Novel asymmetric synthesis of atropisomeric 6-aryl pyrazinones via an unusual chirality transfer process

Article abstract of DOI:10.1021/jo9811930

Cyclization of (S,S)-α-[(1-phenylethyl)amino]-α-(2- iodophenyl)acetonitrile with (COCl)₂ in toluene or chlorobenzene afforded the atropisomeric pyrazinone (aS,S) 6-(2-α-iodophenyl)-3,5-dichloro-1(1- phenylethyl)-2(1H)-pyrazinone in 57% yield. With smaller ortho substituents (F, Cl, CH₃, CF'3, OCH₃) on the aromatic ring, mixtures of atropisomers were obtained from the cyclization reaction. All of the individual atropisomers prepared were stable at room temperature. All but the o-fluoro- substituted atropisomers were stable at elevated temperatures. This paper describes a stereoselective synthesis of pyrazinones and suggests a mechanism for formation via an interesting transfer of chirality.

Full text of DOI:10.1021/jo9811930

J . Org. Chem. 1999, 64, 93-100
93
Novel Asym m etr ic Syn th esis of Atr op isom er ic 6-Ar yl P yr a zin on es
via a n Un u su a l Ch ir a lity Tr a n sfer P r ocess
J ohn Tulinsky,1 B. Vernon Cheney, Stephen A. Mizsak, William Watt, Fusan Han,
Lester A. Dolak, Thomas J udge, and Ronald B. Gammill*,†
Structural, Analytical and Medicinal Chemistry, Pharmacia & Upjohn, Inc.,
Kalamazoo, Michigan 49001-0199
Received J une 18, 1998
Cyclization of (S,S)-R-[(1-phenylethyl)amino]-R-(2-iodophenyl)acetonitrile with (COCl)2 in toluene
or chlorobenzene afforded the atropisomeric pyrazinone (aS,S) 6-(2-R-iodophenyl)-3,5-dichloro-1-
(1-phenylethyl)-2(1H)-pyrazinone in 57% yield. With smaller ortho substituents (F, Cl, CH3, CF3,
OCH3) on the aromatic ring, mixtures of atropisomers were obtained from the cyclization reaction.
All of the individual atropisomers prepared were stable at room temperature. All but the o-fluoro-
substituted atropisomers were stable at elevated temperatures. This paper describes a stereose-
lective synthesis of pyrazinones and suggests a mechanism for formation via an interesting transfer
of chirality.
In tr od u ction
pounds, efficient and expedient methods of synthesis
remain the gatekeepers for the delineation of new and
interesting chemical and biological properties of such
molecules.
Chiral atropisomeric compounds are of considerable
interest due to their presence in a number of biologically
active natural and synthetic products and their utility
as directing groups in asymmetric synthesis.2 The inter-
esting physical, chemical, and biological properties of
these compounds is reflected and to a vast extent con-
trolled by the relative conformation of the two aromatic
rings.2,3 In principle, heteroaryl-aryl atropisomers3b offer
certain opportunities for structural and physical change
that are not available in the more classical biphenyl
systems. For example, heteroatoms provide an opportu-
nity to probe a wide variety of geometric and electronic
situations unavailable in the biphenyl system.3 Despite
many recent advances in the asymmetric synthesis of
chiral biphenyl and heteroaryl-aryl atropisomeric com-
Pyrazinones 1 and 2 are recently discovered ligands
at Pharmacia & Upjohn that bind to a new site on the
1
GABAA/chloride ionophore complex.6a,b In the H NMR
spectra of pyrazinones 1 and 2, the methylene hydrogens
(3) (a) For reviews and papers discussing axially chiral heteroaro-
matics, see: Gallo, R.; Roussel, C.; Berg, U. In Advances in Heterocyclic
Chemistry; Katritzky, A. R., Ed.; Academic Press: San Diego, 1988;
Vol. 43, p 173. Roussel, C.; Adjimi, M.; Chemlal, C.; Djafri, A. J . Org.
Chem. 1988, 53, 5076. Dogan, I.; Pustet, N.; Mannschreck, A. J . Chem.
Soc., Perkin Trans. 2 1993, 1557. (b) For specific examples and leading
references of heterocyclic systems, see: for N-aryl-4-thiazoline-2-
thiones, Roussel, C.; Djafri, A. New J . Chem. 1986, 10, 399. For the
anaesthetic ketamine, the tranquilizer etaqualone, and the anthelm-
intic praziquantel: Mannschreck, A.; Koller, H.; Wernick, R.; Kontakte,
Darmstadt 1985, 1, 40. For examples of flavins: Shinkai, H.; Nakao,
H.; Kuwahara, I.; Miyamoto, M.; Yamaguchi, T.; Manabe, O. J . Chem.
Soc., Perkin Trans. 1 1988, 313. For N-aryl-4-pyridones: Mintas, M.;
Orhanovic, Z.; J akopcic, K.; Koller, H.; Stuhler, G.; Mannschreck, A.
