Radical ipso-substitution of a carbon-fluorine bond leading to fluoro-7-azaindolines and fluoro-7-azaindoles

Article abstract of DOI:10.1021/ol500985f

Rare examples of a synthetically useful radical ipso-substitution of a carbon-fluorine bond are reported, leading to highly functionalized 5,6-difluoro-7-azaindolines. An unexpected hydrogen atom translocation and fragmentation with loss of molecular nitrogen and formation of a nitrile were observed in the case of an N-benzyl-tetrazole derivative.

Full text of DOI:10.1021/ol500985f

Letter
pubs.acs.org/OrgLett
Radical ipso-Substitution of a Carbon−Fluorine Bond Leading to
Fluoro-7-azaindolines and Fluoro-7-azaindoles
Zhibo Liu, Ling Qin, and Samir Z. Zard*
̀
Laboratoire de Synthese Organique, UMR 7652 CNRS/Ecole Polytechnique, 91128 Palaiseau Cedex, France
S
* Supporting Information
ABSTRACT: Rare examples of a synthetically useful radical
ipso-substitution of a carbon−fluorine bond are reported,
leading to highly functionalized 5,6-difluoro-7-azaindolines. An
unexpected hydrogen atom translocation and fragmentation
with loss of molecular nitrogen and formation of a nitrile were
observed in the case of an N-benzyl-tetrazole derivative.
zaindoles and related derivatives have attracted much
Scheme 1. An Unexpected Demethylation
A
synthetic effort in recent years.1 These molecular scaffolds
are akin to the enormously important indoles and constitute
particularly privileged structures in medicinal chemistry.2 While
aza-analogues of indole-based active compounds usually exhibit
improved solubility and bioavailability, synthetic routes to these
derivatives are often much more troublesome than in the indole
series. In fact, the classical indole syntheses are not easily
transposed to their aza analogues, either because of the limited
availability of the requisite precursors or because of reactivity
problems. Thus, when applied to the pyridines, the powerful
and widely used Fischer indole synthesis often results in poor
yields and sometimes even completely fails.3 In addition, very
few pyridylhydrazines are commercially available.
As part of our more general exploration of the potential of
the xanthates for constructing heteroaromatic structures,4,5 we
recently described a radical based approach for fusing various
sized rings around the pyridine and pyrimidine nuclei.6 In the
course of this work, we made a surprising observation while
studying a radical ring closure leading to fluoroazaindolines.7
We found that treatment of xanthate 1 with lauroyl peroxide
resulted in the formation of an azaindoline with the loss of a
fluorine atom and the unexpected loss of a methyl group on the
4-dimethyamino substituent to give difluoroazaindoline 2 as the
major isolated product (Scheme 1).
The reaction was not very clean, and the yield was variable
but generally modest (18−40%). A plausible mechanism for
this unusual transformation is briefly outlined in Scheme 1.
Radical 3 arising from fragmentation of the xanthate undergoes
ring closure into 4. This intermediate is reluctant to expel a
fluorine atom and prefers to react with the peroxide by electron
transfer to furnish the corresponding cation 5, which may be
drawn as mesomer 6 and which can aromatize into final
product 2 via intermediates 7 and 8 through loss of a hydrogen
fluoride and formaldehyde.
the iminium intermediate corresponding to 7 in a synthetically
more useful way.
Compound 11a was therefore prepared by lauroyl peroxide
(DLP) initiated radical addition of xanthate 10a to carbamate-
protected 2-N-allylamine pyridine 9a, a substrate easily
accessible from commercially available and cheap pentafluoro-
pyridine (Scheme 2).8 Upon treatment with stoichiometric
amounts of di-tert-butyl peroxide (DTBP), pyrrolidine
substituted azaindoline 12a was produced in a small yield
(17%), along with an even smaller yield of the corresponding
pyrrole 13a (10%; see discussion of mechanism below). The
addition of 2,6-lutidine resulted in a much cleaner reaction and
significantly improved the yield of both pyrroline 12a and
pyrrole 13a to a 56% combined yield. The lutidine most likely
In an attempt to gain a better understanding of the
mechanism and perhaps to expand the scope of the process,
we examined the behavior of other 4-amino-substituted
fluoropyridines. In particular, by studying cyclic amine
substituents we hoped to avoid the hydrolysis step and capture
Received: April 3, 2014
Published: April 28, 2014
© 2014 American Chemical Society
2704
dx.doi.org/10.1021/ol500985f | Org. Lett. 2014, 16, 2704−2707

