Diastereoselective hydrogenation of pyrazine derivatives: An alternative method of preparing piperazine-(2S)-carboxylic acid

Article abstract of DOI:10.1006/jcat.2002.3599

The diastereoselective hydrogenation of a chiral pyrazine derivative was used for the stereoselective preparation of piperazine-2-carboxylic acid, which is an important chiral building block. The study was focused on the diastereoselective hydrogenation with various noble metal catalysts (Pd, Pt, Rh, Ru) on different supports. It was found that intramolecular cyclization of the substrate takes place during the hydrogenation, forming an unsaturated diketopiperazine derivative. This intermediate was further hydrogenated to a mixture of saturated heterocyclic diastereomers. The influence of the reaction conditions (temperature, pressure of hydrogen, and type of solvent) on the diastereoselectivity was also studied. The highest diastereoselectivity (79%) was reached with 10% Pd/C and with water as solvent. The desired molecule of piperazine-2-carboxylic acid was finally obtained by acidic hydrolysis of the diastereomeric diketopiperazine adduct.

Full text of DOI:10.1006/jcat.2002.3599

Journal of Catalysis 208, 404–411 (2002)
doi:10.1006/jcat.2002.3599
Diastereoselective Hydrogenation of Pyrazine Derivatives: An Alternative
Method of Preparing Piperazine-(2S)-Carboxylic Acid
Pavel Kukula and Roel Prins1
Laboratory for Technical Chemistry, Swiss Federal Institute of Technology (ETH), CH-8093 Zurich, Switzerland
Received November 16, 2001; revised February 26, 2002; accepted February 26, 2002
thesis of chiral N-heterocycles is of interest, because they
The diastereoselective hydrogenation of a chiral pyrazine deriva- are widely used in the life-science industry. During the last
tive was used for the stereoselective preparation of piperazine-2- decade, nonproteinogenic amino acids and their derivatives
carboxylic acid, which is an important chiral building block. The
study was focused on the diastereoselective hydrogenation with
various noble metal catalysts (Pd, Pt, Rh, Ru) on different supp-
orts. It was found that intramolecular cyclization of the substrate
received particular attention (9). One of the most important
molecules is piperazine-(2S)-carboxylic acid because it is
widely used as a chiral building block in a number of bioac-
tive compounds, e.g., Merck’s HIV protease inhibitor Indi-
takes place during the hydrogenation, forming an unsaturated
navir (Crixivan) (10), an N-methyl- -aspartate antagonist
D
diketopiperazine derivative. This intermediate was further hydro-
genated to a mixture of saturated heterocyclic diastereomers. The
influenceofthereactionconditions(temperature, pressureofhydro-
gen, and type of solvent) on the diastereoselectivity was also studied.
The highest diastereoselectivity (79%) was reached with 10% Pd/C
(
11), various matrix-metalloproteinase inhibitors (12–14),
and cardioprotective nucleoside transport blockers (15).
Much work has, therefore, been devoted to the preparation
of the enantiomerically pure piperazine-(2S)-carboxylic
and with water as solvent. The desired molecule of piperazine-2- acid (16–20) and its derivatives (21–29) (mainly N-tert-
carboxylic acid was finally obtained by acidic hydrolysis of the di- butyl-2-piperazinecarboxamide—the Indinavir precursor).
astereomeric diketopiperazine adduct.
ꢀc 2002 Elsevier Science (USA)
Current procedures involve the classical resolution of the
racemate by fractional crystallization of diastereomeric
salts (16, 23, 28, 29), the asymmetric synthesis and homoge-
neouscatalytichydrogenationwithrhodiumcomplexes(18,
Key Words: diastereoselective hydrogenation; pyrazine; proline
auxiliary; optically active piperazine 2-carboxylic acid; nonpro-
teinogenic amino acids.
24–27), and recently developed biocatalytic processes us-
ing the kinetic resolution of the piperazine-2-carboxamide
racemate by amidases (17, 19, 20).
INTRODUCTION
We report a new alternative for preparing piperazine-
2S)-carboxylic acid by diastereoselective hydrogenation of
The synthesis of optically pure substances is becoming
increasingly important, and it is an essential part of many
chemical procedures, especially in the production of phar-
maceuticals, agrochemicals, and fragrance and flavor sub-
stances. Several different approaches are applied in the pro-
duction of optically pure chiral substances (1), of which
one is diastereoselective hydrogenation (2), in which a co-
valent bond between the prochiral substrate and the chiral
molecule (auxiliary) is formed. Thus, the reactant itself rep-
resentsthesourceofchiralityandinducestheformationofa
new stereogenic center. Several reviews (2, 3) have already
been published on this topic, and various applications have
been described (4–8).
(
chirally modified pyrazine-2-carboxylic acid. Up to now, the
hydrogenationofasubstitutedpyrazineringhasbeeninves-
tigated by several authors (30–37) using noble metal cata-
lysts such as palladium, platinum, and rhodium supported
on charcoal. In these cases, hydrogenation does not proceed
easily because the metal catalyst is poisoned by secondary
or tertiary nitrogen atoms in the product molecules. There-
fore, more severe reaction conditions and longer reaction
times are needed to complete the hydrogenation. The reac-
tions are sometimes carried out in acidic media to prevent
poisoning of the catalyst (33, 34). Palladium on charcoal
is the most widely used catalyst for the hydrogenation of
various pyrazine derivatives (30–32, 35–37). For example,
if a large amount of catalyst is used, pyrazine-2-carboxylic
acid can be hydrogenated to piperazine-2-carboxylic acid
in the form of the sodium or potassium salt in aqueous
A large number of biologically active molecules contain
chiral heterocyclic components, which can be prepared by
stereoselective catalytic hydrogenation of the correspond-
ing substituted aromatic compounds. Stereoselective syn-
◦
1
medium already at 50 C and at atmospheric pressure of
To whom correspondence should be addressed. Fax: 41-1-6321162.
E-mail: prins@tech.chem.ethz.ch.
◦
hydrogen (32). More stringent reaction conditions (85 C
404
0
ꢀ
021-9517/02 $35.00
c 2002 Elsevier Science (USA)
All rights reserved.

