Radical arylaminomethylation of unactivated alkenes

Article abstract of DOI:10.1021/ol901055j

Xanthates derived from open chain or cyclic N-chloromethylanllldes are capable of adding to various unactlvated alkenes to give adducts which, In suitable cases, can be made to undergo ring closure onto the aromatic ring. This flexible arylamlnomethylatlon of alkenes allows the rapid synthesis of open chain or polycycllc aniline derivatives.

Full text of DOI:10.1021/ol901055j

ORGANIC
LETTERS
2009
Vol. 11, No. 13
2844-2847
Radical Arylaminomethylation of
Unactivated Alkenes
Fre´de´ric Lebreux, Be´atrice Quiclet-Sire, and Samir Z. Zard*
Laboratoire de Synthe`se Organique, CNRS UMR 7652, Ecole Polytechnique,
91128 Palaiseau Cedex (France)
zard@poly.polytechnique.fr
Received May 13, 2009
ABSTRACT
Xanthates derived from open chain or cyclic N-chloromethylanilides are capable of adding to various unactivated alkenes to give adducts
which, in suitable cases, can be made to undergo ring closure onto the aromatic ring. This flexible arylaminomethylation of alkenes allows
the rapid synthesis of open chain or polycyclic aniline derivatives.
We recently found that it was possible to accomplish the
aminomethylation of unactivated alkenes through the use of
the degenerative radical xanthate transfer reaction.1 While
xanthates such as 1 (succinimido- or phthalimido-residues)
underwent the desired radical addition very efficiently to give
adducts 2 (Scheme 1, eq 1), the reaction of xanthate 4,
3 by a more significant contribution of canonical forms 3b
and 3c, giving the radical a greater “allylic character” (cf.
structure 3d), as compared to those derived from simple
lactams, where a second carbonyl group is missing. The
relative stabilities of the starting and adduct radicals in the
xanthate addition process are a key element for success.2
Normally, and absent special polar effects, the starting radical
has to be more stable than the adduct radical. This is the
case of radical 3, which appears to be more stable than the
simple secondary radical resulting from the addition to an
ordinary alkene.
Scheme 1
To extend the synthetic utility of this process, we examined
the case of isatin derivatives, where the fusion of an aromatic
ring adjacent to the amide nitrogen and the presence of the
extra carbonyl could still allow a similar but now a
“vinylogous” type stabilization, as shown in canonical
structures 5a-c in Scheme 1. We hoped that such stabiliza-
tion would be sufficient to allow control of the radical
addition process. The required xanthate precursor 6 is readily
(1) Quiclet-Sire, B.; Zard, S. Z. Org. Lett. 2008, 10, 3279.
(2) For reviews of the xanthate transfer chemistry, see: (a) Zard, S. Z.
Angew. Chem., Int. Ed. Engl. 1997, 36, 672. (b) Zard, S. Z. Radicals in
Organic Synthesis; Renaud, P., Sibi, M. P., Eds.; Wiley-VCH: Weinheim,
2001; Vol. 1, p 90. (c) Quiclet-Sire, B.; Zard, S. Z. Chem.sEur. J. 2006,
12, 6002. (d) Quiclet-Sire, B.; Zard, S. Z. Top. Curr. Chem. 2006, 264,
201. (e) Zard, S. Z. Org. Biomol. Chem. 2007, 5, 205.
derived from a pyrrolidone and not an imide, produced
mostly oligomers. The successful addition in the case of
xanthates 1 was attributed to the extra stabilization of radicals
10.1021/ol901055j CCC: $40.75
Published on Web 06/09/2009
 2009 American Chemical Society