Organic Letters
Letter
cyclized radical (e.g., 4 and 14a,b) and β-scission of the strong
carbon−fluorine bond (a more detailed discussion of this step
is given below). Both steps are relatively slow. In the presence
of an electron-releasing amine, as in substrates 1, 11a, 11b, and
18a, the oxidation of the intermediate cyclized radical is slightly
faster than elimination of a fluorine atom. This gives rise, via
the resulting iminium species, to demethylation in the case of 1,
to aromatization to pyrrole in the case of 11a, or to
decomposition products in the case of 11b and 18a.
Scheme 2. Synthesis of 7-Azaindolines
Acetylation of the cyclopropylamine in 18b slows down the
electron transfer to the peroxide allowing the fragmentation of
the carbon−fluorine bond to overtake the oxidation step. As a
consequence, the yield of azaindoline 19b increases nearly 4-
fold over 19a and becomes synthetically useful, especially when
combined with that of indole 20b. The formation of the latter is
apparently due to further oxidation of indoline 19b by the
peroxide. Blank experiments indicated that this is indeed
possible, even if complete conversion to the indole could not be
accomplished cleanly. Furthermore, re-examination of the
crude NMR spectra indicated that indoles were formed in
small amounts in all previous cyclization reactions but were
missed because of the relative complexity of the mixtures.
The beneficial effect of an electron-withdrawing group on the
nitrogen on limiting the formation of side products was next
exploited to access variously substituted difluoroazaindolines.
The starting materials 18 are shown in Scheme 4, along with
the corresponding indolines and indoles 19 and 20 (Xa =
−SC(S)OEt and −NPhth = phthalimido, throughout; the
yield of adducts 18 is also given throughout, even though the
precursors, xanthates 10 and alkenes 17, are not shown). A
mesyl group may be present on the nitrogen in the 4-position
of the pyridine ring with no significant effect on the outcome
(18d). In the case of xanthate 18f, we observed, in addition to
indoline 19f and indole 20f, a small amount of the tetralone
(not shown) derived by the competing addition of the
intermediate radical to the fluorobenzene ring.9
acts by neutralizing any hydrogen fluoride produced and limits
its destructive effect. With the higher piperidine homologue
11b, the main product was azaindoline 12b (33%). In this case,
the corresponding iminium 15b cannot evolve into a stable
aromatic structure and this pathway does not lead to
identifiable products.
In a further attempt to divert the reaction from its initial
course, we examined the transformation of adducts 18a and
18b, prepared from cyclopropylamine derivative 17a and its
acetylated analogue 17b (Scheme 3; yields for 17a and 17b are
Scheme 3. 7-Azainolines and 7-Azaindoles
We also found that aniline substituents (18h,i) gave
comparable yields of cyclization products without the need
for acetylation or mesylation and, curiously, the amount of the
accompanying indoles was much lower than in the case with
aliphatic amide substituents. With a phenoxy (18j) and
heteroaromatic substituents, such as a pyrrole, an imidazole, a
triazole, and even an indole, on the 4-position (18k,n), the
formation of the corresponding 7-azaindole was again
significant.
The tetrazole adduct 18o did not behave as expected. We
were surprised to find a major side product, where the benzyl
tetrazole motif had completely disappeared and which proved
to be nitrile 21 (Scheme 5). This compound results most
probably from a rare 1,6-hydrogen atom shift to give benzylic
radical 23 followed by rupture of the tetrazole ring and
expulsion of molecular nitrogen and iminyl radical 24. This
radical fragmentation appears to be unprecedented. In any case,
the translocation of radical 22 reflects the inherent sluggishness
of the cyclization step and is a testimony to the rather unique
nature of the radical chemistry of xanthates that indeed allows
such a difficult process to take place.
Mechanistically, the ring-forming step deserves some addi-
tional comments. Prior to our work, radical ipso-substitutions of
an aromatic or heteroaromatic fluorine atom were extremely
rare, and the possibility of homolysis of a strong C−F bond was
almost never invoked.10 We have argued for its occurrence in
our initial disclosures,7 and while incontrovertible proof is still
overall from pentafluoropyridine). The former, 18a, gave rise
disappointingly to a complex mixture, from which azaindoline
19a was isolated in a very poor yield (12%). In stark contrast,
the reaction of the latter was much cleaner and furnished
azaindoline 19b in 46% yield. Furthermore, a small amount of
the corresponding azaindole 20b could also be isolated in 18%
yield.
From a mechanistic standpoint, a plausible rationalization for
the apparent erratic behavior of these various substrates may be
proposed. It hinges on the competition between oxidation of
2705
dx.doi.org/10.1021/ol500985f | Org. Lett. 2014, 16, 2704−2707