DIASTEREOSELECTIVE HYDROGENATION OF PYRAZINE DERIVATIVES
405
and 2 MPa) were required during the hydrogenation of tal structure of the hydrogenation product was determined
pyrazine-2-carboxylic acid amide (17). on a single-crystal four-circle X-ray diffractometer Syntex
The asymmetric hydrogenation of substituted pyrazine P21. The diastereoselectivity of hydrogenation was deter-
derivatives has been studied using homogeneous cat- mined by GC and the diastereomeric excess (d.e.) was defi-
alysts. The stereoselective hydrogenation of 1,4,5,6- nedas:d.e. = |([R,S-diastereomer] − [S,S-diastereomer])|/
tetrahydropyrazine-2-carboxylic acid derivatives, with pro- ([R,S-diastereomer] + [S,S-diastereomer]) × 100. The en-
tected nitrogens and a substituted carboxylic group, to the antioselectivity of hydrolysis and the optical purity of
corresponding piperazine derivatives using chiral homo- piperazine-2-carboxylic acid were determined by polarime-
geneous rhodium complexes was described (24–27). The try (Perkin–Elmer 241) at 589 nm and by chiral HPLC
best results (96% yield, 99% e.e.) were obtained with the (Agilent 1100 Series LC system, Macherey–Nagel column
[(R)-BINAP(COD)Rh]TfO catalyst (26). The diastereose- Nucleosil 120-3 C18, mobile phase: water : methanol (3 : 1)
lective hydrogenation of pyrazine derivatives has not been with 2 mM N,N-dimethyl-
attemptedyet. Tunglerandcoworkers(38, 39)madethefirst CuSO4 · 5H2O).
attempts at hydrogenating aromatic N-containing hetero-
L
-phenylalanine and 1 mM
cycles diastereoselectively. They carried out the hydrogena-
Synthesis of the Pyrazine-2-(methyl-(S)-
prolinecarboxamide) Substrate
tion of pyridine and pyrrole derivatives using noble metal
catalysts (Pd, Rh, Ru, and Pt) and (S)-proline methylester
as the chiral auxiliary. Some of the obtained diastereoselec-
tivities were high, e.g., the hydrogenation of N-picolinoyl-
Pyrazine-2-(methyl-(S)-prolinecarboxamide) was pre-
pared by coupling the carboxylic group of pyrazine-2-
carboxylic acid with the amino group of the methyl es-
ter of (S)-proline hydrochloride according to the following
procedure: Pyrazine-2-carboxylic acid (3.75 g, 30 mmol,
(
S)-proline methylester over Pd/C resulted in 79% d.e. and
the hydrogenation of N-nicotinoyl-(S)-proline methylester
in 94% d.e. Because of these promising results, we investi-
gated the stereoselective formation of a new chiral center
during the hydrogenation of pyrazine derivatives using the
same chiral auxiliary.
1
1
0
3
eq.) and 1-hydroxybenzotriazol, HOBT (4.85 g, 36 mmol,
.2 eq.), were dissolved in 60 ml of chloroform and cooled to
C. The hydrochloride of the (S)-proline methylester (5 g,
◦
0 mmol, 1 eq.), triethylamine (4.2 ml, 30 mmol, 1 eq.), and
This paper describes the diastereoselective hydrogena-
tion of chirally modified pyrazine-2-carboxylic acid. The
aim of the study was to prepare piperazine-(2S)-carboxylic
acid stereoselectively. Pyrazine-2-carboxylic acid, which is
commercially available in large amounts, was the starting
material. The source of chiral information was introduced
by means of a very common compound, the natural amino
acid (S)-proline. The hydrogenation of pyrazine, modified
by the proline auxiliary, was carried out using various het-
erogeneous noble metal catalysts; the influence of the re-
action conditions on diastereoselectivity of hydrogenation
was investigated. Finally, the procedure for the deprotec-
tion of the chiral auxiliary from the product molecule was
tested.
chloroform (40 ml) were added at regular intervals to the
ꢁ
cooled, stirred solution. N,N -Dicyclohexylcarbodiimide,
DCC (6.2 g, 30 mmol, 1 eq.), was dissolved in 20 ml of chlo-
roform and slowly added to the mixture. The cooling bath
was removed and the mixture was stirred overnight. The
formed dicyclohexylurea was filtered off, chloroform was
evaporated, and ethylacetate was added to the residue. Af-
ter the remaining dicyclohexylurea was filtered off, the or-
ganic phase was washed with a diluted HCl–water solution
and an aqueous NaHCO3 solution and dried over MgSO4.
The ethylacetate was evaporated to dryness in vacuo and
6.1 g of crystalline, slightly yellow pyrazine-2-(methyl-(S)-
prolinecarboxamide) was obtained (yield 85%).
Diastereoselective Hydrogenation
EXPERIMENTAL
The diastereoselective hydrogenation of pyrazine-2-
All the organic chemicals, with the exception of racemic (methyl-(S)-prolinecarboxamide) was carried out using
piperazine-2-carboxylic acid (Aldrich), were purchased various noble metal catalysts on different supports: 10 wt%
from Fluka. The organic solvents were used without further Pd/C (Fluka, Cat. No. 75990), 5 wt% Pt/C (Fluka, Cat. No.
purification. The reaction course of hydrogenation was 80982), 5 wt% Ru/C (Fluka, Cat. No. 84031), 5 wt% Rh/C
monitored by gas chromatography using an HP 5890 Series (Fluka, Cat. No. 83711), 5 wt% Rh/Al2O3 (Fluka, Cat. No.
II Plus instrument equipped with an HP-1 capillary column. 83720), and Rh black, 99 wt% (Johnson Matthey, Cat. No.