Tetrahedron 1985, 41, 229. For N-arylpyrroles: Vorkapic-Furac, J .;
Mintas, M.; Burgemeister, R.; Mannschreck, A. J . Chem. Soc., Perkin
Trans. 2 1989, 713. For the anticonvulsant and hypnotic methaqua-
lone: Mannschreck, A.; Koller, H.; Stuhler, G.; Davies, M. A.; Traber,
J . Eur. J . Med. Chem. 1984, 19, 381. For N-aryl-2(1H)-quinolones and
N-aryl-6(5H)-phenanthridones: Mintas, M.; Mihaljevic, N.; Koller, H.;
Schuster, D.; Mannschreck, A. J . Chem. Soc., Perkin Trans. 2 1990,
619. (c) For a recent report on the atroposelective thermal reactions of
axially twisted amides and imides, see: Curran, D. P.; Qi, H.; Geib, S.
J .; DeMello, N. C. J . Am. Chem. Soc. 1994, 116, 3131.
(4) For references leading to general discussions, see: Wainer, I.
W. Drug Stereochemistry: Analytical Methods and Pharmacology; M.
Dekker, Inc.: New York, 1993. Ariens, E. J . Trends Pharm. Sci. 1993,
14, 68. Williams, K. M. Adv. Pharmacol. 1991, 22, 57. Ariens, E. J .
Med. Res. Rev. 1986, 6, 451. Lam, Y. W. Pharmacotherapy 1988, 8,
147. For examples of enantiomers, see: Froimowitz, M.; Pick, C. G.;
Pasternak, G. W. J . Med. Chem. 1992, 35, 1521 and references therein
(CNS related). Macor, J . E.; Blake, J .; Fox, C. B.; J ohnson, C.; Koe, B.
K.; Level, L. A.; Morrone, J . M.; Ryan, K.; Schmidt, A. W.; Schulz, D.
W.; Zorn, S. H. J . Med. Chem. 1992, 35, 4503. Ariens, E. J . Trends
Pharm. Sci. 1986, 200. Hoyer, D. Trends Pharm. Sci. 1986, 227.
Lehmann, P. A. Trends Pharm. Sci. 1986, 281.
† Current Address: Central Research Division, Pfizer, Inc., Groton,
CT 06340.
(1) Postdoctoral Research Scientist 1993-95. Current address:
Discovery Research, Cell Therapeutics, Inc., 201 Elliott Ave W., Seattle,
WA 98119.
(2) The term atropisomerism refers to optical isomerism resulting
from restricted rotation about a single bond. Biphenyls represent but
one example of this type of isomerism. (a) For a general discussion of
atropisomeric compounds, see: Oki, M. In Topics in Stereochemistry;
Allinger, N. L., Eliel, E. L., Wilen, S. H., Eds.; J ohn Wiley and Sons:
New York, 1983. Eliel, E. L. Stereochemistry of Carbon Compounds;
McGraw-Hill: New York, 1995. For recent asymmetric syntheses of
atropisomeric compounds, see: Lipshutz, B. H.; Liu, Z.; Kayer, F.
Tetrahedron Lett. 1994, 5567. Feldman, K. S.; Ensel, S. M.; Minard,
R. D. J . Am Chem. Soc. 1994, 116, 1742. Nelson, T. D.; Meyers, A. I.
J . Org. Chem. 1994, 59, 2577, 2655. Suzuki, T.; Hotta, H.; Hattori, T.;
Miyano, S. Chemistry Lett. 1990, 807. For a fairly recent review on
the subject of biaryl synthesis, see: Bringmann, G.; Walter, R.; Weirich,
R. Angew. Chem., Int. Ed. Engl. 1990, 29, 977. (b) For leading
references regarding the biological properties of atropisomeric com-
pounds, see: Govindachari, T.; Nagarajan, K.; Parthasarathy, P. C.;
Rajagopalan, T. G.; Desai, H. K.; Kartha, G.; Chen, S. K.; Nakanishi,
K. J . Chem. Soc., Perkin Trans. 1 1974, 1413. Bringmann, G. In The
Alkaloids; Academic Press: New York, 1986; Vol. 29, p 141. Huang,
L.; Si, Y.-K.; Snakzke, G.; Zheng, D.-K.; Zhou, J . Collect. Czech. Chem.
Commun. 1988, 53, 2664. Uchida, I.; Ezaki, M.; Shigematsu, N.;
Hashimoto, M. J . Org. Chem. 1985, 50, 1341. Kannan, R.; Williams,
D. H. J . Org. Chem. 1987, 52, 5435. Kupchan, S. M.; Britton, R. W.;
Ziegler, M. F.; Gilmore, C. J .; Restivo, R. J .; Bryan, R. F. J . Am. Chem.
Soc. 1973, 95, 1335. Tomioka, K.; Ishiguro, T.; Koga, K. Tetrahedron
Lett. 1980, 2973. (c) For an excellent discussion on the use of
stereochemistry as a probe of pharmacological receptors, see: Casy,
A. F. The Steric Factor in Medicinal Chemistry. Dissymmetric Probes
of Pharmacological Receptors; Plenum Press: New York, 1993.
(5) Colebrook, L. D.; Giles, H. G.; Can. J . Chem. 1975, 53, 3431.
Bringmann, G.; Hartung, T.; Gobel, L.; Schupp, O.; Peters, K.; von
Schnering, H. G. Liebigs Ann. Chem. 1992, 769.
