Diastereoselective hydrogenation of a tricyclic α,β -dehydrodipeptide

Article abstract of DOI:10.1016/S0021-9517(03)00038-1

An unsaturated diketopiperazine derivative with a tricyclic α, β-dehydrodipeptide structure was isolated as a reaction intermediate in the hydrogenation of pyrazine-2-(methyl-(S)-prolinecarboxamide). The diastereoselective hydrogenation of this dehydrodipeptide was studied using various noble metals (Pd, Pt, Rh, and Ru) supported on charcoal. The hydrogenation over Pd, Rh, and Ru catalysts proceeded with a high diastereoselectivity (71-79%), and the diastereomer with the (S)-configuration on both chiral carbon atoms was formed preferentially. The reaction rates and diastereoselectivities of the hydrogenation over the Pd, Rh, and Ru catalysts were similar, while the platinum catalyst was much less active and selective (48% d.e.). The obtained results were compared with those of the hydrogenation of pyrazine-2-(methyl-(S)-prolinecarboxamide); two different pathways for the hydrogenation of this molecule were suggested. In one path, cyclization already occurs after hydrogenation to the tetrahydropyrazine molecule, and in the other path cyclization occurs after full hydrogenation of the pyrazine molecule.

Full text of DOI:10.1016/S0021-9517(03)00038-1

Journal of Catalysis 217 (2003) 240–244
www.elsevier.com/locate/jcat
Research Note
Diastereoselective hydrogenation of a tricyclic α, β-dehydrodipeptide
∗
Pavel Kukula and Roel Prins
Laboratory for Technical Chemistry, Swiss Federal Institute of Technology (ETH-Hönggerberg), CH-8093 Zurich, Switzerland
Received 29 November 2002; accepted 17 January 2003
Abstract
An unsaturated diketopiperazine derivative with a tricyclic α,β-dehydrodipeptide structure was isolated as a reaction intermediate in the
hydrogenation of pyrazine-2-(methyl-(S)-prolinecarboxamide). The diastereoselective hydrogenation of this dehydrodipeptide was studied
using various noble metals (Pd, Pt, Rh, and Ru) supported on charcoal. The hydrogenation over Pd, Rh, and Ru catalysts proceeded with a high
diastereoselectivity (71–79%), and the diastereomer with the (S)-configuration on both chiral carbon atoms was formed preferentially. The
reaction rates and diastereoselectivities of the hydrogenation over the Pd, Rh, and Ru catalysts were similar, while the platinum catalyst was
much less active and selective (48% d.e.). The obtained results were compared with those of the hydrogenation of pyrazine-2-(methyl-(S)-
prolinecarboxamide); two different pathways for the hydrogenation of this molecule were suggested. In one path, cyclization already occurs
after hydrogenation to the tetrahydropyrazine molecule, and in the other path cyclization occurs after full hydrogenation of the pyrazine
molecule.

2003 Elsevier Science (USA). All rights reserved.
Keywords: Diastereoselective hydrogenation; Diketopiperazine; Proline auxiliary; Chiral cyclic dehydrodipeptide; Nonproteinogenic amino acids; Cyclic
dipeptide
1
. Introduction
intermediate was formed by an intramolecular N-acylation
of the partially hydrogenated substrate. The intramolecu-
lar N-acylation of various dipeptide esters, resulting in the
formation of a diketopiperazine ring, is a well-known reac-
tion [11–14]. High yields of intramolecular cyclization were
obtained under hydrogenation conditions over palladium on
charcoal [15,16].
The diastereoselective hydrogenationof unsaturated dike-
topiperazine derivatives, which correspond to the cyclic
α, β-dehydrodipeptides (Fig. 1c), was studied by several
authors [17–26] in order to develop a general method for
the stereoselective synthesis of α-amino acids. The chi-
ral information was usually introduced by means of a nat-
ural amino acid, which then plays the role of chiral auxil-
iary. (S)-Proline was used to advantage as a chiral auxiliary
Diketopiperazines corresponding to cyclic dipeptides rep-
resent a large group of compounds which exhibit a wide
range of biological activity [1–3]. They occur widely in na-
ture in the cis-configuration because they are biosynthesized
from proteinogenic L-α-amino acids. A large number of nat-
ural diketopiperazines contain a proline structure, and the
diversity of biological activity of this class of diketopiper-
azines is very impressive [4–9].