The intermediates and products of the reactions were an- 231-125-0). All catalysts were used directly as supplied. Ex-
alyzed by GC–MS (HP 5973 MSD). After the separation periments were carried out in a 60-ml stainless steel auto-
and isolation of the pure compounds, elementary analysis clave (Medimex AG) equipped with a sampling tube and
1
(
automatic analyzer Perkin–Elmer CHN 240C) and H and magnetic gas-inducing impeller. In a typical experiment,
1
3
C NMR spectroscopy (DPX 300 and AVANCE 500 from 50 mg of pyrazine-2-(methyl-(S)-prolinecarboxamide) was
Bruker AG) werecarriedoutfor characterization. Thecrys- dissolved in 30 ml of solvent and 10–40 mg of the catalyst

4
06
KUKULA AND PRINS
was added. The substrate-to-metal molar ratio was kept (15, 17, 28) for the pure dihydrochloride of (S)-piperazine-
constant at S/M ∼ 23. A rather large amount of catalyst with 2-carboxylic acid, but it was not possible to estimate the
respect to substrate was required in order to obtain a rea- exact enantiomeric ratio, since the published data differ
sonable reaction time because of the poisoning effect of slightly. Therefore, the acid was prepared from its hy-
the nitrogen atom in the substrate molecule. Nevertheless, drochloride, and the optical purity of the acid was measured
all the reactions were conducted in the kinetic regime, be- by HPLC. A solution of piperazine-(2)-carboxylic acid di-
cause the reaction rate was proportional to the mass of the hydrochloride in 50 ml of water was passed over a cation ex-
catalyst and not to the stirring speed. The slurry was then change resin (Amberlite IR 120, 20 ml). After neutral wash-
transferred to the autoclave and flushed three times with ing of the exchanger with water, the resin was washed with
hydrogen. The autoclave was pressurized with hydrogen to diluted aqueous ammonia and water. The combined elu-
7
0 bar; and stirring (1000 rpm) and heating to the appro- ates, containing free piperazine-(2)-carboxylic acid (tested
priate temperature were started. Samples of the reaction by TLC and detected with ninhydrin solution), were evapo-
mixture were withdrawn during the reaction and analyzed rated to dryness. The residual solid was recrystallized from
by gas chromatography. After the hydrogenation was com- a water–ethanol mixture, and the crystals were collected by
◦
pleted, the catalyst was filtered off, the solvent was evapo- filtration, and dried at 60 C. The yield was 57 mg (96%) of
rated, and the final product was dried under high vacuum. pure piperazine-(2)-carboxylic acid. The optical purity of
The product was characterized by GC–MS and NMR spec- the acid was measured by HPLC (see above).
troscopy. The yields of the hydrogenation reactions were
approximately 95%.
RESULTS AND DISCUSSION
Hydrolysis of Hydrogenation Product
Preparation and Characterization of the Substrate
The product of total hydrogenation (diketopiperazine
The reaction scheme of the stereoselective prepara-
derivative, 100 mg, 0.48 mmol) was refluxed in 6 M HCl tion of piperazine-(2S)-carboxylic acid (1) is shown in
10 ml) overnight. A white precipitate formed, was fil- Fig. 1. The first step of the synthesis was the coupling of
(
tered off, washed with methanol, and dried under high vac- pyrazine-2-carboxylic acid (2) with the chiral auxiliary (S)-
uum. Organic elemental analysis and NMR spectroscopy proline methylester (3) to form pyrazine-2-(methyl-(S)-
proved that it consisted of the pure dihydrochloride of prolinecarboxamide) (4). This reaction gave a very high
1
3
1
piperazine-(2)-carboxylic acid (93 mg, 96% yield). In or- yield of 85%. The C and H NMR measurements showed
der to check its optical purity the specific optical rota- that at least two conformers of an amide (4) were present.
tion was measured ([α]2 = 5.45 , c = 1.2, H2O). This All the signals of the carbon and hydrogen atoms in the
0
◦
D
1
3
1
value corresponds to the data mentioned in the literature
C and H NMR spectra were double and the ratio of the
O
COOH
+
-
H , Pd/C
HN
DCC
2
N
N
N
+
HOBT
N
N
MeOH
COOCH3
COOCH3
2
3
4
O
O
O
5
6
4a
3a
4
9
3
2
HN
N
HN
N
HN
N
+
H , Pd/C
2
+
+
7
N
N
N
8
a
9
8
1
O
O
O
5
(4aS, 9aS)- 6
(4aR, 9aS)- 7
O
H
N
H+
HN
N
N
N
COOH
H O
2
N
H
COOH
H
O
(
4aS, 9aS)- 6
1
FIG. 1. Reaction scheme of the preparation of piperazine-(2S)-carboxylic acid.

DIASTEREOSELECTIVE HYDROGENATION OF PYRAZINE DERIVATIVES
407
signals for the same hydrogen and carbon atoms was ap-
1
proximately 50 : 50. For instance, the H NMR spectrum
of the substrate contains two doublets of doublets at 4.71
and 5.13 ppm, which were assigned to the proton on the
chiral carbon atom of the proline auxiliary. The ratio of
the signals was 49 : 51. In order to determine the structure
of the conformers, molecular mechanics calculations were
performed using the Hyperchem 5.0 software. Two struc-
tures with minimal energy were found. They differ in the
orientation of the proline auxiliary, in particular the es-
ter group, with respect to the aromatic pyrazine ring. In
one of the conformers the ester group of the proline aux-
iliary is situated near the pyrazine ring, while in the other
conformer it is facing away from the pyrazine ring. This is
FIG. 2. ORTEP drawing of the hydrogenation product (6).
