Application of peptidyl radicals into a new radical cascade leading to unsaturated γ-lactams

Article abstract of DOI:10.1021/jo020602w

Radical cyclization of dipeptides 1a-h proceeds smoothly to give five- and seven-membered rings in good to moderate total yields using Stork's catalytic tin hydride method. A radical is generated on a protecting group and translocated to the peptide moiety. Following a cyclization reaction, the vinyl radical can abstract hydrogen from a benzyl group on an amine, which results in elimination of the protected amine group. Encouraging results have notably been obtained with amino acids other than glycine.

Full text of DOI:10.1021/jo020602w

SCHEME 1. Gen er a l Str a tegy
Ap p lica tion of P ep tid yl Ra d ica ls in to a
New Ra d ica l Ca sca d e Lea d in g to
Un sa tu r a ted γ-La cta m s
Robert Andrukiewicz, Rafał Loska,
Vladimir Prisyahnyuk, and Krzysztof Stalin´ski*
Institute of Organic Chemistry, Polish Academy of Sciences,
Kasprzaka 44/ 52, 01-224 Warsaw, Poland
stalinsk@icho.edu.pl
Received September 16, 2002
Abstr a ct: Radical cyclization of dipeptides 1a -h proceeds
smoothly to give five- and seven-membered rings in good to
moderate total yields using Stork’s catalytic tin hydride
method. A radical is generated on a protecting group and
translocated to the peptide moiety. Following a cyclization
reaction, the vinyl radical can abstract hydrogen from a
benzyl group on an amine, which results in elimination of
the protected amine group. Encouraging results have notably
been obtained with amino acids other than glycine.
tions. In this Note we would like to disclose a promising
protocol for such transformation as generally described
in Scheme 1. Vinyl radicals necessary for this process can
be generated via addition of peptidyl radicals to alkynes.
Following a cyclization reaction, the vinyl radical could
abstract hydrogen from a benzyl group on an amine,
which might result in elimination of the protected amine
group. A driving force for such a process should be
isomerization of the exocyclic double bond to form R,â-
unsaturated γ-lactams. Recently Rancourt reported that
glycinyl radicals were formed using a N-Boc,N-2-bro-
mobenzyl group and a 1-trimethylsilyl alkyne group could
be used to intercept the intermediate radical to form a
vinyl radical and finally functionalized five-membered
heterocycles.9 However, any further elimination process
was not observed. We supposed that alkyl substituents
on a nitrogen atom were necessary for the overall process.
We decided to examine a slightly modified system (N-2-
bromobenzyl,N-methyl group) and for that purpose N,N-
substituted dipeptides 1a -h having a triple bond on a
side chain were chosen as radical precursors.
The dipeptides were prepared by coupling methyl N-3-
(trimethylsilyl)propargyl glycinate with N-2-bromo-ben-
zyl,N-methyl amino acids using TBTU and N-methyl-
morpholine in CH2Cl2. In turn, the N-protected amino
acids were prepared in three steps (40-80% overall
yields) from the corresponding methyl ester hydrochloride
by (1) introduction of the 2-bromobenzyl group with
2-bromobenzaldehyde and anhydrous MgSO4 in CH2Cl2
followed by reduction of the imines with NaBH4, (2)
In recent years, many examples of hydrogen-atom
transfer reactions affording amidocarboxy-substituted
radicals have been reported.1 Such radicals, classified as
captodative, mero, or push-pull systems are stabilized
by the combined action of an electron-releasing amido
substituent and an electron-withdrawing carboxy sub-
stituent.2 Among them, glycinyl ones appeared to be
important intermediates in the preparation of unnatural
R-amino acid derivatives3 and peptidomimetics.4 Deriva-
tives of many other amino acids also form R carbon-
centered radicals in a similar manner, and these react
in an way identical to that of the corresponding glycinyl
radicals. Recent rapid synthetic advances in the forma-
tion of such radicals established protecting/radical trans-
locating groups (PRT) as an alternative way for their
generation.5 In general, a radical is initially generated
in a “protecting group” and then translocated by 1,5-
hydrogen transfer to the desired site before any further
radical event.6 A number of PRT groups are available
now.7 They differ from each other from the standpoint of
protecting and functional groups. However, in most of
the cases they do not undergo elimination during radical
processes and consequently such groups require, if neces-
sary, an additional synthetic step to be removed.8
(6) Generally, other 1,n-H transfers are known, for example. 1,3-H
transfer: Damour, D.; Barreau, M.; Dhaleine, F.; Doerfingler, G.;
Vuilhorgne, M.; Mignani, S. Synlett 1996, 890-892. 1,4-H transfers:
(a) Wallace, T. J .; Gritter, R. J . J . Org. Chem. 1961, 26, 5256. (b)
J ohnson, R. A.; Greene, F. D. J . Org. Chem. 1975, 40, 2186-2192. (c)
Gilbert, B. C.; Parry, D. J .; Grossi, L. J . Chem. Soc., Faraday Trans.