10.1021/jo9811930 CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/15/1998

94 J . Org. Chem., Vol. 64, No. 1, 1999
Tulinsky et al.
Sch em e 1
°C, 1a remained homogeneous by chiral HPLC analysis
(eq 1). These experiments convincingly demonstrated
that pyrazinones such as 1 could be resolved into their
enantiomeric pairs and that those enantiomers are
configurationally stable at 100 °C for several hours. To
assign the absolute stereochemistry of enantiomers 1a
and 1b, the earlier eluting enantiomer (1a ; [R]D -83°)
was coupled with (R)-(+)-R-methylbenzylamine to yield
the 3-amino-substituted pyrazinone 4 (eq 2) in 98% yield.
at N-1 are nonequivalent. This structural information
suggested that these pyrazinones might exist as atropi-
somers due to restricted rotation about the heteroaryl-
aryl carbon-carbon bond.5 This observation, coupled with
the interesting biological activity, presented an op-
portunity to investigate both the chemistry and biology
of the 6-arylpyrazinones, a new class of atropisomers.
In this paper, we confirm the atropisomeric nature of
6-arylpyrazinones and establish the absolute stereochem-
istry of representatives from this class of compounds.
Furthermore, we discuss a simple and highly efficient
asymmetric synthesis and present a mechanistic ratio-
nale for an unusual chirality transfer process.
The absolute configuration of 4 was established by single-
crystal X-ray analysis (see Supporting Information) and
proved to be (aS) about the aryl-heteroaryl bond.10 To
continue our investigation into the stereochemical aspects
and, specifically, the possibility of devising an asymmetric
synthesis of these compounds, we elected to introduce an
asymmetric center into our starting aminonitrile via the
introduction of an R-methylbenzylamine group. We hoped
that introduction of an asymmetric center would provide
us with some insight into the reaction pathway and as a
minimum provide us with easier analysis and isolation
of reaction products (diastereomers vs enantiomers).
Reaction of (R)- or (S)-R-methylbenzylamine hydrochlo-
ride, an aryl aldehyde, and sodium cyanide (Strecker
reaction, see Scheme 2) in aqueous methanol afforded the
corresponding aminonitriles 5 and 7 in quite satisfactory
yield with moderate diastereoselectivity (Table 1).11 In
general, the major diastereomer resulting from the
Strecker reaction (5 or 7) could be purified to homogene-
ity through recrystallization of the corresponding HCl
salts. However, in several cases (6b, 8b, 5g, 7g, 7j, and
8k ), diastereomerically pure aminonitriles were more
difficult to obtain and were used as enriched mixtures
(see Table 1). Cyclization of diastereomerically pure 5a
and 7a , derived from o-iodobenzaldehyde with an excess
Resu lts a n d Discu ssion
Pyrazinones 1 and 2 were readily synthesized in two
steps from commercially available starting materials
(Scheme 1). Strecker reaction of an amine hydrochloride,
an ortho-substituted benzaldehyde, and sodium cyanide
in aqueous methanol afforded the desired racemic ami-
nonitrile 3. Treatment of that aminonitrile hydrochloride
with an excess of oxalyl chloride in chlorobenzene or
toluene at 50° C then provided pyrazinone 1.8 Pyrazino-
nes 1 and 2 were readily resolved into their enantiomeric
counterparts via analytical HPLC using a chiral station-
ary phase (Scheme 1).9 We established that the chiral
axis of the pyrazinone was thermally stable in the
following manner. Toluene solutions of enantiomer 1a
were heated, and aliquots were removed periodically and
analyzed by HPLC. After 21 h at 90 °C and 3 h at 110
(6) (a) Tulinsky, J .; Mizsak, S. A.; Watt, W.; Dolak, L. A.; J udge, T.
M.; Gammill, R. B. Tetrahedron Lett. 1995, 36, 2017. Im, H. K.; Im,
W. B.; J udge, T. M.; Gammill, R. B.; Hamilton, B. J .; Carter, D. B.
Mol. Pharmacol. 1993, 44, 468. In light of their dramatic and direct
influence on chloride flux, it is likely that pyrazinones bind at a site
within the channel domain of the complex. In contrast to barbiturates,
which also modulate the GABAA complex in an allosteric fashion,
concentrations of the pyrazinones greater than that which illicits the
maximum effect on the chloride channel have no further effect on
chloride flux. Thus, the pyrazinones potentially offer a template in
which the pharmacological toxicity associated with barbiturates may
be reduced. (b) Tulinsky, J .; Gammill, R. B. Curr. Med. Chem. 1994,
1, 226. Gammill, R. B.; Carter, D. Annu. Rep. Med. Chem. 1994,
Chapter 5.
(10) Further information regarding the crystal structure of 4 will
be published elsewhere. A ball-and-stick drawing of 4 is provided in
the Supporting Information.