◦
caused by a rotation of the amide bond by 180 . The inter-
conversion between the conformers takes place at higher 3a,6,8a-triaza-cyclopenta[b]naphthalene-4,9-dione (6 and
1
temperature, as was observed from H NMR spectra mea- 7). The ratio of the diastereomers obtained by GC was ap-
◦
13
1
sured in DMSO between 30 and 130 C. The signals of the proximately 8 : 2. The C and H NMR spectra also showed
protons on the pyrazine ring and the signals of the proton two sets of signals at roughly the same ratio. The diastere-
◦
at the chiral carbon atom started to merge at 100 C and at omers were purified by crystallization from toluene; one
1
◦
30 C there were only single signals for these protons in the of them crystallizes preferentially from the toluene solu-
spectrum.
tion. Thus, the first of the formed crystals, which contained
only one diastereomer, was collected and subjected to a
structural analysis by NMR and X-ray diffraction (see Ap-
pendix). Figure 2 shows the ORTEP drawing of the crystal
Diastereoselective Hydrogenation
The diastereoselective hydrogenation of pyrazine-2- structure of this diastereomer obtained by XRD. It is ap-
(
methyl-(S)-prolinecarboxamide) (4) was initially studied parent that the hydrogens at both chiral centers (C-4a and
using the palladium catalyst and methanol as the solvent. C-9a) are oriented in the same direction, toward the plane
During these preliminary experiments it was found that the of projection, and that the configuration on both chiral cen-
1
13
reaction proceeds via the partly hydrogenated intermedi- ters is (4aS, 9aS). H and C NMR experiments confirmed
1
ate (5) to a mixture of saturated diastereomers (6 and 7) the presence of only one diastereomer, but additional H–
1
(
Fig. 1). The intermediate was probably formed by the in-
H COSY, NOESY, and H,C–COSY experiments had to
tramolecular N-acylation of the partly hydrogenated sub- be carried out in order to resolve the spectra and prove the
strate. At the same time, a methanol molecule was formed structure of the favored diastereomer. The NOESY exper-
as the product of splitting the ester group of the proline iment should confirm the configuration on the C-4a carbon
auxiliary. The intramolecular cyclization was probably very by the expected interaction between two protons on the chi-
fast, since we did not observe the formation of a partly hy- ralcarbonatomsC-4aandC-9a. Thedistancebetweenthese
drogenated substrate or of a noncyclic saturated product. protons, as obtained from the crystal structure, was only ap-
In order to characterize the intermediate, the reaction was proximately 0.3 nm. Nevertheless, no NOE interaction for
stopped at close to maximum concentration, and the inter- these two protons was found. The resolved NMR spectra
mediate was separated from the reaction mixture by col- with assigned H and C atoms are given in the Appendix.
umn chromatography (3 MeOH : 1 CHCl3). The structure
After the preliminary experiments with the palladium
13
1
of the intermediate was confirmed by C and H NMR catalyst, the effect of other noble metal catalysts (Pt, Rh,
measurements. As can be seen from the 13C NMR data and Ru) supported on different supports was studied. The
(
see Appendix), there were two tertiary and three quater- experiments were carried out under the same reaction con-
nary carbon atoms. Two of the quaternary carbons belong ditions, and the results are summarized in Table 1. The
to the carbonyl groups, while the third must belong to the activities of the catalysts are reported as integral turnover
unsaturated carbon atom C-4a at the diketopiperazine ring, frequency at 50% conversion of the substrate. Because the
since the signal for this carbon atom was not present in the reaction of the substrate is complex and includes hydro-
DEPT spectrum. Moreover, the presence of only two ter- genation as well as ring formation we also present the re-
tiary carbons and their chemical shifts indicate the position action time, defined as time at which 100% conversion of
of a double bond in the tetrahydropyrazine ring of the in- the intermediate (5) was reached. The main by-products
termediate (5) (Fig. 1).
found in the reaction mixture were identified by GC–MS
The main products of the total hydrogenation were as two saturated diastereomers with the same structure
two diastereomers: (4aS, 9aS)- and (4aR, 9aS)-octahydro- as the product but methylated at the secondary nitrogen

4
08
KUKULA AND PRINS
TABLE 1
Comparison of Different Noble Metal Catalysts during the
modified by the chiral reaction intermediates or products.
These molecules are strongly adsorbed on the surface with
their nitrogen atoms and can thus influence the mode of
adsorption of the substrate and the diastereoselectivity of
the hydrogenation.
Hydrogenation of Pyrazine-2-(methyl-(S)-prolinecarboxamide) in
◦
Methanol at 80 C and 8 MPa
TOF at Reaction Maximum content Content of
A lower temperature usually favors higher diastereose-
lectivity. Therefore, the influence of the reaction temper-
ature on diastereoselectivity was studied at temperatures
5
0% conv.
time
(h)
of intermediate by-products d.e.
−
1
Catalyst
(h
)
(%)
(%)
(%)
◦
Pd/C
Rh/C
Pt/C
Ru/C
Rh/Al2O3
Rh-black
104.1
62.7
46.8
12.9
33.9
7.3
1.3
2.2
4.5
14.8
5.5
3.0
75
31
17
44
36
16
3
<1
8
4
<1
9
67
47
24
44
46
43
between 50 and 85 C with platinum and rhodium catalysts.
It was found that the reaction temperature considerably af-
fected the reaction rate but had no significant influence on
diastereoselectivity. This is probably due to the high con-
formational stability of the intermediate in the studied tem-
perature region. A similar effect was observed for hydrogen
pressure. In the region of 5–10 MPa, no significant changes
in diastereoselectivity were observed, and the higher hy-
drogen pressure resulted only in a higher reaction rate.
Solvents are also known to play a role in determining
the stereoselectivity in hydrogenation reactions. Therefore,
the influence of solvents other than methanol on the hydro-
genation of the substrate was investigated. The results of
experiments carried out with the Pd/C catalyst and with
solvents of different polarity are shown in Table 2. The
type of solvent mostly affected the TOF of the reaction
and the formation of the intermediate. There were no
drastic differences in diastereoselectivity with different sol-
vents. The reaction rate was lower for solvents with lower
polarities. This may be caused not only by the lower hydro-
genation activity but also by the lower rate of intramole-
cular cyclization, which is influenced by the reaction
medium. Protic solvents such as water facilitate the intra-
molecular cyclization of the substrate by the protonation
of the carbonyl group of the proline auxiliary. Therefore,
the maximum content of the intermediate was higher with
solvents of higher polarity. The higher value obtained with
atom (N-6). Their ratio corresponds roughly to the ratio of
the diastereomers (6 and 7) and thus they were probably
formed by the reaction of the products with the methanol
solvent (experiments with other solvents see below). The
most active catalyst was 10% Pd/C, with which the highest
content of intermediates as well as the highest diastereos-
electivity were reached. The activity, the maximum con-
centration of the intermediate, and the diastereoselectiv-
ity were considerably lower with the other catalysts. The
lowest diastereoselectivity as well as the lowest concentra-
tion of the intermediate were obtained with the platinum
catalyst. The ruthenium on active carbon catalyst was the
least active catalyst. Nevertheless, the diastereoselectivity
obtained with ruthenium was the same as with rhodium
on carbon. The influence of the support was tested with
the rhodium catalyst; it was found that it has a minimum
effect on diastereoselectivity, and a considerable effect on
TOF. The lower the total surface area of the catalyst, the
lower the TOF of the reaction. The type of catalyst support
probably affects both hydrogenation and intramolecular
cyclization.