Our study of the diastereoselective hydrogenation of
pyrazine derivatives [10] revealed the stereoselective for-
mation of a new diketopiperazine derivative (Fig. 1a). This
molecule corresponds to a cyclic dipeptide and consists
of one proteinogenic (L-proline) and one nonproteinogenic
(
piperazine-2-carboxylic acid) amino acid. The structure of
(
Fig. 1d) by several authors [19,20,22,25,26]. The palladium
this molecule is described in detail in our previous pa-
per [10]. Moreover, we found that the hydrogenation of
pyrazine-2-(methyl-(S)-prolinecarboxamide) proceeds via
an unsaturated diketopiperazine intermediate, which has the
structure of a cyclic α, β-dehydrodipeptide (Fig. 1b). This
catalyst on charcoal was used most frequently and the ob-
tained diastereoselectivities varied from 20 to 90%. Deuter-
ation experiments of substituted diketopiperazine (R1 =
1
-methylethyl, R2 = phenyl, R3 = H) with different protect-
ing groups (R4 and R5) showed that the presence and type of
the protecting group strongly influence the stereoselectivity
of the reaction and that the diastereomeric excess can even
be reversed using a different protecting group [18].
*
Corresponding author.
E-mail address: prins@tech.chem.ethz.ch (R. Prins).
0021-9517/03/$ – see front matter  2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0021-9517(03)00038-1

P. Kukula, R. Prins / Journal of Catalysis 217 (2003) 240–244
241
2
.1. Synthesis of the pyrazine-2-(methyl-(S)-prolinecarbox-
amide)—α, β-dehydrodipeptide precursor
The description of the synthesis of pyrazine-2-(methyl-
(
S)-prolinecarboxamide) is given elsewhere [10]. There
was one difference; i.e., the precursor of the carboxamide,
the methyl ester of (S)-proline, was not purchased but
prepared by esterification with thionyl chloride in methanol.
Similarly, a racemic proline methyl ester was prepared.
Moreover, we carried out all further reaction steps using
the racemic proline auxiliary in order to prove that no
racemization takes place. The racemic samples were used
as chromatographic standards for chiral HPLC separation.
Fig. 1. Structures of diketopiperazine derivatives a–d.
2
.2. Preparation of tricyclic α, β-dehydrodipeptide—
The substrate for the hydrogenation
The intramolecular cyclization and the hydrogenation
of pyrazine-2-(methyl-(S)-prolinecarboxamide) were dis-
cussed simultaneously in our previous paper [10]. This re-
search note focuses on a separate study of the diastereos-
elective hydrogenation of the cyclic α, β-dehydrodipeptide
The tricyclic α, β-dehydrodipeptide was prepared by par-
tial hydrogenation of pyrazine-2-(methyl-(S)-prolinecarbo-
xamide). The reaction was carried out in a 60-ml stainless-
steel autoclave (Medimex AG) equipped with a sampling
tube and magnetic gas-inducing impeller. In a typical exper-
iment, 1 g of pyrazine-2-(methyl-(S)-prolinecarboxamide)
was dissolved in 30 ml methanol, and 350 mg of 10% Pd/C
(
Fig. 1b), which was isolated from the reaction mixture dur-
ing the hydrogenation of pyrazine-2-(methyl-(S)-prolinecar-
boxamide). The hydrogenation was carried out using vari-
ous heterogeneous noble metal catalysts, and the influence
of the nature of the catalyst on the reaction rate and the
diastereoselectivity of the hydrogenation was investigated.
The aim of the study was to determine the reaction rate
and the diastereoselectivity of the hydrogenation step and
to compare them with those of the simultaneous reaction.
These results should provide us with greater insight into the
mechanism of the hydrogenation of pyrazine-2-(methyl-(S)-
prolinecarboxamide).
◦
was added. The reaction was carried out at 50 C and 5 MPa
hydrogen pressure. The reaction was monitored by analyz-
ing the samples withdrawn from the reaction mixture dur-
ing hydrogenation. The reaction was stopped at about the
maximum concentration of the unsaturated tricyclic α, β-de-
hydrodipeptide intermediate. The catalyst was filtered off,
the solvent evaporated, and the raw reaction mixture sep-
arated by column chromatography (methanol:chloroform,
3:1). The isolated tricyclic α, β-dehydrodipeptide was puri-
fied by recrystallization from methanol, and its crystal struc-
ture was determined by X-ray diffraction analysis.