available from N-chloromethylisatin,3 and this hypothesis
could be easily tested. We were, however, rapidly disap-
pointed.
Scheme 2
When xanthate 6 and allyl acetate were treated with lauroyl
peroxide, a complex mixture was obtained, as indicated by
TLC, from which only a small yield of the desired adduct 7
could be isolated by chromatography. More puzzling, the
NMR spectrum of the crude mixture was very poorly
resolved and exhibited broad signals suggesting the presence
of paramagnetic impurities in the sample. The most reason-
able explanation was that some radical addition was taking
place on the activated ketone group of the isatin, leading to
highly stabilized and perhaps even persistent radicals such
as 8.
Although this unexpected observation dashed our hopes
of extending our radical process to a naked isatin structure,
it raised the question of the actual need for the presence of
the second carbonyl group. The aromatic ring itself may
provide the necessary extra stabilization and thus replace
conveniently one of the carbonyl groups. The ketone group
in the isatin xanthate derivative 6 was therefore protected
as the corresponding ketal 9. Pleasingly, with the deleterious
effect of the reactive ketone removed, the desired radical
addition with allyl cyanide proceeded smoothly to give the
expected adduct 11a in 62% yield. Clearly, a phenyl group
can indeed act as a surrogate for the carbonyl, and intermedi-
ate radical 10 is sufficiently stabilized to allow control of
the degenerative xanthate transfer process.
In the same way, adducts 11b,c were prepared by reaction
with methyl 10-undecenoate and N-acetyl-N-allyl-4-bromoa-
niline, respectively. In the case of the last adduct, 11c, further
treatment with stoichiometric amounts of lauroyl peroxide
in refluxing chlorobenzene furnished indoline 12 in 68%
yield.4 The cyclization could also be performed in refluxing
ethyl acetate, but the yield was slightly lower (64%). No
ring closure on the aromatic ring of the isatin portion was
observed.
One notable advantage of this chemistry is its tolerance
for the presence of aromatic iodides. This is illustrated by
the normal reactivity displayed by 5-iodoisatin xanthate 13
(Scheme 2).5 Additions to allyl acetate, allyl trimethyl silane,
vinyl acetate, dimethyl vinylphosphonate, tridecafluo-
rooctene, and N-Boc allylamine proceeded uneventfully to
give the corresponding adducts 15a-f in generally good
yield. Addition to Boc-protected allylhydrazine provides a
direct access to complex hydrazine 15g.6 Such a compound
would be quite tedious to make by more conventional
chemistry. The presence of the iodine atom in all these
derivatives opens the way to combining the radical addition
with some of the most powerful transition metal catalyzed
processes, such as the Heck, Suzuki, Sonogashira, and
Buchwald-Hartwig reactions. Numerous novel isatin deriva-
tives may thus be accessed, some of which could have useful
pharmacological properties.7
The successful addition of the isatin derived xanthates gave
rise to yet another question: is the flat bicyclic structure, with
a maximized orbital overlap, necessary? In other words, can
the xanthate transfer be extended to open chain anilides or
to ones containing a more flexible ring such as benzazepi-
none? It must be remembered that in the present context the
differences in energy between radicals useful in the xanthate
transfer process, such as 3, and inappropriate radicals, such
as those derived from 4, are quite small, and the crossover
point, where oligomer formation becomes a serious problem,
is easily reached.
We were gratified to find that, even though somewhat less
effective than the isatin derivatives, xanthates 19-21,
prepared from the corresponding anilines, underwent reason-
ably smooth additions to various alkenes. The examples are
assembled in Scheme 3. The superiority of the isatin-derived
xanthates is best appreciated by comparing the yields of
addition to vinyl pivalate, an olefin much more prone to
polymerization than the other alkene partners. Thus, while
with isatin xanthate 13 the expected adduct 15c was obtained
in 53% yield (Scheme 2), extensive oligomerization was
observed with xanthate 20 as the reaction does not stop at
(7) For some recent, medicinally oriented work on isatin derivatives,
see: (a) Diaz, P.; Phatak, S. S.; Xu, J.; Astruc-Diaz, F.; Cavasotto, C. N.;
Naguib, M. J. Med. Chem. 2009, 52, 433. (b) Chu, W.; Rothfuss, J.; Chu,
Y.; Zhou, D.; Mach, R. H. J. Med. Chem. 2009, 52, 2188. (c) Hyatt, J. L.;
Moak, T.; Hatfield, M. J.; Tsurkan, L.; Edwards, C. C.; Wierdl, M.; Danks,
M. K.; Wadkins, R. M.; Potter, P. M. J. Med. Chem. 2007, 50, 1876. (d)
Vine, K. L.; Locke, J. M.; Ranson, M.; Pyne, S. G.; Bremner, J. B. J. Med.
Chem. 2007, 50, 5109. (e) Pirrung, M. C.; Pansare, S. V.; Sarma, K. S.;
Keith, K. A.; Kern, E. R. J. Med. Chem. 2005, 48, 3045.
(3) El Ghammarti, S.; Rigo, B.; Mejdi, H.; Henichart, J.-P.; Couturier,
D. J. Heterocycl. Chem. 1998, 35, 555.
(4) Ly, T.-M.; Quiclet-Sire, B.; Sortais, B.; Zard, S. Z. Tetrahedron Lett.
1999, 40, 2533.
(5) Garden, S. J.; Torres, J. C.; Souza Melo, S. C.; Lima, A. S.; Pinto,
A. C.; Lima, E. L. S. Tetrahedron Lett. 2001, 42, 2089.
(6) Quiclet-Sire, B.; Sortais, B.; Zard, S. Z. Synlett 2002, 903.
Org. Lett., Vol. 11, No. 13, 2009
2845