The results show that there is a relationship between the
final diastereoselectivity and the maximum concentration
of the intermediate. The higher the concentration of the
intermediate during the reaction, the higher the diastere-
oselectivity. This phenomenon is not unexpected, because
the structure of the cyclic intermediate is much more rigid
than that of the substrate molecule. Therefore, the adsorp-
tion of the intermediate from one of its two diastereotopic
faces on the catalyst is preferred and, thus, the formation
of only one diastereomer is favored.
The course of the hydrogenation reaction and the devel-
opment of diastereoselectivity during the reaction over the
palladium catalyst are shown in Fig. 3. The diastereoselec-
tivity first increases with the conversion of the substrate and
then remains constant. A possible explanation is that this is
caused by the total hydrogenation of the substrate followed
by fast cyclization of the saturated product, which proceeds
with a lower d.e. and can occur at the beginning of the re-
100
90
70
6
0
0
80
70
60
50
5
substrate (2)
40
intermediate (5)
e.
(
(
4aS, 9aS)-diastereomer (6)
4aR, 9aS)-diastereomer (7)
d.
3
0
0
40
30
20
10
0
d.e.
2
10
0
0
20
40
60
80
100
Time [min]
FIG. 3. Reaction course of pyrazine-2-(methyl-(S)-prolinecarboxa-
◦
mide) hydrogenation. Catalyst, 10% Pd/C; solvent methanol, T = 80 C,
action. A more probable explanation is that the catalyst is p(H ) = 8 MPa, d.e. = 67%.
2

DIASTEREOSELECTIVE HYDROGENATION OF PYRAZINE DERIVATIVES
409
TABLE 2
This proved that racemization of the hydrogenation prod-
ucts did not occur during the acidic hydrolysis. The yield of
the hydrolysis and subsequent transformation of the dihy-
drochloride to the free base was approximately 92%.
Effect of the Solvent on the Hydrogenation of Pyrazine-2-
methyl-(S)-prolinecarboxamide) over a Pd/C Catalyst at 80 C and
◦
(
8
MPa
TOF at
Reaction Maximum content Content of
CONCLUSION
5
0% conv.
time
(h)
of intermediate
(%)
by-products d.e.
−
1
Solvent
(h
)
(%)
(%)
A new method for the stereoselective preparation of
piperazine-(2S)-carboxylic acid was discovered. It consists
of three simple steps: coupling of the precursor with a chiral
auxiliary, heterogeneous hydrogenation, and acidic hydrol-
ysis. All the steps were completed with a yield higher than
Water
MeOH
EtOH
264.4
104.1
98.6
46.0
23.9
31.8
25.4
17.0
0.8
1.3
3.7
5.4
11.1
9.6
76
75
73
62
54
69
38
37
—
3
5
—
—
—
9
79
67
71
64
67
65
68
74
2
-PrOH
t-BuOH
DMF
THF
85%. The diastereoselectivity obtained in the hydrogena-
6.8
7.8
tion of the substrate over a palladium catalyst was consid-
erably higher than over Rh, Ru, and Pt catalysts. The better
performance of palladium is attributed to the formation of a
larger amount of intermediate, which has a greater molec-
ular rigidity and thus facilitates the differentiation of the
two diastereotopic faces on the metal surface of the cata-
lyst. The separation of the two diastereomers from the hy-
drogenation product by crystallization allows an increase
in the d.e. value to almost 100% diastereoselectivity. Since
the hydrolysis of the hydrogenation product proceeds with-
out racemization, the preparation of enantiomerically pure
piperazine-(2S)-carboxylic acid is attainable.
The described procedure for the stereoselective prepa-
ration of piperazine-2-carboxylic acid has the following ad-
vantages: inexpensive starting materials, heterogeneous hy-
drogenation (easy separation of the catalyst), the possible
separation of the diastereomeric hydrogenation products,
and, last but not least, the recycling of the chiral auxiliary
EtOAc
—
DMF may be attributed to the fact that it can stabilize the
protonated transition state of cyclization. The highest re-
action rate as well as the highest diastereoselectivity were
obtained with water. The main advantages of water are its
high polarity, proton transfer ability, and good solubility
with all reaction components.
Some by-products were formed during the reaction with
methanol, ethanol, and tetrahydrofuran. As mentioned
above, the by-products of the reaction with methanol were
N-methylated diastereomeric products. GC–MS showed
that the by-products of the reaction with ethanol were also
saturated final diastereomers with N-ethylated secondary
nitrogen atoms (N-6). No N-methylated by-products were
observed in the reaction with ethanol. This supports the as-
sumption that the by-products are formed by the reaction
with the solvent and not with the methanol cleaved from
the substrate. The by-products formed during the reaction
with THF were not identified.
(
S)-proline.
APPENDIX
NMR, XRD, GC–MS, and elemental analysis data of the
Deprotection of the Chiral Auxiliary and Preparation
of Piperazine-2-carboxylic Acid
reaction components.