2
. Experimental
2
.3. Diastereoselective hydrogenation
All the organic chemicals were purchased from Fluka.
The diastereoselective hydrogenationof tricyclic α, β-de-
The same analytical instrumentation and methods were
used as in the study of the hydrogenation of pyrazine
derivatives [10]. The NMR, XRD, GC-MS, and elemen-
tal analysis data of the reaction components are published
ibidem [10]. The crystal structure of the prepared tri-
cyclic α, β-dehydrodipeptide was determined on a single
crystal 4-circle X-ray diffractometer (Picker) with an up-
grade from Stoe. The diastereoselectivity of hydrogena-
tion was determined by GC and the diastereomeric ex-
cess (d.e.) was defined as: d.e. = ([R, S-diastereomer] −
hydrodipeptide was carried out using the same noble metal
catalysts as in our previous work [10]. The same types of
Pd/C, Pt/C, Ru/C, Rh/C, and Rh black were used. To com-
pare the data of both studies, we used the same equipment,
procedures, and reaction conditions. The substrate-to-metal
molar ratio was also kept at the same level S/M ∼ 23.
3
. Results and discussion
[
S, S-diastereomer])/([R, S-diastereomer] + [S, S-diastere-
omer]) × 100. The enantiopurity of all the compounds and
the enantioselectivity of all the reaction steps were con-
firmed by a chiral HPLC (Agilent 1050 Series LC system,
Daicel chiral column Chiracel OD-H, mobile phase:hexane:
3.1. Preparation of the tricyclic α, β-dehydrodipeptide
The first step in the synthesis of the substrate to study
the diastereoselective hydrogenationwas the preparation of a
chiral auxiliary, the methylester of (S)-proline (1). A simple
esterification with thionyl chloride and methanol provided a
2
-propanol with a ratio from 90:10 to 70:30 used for the
analyses of different compounds).

242
P. Kukula, R. Prins / Journal of Catalysis 217 (2003) 240–244
Fig. 2. Reaction scheme of the synthesis of the cyclic α,β-dehydrodipeptide and its diastereoselective hydrogenation.
9
2
5% yield. The second step was the coupling of pyrazine-
-carboxylic acid (2) with the (S)-proline methyl ester
triazacyclopenta[b]naphthalene-4,9-dione (5 and 6) (Fig. 2).
One of the diastereomers prevails in the reaction mixture.
This diastereomer was isolated and its configuration on
the chiral carbon atoms C-4a and C-9a was assigned to
(4aS, 9aS) in our previous work [10]. Furthermore, we car-
ried out the chiral separation of the final diastereomers,
which confirmed that no racemization of the proline moiety
took place during the diastereoselective hydrogenation.
Table 1 presents the effect of different noble metal cata-
lysts (Pd, Rh, Pt, and Ru) on the hydrogenation of the dehy-
drodipeptide (4). The experiments were carried out under the
same reaction conditions as during the hydrogenation of the
pyrazine derivative (3) in order to facilitate a comparison of
the results of both studies. The activities of the catalysts are
reported as integral turnover frequency at 50% conversion of
the substrate. It is important to remember that, while the re-
action of the pyrazine derivative is a complex reaction and
includes hydrogenation as well as ring formation, the hydro-
genation of the dehydrodipeptide (4) is a simple one-step re-
action. The reaction time, defined as the time at which 100%
conversion of (4) was reached, is also presented. The main
by-products consist of two saturated diastereomers with the
same structure as the product but methylated at the sec-
ondary nitrogen atom (N-6). Their ratio corresponds roughly
to the ratio of the diastereomers (5 and 6), and they were
formed by the reaction of the products with the solvent [10].
(
(
1) to form pyrazine-2-(methyl-(S)-prolinecarboxamide) (3)
Fig. 2). This reaction also gave a very high yield of 85%.