1 1987, 77-83. (d) J ournet, M.; Malacria, M. Tetrahedron Lett. 1992,
33, 1893-1896. 1,6-H transfers: (a) Gross, A.; Fensterbank, L.; Bogen,
S.; Thouvenot, R.; Malacria, M. Tetrahedron Lett. 1997, 53, 13797-
13810. (b) Nedelec, J . Y.; Lefort, D. Tetrahedron 1975, 31, 411-417.
(c) Curran, D. P.; Yu, H. S.; Liu, H. T. Tetrahedron 1994, 50, 7343-
7366. 1,7-H transfers: (a) Curran, D. P.; Somayajula, K. S.; Yu, H.
Tetrahedron Lett. 1992, 33, 2295-2298. (b) Denenmark, D.; Winkler,
T.; Waldner, A.; De Mesmaeker, A. Tetrahedron Lett. 1992, 33, 3613-
3616. (c) Curran, D. P.; Xu, J .; Lazzarini, E. J . Chem. Soc., Perkin
Trans. 1 1995, 3049-3059.
In this context, we envisioned that it would be useful
to test a group that once used to activate the peptidyl
position could be eliminated after radical transforma-
* To whom correspondence should be addressed. Fax: 22-632-66-
81.
(1) Easton, Ch. J . Chem. Rev. 1997, 97, 53-82.
(2) (a) Viehe, H. G.; J anousek, Z.; Mere´nyi, R.; Stella, R. Acc. Chem.
Res. 1985, 18, 148-154. (b) Balaban, A. T. Rev. Roum. Chim. 1971,
16, 724. (c) Katritzky, A. R.; Soti, F. J . Chem. Soc., Perkin Trans. 1
1974, 1427.
(3) Renaud, P.; Giraud L. Synthesis 1996, 913-926.
(4) (a) Wyss, C.; Batra, R.; Lehmann, C.; Sauer, S.; Giese, B. Angew.
Chem., Int. Ed. Engl. 1996, 35, 2529-2531. (b) Sauer, S.; Staehelin,
C.; Wyss, C.; Giese, B. Chimia 1997, 51, 23-24.
(5) Curran, D. P.; Xu, J . J . Am. Chem. Soc. 1988, 110, 5900-5902.
(7) Curran, D. P.; Xu, J . J . Am. Chem. Soc. 1996, 118, 3142-3147
and references therein.
(8) Curran, D. P.; Yu, H. Synthesis 1992, 123.
(9) Rancourt, J .; Gorys, V.; J olicoeur, E. Tetrahedron Lett. 1998, 39,
5339-5342.