Themixtureofbothdiastereomers77(6) : 23(7)obtained
after hydrogenation was subjected to acidic hydrolysis with
aqueous hydrochloric acid to produce the hydrochlorides
Pyrazine-2-(methyl-(S)-prolinecarboxamide) (4)
1
◦
H NMR (500 MHz, CDCl3, 25 C): δ (ppm) 1.95–2.37
of piperazine-(2)-carboxylic acid and (S)-proline (Fig. 1). (m, 4H, 2 × CH2), 3.65 and 3.78 (s, 3H, OCH3), 3.80–
The piperazine-(2)-carboxylic acid dihydrochloride formed 4.03 (m, 2H, CH2), 4.71 (dd, J = 4.5 Hz, 8.7 Hz, 0.5H)
a precipitate in the reaction mixture and could thus be eas- and 5.13 (dd, J = 3.1 Hz, 8.7 Hz, 0.5H), 8.44 and 8.57 (dd,
ily separated from the (S)-proline hydrochloride, which J = 1.5 Hz, 2.5 Hz, 1H, CH == N), 8.63 and 8.66 (d, J = 2.5
remained dissolved in the acidic aqueous solution. This Hz, 1H, CH == N), 9.19 and 9.29 (d, J = 1.5 Hz, 1H, CH == N).
1
3
◦
advantageous separation of the two amino acids was very
C NMR (125 MHz, CDCl3, 25 C): δ (ppm) 21.9 and 25.5
sensitive to changes in the concentration and pH of the so- (CH2), 28.8 and 31.9 (CH2), 48.4 and 49.7 (CH2N), 52.2 and
lution. Since the determination of the optical purity of the 52.4 (OCH3), 60.3 and 61.4 (CH), 141.5 and 142.3 (CH == N),
dihydrochloride of piperazine-(2)-carboxylic acid by po- 145.7 and 145.8 (CH == N), 146.3 and 146.6 (CH == N), 147.8
larimetry was not very accurate (see experimental part), and 148.4 (CH == N), 163.8 and 164.2 (CON), 172.4 and 173.1
1
3
the free base of piperazine-2-carboxylic acid was prepared, (COO). C DEPT NMR: δ (ppm) 21.9, 25.5, 28.8, 31.9, 48.4,
and its optical purity was determined by the chiral separa- 49.7(allCH2);52.2, 52.4, 60.3, 61.4, 141.5, 142.3, 145.7, 145.8,
tion (HPLC) of the enantiomers. The ratio (77 : 23) of the 146.3, 146.6, 147.8, 148.4 (CH & CH3). MS (EI, 70 eV): 235
+
enantiomersofpiperazine-2-carboxylicaciddidnotchange. (M , 12), 176 (95), 128 (100), 106 (86), 79 (90), 68 (30), 52

4
10
KUKULA AND PRINS
(
42), 41 (26), 28 (14). Anal. Calc. for C11H13N3O3 (235.24):
TABLE 3
C 56.16, H 5.57, N 17.86; found: C 55.87, H 5.65, N 17.76.
Crystal Data and Structure Refinement of (4aS, 9aS)-Octahydro-
3
a,6,8a-triaza-cyclopenta[b]naphthalene-4,9-dione (6)
Piperazine-(2)-carboxylic Acid Dihydrochloride
Empirical formula
C10H15N3O2 · H2O
227.27
293(2) K
0.71073 A˚
Monoclinic
P2(1)
1
◦
H NMR (300 MHz, D2O, 25 C): δ (ppm) 2.75–2.87 (m, Formula weight
2
1
H), 2.93 (dd, 1H), 3.05 (dd, 1H), 3.17 (dt, 1H), 3.32 (dd, Temperature
∗
13
◦
Wavelength
Crystal system
Space group
H), 3.48 (dd, 1H, C H). C NMR (75 MHz, D2O, 25 C):
∗
δ (ppm) 41.65, 41.90, 44.98 (CH2), 56.94 (C H), and 173.25
◦
(
CO). M.p. 247–253 C (with decomposition). Anal. Calc.
Unit cell dimensions
a = 8.77(6) A˚ , alpha = 90 deg.
for C5H10N2O2 · 2HCl (203.07): C 29.57, H 5.96, N 13.80;
b = 6.82(4) A˚ , beta = 94.8(5) deg.
A˚ , gamma = 90 deg.
found: C 29.67, H 5.85, N 13.87.
c = 9.07(7)
540(6) A˚ 3
Volume
Z
2
Piperazine-(2)-carboxylic Acid (1)
3
Density (calculated)
Absorption coefficient
F(000)
Crystal size
Theta range for data collection
Index ranges
Reflections collected
Independent reflections
Absorption correction
Refinement method
1.397 Mg/m
0.104 mm
244
−
1
◦
M.p. 270–273 C (with decomposition). Anal. Calc. for
C5H10N2O2 (130.15): C 46.14, H 7.74, N 21.52; found: C
0.1 × 0.1 × 0.1 mm
4
6.09, H 7.64, N 21.56.
2.25–20.53 deg.
−8 ≤ h ≤ 8, −6 ≤ k ≤ 6, −8 ≤l ≤ 8
2153
1
,2,3,7,8,9a-Hexahydro-6H-3a,6,8a-triaza-
1080 [R(int) = 0.0293]
cyclopenta[b]naphthalene-4,9-dione (5)
None
Full-matrix least-squares on F2
1080/3/152
1
◦
H NMR (300 MHz, CDCl3, 25 ): δ (ppm) 1.86–2.11 (m,
Data/restraints/parameters
3
1
H), 2.34–2.45 (m, 1H), 3.05 (ddd, J = 3.4 Hz, 10.2 Hz,
Goodness-of-fit on F2
0.945
3.0 Hz, 1H), 3.30 (dt, 1H), 3.45–3.74 (m, 3H), 4.10 (dd,
Final R indices [I > 2sigma(I)]
R1 = 0.0620, w R2 = 0.1465
R1 = 0.0677, w R2 = 0.1516
0.35(4)
0.221 and −0.213 e.A˚
J = 6.3 Hz, 9.6 Hz, 1H), 4.57 (d, J = 12.6 Hz, 2H), 6.01 (d, R indices (all data)
1
3
◦
Extinction coefficient
J = 5.9 Hz, 1H). C NMR (75 MHz, CDCl3, 25 C): δ (ppm)
−3
∗
Largest diff. peak and hole
2
2.1, 29.0, 37.6, 40.2, 44.7 (d, CH2), 59.0 (t, C H), 105.9 (q,
C–N), 125.6 (t, CH–N), 159.6, and 162.7 (q, C == O). MS (EI,
+
7
0 eV): 207 (M , 90), 178 (23), 151 (84), 110 (16), 82 (47), 70
(
(
100), 54 (37), 41 (41), 28 (68). Anal. Calc. for C10H13N3O2
207.23): C 57.96, H 6.32, N 20.28; found: C 57.79, H 6.43,
N 20.16.