The products of the esterification as well as of the cou-
pling were subjected to a chiral HPLC. The comparison with
their corresponding racemic mixtures proved that both re-
actions proceeded without racemization and that the enan-
tiopurity of (1) and (3) was higher than 99%. The tricyclic
α, β-dehydrodipeptide (4) was then obtained by partial hy-
drogenation of (3) followed by intramolecular cyclization.
At the same time, a methanol molecule was formed as the
product of splitting the ester group of the proline auxiliary.
The intramolecular cyclization was probably fast, because
we did not observe the formation of a partially hydrogenated
substrate or of a noncyclic saturated product. The main by-
product of this reaction had a similar mass spectrum as the
cyclic dehydrodipeptide (4), but the molecular weight was
lower by two masses. Therefore, we assume that it has the
same tricyclic dehydrodipeptide structure (4) but with an ad-
ditional double bond, between N(6)–C(7) or C(7)–C(8), in
the tetrahydropyrazine ring. The concentration of this by-
product in the reaction mixture was below 10%. The reaction
was carried out under milder reaction conditions than the di-
astereoselective hydrogenation of (1) to (5) and (6) in order
to stop the reaction close to the maximum concentration of
the dehydrodipeptide (4). After the separation and recrystal-
1
3
1
lization of (4), its structure was confirmed by C and H
NMR measurements. The chiral HPLC proved that this re-
action, too, proceeded without racemization. Moreover, an
X-ray diffraction analysis of a prepared single crystal was
carried out. The structure is almost flat and very similar to
the structure of the final hydrogenation product [10].
Table 1
◦
Hydrogenation of α,β-dihydrodipeptide (4) in methanol at 80 C and
8
MPa over different noble metal catalysts
Catalyst
TOF at 50%
conversion (h
Reaction time
(h)
By-products
(%)
d.e.
(%)
−
1
)
Pd/C
Pd/C
Rh/C
Pt/C
Ru/C
Rh-black
36.0
41.6
43.4
11.7
39.4
16.3
1.1
0.8
0.7
7.8
0.8
2.8
<3
<1
<2
13
<2
<5
79
83
77
48
71
75
a
3
.2. Diastereoselective hydrogenation
The diastereoselective hydrogenation of the α, β-dehy-
drodipeptide (4) proceeds to a mixture of two saturated di-
astereomers, (4aS, 9aS)- and (4aR, 9aS)-octahydro-3a,6,8a-
a
Water as solvent.

P. Kukula, R. Prins / Journal of Catalysis 217 (2003) 240–244
243
Table 1 shows that the activity and the diastereoselectiv-
ity of Pd/C, Rh/C, and Ru/C were similar during the hy-
drogenation of the dehydrodipeptide (4). The highest TOF
as well as the shortest reaction time were obtained with
the rhodium catalyst; diastereoselectivity decreased slightly
in the following order: Pd/C > Rh/C > Ru/C. The lowest
activity as well as the lowest diatereoselectivity were ob-
tained with the platinum catalyst. The highest content of
by-products was also observed with the platinum. The dif-
ference between the rhodium catalyst supported on charcoal
and unsupported rhodium black consisted mainly of a lower
reaction rate obtained with unsupported rhodium. The con-
tent of by-products increased slightly and the diastereoselec-
tivity remained almost the same with the rhodium black.
The diastereoselectivity obtained during the hydrogena-
tion of the dehydrodipeptide (4) was significantly higher
over all the catalysts than that obtained during the hydro-
genation of the pyrazine derivative (3). This is probably be-
cause the reaction does not proceed 100% via the dehy-
drodipeptide intermediate (4). The pyrazine derivative may
be hydrogenated first to the tetrahydropyrazine derivative.
Part of this will undergo cyclization to (4), while the other
part will be fully hydrogenated before cyclization takes
place. The stereoselectivity of the C(4)–C(4a) bond will de-
pend on the conformational freedom of the intermediate that
precedes the final hydrogenation step. The stereoselectivity
will be higher when the final hydrogenation occurs in the
rigid dehydrodipeptide (4) structure. This dehydrodipeptide
will preferentially adsorb with one of its two diastereotopic
faces on the catalyst and, thus, the formation of only one di-
astereomer is favored. As a consequence, the final diastere-
oselectivity in the hydrogenation of the pyrazine derivative
imately four times lower than the TOF during the hydrogena-
tion of the pyrazine derivative; the reaction time was almost
double.
