to 2-ethyleneaziridines by way of stereochemical inversion.
In conjunction with deuterium-labeling experiments, these results
indicate that the accepted mechanism6 for ring closures of this type
is flawed.
The reaction of 2-bromoallylamines with sodium amide in liquid
ammonia is known to provide 2-methyleneaziridines in high yields.7
As part of studies aimed at delineating the synthetic scope of these
highly strained heterocycles,8 we became interested in establishing
if this reaction could be extended to the synthesis of derivatives
bearing stereochemically defined substituents on the exocyclic
double bond.9 To this end, we set about the synthesis of (E)- and
(Z)-2-ethyleneaziridines 1. Derivatives containing either a simple
benzyl (R ) CH2Ph) or chiral R-methylbenzyl (R ) (S)-CHMePh)
group on nitrogen were chosen. Four cyclization precursors (E)-
and (Z)-2a,b were made by treatment of the (E)- and (Z)-isomers
of 2-bromobut-2-en-1-ol10 with methanesulfonyl chloride and further
alkylation with either benzylamine or (S)-R-methylbenzylamine.
The stereochemistry about the double bond within (E)-2a,b and
(Z)-2a,b was confirmed by NOE difference experiments. Treatment
of (E)- and (Z)-2a,b with sodium amide in liquid ammonia provided
2-ethyleneaziridines (E)- and (Z)-1a,b in moderate to excellent
yields (Scheme 1).11 Several aspects of these cyclizations are of
special interest. First, the reactions are highly stereoselective with
the aziridines being obtained as essentially single isomers about
the double bond.12,13 Second, lower yields were observed for the
formation of (E)-1a,b as alkyne formation via antiperiplanar E2-
elimination of HBr from (Z)-2a,b was competitive. Finally, and
most significantly, the reactions proceed by net inversion at the
alkene carbon atom undergoing substitution.
The stereochemistry of (Z)-1a and (Z)-1b have been unambigu-
ously established by derivatization and subsequent crystallographic
analysis. Lithiation of (Z)-1b, and separately (Z)-1a, with sec-BuLi
(THF, 5 h, -78 °C) and further reaction with benzophenone
provided (Z)-3 and (Z)-4 in 83% and 62% yields, respectively
(Figure 1). Importantly, in a control experiment, lithiation/proto-
nation of (Z)-1a was shown to occur without isomerization of the
double bond. Single crystals of 3 and 4 grown from diethyl ether
were analyzed by single-crystal X-ray diffraction, which revealed
that they both possess the (Z)-stereochemistry. For 3, the major
diastereomer (crude dr 93:7) possessed the expected (S)-configu-
ration at C-3.14 Lithiation of (E)-1a and alkylation with benzophe-
none produced the (E)-isomer of 4.
Bottini and Olsen produced evidence indicating that the formation
of methyleneaziridines proceeds primarily if not exclusively by an
elimination-addition mechanism (Scheme 2).6 Performing the
reaction using tritium-enriched ammonia as solvent, they cyclized
amine 5 to methyleneaziridine 6 (R ) Pr) which was tritium labeled
exclusively at C-3. In further crucial experiments, they reported
that the hydrogens bonded to carbon of 5 and 6 do not undergo
exchange with the solvent under the reaction conditions. Thus, they
concluded that the reaction was proceeding through aziridinyl anion
9 by way of allenes 7 and 8. The observation that (E)- and (Z)-2a
yield different aziridine products is inconsistent with this elimina-
tion-addition mechanism as both these substrates would be
expected to yield the same allene intermediate.
To further investigate the mechanism, deuterated 2-bromoally-
lamine 10 (95% D incorporation) was prepared in two steps from
ethyl 2-bromoacrylate [(i) LiAlD4, AlCl3, Et2O, 75%; (ii) MsCl,
Et3N, 0 °C, THF, then (S)-PhCHMeNH2, 69%]. Treatment of 10
with sodium amide (1.25 equiv) for 5 min at -33 °C yielded (S)-
11 (92% D) (Scheme 3).15 Thus, anion 9 cannot be an intermediate
in the production of methyleneaziridines. Performing the cyclization
† University of Warwick.
‡ GlaxoSmithKline, Tonbridge.
§ University of St. Andrews.
9
6868
J. AM. CHEM. SOC. 2004, 126, 6868-6869
10.1021/ja0482684 CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
Acknowledgment. We thank the Engineering and Physical
Sciences Research Council and GlaxoSmithKline for financial
support of this work.
Supporting Information Available: Crystallographic data of 3 and
4 (CIF) experimental procedures and spectroscopic data for 1-4 and
10-12 (PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
Figure 1. Derivatives produced from (Z)-ethyleneaziridines used to confirm
the alkene geometry by X-ray crystallography.
Scheme 2. Original Ring Closure Mechanism Proposed by Bottini
and Olsen6
References
(1) (a) Katritzky, A. R.; Brycki, B. E. Chem. Soc. ReV. 1990, 83-105. (b)
Rossi, R. A.; Pierini, A. B.; Pen˜e´n˜ory, A. B. Chem. ReV. 2003, 71-167.
(2) (a) Rappoport, Z. Acc. Chem. Res. 1981, 14, 7-15. (b) Rappoport, Z.
Acc. Chem. Res. 1992, 25, 474-479.
(3) (a) Clarke, T. C.; Kelsy, D. R.; Bergman, R. G. J. Am. Chem. Soc. 1972,
94, 3626-3627. (b) Summerville, R. H.; Senkler, C. A.; Schleyer, P. v.
R.; Dueber, T. E.; Stang, P. J. J. Am. Chem. Soc. 1974, 96, 1100-1110.
(c) Summerville, R. H.; Schleyer, P. v. R. J. Am. Chem. Soc. 1974, 96,
1110-1120.
