Nucleophilic substitution by Grignard reagents on sulfur mustards

Article abstract of DOI:10.1021/jo0489869

With proper activation of the leaving group, sulfur mustards react with Grignard reagents with neighboring group participation of the sulfur atom. 2,6-Dichloro-9-thiabicyclo[3.3.1]nonane is especially useful in this regard, providing clean reactivity with organomagnesium nucleophiles on a topologically constrained scaffold.

Full text of DOI:10.1021/jo0489869

SCHEME 1
Nu cleop h ilic Su bstitu tion by Gr ign a r d
Rea gen ts on Su lfu r Mu sta r d s
Antonella Converso, Pierre-Lo¨ıc Saaidi,1
K. Barry Sharpless, and M. G. Finn*
magnesium, cuprate, and silyl reagents.4 The only known
example with 1 is a brief report of the use of trialkyl-
aluminum compounds.5 It should be noted that by “an-
chimeric assistance” we do not mean the tethering of a
reactive organometallic to the electrophile by means of
a pendant heteroatom,6 which renders the C-C bond-
forming event intramolecular and has occasionally been
given the anchimeric assistance label. Rather, we refer
to the stereochemically distinct double inversion process,
in which the electrophile is activated by intramolecular
interaction with a nucleophilic component and is subse-
quently captured by external nucleophile. We report here
that 1 engages in clean substitution with retention in
reactions with a variety of Grignard reagents, presum-
ably via the neighboring group participation of sulfur as
shown in Scheme 1. Precedents include the chemistry of
â-chloro-â-alkoxythioether compounds derived from the
addition of arylsulfenyl chlorides to vinyl ethers,4 ap-
plications to the synthesis of C-glycosides,7 and substitu-
tion reactions of N-(R-haloalkyl)benzotriazoles.8
Alkyl- and aryllithium reagents were found to promote
elimination in preference to substitution, although di-
substituted products from 1 were observed in trace
amounts (Scheme 2A). Grignard reagents, which are
more polarizable than lithium compounds and are there-
fore expected to be less prone toward proton abstraction,
gave intractable mixtures at 0 °C in ethereal solvents.
Such mixtures contained <10% of the desired compounds,
along with olefins derived from elimination and other
compounds. However, the use of Grignard reagents at
room temperature or above provided disubstitution of 1
with retention of configuration in good to excellent yields
for aromatic, acetylenic, and primary alkylmagnesium
bromides (Scheme 2B, Table 1). All compounds described
here were prepared in racemic form.
Department of Chemistry and the Skaggs Institute for
Chemical Biology, The Scripps Research Institute,
10550 North Torrey Pines Road, La J olla, California 92037
mgfinn@scripps.edu
Received J une 17, 2004
Abstr a ct: With proper activation of the leaving group,
sulfur mustards react with Grignard reagents with neigh-
boring group participation of the sulfur atom. 2,6-Dichloro-
9-thiabicyclo[3.3.1]nonane is especially useful in this regard,
providing clean reactivity with organomagnesium nucleo-
philes on a topologically constrained scaffold.
The chemistry of 2,6-dichloro-9-thiabicyclo[3.3.1]nonane,
1, provides a good illustration of the power of anchimeric
assistance in organic chemistry.2,3 The reactivity of this
compound is enhanced by the central heteroatom, which
engages the â-chlorinated centers to facilitate nucleo-
philic substitution via a highly reactive episulfonium ion
(Scheme 1). The double inversion process preserves
stereochemistry in substitution reactions with a broad
range of heteroatom nucleophiles.
Compound 1 and its disubstituted derivatives 2 are
chiral, C2-symmetric structures. The C-Cl/Nuc axes
describe two vectors with a dihedral angle of 84°, and
the electrophilic carbon centers are quite sterically
hindered. The latter point has been most clearly dem-
onstrated by the fact that chloride in 1 cannot be
displaced by NMe3 or NEt3, whereas the weaker but less
sterically demanding nucleophile pyridine forms very
stable adducts.3 The parent dichloride and several di-
substituted derivatives can be prepared in enantiomeri-
cally enriched form either by chiral HPLC or by substi-
tution of a diastereomerically enriched bis(brucine)
adduct.3 The system therefore lends itself to the display
of functionality in chiral form, as may be found in chiral
ligands for metals, chiral acids or bases, and components
of chiral polymers.