TABLE 4
Bond Lengths ( A˚ ) and Angles (Deg) of (4aS, 9aS)-Octahydro-
a,6,8a-triaza-cyclopenta[b]naphthalene-4,9-dione (6)
(
4aS, 9aS)-Octahydro-3a,6,8a-triaza-
3
cyclopenta[b]naphthalene-4,9-dione (6)
1
◦
O(9)–C(9)
N(8a)–C(9)
N(8a)–C(8)
N(8a)–C(4a)
1.226(9)
1.364(9)
1.461(11)
1.485(10)
1.492(11)
1.512(11)
1.246(9)
1.331(9)
1.457(11)
N(3a)–C(9a)
C(9a)–C(9)
C(9a)–C(1)
C(5)–N(6)
N(6)–C(7)
C(7)–C(8)
C(3)–C(2)
C(2)–C(1)
1.456(9)
H NMR (500 MHz, C6D6, 25 C): δ (ppm) 0.76 (s, H-6),
.94–1.04 (m, H-2-a), 1.09–1.16 (m, H-2-b), 1.53–1.61 (m,
1.482(11)
1.496(11)
1.440(11)
1.437(11)
1.525(12)
1.521(12)
1.528(13)
0
H-1-a), 1.97–2.02 (m, H-1-b), 2.04–2.10 (m, H-8-a), 2.31–
2
.40 (m, 2H, H-7-a, H-8-b), 2.46 (dd, JH-5-a,H-4a = 10.8 Hz, C(4a)–C(4)
C(4a)–C(5)
JH-5-a,H-5-b = 12.5Hz, H-5-a), 3.01–3.05(m, H-3-a), 3.23(dd,
C(4)–O(4)
C(4)–N(3a)
N(3a)–C(3)
J = 6.3 Hz, J = 10.5 Hz, H-9a), 3.32–3.38 (m, H-3-b), 3.47
(
(
dd, JH-4a,H-5-b = 4.0 Hz, JH-4a,H-5-a = 10.8 Hz, H-4a), 3.58
ddd, JH-5-b,H-7-b = 1.5 Hz, JH-4a,H-5-b = 4.0 Hz, JH-5-b,H-5-a =
C(9)–N(8a)–C(8)
C(9)–N(8a)–C(4a)
118.1(5)
120.3(5)
116.5(6)
111.2(6)
109.7(5)
109.4(5)
122.0(6)
120.9(6)
117.1(6)
125.3(5)
121.9(6)
112.8(5)
113.2(6)
N(3a)–C(9a)–C(1)
C(9)–C(9a)–C(1)
O(9)–C(9)–N(8a)
O(9)–C(9)–C(9a)
N(8a)–C(9)–C(9a)
N(6)–C(5)–C(4a)
C(7)–N(6)–C(5)
N(6)–C(7)–C(8)
N(8a)–C(8)–C(7)
N(3a)–C(3)–C(2)
C(3)–C(2)–C(1)
C(9a)–C(1)–C(2)
102.5(6)
116.3(6)
121.7(6)
121.8(6)
116.4(5)
111.0(4)
109.8(6)
108.6(5)
111.7(4)
103.1(5)
103.0(5)
103.7(6)
1
2.5 Hz, H-5-b), 4.34–4.38 (m, JH-7-b,H-5-b = 1.5 Hz, H-7-b).
1
3
◦
C NMR (125 MHz, CDCl3, 25 C): δ (ppm) 21.78 (t, C-2), C(8)–N(8a)–C(4a)
2
4
1
9.79 (t, C-1), 42.14 (t, C-3), 44.83 (t, C-8), 45.23 (t, C-7), N(8a)–C(4a)–C(4)
N(8a)–C(4a)–C(5)
C(4)–C(4a)–C(5)
O(4)–C(4)–N(3a)
O(4)–C(4)–C(4a)
8.93(t, C-5), 57.63(d, C-9a), 58.49(d, C-4a), 163.10(s, C-9),
67.17 (s, C-4). MS (EI, 70 eV): 209 (M , 85), 192 (15), 154
32), 139 (60), 124 (12), 112 (12), 86 (68), 70 (100), 56 (49),
+
(
5
5 (54), 42 (43), 41 (47), 28 (49). Anal. Calc. for C10H15N3O2 N(3a)–C(4)–C(4a)
(
209.25): C 57.40, H 7.23, N 20.08; found: C 57.19, H 7.24, C(4)–N(3a)–C(3)
C(4)–N(3a)–C(9a)
C(3)–N(3a)–C(9a)
N(3a)–C(9a)–C(9)
N 19.93. Crystal data and structure refinement, and bond
length and valence angles of (6), are summarized in Tables 3
and 4.

DIASTEREOSELECTIVE HYDROGENATION OF PYRAZINE DERIVATIVES
411
(
4aR, 9aS)-Octahydro-3a,6,8a-triaza-
cyclopenta[b]naphthalene-4,9-dione (7)
14. Cheng, M., De, B., Pikul, S., Almstead, N. G., Natchus, M. G.,
Anastasio, M. V., McPhail, S. J., Snider, C. E., Taiwo, Y. O., Chen,
L., Dunaway, C. M., Gu, F., Dowty, M. E., Mieling, G. E., Janusz,
M. J., and Wang-Weigand, S., J. Med. Chem. 43, 369 (2000).
The 13C NMR spectrum of the (4aR, 9aS)-diastereomer
15. Bruce, M. A., St. Laurent, D. R., Poindexter, G. S., Monkovic, I.,
was obtained from a sample in which the concentra-
tion was about 20%. In comparison with the (4aS, 9aS)-
diastereomer, all the signals were slightly shifted. C NMR
Huang, S., and Balasubramanian, N., Synth. Commun. 25, 2673
(
1995).
6. Stingl, K., Kottenhahn, M., and Drauz, K., Tetrahedron Asymmetry 8,
79 (1997).