10.1021/jo020602w CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/28/2003
1552
J . Org. Chem. 2003, 68, 1552-1554

SCHEME 2. Ra d ica l Cycliza tion s of Dip ep tid es
SCHEME 3. Mech a n istic P a th w a y
1a -h
TABLE 1. Yield s of Isola ted P r od u ctsa
R1
R2
2 (%)
3 (%)
4 (%)
1a
1b
1c
1d
1e
1f
H
Me
Me
Et
CH2Ph
i-Pr
i-Bu
Me
Me
Bn
Me
Me
Me
Me
4
17
8
14
17
22
36
23
90
68
39
45
37
32
32
39
9
49
11
10
13
11
1g
1h
CH2CH2CH2
a
Consistent spectral data (IR, MS,1H, and 13C NMR) and correct
HRMS were obtained for all compounds.
radical chlorination of derivatives of valine and sarcosine
where reaction occurred at other than amidocarboxy-
positions.11
reductive methylation of the nitrogen atom using form-
aldehyde, and (3) acidic hydrolysis of the N,N-protected
methyl esters with HClaq. The cyclization reactions were
initially conducted under different concentration condi-
tions with n-tributyltin hydride (0.01-0.05 M). After
some experimentation, it was found that Stork’s catalytic
procedure10 gave better results. Isolated yields were
determined in these reactions (Scheme 2, Table 1).
Noncatalytic procedures lowered yields slightly and did
not significantly improve reaction time. Reaction time
varied from one to several days. In the case of 1a , the
reaction proceeded very cleanly in high yield (90%). In
contrast, the attempted cyclization reactions with 1b-h
gave products 3b-h in lower yields. At ∼80 °C, the yields
were in the vicinity of 35% (except 3b, 68%) under either
set of the conditions. Two types of side products were
observed. One of them reflected direct reduction of
radicals before any further cyclization (2a -h ). At this
point it is not clear if it comes from direct reduction of
the aryl radical prior to 1,5-hydrogen shift or takes place
after it. Labeling experiments are currently underway.
The second type of products resulted from another
possible 1,5-hydrogen transfer from the N-methyl group
followed by 7-exo cyclization (4b-h ). Generally, ami-
docarboxy-substituted radicals are considerable more
stable and they should be formed preferentially, but
hydrogen-transfer processes may afford less stable prod-
ucts if a radical character in the reaction transition state
is important. For example, Easton reported regioselective
A reasonable mechanism for the reaction of 1c is
illustrative of all of the substrates, and this is depicted
in Scheme 3. Bromine abstraction to generate the aryl
radical is followed by a competition among direct reduc-
tion (leading to 2c), and 1,5-hydrogen transfer (leading
to 3b or 4c). Fortunately, the hydrogen transfer reactions
of aryl radicals are sufficiently rapid to allow the follow-
ing cyclizations.12 From two possible modes of cyclization
5-exo one is largely favored over a 7-exo ring-closure (6).
However, they are independent from each other. Once
formed, radical 5 can smoothly undergo 5-exo cyclization
to afford vinyl radical 7. This, in turn can be reduced;
however, in practice another 1,5-hydrogen transfer takes
place, which is at the heart of the whole system (8).
Similar intramolecular 1,5-hydrogen transfers producing
R-amino radicals have been reported.13 The next step
presumably involves â-fragmentation to form radical 9.
This in turn could follow a two-step process: reduction
and subsequent isomerization of the exocyclic double
bond or more preferentially allylic isomerization to a
more stable conjugated system and finally its reduction
with the tin hydride to 3b.
(11) (a) Bowman, N. J .; Hay, M. P.; Love, S. G.; Easton, Ch. J . J .
Chem. Soc., Perkin Trans. 1 1988, 259-264. (b) Hay, M. P.; Love, S.
G.; Easton, Ch. J . J . Chem. Soc., Perkin Trans. 1 1988, 265-268.
(12) Aryl radical/solvent reactions are quite fast; therefore rates of
the cyclization should be at least ∼106.
(13) (a) Murakami, M.; Hayashi, M.; Ito, Y. J . Org. Chem. 1992,
57, 794-795. (b) Undheim, K.; Williams, L. J . Chem. Soc., Chem.
Commun. 1994, 883-884.
(10) Stork, G.; Sher, P. M. J . Am. Chem. Soc. 1986, 108, 303-304.