In contrast, the TOFs of the hydrogenation of the di-
hydrodipeptide over Ru/C and Rh black were higher than
during the hydrogenation of the pyrazine derivative. For
these catalysts, the hydrogenation of the pyrazine derivative
seems to be inhibited by the substrate itself or by partially
hydrogenated intermediates. A possible explanation is that
the cyclization reaction is slower with these catalysts and
that the hydrogenation of a noncyclic partially hydrogenated
pyrazine derivative is more difficult to achieve than the hy-
drogenation of the dehydrodipeptide.
In the previous study we found that water as a solvent has
a beneficial influence on the reaction rate as well as on the
diastereoselectivity. Therefore, we tested water as a solvent
also during the hydrogenation of the dehydrodipeptide over
the palladium catalyst. The TOF of the reaction increased
slightly, and the reaction time was shorter when water was
used as a solvent (Table 1). The diastereoselectivity was
also somewhat higher than with methanol and almost no
by-products were formed. The slight difference in TOF
compared with the large change during the hydrogenation
of the pyrazine derivative indicates that the solvent mainly
influences the cyclization reaction and that its influence on
the hydrogenation is much weaker.
4
. Conclusion
The study of the diastereoselective hydrogenation of the
cyclic α, β-dehydrodipeptide showed that high diastereose-
lectivity can be obtained over noble metal catalysts. Pd/C,
Rh/C, and Ru/C were the most active as well as the most
selective catalysts for this reaction. The activity of Pt/C and
Rh black catalysts was much lower. The lowest d.e. as well
as the largest amount of by-products were obtained with the
Pt/C catalyst. This corresponds to our former results of the
hydrogenation of the pyrazine derivative. The reaction rates
and the diastereoselectivities obtained during the hydrogena-
tion of the pyrazine derivative over Pd/C were considerably
higher than over the other catalysts. On the other hand, sim-
ilar reaction rates and diastereoselectivities were obtained
with Pd/C, Rh/C, and Ru/C during the hydrogenation of
dehydrodipeptide. The main reason for this might be that the
cyclization reaction on palladium was much faster than on
the other catalysts and that a greater molecular rigidity of the
dehydrodipeptide facilitates the differentiation of the two di-
astereotopic faces on the metal surface of the catalyst and
thus increases diastereoselectivity.
(
3) will depend on the rate of further hydrogenation ver-
sus the rate of cyclization of the tetrahydropyrazine deriv-
ative. This is also supported by our former results [10],
which show that there is a relationship between the final
diastereoselectivity and the maximum concentration of the
dehydrodipeptide intermediate. Moreover, the diastereose-
lectivity was increasing at the beginning of the hydrogena-
tion of the pyrazine derivative [10]. This may also indicate
that the pyrazine substrate (3) was fully hydrogenated be-
fore cyclization could occur. On the other hand, the diastere-
oselectivity remained constant during the hydrogenation of
the dehydrodipeptide (4). Therefore, the explanation that an
increased d.e. results from the modification of the catalyst
surface with the chiral reaction intermediates or products,
which are strongly adsorbed on the surface and influence the
mode of adsorption of the substrate, seems unlikely.
A comparison of the TOF data obtained during the hydro-
genation of the pyrazine derivative [10] with the TOF data of
the hydrogenation of the cyclic α, β-dehydrodipeptide over
Pd/C, Rh/C, and Pt/C (Table 1) shows that the latter TOF
data are lower. This was expected, because the hydrogena-
tion of a double bond conjugated with a carbonyl group is
more difficult than the partial hydrogenation of a pyrazine
ring [27]. The lowest TOF, obtained with Pt/C, was approx-
A comparison of the results of both studies suggests
that the hydrogenation of the pyrazine derivative only
partly proceeds via the dehydrodipeptide and that the ratio
between the reaction rates of cyclization and hydrogenation
influences the final diastereoselectivity.

244
P. Kukula, R. Prins / Journal of Catalysis 217 (2003) 240–244
Acknowledgments
[8] C. Takahashi, A. Numata, I. Ito, E. Matsumura, H. Araki, H. Iwaki, K.
Kushida, J. Chem. Soc., Perkin Trans. 1 (1994) 1859.