13
1
◦
(
125 MHz, CDCl3, 25 C): δ (ppm) 21.68 (t, C-2), 29.86 (t,
9
C-1), 43.10 (t, C-3), 45.06 (t, C-8), 45.31 (t, C-7), 49.40 (t, 17. Eichhorn, K., Roduit, J.-P., Shaw, N., Heinzmann, K., and Kiener, A.,
Tetrahedron Asymmetry 8(15), 2533 (1997).
8. Fuchs, R., and Roduit, J.-P., U.S. Patent 5,886,181 (1996).
9. Kiener, A., Roduit, J.-P., Kohr, J., and Shaw, N., European Patent
C-5), 58.83 (d, C-9a), 62.58 (d, C-4a), 162.51 (s, C-9), 164.16
1
1
(
s, C-4).
0
686 698 (1995).
0. Kiener, A., Roduit, J.-P., and Heinzmann, K., WO Patent 96/35775
1996).
1. Brieden, W., Roduit, J.-P., and Fuchs, R., U.S. Patent 6,111,106
1999).
CCDC 181774 contains the supplementary crystallo-
2
2
graphic data for this paper. 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:+441223336033;e-mail:deposit@ccdc.cam.ac.uk). 22. Watanabe, N., Sadao, A., and Kuranishi, H., U.S. Patent 5,922,873
(
1998).
2
2
3. Murakami, H., Satoh, S., Tobiama, T., Sakai, K., and Nohira, H., U.S.
Patent 5,792,869 (1995).
4. Rosen, K., Pye, P. J., Volante, R. P., and Reider, P. J., U.S. Patent
5,977,364 (1999).
ACKNOWLEDGMENTS
The authors thank Prof. Volker Gramlich and Felix Bangerter for per-
forming the XRD and NMR measurements, respectively.
25. Kuwano, R., and Ito, Y., J. Org. Chem. 64, 1232 (1999).
26. Rosen, K., Weissman, S. A., Sager, J., Reamer, R. A., Askin, D.,
Volante, R. P., and Reider, P. J., Tetrahedron Lett. 36, 6419 (1995).
REFERENCES
27. Rosen, K., Pye, P. J., Volante, R. P., DiMichele, L. M., and Reider, P.
J., Tetrahedron Lett. 39, 6823 (1998).
1
2
3
4
. Sheldon, R. A., “Chirotechnology.” Decker, New York, 1993.
. Besson, M., and Pinel, C., Topics Catal. 5, 25 (1998).
. Tungler, A., and Fodor, K., Catal. Today 37, 191 (1997).
. Besson, M., Gallezot, P., Neto, S., and Pinel, C., Tetrahedron Asymme-
try 11, 1809 (2000).
28. Aebischer, B., Fray, P., Haerter, H. P., Herrling, P. L., Mueller, W.,
Olverman, H. J., and Watkins, J. C., Helv. Chim. Acta 72, 1043 (1989).
29. Shiraiwa, T., Shinjo, K., and Korokawa, H., Bull. Chem. Soc. Jpn. 64,
3251 (1991).
30. Mager, H. I. X., and Berends, W., Recl. J. R. Neth. Chem. Soc. 78, 109
(1959).
31. Felder, E., Maffei, S., Pietra, S., and Pitr e` , D., Chimia 13, 263 (1959).
32. Felder, E., Maffei, S., Pietra, S., and Pitr e` , D., Helv. Chim. Acta 43, 888
(1960).
33. Melloni, P., Torre, A. D., Munari, S. D., Meroni, M., and Tonani, R.,
Gazz. Chim. Ital. 115, 159 (1985).
34. Sahu, D. P., Indian J. Chem. Sect. B 37, 1149 (1998).
35. Landau, R. N., Singh, U., Gortsema, F., Sun, Y., Gomolka, S. C.,
Lam, T., Futran, M., and Blackmond, D. G., J. Catal. 175, 201 (1995).
36. Bigge, C. F., Hays, S. J., Novak, P. M., Drummond, J. T., Johnson, G.,
and Bobovski, T. P., Tetrahedron Lett. 30, 5193 (1989).
5
6
. Ranade, V. S., Consiglio, G., and Prins, R., Catal. Lett. 58, 71 (1999).
. Gertosio, V., Santini, C. C., Basset, J. M., Bayard, F., Buendia, J., and
Vivat, M., J. Mol. Catal. 142, 141 (1999).
7
8
9
. Tungler, A., F u¨ rcht, A., Karancsi, Zs. P., Toth, G., Math e´ , T.,
Hegedus, L., and Sandi, A., J. Mol. Catal. 139, 239 (1999).
. Mateus, C. R., Almeida, W. P., and Coelho, F., Tetrahedron Lett. 41,
2
533 (2000).
. Petersen, M., and Sauter, M., Chimia 53, 608 (1999).
1
0. Askin, D., Eng, K. K., Rossen, K., Purick, R. M., Wells, R. P., Volante,
R. P., and Reider, P. J., Tetrahedron Lett. 35, 673 (1994).
1. Bigge, C. F., Johnson, G., Ortwine, D. F., Drummond, J. T., Retz, D. M.,
1
Brahce, L. J., Marcoux, F. W., and Probert, A. W., J. Med. Chem. 35, 37. Hu, L.-Y., Ryder, T. R., Nikam, S. S., Millerman, E., Szoke, B. G., and
1
371 (1992).
Rafferty, M. F., Bioorg. Med. Chem. Lett. 9, 1121 (1999).
38. Hegedus, L., Hada, V., Tungler, A., Math e´ , T., and Szepesy, L., Appl.
Catal. 201, 107 (2000).
1
1
2. Neya, M., Yamazaki, H., Kayakiri, N., Satto, K., and Oku, T., WO
Patent 98/27069 (1998).
3. De, B., Natchus, M. G., Pikul, S., Almstead, N. G., Matthews, R. S.,
Taiwo, Y. O., and Cheng, M., WO Patent 98/08825 (1998).
39. Hada, V., Tungler, A., and Szepesy, L., Appl. Catal. 210, 165
(2001).