(11) (a) Stout, D. M.; Black, L. A.; Matier, W. L. J . Org. Chem. 1993,
48, 5369. A single recrystallization of the hydrochloride salt of the
aminonitrile was generally sufficient to provide the pure (R,R) or (S,S)
diastereomer. The absolute stereochemistry of the aminonitrile was
established via single-crystal X-ray crystallography. A representative
description of the Strecker reaction can be found in the Experimental
Section. (b) In the stereochemical designation of the aminonitriles, the
first center listed relates to the R-methylbenzyl center and the second
to the nitrile-bearing carbon. In the pyrazinones, the first center
designated relates to the chiral axis and the second to the R-methyl-
benzyl center. The prefix a denotes axial chirality. This can lead to
confusion in the interpretation of the data if not recognized. (b) For a
leading reference on the epimerization of related aminonitriles, see:
Subramanian, P. K.; Woodard, R. W. Synth. Commun. 1986, 16, 337.
Patel, M. S.; Worsley, M. Can. J . Chem. 1970, 48, 1881.
(7) All new compounds had physical and analytical properties (1H
and 13C NMR, IR, UV, MS, HRMS, and combustion analysis) consistent
with their assigned structures.
(8) Vekemans, J .; Pollers-Wieers, C.; Hoornaert, G. J . Heterocycl.
Chem. 1983, 20, 919.
(9) Separation by chiral HPLC was achieved on a (R,R) Whelk-01
column (0.46 × 25 cm); elution with 30% 2-propanol in hexane, 1.5
mL/min, UV monitor at 345 nm.

Synthesis of Atropisomeric 6-Aryl Pyrazinones
J . Org. Chem., Vol. 64, No. 1, 1999 95
Sch em e 2
Ta ble 1. Da ta for Am in on itr ile Str u ctu r es 5-8
was established by X-ray crystallography (Supporting
Information). The diastereomer of 11a (i.e., 12a ) was
isolated from the cyclization of a mixture of aminonitrile
diastereomers (7a and 8a ). An X-ray of pyrazinone 12a
was also obtained (Supporting Information), providing
further structural information about axially chiral pyrazi-
nones.14 These initial experiments with aminonitrile
substrates containing a chiral auxiliary established that
an efficient method for asymmetric synthesis of these
pyrazinones was possible.
Cyclization of the diastereomerically pure (S,S) and
(R,R) aminonitriles 5b and 7b derived from o-bromoben-
zaldehyde likewise proceeded with high stereoselectivity
(see Table 2) to provide diastereomerically pure pyrazi-
nones (aS,S) 9b and (aR,R) 11b. We suspected from the
above results that the nitrile-bearing carbon in the
aminonitrile was transferring its configuration to the
chiral axis in the pyrazinone. Indeed, cyclization of the
(S,R) bromoaminonitrile 6b (90% de; eq 5) afforded the
entry
Ar
aminonitrilea
% yield
a
a
b
b
b
b
c
d
d
e
e
f
2-iodophenyl
2-iodophenyl
5
7
5
6
7
8
7
5
7
5
7
7
5
7
5
7
7
5
7
8
81
54
56
2-bromophenyl
2-bromophenyl
2-bromophenyl
2-bromophenyl
2-chlorophenyl
2-fluorophenyl
2-fluorophenyl
2-methylphenyl
2-methylphenyl
2-trifluoromethyl
2-methoxyphenyl
2-methoxyphenyl
2,3-difluorophen
3-methylphenyl
1-naphthyl
15 (90% de)
52
14 (88% de)
52
60
62
54
40
48
g
g
h
i
27 (88% de)
23 (92% de)
46
33
j
54 (93% de)
48 (98% de)
51
k
k
k
1-thienyl
1-thienyl
1-thienyl
4 (95% de)
a
All aminonitriles were homogeneous by 1H NMR (300 MHz)
unless otherwise stated.
of oxalyl chloride in chlorobenzene at 70 °C for 14 h,
provided pyrazinones (aS,S) 9a and (aR,R) 11a respec-
tively, in 57% and 56% yields (eqs 3 and 4). Conducting
corresponding (aR,S) pyrazinone 10b
(12) Charton, M. J . Org. Chem. 1977, 42, 2528. Roussel, C.; Ayada,
D. Nouv. J . Chim. 1986, 10, 399.
(13) In the case of biphenyls, steric size of the ortho substituents as
measured by barriers to rotation match increased/decreased selectivity
in pyrazinone synthesis. For examples in the biphenyl series, see: Ling,
C. CH. K.; Harris, M. H. J . Chem. Soc. 1964, 825. Oki, M.; Yamamoto,
G. Bull. Chem. Soc. J pn. 1971, 44, 266. Adams, R. Rec. Prog. Chem.
1949, 91.
(14) The (aS,R) pyrazinone diastereomer 12a was obtained from the
cyclization of a 4:1 mixture of the aminonitriles 8a and 7a . That
reaction afforded a 4:1 mixture of the (aS,R) and (aR,R) pyrazinones
12a and 11a in 24% yield. Column chromatography was sufficient to
provide each diastereomer as a pure compound. The absolute stereo-
chemistry of the (aR,R) and (aS,R) diastereomers was confirmed
through single-crystal X-ray crystallography. Full details of the X-ray
data for 11a and 12a will be published elsewhere. Ball-and-stick
drawings of 11a and 12a are provided in the Supporting Information.
the reaction in a sealed tube (see Experimental Section)
gave a slight improvement in the yield. The relative and
absolute stereochemistry of the (aR,R) pyrazinone 11a

96 J . Org. Chem., Vol. 64, No. 1, 1999
Tulinsky et al.
