A Polar Effects Controlled Enantioselective 1,2-Chlorine Atom Migration via a Chlorine-Bridged Radical Intermediate

Full text of DOI:10.1021/ja011129r

Published on Web 02/13/2002
A Polar Effects Controlled Enantioselective 1,2-Chlorine Atom Migration via a
Chlorine-Bridged Radical Intermediate
Eng Wui Tan,* Bun Chan, and Allan G. Blackman
Department of Chemistry, UniVersity of Otago, P.O. Box 56, Dunedin, New Zealand
Received May 7, 2001; Revised Manuscript Received September 7, 2001
Previously, radical rearrangements involving chlorine atom
migration have been attributed to the formation of more stable
radical intermediates.1 Here, we report strong evidence of an
enantioselective 1,2-chlorine atom migration governed by polar
substituent effects. In previous work,2 we described experimental
and theoretical studies on the stereoselective N-bromosuccinimide
Figure 1. Stereoselective radical bromination of methyl (S)-2-chloro-3-
phenylpropanoate 1a via a chloro-bridged intermediate.
(NBS) radical bromination of R-chlorohydrocinnamic acid deriva-
tives (for example, Figure 1). The stereoselectivity was observed
to be dependent upon the nature of the acid derivative and was
partly attributed to the conformer distribution of the radical
intermediate. A chlorine-bridged radical was proposed to account
for the preferred conformation of the reaction intermediate.
To gain a deeper understanding of the nature of the bridging
chlorine radical, the bromide 2a was treated with tributyltin hy-
dride, as it was expected that the tributyltin radical would pref-
erentially abstract bromine to chlorine and that the tin reaction
would show contrasting polar effects to that of NBS bromination3
and hence provide us with greater insight into the bridged chlorine
intermediate.
The reaction of 2a with tributyltin hydride gave a mixture of
three products (Figure 2, entry a). Compounds 1a and 4a were
expected from reduction and â-scission processes, respectively.
However, one of the products, 3a, showed a chlorine migration
from C-2 to C-3. This atom migration implies a radical rearrange-
ment from a benzylic radical to a heteroallylic radical, which is
presumably thermodynamically disfavored.4 To further investigate
the nature of this rearrangement, compounds 2a-j were synthesized.
The substrates 2a-j were prepared by radical bromination of
Figure 2. Reactions between carboxylic acid derivatives 2a-j and
the corresponding R-chlorodihydrocinnamic acid derivatives with
NBS1 except for compounds 2b, 2c, and 2e. The R,â-dichlorodi-
hydrocinnamate 2c was prepared by radical chlorination of 1c
employing sulfuryl chloride. The R-bromo-â-chlorodihydrocinna-
mate 2b was made by treating methyl cinnamate with NBS and
HCl in acetone/water, followed by reaction of the bromohydrin with
PCl5. Treatment of methyl p-methoxycinnamate with NBS and HCl
in acetone/water gave compound 2e directly. The reduction of com-
pounds 2a-j (Figure 2) was carried out at room temperature with
1 mol equiv of tributyltin hydride in benzene for 2 h, with initiation
by irradiation with a 160 W mercury gas discharge lamp. The crude
reaction mixtures were evaporated under reduced pressure, and the
residues were analyzed using 1H NMR spectroscopy. By comparing
the integration of the signals corresponding to the R- and â-protons
of compounds 1-4 to that of the n-butyl signals of tributyltin, used
as an internal standard, it was determined that for each case the
total recovery was greater than 90%. The relative proportions of
the compounds were also determined from the integration of the
R- and â-proton signals and are shown in Figure 2.
tributyltin hydride gave reduction products 1a-j and 3a-j, and elimination
products 4a-j. *Compounds 2b and 2e were racemic mixtures of (2R*,3S*)
configuration. Products 1b and 3b were also racemic mixtures. An
enantiomerically pure sample of compound 2b would be expected to produce
1b and 3b with stereochemistry that is inverted to that shown in the scheme.
