Imidazolopyrazinones as potential antioxidants

Article abstract of DOI:10.1016/S0960-894X(01)00445-0

A series of imidazolopyrazinones 3, substituted at C-2, and C-2/C-6, has been prepared. The compounds behaved as quenchers of superoxide anion. The more active compounds are structurally related to coelenterazine, a natural substrate of marine bioluminesc

Full text of DOI:10.1016/S0960-894X(01)00445-0

Bioorganic & Medicinal Chemistry Letters 11 (2001) 2305–2309
Imidazolopyrazinones as Potential Antioxidants
Ingrid Devillers,a Georges Dive,c Catherine De Tollenaere,a Benedicte Falmagne,a
Bertrand de Wergifosse,b Jean-Francois Reesb and Jacqueline Marchand-Brynaerta,*
a
´
Unite de Chimie organique et medicinale, Universite catholique de Louvain,
Baˆtiment Lavoisier, place L. Pasteur 1, B-1348Louvain-la-Neuve, Belgium
´
´
b
´
Unite de Biologie animale, Universite catholique de Louvain, Batiment Carnoy,
´
ˆ
place Croix du Sud 4, B-1348Louvain-la-Neuve, Belgium
c
´
´
´
`
ˆ
`
Centre d’Ingenierie des Proteines, Universite de Liege, Batiment B6, B-4000 Sart Tilman (Liege), Belgium
Received 16 October 2000; accepted 13 June 2001
Abstract—A series of imidazolopyrazinones 3, substituted at C-2, and C-2/C-6, has been prepared. The compounds behaved as
quenchers of superoxide anion. The more active compounds are structurally related to coelenterazine, a natural substrate of marine
bioluminescence. Theoretical parameters based on Hartree–Fock instabilities have been examined. # 2001 Elsevier Science Ltd. All
rights reserved.
Partially reduced derivatives of oxygen, which are pro-
duced in aerobic organisms as part of normal physio-
logical and metabolic processes, are toxic species since
they can oxidize numerous biomolecules leading to tis-
sue injury and cell death.1 Such reactive oxygen species
(ROS), produced in excessive concentrations or in
wrong locations, cause an oxidative stress associated
with a variety of disease states in humans.2ꢀ4 Thus,
when the natural protective systems towards ROS is
running over, exogeneous antioxidative compounds
have to be delivered. The research of new antioxidants
as potential drugs is an active field of medicinal
chemistry.5ꢀ7 The compounds usually involve N-hetero-
cycle and/or phenol moieties as radical scavengers.
produces carbon dioxide and coelenteramide in an exci-
ted state which deactivates by emission of light.
The sensitivity of CLZ towards oxygen and ROS led us
to consider this heterocyclic system as a potential lead in
medicinal chemistry for the discovery of new anti-
oxidants.16 In this paper, a series of simple imidazolo-
pyrazinones, structurally related to CLZ, has been
prepared and evaluated in a standard test towards
superoxide anion, in view to assess the interest of this
class of compounds. Theoretical evaluation also sug-
gested a high antioxidative potential for imidazolopyr-
azinone compounds.
Imidazolopyrazinones 3a–l (Scheme 1, Table 1) were
readily obtained by condensing 2-aminopyrazines 1 with
a-keto-aldehydes (or the derived acetals) 2, according to
known procedures.16ꢀ20 5-Aryl-2-aminopyrazines (R1¼H)
resulted from a Suzuki-like coupling21 of 5-bromo-2-
aminopyrazine22 with arylboronic acids (phenylboronic
acid and 4-methoxyphenylboronic acid). The depro-
tection of 5-(4-methoxyphenyl)-2-aminopyrazine into
5-(4-hydroxyphenyl)-2-aminopyrazine was realized with
sodium ethanethiolate in DMF at 100 ꢁC.23 The diethyl
acetal of benzyl glyoxal (R3=PhCH2) was prepared by
substitution of 1-(diethoxyacetyl) piperidine with benzyl-
magnesium bromide.24
Recently, we demonstrated that coelenterazine (CLZ;
Scheme 1, Table 1) shows high antioxidative properties
in cells submitted to oxidative stress induced by t-butyl
hydroperoxide, and inhibits the oxidation of linoleate
initiated by azoperoxyl radicals.8ꢀ10 Coelenterazine is an
imidazolopyrazinone derivative isolated from marine
organisms; this natural compound is the chromophoric
ligand of
a calcium-sensitive photoprotein called
aequorin.11ꢀ13 The molecular mechanism of CLZ biolu-
minescence and chemiluminescence is still a subject of
investigations:14,15 the catalyzed oxidation of CLZ
Compounds 3, isolated as the hydrochloride salts,25
were characterized by NMR spectroscopy. Since no
*Corresponding author. Tel.: +32-10-472740; fax: +32-10-474168;
e-mail: marchand@chim.ucl.ac.be
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00445-0

2306
I. Devillers et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2305–2309
Scheme 1.