Ta ble 2. Da ta for Atr op isom er ic P yr a zin on e Str u ctu r es
mixture of 11d /12d , indicating that partial racemization
had occurred. When similar experiments were carried out
with pyrazinones 11b, 11c, 11e, and 11g, no evidence of
racemization about the asymmetric axis was observed by
NMR. From that data, we concluded that although the
configuration about the asymmetric axis is determined
by the configuration at the nitrile-bearing carbon, the
selectivity is related to the steric bulk of the ortho
substituent. Racemization of the pyrazinone after forma-
tion plays little or no role in determining the final product
ratio in all but the o-F case.
9-12
X or
6-aryl
aminonitrile
(de)a
pyrazinone
(ratio)a
entry
% yieldb
a
a
b
b
b
b
c
d
d
e
e
f
g
g
h
i
j
k
k
I
I
5
7
5
9
11
9
57 (60)
56 (62)
64 (55)
73
68 (54)
43
67
70
62
71 (72)
28 (73)
33
61
78
68
59
45
48
Br
Br
Br
Br
Cl
F
F
CH3
CH3
CF3
OCH3
OCH3
2,3-F
3-CH3
1-naphthyl
1-thienyl
1-thienyl
6 (90%)
7
8 (88%)
7
5
7
5
7
10/9 (8:1)
11
12/11 (18:1)
11/12 (1.4:1)
9/10 (1:1)
11/12 (1:1)
9/10 (4:1)
11/12 (4:1)
11/12 (3.4:1)
9/10 (1.2:1)
11/12 (1:1)
9/10 (1:1)
11/12 (1:1)
11/12 (1:1)
9
All of the 6-aryl pyrazinones studied in which there is
an ortho substituent on the aromatic ring are conforma-
tionally locked at ambient temperature, including the
o-fluoro-substituted case. To help further define the limits
of conformational stability, we replaced the 6-aryl group
with an aromatic heterocycle, thiophene. The required
aminonitriles 5k and 7k were prepared using the normal
Strecker protocol and when submitted to the cyclization
conditions afforded the corresponding 6-thienyl pyrazi-
nones. However, unlike the ortho-substituted aryl cases,
each of the diastereomeric thiophene aminonitriles 5k
7
5 (88%)
7 (92%)
5
7
7
5
7
11
24
a
de’s and product ratios were determined by 1H NMR (300
MHz). If no de or ratio is listed, product was homogeneous by 1H
NMR. Aminonitriles were epimeric at the nitrile-bearing carbon.
1
and 7k gave a single pyrazinone product by H NMR,
b
Yields in parentheses are for reactions carried out in a sealed
9k and 11k , respectively, indicating free rotation about
tube vs a reaction vessel under a positive nitrogen pressure.
(78% de) in 73% yield, further supporting our notion that
the configuration about the chiral axis was indeed
determined by the configuration at the nitrile-bearing
carbon in the acyclic precursor and not by the chiral
auxiliary.
To continue to probe the mechanistic aspects of the
cyclization, we investigated the influence of the ortho aryl
substituent on the stereochemical course of the reaction.
As illustrated in Table 2, cyclization of the diastereo-
merically pure (R,R) aminonitrile 7c containing an o-Cl
substituent afforded a 1.4:1 mixture of pyrazinones
(aR,R) 11c and (aS,R) 12c. With the smaller o-F sub-
stituent (S,S) 5d , cyclization afforded an equimolar
mixture of pyrazinones (aS,S) 9d and (aR,S) 10d . This
was also true for the 2,3-difluoro aminonitrile example
5h (obtained diastereomerically pure from a single re-
crystallization), establishing the absence of a buttressing
effect in the case of o-fluoro substitution.12 Cyclization
of the o-CH3 and o-CF3 aminonitriles (R,R) 7e and (R,R)
7f afforded the corresponding pyrazinones in 4:1 (11e/
12e) and 3.4:1 (11f/12f) ratios, respectively. Aminonitrile
5g, with an o-OCH3 substituent, afforded an equimolar
mixture of pyrazinones 9g and 10g. Cyclization of the
naphthyl aminonitrile 7j also afforded a mixture of
diastereomeric pyrazinones (11j and 12j). Last, cycliza-
tion of the m-methylaryl aminonitrile 7i gave an equimo-
lar mixture of pyrazinones 11i and 12i, supporting the
obvious in that the major influence on the stereochemical
outcome of the reaction was indeed the ortho substituent
and that the influence was steric in nature. The above
trends are also consistent with those seen in the biphenyl
series.13
the aryl-heteroaryl axis in this case. Interestingly, the
thienyl-substituted pyrazinones exhibit 1H NMR char-
acteristics at ambient temperature indicating hindered
rotation on the NMR time scale. The H-1 proton signal
corresponding to the methine of the R-benzyl substituent
was broadened rather than existing as a sharp quartet.
As the temperature was increased, the resolution of
the methine improved and the sharpness of the quartet
improved. At lower temperatures, the methine signal
broadened as expected. The lowest temperature exam-
ined was 5 °C, and at that temperature, the signals have
not decoalesced into those of separate atropisomers.
While qualitative in nature, these results imply that the
thienyl-substituted pyrazinones are a borderline case in
which rotation about the biaryl bond is slow on the NMR
time scale but sufficiently unhindered to prevent physical
resolution of the individual atropisomers.