quinolones 28 from xanthates 27 (R′ ) Me or even H) we
examined a few years ago.8
Scheme 3
The lack of propensity to ring close onto the aromatic ring
is probably due to an unfavorable conformation in this case.
The acetyl on the nitrogen atom was therefore replaced with
a benzoyl to allow the radical another possibility of cycliza-
tion, namely on the aromatic ring of the benzoyl group. This
would lead to the creation of a fused seven-membered ring
benzazepinone structure (cf. 32). Radical cyclizations onto
aromatic rings to produce fused seven-membered rings are
quite rare but nevertheless precedented.9
Xanthate 29a was readily obtained from benzanilide by
the same procedure as for N-acetyl xanthate 19. This xanthate
underwent a fairly smooth radical addition to a number of
typical olefins to give the corresponding addition products
30a-d in satisfactory yields (Scheme 5). Treatment of adduct
the desired monoadduct 23c (Scheme 3). In the latter case,
the difference in stability between the initial radical and the
adduct radical (now stabilized by a three-electron interaction
with the lone pair of the pivalate oxygen) is not sufficient to
allow proper control of the addition process.
Scheme 5
We also noticed that the presence of an electron-donating
group on the aromatic ring, such as the butoxy group in 21,
caused a significant decrease in the efficiency of the addition,
as indicated by the modest yield of adducts 24a and 24b.
Furthermore, when a strong electron-withdrawing group was
present on the aromatic ring, the required N-chloromethyl
precursor could not be prepared by the usual procedure.
In the case of the isatin derivatives, no ring closure on
the aromatic portion of the isatin residue, leading for example
to 25, was observed when the adducts were further exposed
to stoichiometric amounts of peroxide (Scheme 4). The strain
Scheme 4
30b with stoichiometric amounts of lauroyl peroxide in
refluxing chlorobenzene resulted in the formation of tetrahy-
droquinoline 31b in 46% yield and none of benzazepinone
32. Replacement of the acetyl group on the aniline nitrogen
by a benzoyl group resulted thus in a very significant
improvement in the cyclization step, without the competitive
formation of benzazepinone 32. In the same way, substituted
xanthates 29b and 29c were converted into corresponding
tetrahydroquinolines 31e-g in a comparable overall yield,
without isolation of intermediate adducts 30e-g. As for 23c
above, the radical addition to vinyl acetate could not be
controlled, and essentially no 30h was obtained.
In a further example, lactam xanthate 33 not only
underwent the radical addition to N-Boc-allylamine to give
compound 34a in 51% yield, but the corresponding tricyclic
product 35a started forming already at this stage (Scheme
6). Indeed, exposure of purified adduct 34a to stoichiometric
amounts of lauroyl peroxide furnished tricycle 35a in 70%
resulting from the fusion of a five- and six-membered ring
around the aromatic nucleus must certainly raise the activa-
tion barrier of the cyclization. We hoped that the open chain
derivatives would prove to be better substrates. Unfortu-
nately, when compound 22a was treated with lauroyl
peroxide in refluxing chlorobenzene, only a small yield of
desired tetrahydroquinoline 26 was obtained. This contrasted
sharply with the ready formation of the analogous dihydro-
(8) Binot, G.; Zard, S. Z. Tetrahedron Lett. 2005, 46, 7503.
(9) (a) Bacque´, E.; El Qacemi, M.; Zard, S. Z. Org. Lett. 2004, 6, 3671.
(b) Kaoudi, T.; Quiclet-Sire, B.; Seguin, S.; Zard, S. Z. Angew. Chem., Int.
Ed. Engl. 2000, 39, 731, and references cited therein.
2846
Org. Lett., Vol. 11, No. 13, 2009