Organic Letters
Letter
Scheme 4. Synthesis of Substituted 7-Azaindolines
Scheme 5. An Unexpected Fragmentation
slow unimolecular fragmentation with a high entropy term (cf.
ΔG = ΔH − TΔS).
This is further supported by the experiment shown in
Scheme 6, where a 5-fold dilution has essentially no effect on
Scheme 6. Effect of Dilution on the Yield
the yield in the conversion of 18p into indoline 19p and indole
20p. This observation would be incompatible with a slow
bimolecular oxidation or reduction of intermediate radical 25
and leaves open only two unimolecular options: either a simple
homolysis with loss of a fluorine atom or a solvolysis implying
expulsion of a fluoride anion.11 The poor leaving-group ability
of a fluoride anion, especially in a nonpolar solvent which
cannot stabilize an intermediate such as 26, militates strongly
against the latter possibility.
In summary, this approach offers numerous opportunities for
the concise and modular preparation of fluoroindolines and
fluoroindoles with potential applications in medicinal and
phyto-chemistry.12 The diversity in the structures may be
further increased by postmodification of the products, as
illustrated by the conversion of adduct 18q into ethoxy-
substituted indole 27, without separation of intermediates 19q
and 20q and by using IBX to complete the oxidative
aromatization of the N-deacetylated indoline (Scheme 7).13 It
is interesting to note that of the five fluorine atoms in
pentafluoropyridine only one fluorine is left in indole 27.
Finally, a preliminary experiment was performed to explore
unavailable, all our observations point to it. In particular, the
strong effect of temperature (data not shown) is in tune with a
2706
dx.doi.org/10.1021/ol500985f | Org. Lett. 2014, 16, 2704−2707