Reactions to form the pyrazinones from the aminoac-
etonitrile hydrochloride compounds were carried out with
excess oxalyl chloride using either chlorobenzene or
toluene as solvent. Saturation of the system with HCl is
essential for complete reaction. Intermediates produced
during the course of the reaction were too short-lived to
be isolated or characterized. For this reason, we employed
computational methods as a means of gaining insight into
the involvement of HCl and the mechanistic basis for
stereoselectivity. Because the low dielectric constant of
the solvent rules out ionic mechanisms, a series of
reactions were investigated in which molecular HCl
serves as a catalyst. On the basis of the results of those
calculations, which are discussed in some detail hereaf-
ter, we propose the following mechanism to explain the
stereospecific pyrazinone formation and chirality trans-
fer. As shown in Scheme 3, initial acylation of aminoni-
trile 7 with oxalyl chloride provides the oxoamide inter-
To examine the possibility that racemization of the
pyrazinones was occurring after their formation, the
o-fluoro-substituted pyrazinones were examined more
closely. Resolution of 11d and 12d was readily achieved
via preparative HPLC. A solution of 11d in toluene-d8
was heated at 70-75 °C for 20 h, conditions that
1
thermally mimic the cyclization protocol. The H NMR
of the solution after 20 h at 70-75 °C showed a 7/3

Synthesis of Atropisomeric 6-Aryl Pyrazinones
J . Org. Chem., Vol. 64, No. 1, 1999 97
Sch em e 3. P r op osed Mech a n ism for P in n er Sa lt F or m a tion
a
Sch em e 4
a
Barriers in kcal mol-1 are indicated by numbers on the arrows for the transition.
mediate 15. Addition of HCl across the nitrile triple bond
then forms Pinner salt 16,16 which induces two distinct
modes of cyclization leading to either 17b or 17b′ in the
case where X is bromine, as illustrated in Scheme 4.
Addition of HCl to either 17b or 17b′ is believed to give
rise to the corresponding gem-dichloro intermediates 18b
and 18b′.17 Elimination of HCl from those intermediates
then yields diones 19b and/or 20b. Thereafter, reaction
of the diones with oxalyl chloride yields the final prod-
ucts. To gain further insight into stereochemical control
of the reaction, it is necessary to examine the dynamics
between the intermediates that fall on the reaction
coordinate from 16b to 19b or 20b.
yield intermediate 18b/18b′, and finally the process
whereby HCl is lost from 18b/18b′ to yield 19b and/or
20b. Although this method yields highly approximate
values of potential-energy barriers, trends in barrier
height often provide useful insights into the underlying
mechanisms in a series of chemical reactions. In the
initial series of calculations, we wished to determine
which step exerts the greatest control over the final
stereochemical outcome.
Scans of the torsional angle θ governing rotation about
the biaryl bond in 19b (Figure 1) revealed barriers
exceeding 30 kcal mol-1 in the potential-energy curve,
V(θ). In light of these high values, atropisomerism is
clearly established in this intermediate leading to 11b,
and the final step involving the reaction of 19b with
oxalyl chloride need not be further considered. Scheme
4 illustrates two reaction pathways and gives the calcu-
lated barriers for transitions leading to either 19b or 20b.
The unaided cyclization (16 f 17), addition (17 f 18),
and elimination (18 f 19) reactions involve four-electron
Hu¨ckel transition states with extremely high barriers.18
In contrast, the transition-state complexes shown in
Figure 2 are each stabilized by a catalytic HCl molecule
in a six-electron pericyclic process. The magnitude of the
reaction barriers shown in Scheme 4 gives qualitative
insight as to which step controls the stereochemistry.
By means of AM1 molecular orbital calculations, we
examined the cyclization step involving the conversion
of 16b to 17b/17b′, the addition of HCl to 17b/17b′ to
(15) NOE experiments revealed a characteristic interaction in each
diastereomer between the o-hydrogen of the aryl ring and the methine
or methyl hydrogens of the I-methylbenzyl group. This observation
proved to be a convenient means of assigning the stereochemistry of
the diastereomeric pyrazinones.
(16) A simple model for Pinner salt formation using HCl attack on
acetonitrile yields a very high barrier (56.6 kcal mol-1). A much lower
barrier (26.3 kcal mol-1) is found using a second HCl molecule as
catalyst. The calculations suggest that the reaction may avoid the high
entropic cost of a tertiary collision by proceeding through two binary
steps where an acetonitrile HCl complex serves as intermediate. In
any case, Pinner salt formation is suggested to be the rate-determining
step of the reaction on the basis of simulation of the kinetics of the
system (see Supporting Information) using the CHEMKIN II program
(Kee, R. J .; Rupley, F. M.; Miller, J . A. CHEMKIN II: A Fortran
Package for the Analysis of Gas Phase Chemical Kinetics; Sandia
National Laboratories, Livermore, CA, report SAND 89-8009). Al-
though CHEMKIN II has been designed for gas-phase kinetics, it may
be appropriately used in constant-temperature studies with rate
constants determined for condensed fluids.