attempts to resolve the regioisomers were unsuccessful. Of the
compounds 3 that were not contaminated with 1, 3h is crystalline,
while 3f is an oil. Therefore, the stereochemistry of 3h was
determined and used to represent the configuration of 3. Compound
3h was separated from the reaction mixture of 2h,5 and the absolute
configuration was found to be (S) by X-ray crystallography.6 A
similar experiment with racemic 2h gave racemic crystals of 3h.7
These results show that conglomerate crystallization of 3h does
not occur and that the chlorine migration proceeds with a high
degree of stereoselectivity. From a comparison of the optical rotation
of crude 3h to that of a recrystallized sample used for crystal
structure determination, the enantiomeric excess was found to be
at least 85%. Thus, the possibility of rearrangement via nonste-
reoselective pathways, such as elimination of 2 to give the
corresponding cinnamate 4, followed by addition of HCl to give 1
and/or 3, can be excluded. In addition, prolonged reaction of 2a
gave identical results, as shown in entry a of Figure 2, indicating
To elucidate the mechanism of the chlorine migration, the
stereochemistry of product 3 needed to be determined. Unfortu-
nately, in most cases 3 was formed in conjunction with 1, and
9
2078 VOL. 124, NO. 10, 2002 J. AM. CHEM. SOC.
10.1021/ja011129r CCC: $22.00 © 2002 American Chemical Society

C O M M U N I C A T I O N S
that the products from the reduction reactions were stable under
the reaction conditions. The formation of the cinnamic acid
derivatives 4 can be attributed to the loss of a chlorine atom via a
â-scission process during bromine atom abstraction.
nol being electron deficient at the reacting carbon.10 Hence, hydro-
gen delivery to the benzylic carbon is favored over reduction at
the R-carbon.
In conclusion, reduction of various dihalogenated-dihydrocin-
namic acid derivatives 2 by tributyltin hydride gave varying amounts
of the corresponding direct reduction product, cinnamate from elimi-
nation, and a product from radical rearrangement via a 1,2-chlorine
migration. In one representative case, the chlorine migration was
shown to be enantioselective. The reaction mechanism is likely to
involve generation of chlorine-bridged intermediates from the halo-
gen abstraction step, followed by hydrogen atom transfer to either
the â-carbon or the R-carbon to give 1 and 3, respectively. The
extent of chlorine migration was affected by the nature of the hydro-
gen atom donor and electron-withdrawing substituents. Electron-
withdrawing groups appear to favor reduction at the proximate
carbon. This is likely to result from the stabilization of an electron-
rich carbon in the transition state of the hydrogen delivery step.
These results point to a 1,2-chlorine atom migration controlled by
polar substituent effects and suggest a chlorine-bridged radical
where there is substantial interaction between the chlorine sub-
stituent and the vicinal semioccupied orbital of the radical
intermediate.
Reduction of R,â-bromo-chloro regioisomers 2a and 2b gave
effectively the same product distribution. Thus, reaction of 2a and
2b presumably gave the same radical intermediate, or the same
ratio of intermediates, along the reaction path. In addition, reduction
of the R,â-dichloro compound 2c also gave roughly the same
product distribution as 2a and 2b. The low conversion of 2c can
be attributed to the benzylic C-Cl bond in 2c being stronger than
the corresponding C-Br bond in 2a and 2b. Reduction of the free
acid 2g, the acid chloride 2f, and the pentafluorophenyl ester 2h
all gave higher ratios of isomer 3 to isomer 1. In the case of 2f and
2h, no R-chloro product 1 was produced. Conversely, when an
electron-withdrawing nitro group was introduced on to the phenyl
ring (entry d), the reduction gave more of isomer 1 than of isomer
3. However, when the para-substituent was an electron-donating
methoxy group (entry e), only the product from elimination was
formed. In contrast to most of the ester derivatives, reaction with
the amides 2i and 2j gave more of isomer 1 than of isomer 3.
From the relative amounts of compounds 1 and 3, it can be seen
that electron-withdrawing groups tend to favor hydrogen atom
delivery to the proximate carbon. As tin is more electropositive
than carbon, the transition state of halogen abstraction and of
hydrogen atom delivery would be expected to have a charge
separation, with a partial positive charge on tin and a partial negative
charge on carbon.3b An electron-withdrawing group close to the
reacting carbon would stabilize the transition state by delocalizing
the developing partial negative charge and hence increase the rate
of hydrogen atom transfer to that carbon. Thus, these observations
are likely to be a result of polar substituent effects.
Acknowledgment. We thank Ward T. Robinson and Jan
Wikaira for their assistance with the X-ray crystal structure of
compound 3h.
Supporting Information Available: Detailed experimental pro-
cedures for the synthesis of all starting material, tributyltin hydride
reduction, and thiophenol trapping, characterization data of reduction
products, and X-ray crystallography data for 3h (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
References
To further clarify the mechanism of the radical rearrangement,
a trapping experiment was conducted employing thiophenol8 as an
alternative hydrogen atom donor to tributyltin hydride. Thus, 1 mol
equiv of thiophenol was mixed with 2a, and tributyltin hydride (1
mol equiv) was introduced to the mixture slowly as the reaction
(1) (a) Urry, W. H.; Eiszner, J. E. J. Am. Chem. Soc. 1952, 74, 5822. (b)
Benson, H. L., Jr.; Willard, J. E. J. Am. Chem. Soc. 1961, 83, 4672. (c)
Benson, H. L., Jr.; Willard, J. E. J. Am. Chem. Soc. 1966, 88, 5689. (d)
Skell, P. S.; Shea, K. J. In Free Radicals; Kochi, J. K., Ed.; Wiley: New
York, 1973; Vol. 2, p 809.