(4) (a) Ochiai, M.; Oshima, K.; Masaki, Y. J. Am. Chem. Soc. 1991, 113,
7059-7061. (b) Okuyama, T.; Ochiai, M. J. Am. Chem. Soc. 1997, 119,
4785-4786. (c) Ochiai, M.; Yamamoto, S.; Sato, K. Chem. Commun.
1999, 1363-1364. (d) Ochiai, M.; Yamamoto, S. Chem. Commun. 2002,
2802-2803. (e) Okuyama, T.; Lodder, G. AdV. Phys. Org. Chem. 2002,
37, 1-56.
Scheme 3. Cyclization Studies Using Deuterium-Labeled
Substrates
(5) Bach, R. D.; Baboul, A. G.; Schlegel, H. B. J. Am. Chem. Soc. 2001,
123, 5787-5793. Kim, C. K.; Hyun, K. H.; Kim, C. K.; Lee, I. J. Am.
Chem. Soc. 2000, 122, 2294-2299 and references therein.
(6) Bottini, A. T.; Olsen, R. E. J. Am. Chem. Soc. 1962, 84, 195-199.
(7) (a) Pollard, C. B.; Parcell, R. F. J. Am. Chem. Soc. 1951, 73, 2925-
2927. For the correct structural assignments, see: (b) Ettlinger, M. G.;
Kennedy, F. Chem. Ind. 1956, 166-167. (c) Bottini, A. T.; Roberts, J. D.
J. Am. Chem. Soc. 1957, 79, 1462-1464.
(8) (a) Hayes, J. F.; Shipman, M.; Twin, H. J. Org. Chem. 2002, 67, 935-
942. (b) Pre´vost, N.; Shipman, M. Org. Lett. 2001, 3, 2383-2385. (c)
Hayes, J. F.; Shipman M.; Twin, H. Chem. Commun. 2001, 1784-1785.
(9) All previous routes to 2-methyleneaziridines have led to products in which
both substituents on the exocyclic double bond are identical. For example,
the synthesis of 2-isopropylidineaziridines is known: Wijnberg, J. B. P.
A.; Wiering, P. G.; Steinberg, H. Synthesis 1981, 901-903.
(10) For the preparation of (E)-2-bromobut-2-en-1-ol, see: (a) Hiskey, C. L.;
Slates, H. L.; Wendler, N. L. J. Org. Chem. 1956, 21, 429-433. For (Z)-
2-bromo-but-2-en-1-ol, see: (b) Loh, T.-P.; Cao, G.-Q.; Pei, J. Tetrahedron
Lett. 1998, 39, 1453-1456.
(11) Sodium amide was made in situ by adding sodium metal to liquid ammonia
at -33 °C containing a small quantity (2.5 mol %) of an Fe(III) source.
Identical results were obtained using either Fe(NO3)3‚9H2O or FeCl3.
(12) Substrates (E)- and (Z)-2a,b contained traces of the double bond isomer
(e8%). This was relayed into the aziridines which also contained a small
quantity of the stereoisomer (e9%). See Supporting Information.
for a longer time using excess sodium amide (2.5 equiv, 1 h, -78
°C) resulted in conversion to (S)-12 wherein virtually all the
deuterium had been lost (12% D). Consistent with this finding,
resubjection of deuterated methyleneaziridine (S)-11 (92% D) to
the cyclization conditions resulted in clean conversion to nondeu-
terated (S)-12 (2.5% D). These results indicate that this methyl-
eneaziridine, formed quickly under the cyclization conditions (t1/2
≈ 10 s at -78 °C using 2.5 equiv of NaNH2), undergoes slow
reversible exchange with the solvent by way of deprotonation at
C-3 by the excess sodium amide.16
The question remains what is the mechanism of ring closure?
Our evidence discounts the earlier proposal of an elimination-
addition reaction. Furthermore, stereochemical inversion seems to
rule out the possibility of an addition-elimination process wherein
retention or stereoconvergence would be expected.2 Substitution
with inversion by in-plane σ-attack from the backside of the C-Br
bond fits all the available experimental data. It remains to be seen
if the constraints imposed on the reaction trajectory by the formation
of a three-membered ring are important in encouraging this pathway.
To conclude, a rare example of substitution at a vinylic carbon
proceeding with inversion has been unearthed. Efforts to explore
the utility of the geometrically defined 2-ethyleneaziridines in
synthesis will form the basis of future studies.
(13) No appreciable racemisation occurred during the formation of (Z)-1b (90%
ee) and (E)-1b (94% ee) as determined by chiral shift NMR experiments
using (S)-2,2,2-trifluoro-1-(9-anthryl)ethanol. Racemic materials made and
used as controls.
(14) Hayes, J. F.; Pre´vost, N.; Prokes, I.; Shipman M.; Slawin, A. M. Z.; Twin,
H. Chem. Commun. 2003, 1344-1345.
(15) The level of deuterium incorporation was measured using 1H NMR
spectroscopy and is the average of three experimental runs.
(16) It is unclear why Bottini and Olsen (ref 6) were unable to detect exchange
with the solvent under their experimental conditions. However, we suspect
that an insufficient amount of base was added. This conclusion was reached
on the basis of the following observations. First, for the exchange
experiment, they used just a catalytic quantity (0.26 equiv) of commerical
sodium amide, whereas cyclisations were conducted using NaNH2 made
in situ from sodium metal and FeCl3. In our experience, the activity of
commercial material is far inferior and often necessitates the use of a
large excess of reagent. Second, 6 used in the experiment was contaminated
with 7% of the corresponding alkyne which would have consumed a
significant proportion of the added base.
JA0482684
9
J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004 6869