(17) The calculated barriers exceed 50 kcal mol-1 for HCl attack on
17b/17b′ to give 19b/19b′ in a direct tautomerization reaction, whereas
those for HCl addition to 17b/17b′ are less than 20 kcal mol-1. As a
result, the low-energy pathways lead to production of the gem dichloro
intermediates 18b/18b′.
(18) Simons, J . Energetic Principles of Chemical Reactions; J ones
and Bartlett: Boston, 1983; pp 65-68.

98 J . Org. Chem., Vol. 64, No. 1, 1999
Tulinsky et al.
F igu r e 1. Rotation about the biaryl bond of 19b leads to notable distortion of the pyrazino ring as Br sweeps past the
R-methylbenzyl and chlorine substituents when θ is near 0° and 180°, respectively. A sharp drop in energy occurs as the system
relaxes when the bulky ortho substituents on each ring push beyond one another. The shape of the barrier is dependent upon the
direction of rotation, as shown by the solid (increasing θ) and dashed (decreasing θ) curves.
of the catalytic HCl molecule in the transition-state
complex formed from 18b′. V(θ) exhibits a barrier to
rotation between 18b′ and 18b that is approximately 11
kcal mol-1 smaller than the reaction barrier. This evi-
dence suggests that rotation of the o-bromophenyl ring
occurs much more rapidly than elimination of HCl. For
this reason, a near-equilibrium population of the more
reactive 18b rotamer will be maintained during the
course of the reaction. As a result, these AM1 calculations
predict that kinetic factors in the 18b′ f 19b step govern
the production of the observed o-Br substituted atropi-
somer.
In this mechanism, the kinetics of ring closure are
adversely affected by factors related to the entropy of
activation because rotational freedom of three single
bonds is sacrificed in the formation of the transition-state
complexes. As the high entropy of activation in this step
slows the overall reaction, it undoubtedly contributes to
the lack of observable concentrations of dione intermedi-
ates, which appear to react as rapidly as they are formed.
Catalysis in the HCl-addition step requires two HCl
molecules, and this appears to involve an entropically
unfavorable tertiary collision. However, the calculations
reveal that the reaction is most likely a binary process
F igu r e 2. Transition-state complexes in the favored pathways
for the reactions: TS1 corresponds to the cylclization step 16b
involving HCl attack on the stable complex 17‚HCl. A
kinetic simulation of the reaction taking into account
these entropic factors and scaled AM1 barriers is pre-
sented as Supporting Information. The simulation sup-
ports the proposed mechanism by providing quantitative
estimates of reactant, intermediate, and product concen-
trations that are consistent with experimental observa-
tions.
f 17b; TS2 is associated with the HCl-addition step 17b f
18b; and TS3 corresponds to the HCl-elimination step 18b f
19b. Each reaction is catalyzed by HCl. Arrows represent
motion of atoms along the reaction coordinate leading to
products.
Values of the barriers are 14.8 kcal mol-1 for cyclization
of 16b, 17.3 kcal mol-1 for HCl addition to 17b′, and 34.1
kcal mol-1 for HCl elimination from 18b′. Every step is
predicted to be strongly exothermic. Scans of θ indicate
that rotamers 17b′ and 18b′ in pathway B are actually
favored in both cases. Furthermore, the low-energy
barriers associated with transition-state complexes TS1
and TS2 lead to formation of the conformers in pathway
B that ultimately would produce intermediate 20b. On
the other hand, passage of the HCl-elimination barriers
favors the route through TS3 toward 19b over the one
toward 20b by more than 4 kcal mol-1. The bulk of the
o-Br ring substituent clearly interferes with the approach
The conclusions drawn from the o-Br case were further
examined by performing calculations on the 18 f 19 HCl-
elimination step in the reaction sequences initiated with
the (R,R) 7 stereoisomer of other aminonitriles in the
series under consideration. The difference in the pyrazi-
none ratio obtained in the reaction undertaken with 8b
was also investigated. The results of these calculations
are summarized in Table 3. In those cases where a
dominant atropisomer was observed, the calculations do
give a correct prediction based on the difference in the
2
value of ∆Hf (18 f 19) for the two pathways. The small

Synthesis of Atropisomeric 6-Aryl Pyrazinones
J . Org. Chem., Vol. 64, No. 1, 1999 99
Ta ble 3. AM1 Da ta a for Ster eo-Con tr ollin g Step in
Syn th esis of Atr op isom er ic P yr a zin on es
interaction is reduced or eliminated, and mixtures of
products are obtained. This model is consistent with the
observed trends in selectivity. The selectivity is appar-
ently based on the steric bulk and partial electronic
charge of the ortho substituent, both of which impact the
energy of the transtition-state complex. The electronic
nature of the aromatic ring appears to have little if any
effect on the observed ratio of atropisomers.