complete data for imidazolopyrazinones26 have been
found in literature, we collected the spectroscopic values
of 3a–l in Tables 2 and 3 (spectra recorded respectively
at 500 and 125 MHz). Typically, in the 1H spectra
(Table 2), H-5 appeared at 8.2–8.5 d when R3 is an alkyl
group (R3=CH3, CH2Ph), and at 7.8–8.1 d when R3 is
phenyl. Similarly, H-8 was slightly shielded when R3 is
phenyl (8.3–8.5 d) comparatively to the chemical shift
found for this proton when R3 is an alkyl group (8.6–8.9
d). The 13C spectra (Table 3) showed the signal attrib-
uted to the carbonyl function (C-3) around 148 d when
R3 is phenyl, and at 136–142 d when R3 is an alkyl
group (R3=CH3, CH2Ph). This is consistent with pre-
vious data obtained by X-ray diffraction analysis of
crystals:16 2-phenylimidazolopyrazinone 3b appeared in
the ketone form, while 2-methylimidazolopyrazinone 3a
was stabilized in the enol tautomeric form. In structures
3d–l, the chemical shifts of C-2 and C-9, two atoms
making part of the conjugated system which extends
from N-7 to C-3,16 are very similar. On the other hand,
C-6 and C-8 appear to be influenced by the nature (aryl
or alkyl) of the substituent R3.
The chemical reactivity of CLZ towards ROS has been
correlated to the rate constant of its reaction with
superoxide anion using the hypoxanthine–xanthine
oxidase system.27,28 The same test was used for the eva-
luation of the synthetic imidazolopyrazinones in the
presence of MeO–CLZ (Scheme 1, Table 1) as the com-
petitor. The rate constant of this luminescent reference16
has been previously determined by using TroloxR, a
water-soluble derivative of vitamin E,29 as the known
competitor (Table 4). Thus, compounds 3a–l were reac-
Table 1. Imidazolopyrazinone derivatives
ꢂ
ꢀ
ted with O2 in the presence of MeO–CLZ in different
concentrations;30 the intensities of light emission were
measured at 380 nm, and the reaction rate constants
were calculated from the following equation:31ꢀ33
Compd
R1
R2
R3
Ref
CLZ
MeO–CLZ
p-HO-Ph
p-MeO-Ph
H
CH2Ph
CH2Ph
H
H
H
H
H
H
H
H
H
H
H
CH2-Ph-p-OH
17–19
8,10
16
16
20
20
20
20
20
20
20
20
20
CH2-Ph-p-OH
Me
Ph
3a
3b
3c
3d
3e
3f
3g
3h
3i
H
H
Ph
Ph
Ph
Io=Ic ¼ 1 þ kc=ki  ½MeOꢀCLZꢁ=½3ꢁ
CH2Ph
Me
Ph
CH2Ph
Me
Ph
CH2Ph
Me
Ph
where Io=luminescence measured without MeO–CLZ,
Ic=luminescence measured in the presence of MeO–
CLZ (competitor), kc=rate constant of MeO–CLZ
(competitor), ki=rate constant of the tested compound
3, [MeO–CLZ]=concentration of competitor and
[3]=concentration of tested compound.