(4) Smoliakova, I. P.; Caple, R.; Brenny, J . W.; Smit, W. A.;
Kryschenko, Y. K.; Shashkov, A. S.; Chizhov, O. S.; Krimer, M. Z.;
Morar, G. V.; Kalyan, Y. B. Synlett 1995, 275-276 and references
therein.
(5) Tolstikov, G. A.; Kantyukova, R. G.; Spirikhin, L. V. Zh. Org.
Kim. 1980, 16, 1408-1418.
(6) (a) Examples include the following: (i) Fuji, K.; Tanaka, K.; Ahn,
M.; Mizuchi, M. Chem. Pharm. Bull. 1994, 42, 957-959. (ii) Calo, V.;
De Nitti, C.; Lopez, L.; Scilimati, A. Tetrahedron 1992, 48, 6051-6058.
(iii) Oliva, M.; Safont, V. S.; Andres, J .; Castillo, R.; Moliner, V. Int. J .
Quantum Chem. 1997, 65, 719-728. (b) An ambiguous case is provided
by: Marot, C.; Philipp, C.; Rollin, P. Tetrahedron Lett. 1992, 33, 4575-
4578. It is not clear which phenomenon - tethered reactivity or true
anchimeric assistance - is operative in the examples described.
(7) (a) Smoliakova, I. P.; Han, M. M.; Gong, J . C. Tetrahedron 1999,
55, 4559-4572. (b) Han, M. M.; Smoliakova, I. P.; Koikov, L. N.
Carbohydr. Res. 2000, 323, 202-207. (c) Smoliakova, I. P. Curr. Org.
Chem. 2000, 4, 589-608. (d) Liu, H.; Smoliakova, I. P. Tetrahedron
2001, 57, 2973-2980.
To realize certain examples of these types of structures,
new carbon-carbon bonds must be made with mustard-
type electrophiles such as 1. The literature contains very
few well-characterized examples of anchimeric assistance
in substitution by such activated carbon nucleophiles as
(1) Present address: Laboratoire de Chimie, UMR-5182 CNRS/ENS,
Ecole Normale Supe´rieure de Lyon, France.
(2) Weil, E. D.; Smith, K. J .; Gruber, R. J . J . Org. Chem. 1966, 31,
1669-1679.
(3) Converso, A.; Burow, K.; Marzinzik, A.; Sharpless, K. B.; Finn,
M. G. J . Org. Chem. 2001, 66, 4386-4392.
(8) Katritzky, A. R.; Abdel-Fattah, A. A. A.; Tymoshenko, D. O.;
Belyakov, S. A. Synthesis 1999, Supplment, 1437-1440.
10.1021/jo0489869 CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/14/2004
7336
J . Org. Chem. 2004, 69, 7336-7339

SCHEME 2
The conformation of aryl groups in 2b, in which the
o-methyl substituents are found on the “sulfur” side of
the structure, is consistent with molecular mechanics
calculations showing this to be the most stable rotameric
form by approximately 5 kcal/mol. Rotation about the
C-aryl bond is restricted, with severe steric interactions
between the aromatic ortho positions and the axial C3-H
and equatorial C8-H atoms of the thiabicyclo[3.3.1]-
nonane core. To escape these interactions, the ortho
substituent of 2-substituted aryl fragments must be
placed away from the thiabicyclononane core, and such
compounds are locked at room temperature with respect
to rotation about the C2-aryl bond axis. Compound 2c,
derived from 1-bromo-2-methylnaphthalene, displays the
expected hindered rotation about the equivalent C2
positions in solution, giving rise to a broadened NMR
spectrum. Heating to 70 °C causes the signals to sharpen
to a single set, indicative of faster rotation around these
bonds on the NMR time scale at elevated temperature.
The stereochemical control exemplified by 2b and the
hindered bond rotation of 2c are potentially useful in the
design of chiral ligands and polymeric materials.