9] S.D. Bull, S.G. Davies, R.M. Parkin, F. Sanchez-Sancho, J. Chem.
Soc., Perkin Trans. 1 (1998) 2313.
[
The authors thank Prof. Volker Gramlich and Mauro
Perseghini for performing the XRD and chiral HPLC mea-
surements, respectively.
[
10] P. Kukula, R. Prins, J. Catal. 208 (2002) 404.
[11] G.G. Smith, R.C. Evans, R. Baum, J. Am. Chem. Soc. 108 (1986)
327.
7
[
[
[
[
12] D.M. Harrison, R.B. Sharma, Tetrahedron 49 (1993) 3165.
13] J. Savrda, L. Chertanova, M. Wakselman, Tetrahedron 50 (1994) 5309.
14] H. Tomiyasu, S. Kimura, Y. Imanishi, Helv. Chim. Acta 65 (1982) 775.
15] M.M. Lenman, S.L. Ingham, D. Gani, J. Chem. Soc., Chem. Commun.
(1996) 85.
Appendix
Supplementary crystallographic data are available at the
Cambridge Crystallographic Data Centre (CCDC) (referen-
ce number CCDC 195739). These data can be obtained free
of charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or
from the CCDC, 12 Union Road, Cambridge CB2 1EZ, UK;
fax: +44 1223 336033; E-mail: deposit@ccdc.cam.ac.uk).
[16] T.A. Gudasheva, N.I. Vasilevich, R.U. Ostrovskaya, S.S. Trofimov,
T.A. Voronina, Pharm. Chem. J. (Engl. transl.) 30 (1996) 562.
[
[
[
[
[
17] M. Oba, S. Nakajima, K. Nishiyama, J. Chem. Soc., Chem. Commun.
(1996) 1875.
18] M. Oba, T. Terauchi, Y. Owari, Y. Imai, I. Motoyama, K. Nishiyama,
J. Chem. Soc., Perkin Trans. 1 (1998) 1275.
19] T. Kanmera, S. Lee, H. Aoyagi, N. Izumiya, Tetrahedron Lett. 46
(1979) 4483.
20] H. Aoyagi, F. Horike, A. Nakagawa, S. Yokote, N. Park, Y. Hashimoto,
T. Kato, N. Izumiya, Bull. Chem. Soc. Jpn. 59 (1986) 323.
21] N.G. Park, S. Lee, H. Maeda, H. Aoyagi, T. Kato, Bull. Chem. Soc.
Jpn. 62 (1989) 2315.
References
[1] C. Prasad, Peptides 16 (1995) 151.
[
2] J.P. Milne, A.L. Hunt, K. Rostoll, J.J. Van der Walt, C.J. Graz,
J. Pharm. Pharmacol. 50 (1998) 1331.
[22] B.W. Bycroft, G.R. Lee, J. Chem. Soc., Chem. Commun. (1975) 988.
[23] K. Bergmann, R. Tietzman, J. Biol. Chem. 155 (1944) 535.
[24] H. Poisel, U. Schmidt, Chem. Ber. 106 (1973) 3408.
[25] S.D. Bull, S.G. Davies, S.W. Epstein, M.A. Leech, J.V.A. Ouzman,
J. Chem. Soc., Perkin Trans. 1 (1998) 2321.
[26] S.D. Bull, S.G. Davies, S.W. Epstein, J.V.A. Ouzman, J. Chem. Soc.,
Chem. Commun. (1998) 659.
[27] R.N. Landau, U. Singh, F. Gortsema, Y. Sun, S.C. Gomolka, T. Lam,
M. Futran, D.G. Blackmond, J. Catal. 175 (1995) 201.
[3] C.J.M. Graz, G.D. Grant, S.C. Brauns, A. Hunt, H. Jamie, J.P. Milne,
J. Pharm. Pharmacol. 52 (2000) 75.
[4] C. Cui, H. Kakeya, H. Osada, Tetrahedron 52 (1996) 12651.
[5] H. Hayashi, S. Murao, M. Arai, Chem. Express 6 (1991) 989.
[6] G. Deffieux, M.-J. Filleau, R. Baute, J. Antibiot. 31 (1978) 1106.
[7] P. Waring, R.D. Eichner, A. Mullbacher, Sjaarda, J. Biol. Chem. 263
(1988) 18493.