p-MeO-Ph
p-MeO-Ph
p-MeO-Ph
p-HO-Ph
p-HO-Ph
p-HO-Ph
3j
3k
3l
H
CH2Ph
20
Table 3. 13C NMR data (d)
Table 2. 1H NMR data (d)
Compd
Solvent
C-2
C-3
C-5
C-6
C-8
C-9
Compd
Solvent
H-5
H-6
H-8
J (Hz)
3a
CD3OD
135.9 142.8 116.5 122.4 133.6 130.3
3a
CD3OD
DMSO-d6
CD3OD
DMSO-d6
CD3OD
8.09
8.30
8.29
7.87
8.19
7.65
7.73
7.70
7.22
7.70
8.78
8.99
8.86
8.47
8.77
3J5ꢀ6=5.5; 5J5ꢀ8=1.1
4J6ꢀ8=0.5
DMSO-d6 137.4 142.9 114.4 119.1 130.1 128.8
CD3OD 139.0 148.2 113.4 116.1 127.1 129.0
DMSO-d6 139.4 149.9 112.5 115.5 125.7 129.1
3b
3c
3b
3c
3J5ꢀ6=5.5; 5J5ꢀ8=0.8
4J6ꢀ8=0.5
CD3OD
140.3 144.5 116.2 121.2 132.6 130.7
3J5ꢀ6=5.5; 5J5ꢀ8=0.9
4J6ꢀ8=0.5
3d
3e
3f
CD3OD 123.9 139.9 112.4 141.2 135.8 128.5
DMSO-d6 128.7 148.0 109.5 129.2 128.0 129.3
CD3OD
3d
3e
3f
3g
3h
3i
3j
3k
3l
CD3OD
DMSO-d6
CD3OD
CD3OD
DMSO-d6
CD3OD
DMSO-d6
DMSO-d6
CD3OD
8.30
8.15
8.24
8.41
7.89
8.57
8.48
7.88
8.42
—
—
—
—
—
—
—
—
—
8.76
8.44
8.61
8.88
8.30
8.96
8.90
8.33
8.84
5J5ꢀ8=1.1
124.3 141.4 112.5 141.2 134.8 129.3
5J5ꢀ8=0.8
5J5ꢀ8=1.1
3g
3h
3i
CD3OD
126.9 140.0 111.2 141.0 134.7 128.8
DMSO-d6 128.7 148.7 108.3 129.5 127.2 129.3
CD3OD 126.2 140.5 111.4 141.0 133.4 129.0
5J5ꢀ8=0.9
3j
3k
3l
DMSO-d6 127.0 137.5 109.6 139.2 133.6 127.0
DMSO-d6 128.6 147.8 108.1 129.2 127.6 129.3
CD3OD
5J5ꢀ8=0.9
127.4 136.5 110.7 141.4 135.0 129.5

I. Devillers et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2305–2309
2307
The results, summarized in Table 4, clearly showed that
all the tested imidazolopyrazinones are more active than
TroloxR; however, two compounds (3e and 3h) could
not be evaluated because they were not totally soluble in
the required concentrations. The presence of an aryl
substituent R1 at position C-6 increased the activity,
comparatively to the C-6 unsubstituted series (com-
pounds 3d, 3g, and 3j). The substituent R3 at position
C-2 appeared to exercise a moderate influence. Accord-
ingly, the p-hydroxybenzyl substituent found in the
natural CLZ is not absolutely required. Compared to
the natural derivatives (CLZ and MeO–CLZ) possess-
ing a benzyl substituent (R2) at position C-8, the
synthetic derivatives 3 devoid of such a substitution are
2 or 3 times more active. Thus, simple imidazolopyr-
azinones, more easily synthesized than CLZ, can be
considered as potential antioxidants.
The Hartree–Fock instabilities38 of compounds 3 are
summarized in Table 5. At the singlet state structure,
the optimization of the wave function gives rise to a
stabilization energy (delta-stable). From this result, a
complete geometry re-optimization generates the rela-
tive energy values at the UHF level (UHF/UHF) in
which alpha and beta electrons occupy different spatial
molecular orbitals. The more the molecule is substituted
by aromatic fragments, the more the stabilization
energy is high. This feature is nicely correlated to the
antioxidant activity for the four compounds bearing a
methyl substituent at position C-2 (R3=CH3), namely
3a, 3d, 3j, and 3g. With a benzyl substituent in that
position (R3=CH2–Ph), additional steric effects could
explain the less clear relation with activity, except
between 3c and 3i. Figure 1 shows the spin density of
the reference compound 3 (R1=R2=R3=H): the iso-
contour of spin density at 0.005 AU (e/A**3) displays
the alternation of positive and negative clouds. The
break induced by the substitution is illustrated with
compound 3j (Fig. 1).39 For both molecules, the highest
spin density is located on the carbon atoms C-2 and C-8.