For the reactions described above, donor solvents were
found to be detrimental to the purity of the final
product: both diethyl ether and tetrahydrofuran, com-
monly used in Grignard preparation, promote elimina-
tion, albeit to a modest extent at room temperature
(e10% of the product mixture). Elimination can be
minimized if the ether or THF is removed by evaporation
prior to addition of 1, but this may occur at the expense
of Grignard reagent solubility and therefore of yield. The
use of dichloromethane10 instead of toluene further
suppressed elimination. However, it proved to be most
convenient to perform the reactions in toluene at 70 °C
without pre-evaporation of ethereal cosolvent. Under
these conditions, diethyl ether is carried away from the
mixture and the low-temperature elimination pathway
is further disfavored. If it is required to perform the
reaction at room temperature, the best performance is
usually obtained in dichloromethane with prior removal
of ether or THF from the Grignard mixture. In most
cases, poorer results are achieved if the Grignard reagent
concentration is less than 0.2 M. Diethynyl compound
2f was isolated in low yield due to the formation of
bridging acetylide dimers and higher oligomers. Second-
ary magnesium bromides gave mixtures of elimination
and reduction products, whereas t-BuMgCl induced
mostly reduction.
TABLE 1. Isola ted Yield s a n d Ch a r a cter istic In fr a r ed
Ba n d s of Disu bstitu ted P r od u cts fr om Rea ction s of 1
w ith Gr ign a r d Rea gen ts a t 70 °C in Tolu en e (Sch em e 2B)
a
b
RMgX
yield, %
product
νC-H
δC-H
PhMgBr
o-tolyl-MgBr
2-Me-1-naphthyl-MgBr
vinyl-MgBr
allyl-MgBr
CH3CtC-MgBr
HCtC-MgBr
n-pentyl-MgBr
MeMgBr
99
98
62
69
90
97
30
99
85
2a
2b
2c
2d
2e
2f
2g
2h
2i
2977
2990
2974
2973
2981
2983
2980
2955
2951
1490
1483
1480
1480
1483
1477
1483
1479
1480
a
“Anomalous” stretching band for axial C3-H and C7-H
bonds, cm-1
.
“Anomalous” bending band for axial C3-H and
b
C7-H bonds, cm-1
.
F IGURE 1. X-ray crystal structures of 2a and 2b. (Middle)
View showing the chair/chair conformation of the bicyclo-
nonane core. (Right) View down the C2 symmetry axis.
Relevant carbon position numbers are shown on the sketch of
2b.
Assignment of retention of configuration, and therefore
firm evidence for a pathway of neighboring group par-
ticipation, was obtained by X-ray crystallography of 2a
and 2b (Figure 1). The trans orientation of sulfur and
substituent centers in both compounds was accompanied
by characteristic coupling constants in the 1H NMR
spectrum and C-H stretching and bending bands in the
IR spectrum at uniquely high frequencies (Table 1) due
to congestion of axial C-H bonds at the 3 and 7 positions
of the chair/chair conformation favored for the 2,6-
disubstituted-9-thiabicyclo[3.3.1] core.2,9 These spectro-
scopic signatures were found in each of the remaining
substitution products 2c-i, allowing us to extend the
assignment of retention beyond the two cases character-
ized in the solid state.
No substitution was observed for several cases in which
the magnesium reagents were prepared from aryl bro-
mides by activation of Mg0 with 1,2-dibromoethane (20%
relative to 1). In these cases, the dibromide adduct 4 was
the only product observed in addition to recovered 1; an
example is shown in Scheme 3. (Activation of metallic
Mg with catalytic amounts of iodine, rather than dibromo-
ethane, works well; compound 2c was prepared in this
manner.) Since the formation of MgBr2 would be expected
from dibromoethane and Mg0, the response of a standard
reaction to the addition of MgBr2 was tested. When 20%
MgBr2 was added to commercially available PhMgBr,
which otherwise performs well, significantly poorer re-
(9) Brown, W. A. C.; Eglinton, G.; Martin, J .; Parker, W.; Sim, G.
A. Proc. Chem. Soc. 1964, 57-58.
(10) Keck, G. E.; Andrus, M. B.; Romer, D. R. J . Org. Chem. 1991,
56, 417-420.
J . Org. Chem, Vol. 69, No. 21, 2004 7337

SCHEME 3
TABLE 2. Rea ction s of Gr ign a r d Nu cleop h iles w ith
2-Ch lor oeth yl Eth yl Su lfid es 5 a n d 6 (Eq 1)
bromide attack with 6. Cyclopentylmagnesium bromide
also deposited bromine on 6, and gave reduction with 5.