J . Org. Chem, Vol. 68, No. 4, 2003 1553

SCHEME 4. Ra d ica l Ca sca d e of Dip ep tid e 1h
SCHEME 5. Rota m er s of r-Am id oca r boxy Ra d ica ls
It has been suggested that the geometry of an initial
rotamer of amidocarboxy-substituted radicals played a
crucial role in deciding the course of radical reactions.15
The radical generated from syn-13 can cyclize, whereas
the radical from anti-13 is topologically prohibited from
cyclizing. It is not clear whether at 80 °C the conforma-
tional barrier can be efficiently traversed within the
lifetime of the radical, providing a cyclization option for
it (Scheme 5). Thus, in the tin hydride reaction direct
reduction of anti-rotamer 13 is observed.
The idea of the self-deprotecting group is already
useful, but further improvements are desirable. This is
perhaps due to relatively slow 1,5-hydrogen transfer in
the case of amino acids other than glycine and another
competing 1,5-H transfer leading to seven-membered
rings. Further work is clearly needed to improve ef-
ficiency of the overall process. This problem constitutes
a potentially serious limitation that, depending on sub-
stituents, may be difficult to overcome in some cases.
However, there is a hope that an alternative would be to
use the “double-barreled”16 o,o-dibromobenzyl group. If
the directly reduced product is formed after abstraction
of the first bromine atom and subsequent reduction, then
this molecule gets a second chance at cyclization when
the second bromine exists. The results of these investiga-
tions will be reported in due course.
The proposed mechanism raised a question about a
source of hydrogen atom in the final 1,5-hydrogen
transfer. One approach should involve hydrogen atom
transfer from the benzylic position; the second one
indicated the N-methyl group as a possible explanation.
The main problem in answering this question was how
to prove it experimentally. Compound 1c (Table 1) served
also as a probe for the existence of a radical in the
benzylic position just before splitting off amine. Indeed,
when dipeptide 1c was heated in t-BuOH under Stork’s
conditions, product 3b was produced along with 4c as a
mixture of two isomers (∼1:1). Moreover, dibenzylamine
was isolated (21%), which provided other evidence con-
cerning the proposed mechanism. The experiment with
1c did not constitute an unambiguous proof, because the
N-methyl group was not present in the substrate. How-
ever, this explanation seems to be reasonable.
A similar conversion was also applicable to proline
derivative 1h . A plausible mechanism for the cyclization
is shown in Scheme 4. A radical is generated on the
protecting group and translocated to the peptide moiety.
Hydrogen transfer from a methylene group competes
with hydrogen transfer from the R position (10 and 11).
It has been reported that radicals of type 10 were formed
preferentially to those of type 11 as a result of the severe
nonbonding interactions associated with planar confor-
mations of the radical 11.14 However, in an intramolecu-
lar process like this is it does not seem to be so obvious;
10 for steric reasons is reduced to 2h . The peptidyl radical
11 closes to 3h via spiro intermediate 12.
In conclusion, although the results here reported are
largely preliminary, the radical cascade provides an
accessible route starting from amino acids to a range of
nitrogen-containing heterocycles in good to moderate
yields when other methods fail.
Ack n ow led gm en t. This work was supported by
Institute of Organic Chemistry, Polish Academy of
Sciences.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures and spectral data for compounds 1-4 (a -h ). This
material is available free of charge via the Internet at
http://pubs.acs.org.
J O020602W
A possible explanation for the lower yields in the case
of amino acid derivatives other than glycine may be also
derived from consideration of rotamers of the dipeptides.
(15) (a) Stork, G.; Mah, R. Heterocycles 1989, 28, 723-727. (b)
Curran, D. P.; Tamine, J . J . Org. Chem. 1991, 56, 2746-2750. (c)
Curran, D. P.; Liu, W.; Chen, C. H.-T. J . Am. Chem. Soc. 1999, 121,
11012-11013. (d) Musa, M. O.; Horner, J . H.; Newcomb, M. J . Org.
Chem. 1999, 64, 1022-1025.
(14) Burgess, V. A.; Easton, C. J .; Hay, M. P. J . Am. Chem. Soc.
1989, 111, 1047-1052.
(16) Wilcox, C. S.; Gaudino, J . J . J . Am. Chem. Soc. 1986, 108, 3102-
3104.
1554 J . Org. Chem., Vol. 68, No. 4, 2003