Theoretical parameters have been recently defined to
characterize antioxidants and to predict antioxidative
activities. Testa et al.34,35 considered ÁHox (relative
adiabatic oxidation potential) and the shape of the
SOMO (singly occupied molecular orbital) as the quan-
tum chemical descriptors. More recently, Haemers et
al.36 correlated antioxidant activity with radical stabili-
zation properties. These parameters are derived from
the analysis of the radical form obtained by elimination
of one electron. In the present study, an analysis of the
propension of the neutral compounds 3 with all their
paired electrons to generate radicalar structures is con-
cerned. This approach is based on the investigation of
Hartree–Fock instabilities. By its fused five-membered
ring, the imidazolopyrazinone compounds break down
the aromaticity and only one Kekule form can be
drawn. This feature has been previously pointed out as
a source of the wave function instability.37 It is related
to the vicinity of a triplet electronic state near the fun-
damental singlet state. Imidazole wave function is stable
and that of pyrazine presents the same weak instability
as benzene and pyridine; but the association of the two
fragments leads to a significant instability. This one can
be seen as a propensity of the molecules to present a
local biradicalar character which could directly be rela-
ted to their reactivity towards superoxide anion.
Thus, imidazolopyrazinones represent a valuable class
of new antioxidants; an aryl substituent R1 at position
Table 5. Theoretical parameters
Compd
Delta-E stable
Delta-E UHF/UHF
Stabilization
3a
3b
3c
3d
3f
3g
3i
16.29
31.18
25.00
28.37
36.96
27.35
35.60
27.42
35.52
27.89
46.48
39.72
44.01
54.75
42.78
53.92
42.87
52.87
11.60
15.30
14.72
15.64
17.79
15.43
18.32
15.45
17.35
3j
3l
Referencea
17.59
30.36
12.77
aReference=compd 3 with R1=R2=R3=H.
ꢀ
Table 4. Reaction rate constants with O2
a
Compd
kiÂ104 (Mꢀ1 sꢀ1
)
krel
Trolox
MeO–CLZ
1.728
9.6Æ0.115
6.0Æ0.1
8.4Æ0.6
2.8Æ0.3
17Æ0.2
n.d.b
6.5Æ0.1
28Æ0.1
n.d.
16Æ0.3
21Æ0.4
10.5Æ0.3
3.3Æ0.2
12Æ0.116
1
5.6
3a
3b
3c
3d
3e
3f
3g
3h
3i
3.5
4.9
1.6
10.0
—
3.8
16.5
—
9.4
12.3
6.2
1.9
7.1
3j
3k
3l
CLZ
akrel ¼ ki=kTrolox
.
bn.d., not determined for solubility reasons.
Figure 1.

2308
I. Devillers et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2305–2309
C-6, in particular a phenol moiety, reinforces this prop-
erty, while the R3 substituent at position C-2 seems to
play a moderate role. The C-8 position, substituted with
a benzyl group (R2=Bz) in the natural CLZ but
unsubstituted in the synthetic derivatives 3a–l
(R2=H), could be used for the anchorage of mole-
cular fragments susceptible to improve the biodis-
ponibility.
22. Nakamura, H.; Takeuchi, D.; Murai, A. Synlett 1995,
1227.
23. Feutrill, G. I.; Merrington, R. N. Aust. J. Chem. 1972, 25,
1719.
24. Hirano, T.; Negishi, R.; Yamaguchi, M.; Chen, F. Q.;
Ohmiya, Y.; Tsuji, F. I.; Ohashi, M. Tetrahedron 1997, 53,
12903.
25. Typical procedures for the preparation of compounds 3:
Method A (R3=Me, Ph): to a 0.5 M solution of 1 (1 equiv)
and methyl- or phenylglyoxal 2 (1.5 equiv) in ethanol, was
added aqueous HCl (37%, 3.6 equiv). The mixture was heated
under argon atmosphere for 4 h at 80 ꢁC, then concentrated in
vacuum. The residue was dissolved in cold methanol and left
overnight at ꢀ18 ꢁC to crystallize. The solid was filtered off
and washed several times with cold methanol, ethyl acetate
and ether. Method B (R3=CH2Ph): to a 0.25 M solution of 1
(1 equiv) and benzylglyoxal 2 (diethyl acetal, 1.3 equiv) in
dioxane–water (2:1, v/v), was added aqueous HCl (37%, 10
equiv). The mixture was heated under argon atmosphere for
4 h at reflux, then concentrated in vacuum. The residue, dis-
solved in methanol, was precipitated by addition of cold ether.