and 43% overall yield. The more rigid benzodiazepinone
analogue 40, obtained in good yield from known lactam 39,10
gave rise somewhat more efficiently to the very rare
structures 42a-d,11 again without isolation of the intermedi-
ate addition products 41a-d.
Scheme 6
These preliminary results greatly extend our initial work
on the radical aminomethylation of unactivated alkenes,
summarized by eq 1 in Scheme 1. The replacement of one
of the carbonyl groups in the original imide 1 with an
aromatic ring maintains sufficient stabilization of the initial
primary carbon radical to allow control of the radical chain
and results in a simple yet powerful and flexible process for
performing useful arylaminomethylations of unactivated
alkenes. As encapsulated in Scheme 7, this approach can
Scheme 7
lead rapidly to open chain or cyclic aniline derivatives
bearing varied and useful functionality, both in the xanthate
and olefinic partners. Aniline-based structures are of con-
siderable importance, especially in medicinal chemistry,12
and many of the compounds described in the present work
would not be readily available by more conventional routes.
Compounds 35e and 42d are striking examples of the
complexity that can rapidly be attained.
yield. It was in fact better to perform the intermolecular
addition and cyclization in the same flask, as shown by the
reaction with allyl acetate. The first adduct 34b was obtained
in 51% yield, whereas pushing the reaction all the way to
the tricyclic structure gave 35b in superior overall yield of
62%. In the same manner, compounds 35c, 35d, and 35d
were prepared in 46%, 70%, and 58% yield, respectively,
using allyl trimethylsilane, vinyl acetate, and vinylidene
carbonate as the olefinic traps. In the last case, the intermedi-
ate xanthate adduct was not even observed by thin-layer
chromatography.
Acknowledgment. We respectfully dedicate this paper to
Professor Rolf Huisgen (Ludwig Maximilians Universita¨t,
Munich). F.L. thanks Ecole Polytechnique for a scholarship.
Supporting Information Available: Experimental pro-
1
cedures, full spectroscopic data, and copies of H and 13C
NMR spectra for all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
OL901055J
It was also pleasing to find that the homologous benza-
zepinone xanthate 36 underwent the same addition-cycliza-
tion, even if the efficacy was slightly lower. Thus, without
isolation of the intermediate adducts 37a and 37b to allyl
and vinyl acetate, respectively, it was possible to obtain the
corresponding fused tricyclic derivatives 38a and 38b in 55%
(10) Dai-Il, J.; Tae-Wonchoi, C.; Yun-Young, K.; In-Shik, K.; You-
Mi, P.; Yong-Gyun, L.; Doo-Hee, J. Synth. Commun. 1999, 29, 1941.
(11) A search of Beilstein returned only 15 structures. See: Katritzky,
A. R.; Abonia, R.; Yang, B.; Qi, M.; Insuasty, B. Synthesis 1998, 1487.
(12) For example, tricyclic amides closely related to compounds 35 have
been found to be potent systemic fungicides. See: Bass, R. J.; Koch, R. H.;
Richards, H. C.; Thorpe, J. E. J. Agric. Food Chem. 1981, 29, 576.
Org. Lett., Vol. 11, No. 13, 2009
2847