A mixture of methyl and bromine substitution products
was observed for MeMgBr with 6. It therefore appears
that, without anchimeric assistance (compound 6), treat-
ment with Grignard reagents results largely in substitu-
tion by bromide and not alkyl. When an episulfonium
intermediate is accessible (as with 5), alkyl attack and
elimination are more favored. The clean reactivity of
PhMgBr may result from its weaker basicity, more Lewis
acidic Mg2+ center, or different aggregation state com-
pared to alkylmagnesium reagents.
entry
R1
R2
R3
product
yield, %
1
2
3
4
5
6
7
Et
Et
Et
Ph
Ph
Ph
Ph
n-pentyl
cyclopentyl
Ph
n-pentyl
cyclopentyl
Ph
n-pentyl
H
Ph
Br
Br
Ph
Br, Me
7
8
9
10
10
11
10, 12
26
91
82
60
51
98
70, 30
Me
The nature of the optimal solvent is obviously different
for the Grignard reactions of 1 (best performed in toluene)
and its wide-ranging substitution chemistry with hetero-
atom nucleophiles (done in water or acetonitrile).3 While
the distribution of potential magnesium species (RMgBr,
MgR2, and oligomeric variations) is affected by the
reaction conditions,12 we suggest that the solvent exerts
its most important influence on the electrophile.13 In a
nonpolar medium, the reactive episulfonium ion inter-
mediate that drives the chemistry of this system is highly
disfavored. This charged species is of course destabilized
by a mismatch with a nonpolar environment, but more
importantly, aprotic solvents provide no protic solvation
to the outgoing chloride ion.14 Lewis acidic magnesium
complexes can overcome this problem by assisting the
departure of chloride to promote episulfonium ion forma-
tion (Figure 2), but only when nucleophilic solvents such
as ether and THF (which might otherwise compete for
binding to the metal ion) are absent. (Lewis acids have
been used for the generation of R-alkoxyepisulfonium ions
from the corresponding â-chlorothioethers.7d) The deliv-
ery of the carbon nucleophile to the episulfonium ion may
occur from within the proposed intermediate ion pair, as
shown in Figure 2, or may occur from a different
organomagnesium center.
sults were obtained (lower and variable substitution
yields and approximately 10% of olefins derived from
elimination of HCl from 1).
These observations suggest that Br-for-Cl substitution
is enhanced by MgBr2. Furthermore, either the bromide
adduct 4 is a much poorer electrophile than 1 toward
Grignard nucleophiles or the organomagnesium byprod-
uct of the Br-for-Cl substitution is a poor nucleophile. The
former possibility was discounted by comparisons be-
tween 1 and 4 in reactions with a standard amine
nucleophile (benzylamine, acetonitrile solution, room
temperature) and with PhMgBr (toluene, 70 °C). In the
first case, 4 was found to be at least 10 times more
reactive than 1, whereas in the second case, both elec-
trophiles were consumed within a very short time. Thus,
4 can be expected to be reactive with bona fide Grignard
nucleophiles. It must therefore be the case that RMgCl
is produced from the reaction of 1, RMgBr, and MgBr2
in an inactive or poorly active form. It is perhaps
worthwhile to note in this context that commercially
available organomagnesium bromides were found to be
superior to chlorides in terms of overall yield and the
formation of unwanted byproducts, whereas iodides gave
little or no substitution. For example, MeMgBr gave 2i
in 85% yield (Table 1) while MeMgCl provided 2i in
approximately 42% yield, along with 7% of unreacted 1
and approximately 4% of material containing the olefinic
products of elimination.
Substitution reactions of simple alkyl halides with
Grignard reagents are surprisingly rare.15 One reason
may be the complex interplay between factors such as
(11) McManus, S. P.; Karaman, R. M.; Sedaghat-Herati, R.; Harris,
J . M. J . Org. Chem. 1995, 60, 4764-4766.
To compare 1 to simple sulfur mustard electrophiles,
a series of reactions was performed on the S-ethyl- and
S-phenyl â-chlorothioether electrophiles 5 and 6 (eq 1,
Table 2). The former is presumed to react with participa-
tion of the alkylthioether group, while the latter is known
to eschew anchimeric assistance in favor of direct sub-
stitution of halide nucleophiles.11 For both 5 and 6,
PhMgBr gave phenyl substitution in high yield. In
contrast, n-pentyl-MgBr provided alkyl attack with 5, but
(12) RMgBr has been shown to be favored at higher temperatures
and MgR2 is favored by precipitation of MgBr2 in nonpolar solvents.