26. Usami, K.; Isobe, M. Tetrahedron 1996, 52, 12061.
27. Lucas, M.; Solano, F. Anal. Biochem. 1992, 206, 273.
28. Teranishi, K.; Shimomura, O. Anal. Biochem. 1997, 249,
37.
Acknowledgements
This work was supported by the Fonds National de la
Recherche Scientifique (FNRS, Belgium) and the Wal-
loon Government (Convention no. 9713664). I.D. and
B. de W. are fellows of the Fonds pour la Formation a
la Recherche dans l’Industrie et l’Agriculture (FRIA,
Belgium). J.-F.R., G.D. and J.M.-B. are senior research
associates of FNRS.
References and Notes
29. Gotoh, N.; Niki, E. Methods Enzymol. 1994, 233, 154.
30. Typical procedure for the reaction with superoxide anion:
hypoxanthine (HX; first dissolved in 1 N NaOH), xanthine
oxidase (XOD), Trolox (6-OH-2,5,7,8-tetramethylchroman-2-
carboxylic acid; first solubilized in DMSO), and albumin were
purchased from Sigma-Aldrich. All the solutions were made at
25 ꢁC in 50 mM Tris–HCl buffer (pH 7.8) containing EDTA
(0.1 mM). The final concentrations of HX and XOD were 500
mM and 8.25 U/L, respectively. Albumin was added at a final
concentration of 15 mg/L to minimize the inactivation of
XOD. During the stationary phase of the reaction, the light
yields [relative luminescence units (RLU)] were recorded for
200 s, in 96-well plates, with a Microlumat LB96P6 lumin-
ometer (Berthold Inc., Wildbad, Germany). Each well con-
tains a total volume of 200 mL after addition of HX, that is 40
mL of the competitor (25 mM), 40 mL of 3 (from 25 to 250 mM),
15 mL of XOD, 55 mL of buffer, and 50 mL of HX. Back-
ground chemiluminescence (before HX addition) was sub-
tracted from the luminescence signal. Competitive quenching
1. Morrissey, P. A.; O’ Brien, N. M. Int. Dairy J. 1998, 8, 463.
2. Gordon, M. H. Nat. Prod. Rep. 1996, 265.
3. Hollan, S. Haematologia 1995, 26, 177.
4. Gille, G.; Sigler, K. Folia Microbiol. 1995, 40, 131.
5. Nakao, K.; Shimizu, R.; Kubota, H.; Yasuhara, M.;
Hashimura, Y.; Suzuki, T.; Fujita, T.; Ohmizu, H. Bioorg.
Med. Chem. 1998, 6, 849.
6. Parmar, V. S.; Kumar, A.; Prasad, A. K.; Singh, S. K.;
Kumar, N.; Mukherjee, S.; Raj, H. G.; Goel, S.; Errington,
W.; Puar, M. S. Bioorg. Med. Chem. 1999, 7, 1425.
7. Dorey, G.; Lockhart, B.; Lestage, P.; Casara, P. Bioorg.
Med. Chem. Lett. 2000, 10, 935.
8. Rees, J. F.; de Wergifosse, B.; Noiset, O.; Dubuisson, M.;
Janssens, B.; Thompson, E. M. J. Exp. Biol. 1998, 201, 1211.
9. de Wergifosse, B.; Noiset, O.; Dubuisson, M.; Marchand-
Brynaert, J.; Baguet, F.; Rees, J.-F. In Bioluminescence and
Chemiluminescence; Roda, A., Kricka, L., Stanley, P., Eds.;
Wiley: New York, 1999; p 396.
ꢂ
ꢀ
experiments towards O2 were performed between lumines-
cent compounds (CLZ, MeO–CLZ; 5 mM) and increasing
TroloxR concentrations (from 0 to 60 mM). Rate constants
10. Dubuisson, M.; de Wergifosse, B.; Trouet, A.; Baguet, F.;
Marchand-Brynaert, J.; Rees, J.-F. Biochem. Pharmacol. 2000,
60, 471.