Organic Letters
Letter
(5) For reviews on the xanthate radical addition-transfer process, see:
(a) Zard, S. Z. Angew. Chem., Int. Ed. Engl. 1997, 36, 672. (b) Quiclet-
Sire, B.; Zard, S. Z. Chem.Eur. J. 2006, 12, 6002. (c) Quiclet-Sire, B.;
Zard, S. Z. Top. Curr. Chem. 2006, 264, 201. (d) Zard, S. Z. Aust. J.
Chem. 2006, 59, 663. (e) Quiclet-Sire, B.; Zard, S. Z. Pure Appl. Chem.
2011, 83, 519.
Scheme 7. Further Extensions
(6) (a) Bacque,
3671. (b) El Qacemi, M.; Ricard, L.; Zard, S. Z. Chem. Commun. 2006,
42, 4422. (c) Bacque, E.; El Qacemi, M.; Zard, S. Z. Heterocycles 2012,
́
E.; El Qacemi, M.; Zard, S. Z. Org. Lett. 2004, 6,
́
84, 291. (d) Petit, L.; Botez, I.; Tizot, A.; Zard, S. Z. Tetrahedron Lett.
2012, 53, 3220. (e) Laot, Y.; Petit, L.; Zard, S. Z. Chem. Commun.
2010, 46, 5784. (f) Liu, Z.; Qin, L.; Zard, S. Z. Org. Lett. 2012, 14,
5976.
(7) (a) Laot, Y.; Petit, L.; Zard, S. Z. Org. Lett. 2010, 12, 3426.
(b) Laot, Y.; Petit, L.; Tran, N. D. M.; Zard, S. Z. Aust. J. Chem. 2011,
64, 416.
(8) Pentafluoropyridine usually reacts with amines first at the 4-
position then at the 2-position. (a) Hargreaves, C. A.; Sanford, G.;
Slater, R.; Yufit, D. S.; Howard, J. A. K.; Vong, A. Tetrahedron 2007,
63, 5204. (b) Sandford, G.; Slater, R.; Yufit, D. S.; Howard, J. A. K.;
Vong, A. J. Org. Chem. 2005, 70, 7208. (c) Cartwright, M. W.;
Convery, L.; Kraynck, T.; Sandford, G.; Yufit, D. S.; Howard, J. A. K.;
Christopher, J. A.; Miller, D. D. Tetrahedron 2010, 66, 519.
(9) For the synthesis of tetralones, see: Liard, A.; Quiclet-Sire, B.;
Saicic, R. N.; Zard, S. Z. Tetrahedron Lett. 1997, 38, 1759.
̌
(10) (a) Brooke, G. M. J. Fluorine Chem. 1997, 86, 1. (b) Sket, B.;
access to tetrahydro-azaquinolines such as 30 via the
corresponding precursor 29. This entails the more difficult
creation of a six-membered ring. Indeed, the efficiency of the
cyclization step proved much lower, as expected, but further
work is needed to delineate its scope.
Zupan, M.; Zupancic, N.; Pahor, B. Tetrahedron 1991, 47, 5029.
(c) Kobrina, L. S. J. Fluorine Chem. 1989, 82, 301. (d) Claret, P. A.;
Williams, G. H.; Coulson, J. J. Chem. Soc. C 1968, 341.
(11) Solvolyses from radicals are well documented, but require a
good leaving group and a polar solvent: (a) Beckwith, A. L. J.; Crich,
D.; Duggan, P. J.; Yao, Q. Chem. Rev. 1997, 97, 3273. (b) Crich, D.;
Brebion, F.; Suk, D.-H. Top. Curr. Chem. 2006, 263, 1.
(12) (a) Fluorinated Heterocyclic Compounds: Synthesis, Chemistry and
Applications; Petrov, V. A., Ed.; J. Wiley & Sons: Hoboken, NJ, 2009.
́ ́
(b) Wang, J.; Sanchez-Rosello, M.; Acena, J. L.; del Pozo, C.;
̃
Sorochinsky, A. E.; Fustero, S.; Soloshonok, V. A.; Liu, H. Chem. Rev.
2014, 14, 2432.
̌
̌
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures, full spectroscopic data, and copies of
1H and 13C NMR spectra for all new compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
(13) Nicolaou, K. C.; Mathison, C. J. N.; Montagnon, T. J. Am. Chem.
Soc. 2004, 126, 5192.
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: samir.zard@polytechnique.edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Ecole Polytechnique for scholarships to Z.L. and L.Q.
This paper is dedicated with respect to the memory of Peter L.
Pauson (University of Strathclyde).
REFERENCES
(1) For reviews, see: (a) Mer
Joseph, B. Tetrahedron 2013, 69, 4767. (b) Popowycz, F.; Routier, S.;
■
́
our, J.-Y.; Routier, S.; Suzenet, F.;
Joseph, B.; Mer
́
our, J.-Y. Tetrahedron 2007, 63, 1031. (c) Popowycz,
F.; Merour, J.-Y.; Joseph, B. Tetrahedron 2007, 63, 8689. (d) Song, J.
́
J.; Reeves, J. T.; Gallou, F.; Tan, Z.; Yee, N. K.; Senanayake, C. H.
Chem. Soc. Rev. 2007, 36, 1120. (e) Schirok, H. J. Org. Chem. 2006, 71,
5538. (f) Humphrey, G. R.; Kuethe, J. T. Chem. Rev. 2006, 106, 2875.
(g) Cacchi, S.; Fabrizi, G. Chem. Rev. 2005, 105, 2873.
(2) For a review, see: Zhang, A.; Neumeyer, J. L.; Baldessarini, R. J.
Chem. Rev. 2007, 107, 274.
(3) (a) Jeanty, M.; Blu, J.; Suzenet, F.; Guillaumet, G. Org. Lett. 2009,
11, 5142 and references therein. (b) Zhang, Z.; Yang, Z.; Wong, H.;
Zhu, J.; Meanwell, N. A.; Kadow, J. F.; Wang, T. J. Org. Chem. 2002,
67, 6226.
(4) For a review, see: El Qacemi, M.; Petit, L.; Quiclet-Sire, B.; Zard,
S. Z. Org. Biomol. Chem. 2012, 10, 5707.
2707
dx.doi.org/10.1021/ol500985f | Org. Lett. 2014, 16, 2704−2707