(2) (a) Shaw, J. P.; Tan, E. W. J. Org. Chem. 1996, 61, 5635. (b) Wong, L.
S.; Chan, B.; Tan, E. W. Tetrahedron Lett. 2000, 41, 2671.
1
proceeded. Analysis of the crude reaction material by H NMR
(3) For an overview of polar effects in halogenation and tributyltin hydride
reduction see: (a) Russell, G. A. In Free Radicals; Kochi, J. K., Ed.;
Wiley: New York, 1973; Vol. 2, p 275. (b) Blackburn, E. V.; Tanner, D.
D. J. Am. Chem. Soc. 1980, 102, 692.
spectroscopy as described above showed 31% of 2a, 39% of the
R-chloro product 1a, and 30% of the cinnamate 4a. The â-chloro
product 3a was not detected. In addition, another trapping experi-
ment employing the bromo-chloro regioisomer 2b under identical
reaction conditions effectively gave the same result, with 30% of
2b, 42% of 1b, and 28% of 4b.
(4) For a comparison of various R-H bond energies see: Berkowitz, J.;
Ellison, G. B.; Gutman, D. J. Phys. Chem. 1994, 98, 2744.
(5) To obtain the optically active â-chlorodihydrocinnamate 3h, the reduction
of 2h was carried out with 2 mol equiv of tributyltin hydride for a complete
consumption of 2h. The cinnamate 4h was reacted with NBS in acetone
under acidic conditions to give the corresponding bromohydrin, which
was then separated from 3h by column chromatography. Recrystallization
from hexane gave 3h as white crystals.
The results from the trapping experiments suggest that only one
radical intermediate was formed from the halogen abstraction step
and that the same intermediate was formed from halogen abstraction
reactions of 2a and 2b. Thus, in the tributyltin hydride reduction
of compounds 2a-j, a chlorine-bridged radical intermediate was
formed in the halogen abstraction step, and a hydrogen atom can
be delivered to either the â-carbon to give 1 or the R-carbon to
give 3. There have been various electron spin resonance studies of
â-chloroalkyl radicals which have indicated that the chloro sub-
stituent interacts with the radical center.9 In this case, there would
also have to be interaction between the chloro substituent and the
vicinal radical center to account for the chlorine migration. The
proportion of various products was affected by the nature of the
hydrogen atom donor, such that reaction between the chlorine-
bridged intermediate and tributyltin hydride gave both 1a and 3a,
while thiophenol trapped the radical intermediate to give 1a
exclusively. The selectivity of the thiophenol reaction can be
attributed to the transition state of hydrogen delivery from thiophe-
(6) Despite numerous attempts at recrystallization of 3h, from a variety of
solvents, only twinned crystals were produced. However, a poor quality
(R1 ) 19%) crystal structure of 3h was obtained from a multiply twinned
crystal. RLATT was used to extract data from one reciprocal lattice only,
and these were then integrated. The space group P212121 was assigned
on the basis of figures of merit obtained from XPREP, and the structure
was solved using SHELXS-97 and SHELXL-97, respectively. The Flack
absolute structure parameter has a value of -0.1240 ( 0.5032, while the
inverted stereochemistry requires a value of 1.0788 ( 0.5039. Despite
the poor quality of the data, we believe that the absolute configuration of
3h has been determined correctly.
(7) Chan, B.; Tan, E. W.; Blackman, A. G. Acta Crystallogr. 2001, E57, o725.
(8) For a comparison of tributyltin hydride and thiophenol as hydrogen atom
donors, see kinetic results for reactions of tributyltin hydride and
thiophenol with alkyl radicals: (a) Chatgilialoglu, C.; Ingold, K. U.;
Scaiano, J. C. J. Am. Chem. Soc. 1981, 103, 7739. (b) Franz, J. A.; Bushaw,
B. A.; Alnajjar, M. S. J. Am. Chem. Soc. 1989, 111, 268.
(9) (a) Lyons, A. R.; Symons, M. C. R. J. Am. Chem. Soc. 1971, 93, 7330.
(b) Bowles, A. J.; Hudson, A.; Jackson, R. A. Chem. Phys. Lett. 1970, 5,
522. (c) Edge, D. J.; Kochi, J. K. Tetrahedron Lett. 1972, 1341. (d) Edge,
D. J.; Kochi, J. K. J. Am. Chem. Soc. 1972, 94, 6485.
(10) Dolbier, W. R., Jr.; Rong, X. X. Tetrahedron Lett. 1994, 35, 6225.
JA011129R
9
J. AM. CHEM. SOC. VOL. 124, NO. 10, 2002 2079