In addition, the best reaction conditions (high concentration in nonpolar
solvent) lend themselves to the formation of oligomers, which are less
basic than monomeric Grignard reagents. See: Silverman, G. S. In
Handbook of Grignard Reagents; Silverman, G. S., Rakita, P. E., Eds.;
Marcel Dekker Inc.: New York, 1996; pp 307-353.
(13) Krow, G. R.; Lin, G.; Rapolu, D.; Fang, Y.; Lester, W. S.; Herzon,
S. B.; Sonnet, P. E. J . Org. Chem. 2003, 68, 5292-5299.
(14) Gajewski, J . J . J . Am. Chem. Soc. 2001, 123, 10877-10883.
7338 J . Org. Chem., Vol. 69, No. 21, 2004

F IGURE 2. Proposed Lewis acid assisted formation and alkylation of episulfonium ion from 2,6-dichloro-9-thiabicyclo[3.3.1]-
nonane.
solvent, counterion, and leaving group, which contributes
to variable outcomes as described above. Alternative
strategies, such as transition metal catalysis,16 are usu-
ally required. As shown here, anchimeric assistance,
engineered with the proper combination of Lewis acid and
solvent, provides another powerful activating mechanism
for such transformations.
(15) (a) Ganellin, C. R.; Hollyman, D. R.; Ridley, H. F. J . Chem. Soc.
C 1967, 2220-2225. (b) Ring opening of aziridinium electrophiles with
Grignard reagents has been reported, in a two-step sequence rather
than as a case of anchimeric assistance: Draper, R. W.; Hou, D.; Iyer,
R.; Lee, G. M.; Liang, J . T.; Mas, J . L.; Tormos, W.; Vater, E. J .;
Guenter, F.; Mergelsberg, I.; Scherer, D. Org. Proc. Res. Dev. 1998, 2,
175-185.
(16) (a) Frisch, A. C.; Shaikh, N.; Zapf, A.; Beller, M. Angew. Chem.,
Int. Ed. 2002, 41, 4056-4059. (b) Nakamura, M.; Matsuo, K.; Ito, S.;
Nakamura, E. J . Am. Chem. Soc. 2004, 126, 3686-3687. (c) Nagano,
T.; Hayashi, T. Org. Lett. 2004, 6, 1297-1299. (d) Scheiper, B.;
Bonnekessel, M.; Krause, H.; Furstner, A. J . Org. Chem. 2004, 69,
3943-3949. (e) Yang, L.-M.; Huang, L.-F.; Luh, T.-Y. Org. Lett. 2004,
6, 1461-1463. (f) Terao, J .; Ikumi, A.; Kuniyasu, H.; Kambe, N. J .
Am. Chem. Soc. 2003, 125, 5646-5647. (g) Tsuji, T.; Yorimitsu, H.;
Oshima, K. Angew. Chem., Int. Ed. 2002, 41, 4137-4139. (h) Terao,
J .; Watanabe, H.; Ikumi, A.; Kuniyasu, H.; Kambe, N. J . Am. Chem.
Soc. 2002, 124, 4222-4223. (i) Dohle, W.; Lindsay, D. M.; Knochel, P.
Org. Lett. 2001, 3, 2871-2873. (j) Shimizu, R.; Yoneda, E.; Fuchikami,
T. Tetrahedron Lett. 1996, 37, 5557-5560. (k) Hayashi, T.; Konishi,
M.; Kobori, Y.; Kumada, M.; Higuchi, T.; Hirotsu, K. J . Am. Chem.
Soc. 1984, 106, 158-163.
Ack n ow led gm en t. This work was supported by the
National Institute of General Medical Sciences, the
National Institutes of Health (GM 28384), the National
Science Foundation (CHE-9985553), the W. M. Keck
Foundation, and the Skaggs Institute for Chemical
Biology. P.-L.S. thanks the Ecole Normale Supe´rieure
de Lyon for a scholarship. We are grateful to Dr. Michal
Sabat (University of Virginia) for X-ray crystal struc-
tures.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures and characterization data for the products, and
crystallographic details for 2a and 2b. This material is
available free of charge via the Internet at http://pubs.acs.org.
J O0489869
J . Org. Chem, Vol. 69, No. 21, 2004 7339