ꢂ
ꢀ
of non-luminescent compounds (3a–l) towards O2
were
11. Ohmiya, Y.; Hirano, T. Chem. Biol. 1996, 3, 337.
12. Jones, K.; Hibbert, F.; Keenan, M. Trends Biotechnol.
1999, 17, 477.
13. Head, J.-F.; Inouye, S.; Teranishi, K.; Shimomura, O.
Nature 2000, 405, 372.
14. Kuse, M.; Isobe, M. Tetrahedron 2000, 56, 2629, and
references cited therein.
15. Nakamura, H.; Wu, C.; Inouye, S.; Murai, A. J. Am.
Chem. Soc. 2001, 123, 1523.
16. Devillers, I.; de Wergifosse, B.; Bruneau, M.-P.; Tinant,
B.; Declercq, J.-P.; Touillaux, R.; Rees, J.-F.; Marchand-Bry-
naert, J. J. Chem. Soc., Perkin Trans. 2 1999, 1481.
17. Gonzalez-Trueba, G.; Paradisi, C.; Zoratti, M. Anal. Bio-
chem. 1996, 240, 308, and references cited therein.
18. Teranishi, K.; Goto, T. Bull. Chem. Soc. Jpn. 1990, 63,
3132.
19. Keenan, M.; Jones, K.; Hibbert, F. J. Chem. Soc., Chem.
Commun. 1997, 323, and references cited therein.
20. Hirano, T.; Nishibuchi, S.; Yoneda, M.; Tsujimoto, K.;
Ohashi, M. Tetrahedron 1993, 49, 9267.
obtained by competition between them and MeO–CLZ. Plotting
[MeO–CLZ]/[3] versus the Io=Ic values allows the determina-
tion of kc=ki from the slope of the linear relationship linking
these parameters. Each experiment was performed six times.
31. Gotoh, N.; Niki, E. Chem. Lett. 1990, 1475.
32. Suzuki, N.; Suetsuna, K.; Mashiko, S.; Yoda, B.;
Nomoto, T.; Toya, Y.; Inaba, H.; Goto, T. Agric. Biol. Chem.
1991, 55, 157.
33. Akutsu, K.; Nakajima, H.; Katoh, T.; Kino, S.; Fujimori,
K. J. Chem. Soc., Perkin Trans. 2 1995, 1699.
34. Migliavacca, E.; Carrupt, P.-A.; Testa, B. Helv. Chim.
Acta 1997, 80, 1613.
35. Migliavacca, E.; Ancerewicz, J.; Carrupt, P.-A.; Testa, B.
Helv. Chim. Acta 1998, 81, 1337.
36. Rajan, P.; Vedernikova, I.; Cos, P.; Vanden Berghe, D.;
Augustyns, K.; Haemers, A. Bioorg. Med. Chem. Lett. 2001,
11, 215.
37. Dehareng, D.; Dive, G. J. Comp. Chem. 2000, 21, 483.
38. Computational tool: all the geometry optimizations and
instability calculations have been performed at the ab initio
level using the MINI-10 basis set.40,41 As pointed out in the
21. Jones, K.; Keenan, M.; Hibbert, F. Synlett 1996, 509.

I. Devillers et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2305–2309
2309
study of Hartree–Fock instabilities occurring in benzimidazole
derivatives,42 MINI-10 provides results well correlated with
those derived from extended basis sets.
39. Computational tool: the calculations have been performed
at the ab initio level using the 6-31G* basis set (self-consistent
molecular orbital method 25: supplementary functions for
gaussian basis sets)43 and the numerical procedures available
in Gaussian 94 (revision E.2).44
40. Tatewaki, H.; Huzinaga, S. J. Comput. Chem. 1980, 1, 205.
41. Dive, G.; Dehareng, D.; Ghuysen, J.-M. Theoret. Chim.
Acta 1993, 85, 409.
42. Vancampenhout, N.; Dive, G.; Dehareng, D. Int. J.
Quant. Chem. 1996, 60, 911.
43. Frisch, M. J.; Pople, J. A.; Binkley, J. S. J. Chem. Phys.
1984, 80, 3265.
44. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M.
W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith,
T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.;
Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman,
J. B.; Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.;
Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.;
Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.;
Baker, J.; Stewart, J. P.; Head-Gordon, M.; Gonzalez,
C.; Pople, J. A. Gaussian 94, Revision E.2, Gaussian Inc.:
Pittsburgh, PA, 1997.