.
Angewandte
Communications
(55.1 ppm) spectra of 22b. Finally, the reaction of only one
amide group is clearly confirmed by the significant high-field
shift (approx. 100 ppm, d 15N = À351) of one signal in the
15N NMR spectrum (1H,15N HMBC, 4-NCH3 correlates with
4-NCH3), which can be assigned to an amine moiety. The
position of the newly formed CH2 group within the cyclo-
1
sporine ring system was also confirmed by the H,1H COSY
correlations found for a -CH2-CH2- spin system. Such an atom
sequence is only possible in the case of MeN-CH2-CH2-
located in the 3-position (sarcosine) of this ring system. To
demonstrate the practicability of our catalytic protocol, the
reaction was performed on a 5 mmol scale (6.2 g of 22a)
giving a similar yield of isolated product compared to the
small-scale experiment (Scheme 5).
Figure 1. Selective reduction of the tertiary amide bond in peptides.
inverse gated decoupling (IG) NMR spectroscopy, and the
structures of the products were confirmed by one- and two-
dimensional 1H and 13C NMR spectroscopy as well as HRMS.
To show the selective transformation, representative HRMS
and NMR spectra of 20b and 21b are shown in Figures S1–S5.
Finally, we became interested in testing the generality of
our methodology for the selective reduction of pharmaceuti-
cally important cyclic peptides. For this purpose we adopted
the catalyst system to cyclosporine A, which constitutes one
of the most important immunosuppressants today.[24] Hence,
there exists significant interest in the synthesis of new
derivatives. More specifically, cyclosporine A is a cyclic
polypeptide of eleven amino acids containing four secondary
and seven tertiary amide bonds, which is produced conven-
iently on large scale by fermentation.[25–27] Taking the complex
structure of cyclosporine A into account, raises the question if
a selective reduction of a certain amide bond in this cyclic
polypeptide is still possible with our catalyst system. Due to
the challenging analytical setup of macropeptides, the product
formation was monitored by 1H NMR spectroscopy. First we
started with the above conditions (1.5 mol% [Rh(cod)2]BF4;
1.5 mol% dppp; 2 equiv PhSiH3; THF, rt, 46 h), but no
reactivity was observed. To our delight, at higher temperature
(808C) in toluene we detected a new product in the reaction
mixture indicating reduction of an amide function.
The HRMS spectrum of the crude product showed
a characteristic signal at m/z = 1210.85 for [M + Na]+, dem-
onstrating reduction of only one amide moiety. In agreement
with this observation the 1H NMR spectrum showed signals of
a new CH2-group in the range of 2.7 ppm. Due to isolation
problems, we performed similar reactions with the known
acetylated cyclosporine A 22a.[28] This compound is expected
to allow an easier purification due to increased interactions
with the stationary silica phase during chromatography.[28]
Having a reliable work-up strategy in hand, selectivity and
yield of the reduction were improved. Best results (30% yield
of isolated 22b) were obtained using an excess (20 equiv) of
PhSiH3 which was added in two portions. Notably, a further
increase of the amount of silane had no positive impact on the
chemoselectivity. Regarding the catalyst/ligand ratio, the best
reactivity was obtained using a 1:2 ratio. NMR-spectroscopic
investigations of the isolated product showed the reduction of
only one amide group of acetylated cyclosporine A. The
structure of compound 22a and the reaction product 22b were
confirmed by 1H, 13C, and 15N NMR spectroscopy. Thus,
eleven carbonyl signals were found for the amide groups in
22a, whereas for 22b only ten signals were observed.
Furthermore, signals for an additional CH2 group were
found in both the 1H (2.81 and 2.69 ppm) and 13C NMR
Scheme 5. Selective rhodium-catalyzed reduction of acetylated cyclo-
sporine A.
In conclusion, we have established for the first time
a highly selective reduction of amino acid esters and peptides
using rhodium phosphine catalysts and silane under mild
conditions. The optimized catalyst system tolerated a wide
variety of functional groups including secondary amides,
esters, nitrile, thiomethyl, and hydroxy as well as heterocycles,
for example, indole. The convenient procedure (no pressure
or air-sensitive reagents), the operational safety (no highly
reactive metal hydrides), and the mild conditions make this
new reduction protocol attractive. Based on this method-
ology, one can expect that catalytic hydrosilylations should be
able to modify various other peptides, too. By combining the
advantages of automated peptide synthesis or fermentation
processes with the here presented selective reduction, a multi-
tude of novel peptide derivatives for chemical biology studies
as well as potential pharmaceutical applications should be
available. Moreover, we expect our system to be applicable
for the derivatization of other interesting natural products
containing amides without using protecting and deprotecting
steps.[30]
Keywords: amide bond · amino acid · peptide modification ·
rhodium · selective reduction
How to cite: Angew. Chem. Int. Ed. 2015, 54, 12389–12393
Angew. Chem. 2015, 127, 12566–12570
[1] D. M. Nelson, M. M. Cox, Lehninger Principles of Biochemistry,
6th ed., W. H. Freeman, New York, 2012.
[2] P. Vlieghe, V. Lisowski, J. Martinez, M. Khrestchatisky, Drug
[3] a) M. Szelke, B. Leckie, D. M. Hallett, J. J. Sueiras, B. Atrash, F.
Acad. Sci. 1985, 300, 437; c) A. Wlodawer, J. W. Erickson, Annu.
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 12389 –12393