2 b
c
∆Hf
19
∆Hf
aminonitrile
reactant
kinetic
preference
20
19
20
31.2
19.1
19.1
7.2
7a
7b
8b
7c
7d
7e
7f
7i
31.6
29.8
35.6
29.6
30.0
31.4
34.6
29.9
30.5
30.6
35.2
34.1
30.1
33.4
33.5
31.3
37.8
30.0
31.0
30.2
31.9
19.9
19.9
11a . 12a
11b . 12b
12b . 11b
11c . 12c
11d . 12d
11e ∼ 12e
11f . 12f
11i ∼ 12i
7.6
-29.9
-30.6
6.1
5.7
Exp er im en ta l Section
-136.8
4.6
-138.5
4.6
Gen er a l Meth od s. Solvents were the anhydrous grade
purchased from Aldrich Chemical Co., Milwaukee, WI and
were dispensed using standard syringe techniques. All re-
agents were used as received unless otherwise stated. All
nonaqueous reactions were performed in oven-dried (130 °C
for >24 h) glassware under a nitrogen atmosphere unless
otherwise stated. 1H and 13C NMR spectra were recorded on a
Bruker AC 300 NMR spectrometer; infrared spectra, mass
spectra, high-resolution mass spectra, and combustion analy-
ses were performed by the Physical and Analytical Chemistry
unit of Upjohn Laboratories.
In th e p r ep a r a tion of (a R,R)-6-(2-r-ch lor op h en yl)-3-
[(1-p h en yleth yl)a m in o]-5-ch lor o-1-(1-fu r ylm eth yl)-2(1H)-
p yr a zin on e (4), a solution of pyrazinone 1a (372 mg, 1.05
mmol), (R)-(+)-R-methylbenzylamine (0.40 mL, 3.10 mmol),
and Et3N (0.70 mL, 5.02 mmol) in CH2Cl2 (5.0 mL) was stirred
at ambient temperature. After 72 h, the reaction mixture was
diluted with CH2Cl2 (50 mL) and washed with 1% aqueous
HCl (2 × 50 mL) and H2O (50 mL). The CH2Cl2 extract was
dried (MgSO4), filtered, and concentrated in vacuo. The residue
was purified via flash chromatography on silica (eluent, CHCl3/
hexanes 95:5), which gave a colorless syrup. Recrystallization
(EtOAc/hexanes) gave the title compound (457 mg, 99%) as
white crystals.
Th e syn th esis of (R,R)-r-[(1-p h en yleth yl)a m in o]-r-(2-
br om op h en yl)a ceton itr ile h yd r och lor id e (7b) is repre-
sentative of the procedure used for the synthesis of aminoni-
triles5-8listedin Table1.(R)-R-Methylbenzylaminehydrochloride
(12.17 g, 77.20 mmol) and sodium cyanide (3.79 g, 77.33 mmol)
were dissolved in H2O (80 mL). MeOH (80 mL) and 2-bro-
mobenzaldehyde (9.1 mL, 77.95 mmol) were added, and the
mixture was stirred for 16 h. The reaction mixture was diluted
with H2O (ca. 200 mL) and extracted with CH2Cl2 (3 × 100
mL). The combined CH2Cl2 extracts were dried (Na2SO4),
filtered, and concentrated in vacuo to give an off-white solid.
1H NMR (300 MHz, CDCl3) of this crude material shows it to
be a 4.4:1 mixture of diastereomers. The crude solid was
dissolved in Et2O (ca. 300 mL). The addition of methanolic HCl
(40 mL; ca. 2 M) gave 7b as a white crystalline solid (14.008
g, 52%).
7j
7k
32.8
17.7
32.4
17.5
11j ∼ 12j
11k ∼ 12k d
Energies are expressed in kcal mol-1
.
∆Hf denotes a calcu-
a
b
2
lated barrier for acid-catalyzed elimination of HCl from 18. As
indicated in the two subcolumns, the magnitude of this barrier
differs in the reaction pathways leading to products 19 and 20.
The predicted kinetic preference is based on the calculated product
ratio given by 11:12 ) 19:20 ≈ exp(∆∆Hf2/RT) where ∆∆Hf
)
2
2
2
∆Hf (20) - ∆Hf (19). c ∆Hf is the calculated AM1 heat of formation.
d
The calculated torsional barriers about the biaryl bond of 19k
are <9 kcal mol-1, which is in accordance with the observed
rotational effects on an NMR time scale.
barrier differences obtained in cases 7e and 7i-k are also
in accord with the lack of strongly dominant atropiso-
meric products. In the case of the 1-naphthyl substituent
(7j), the near equality of the barriers indicates why a 1:1
ratio of enantiomers is found even though racemization
of the individual atropisomers is predicted to be very
slow. However, the prediction that 11 . 12 in the cases
involving o-Cl (7c), o-F (7d ), and o-CF3 (7f) is clearly out-
of-line with experiment. Apparently, the AM1 procedure
exaggerates barrier differences for halogen-substituted
molecules, especially F and Cl.
Another factor that must be considered to account for
the observed product ratios is the rate of racemization
of 19 following reaction. From differences in the lowest
barrier of the potential curve V(θ) separating 19 from 20,
the substituent effect on the calculated half-life for the
racemization reaction exhibits the trend o-CF3 > o-I >
o-Br > 1-naphthyl > o-CH3 > o-Cl > o-F . m-CH3
.
1-thienyl. The calculated rates suggest that rotational
isomerization occurs rapidly under high-temperature
reaction conditions when X ) o-F. In comparison, the rate
is 20-fold slower when X ) o-Cl. The observed partial
racemization of 11d and failure to racemize 11c under
experimental conditions of time and temperature is in
qualitative accord with the calculated findings. Racem-
ization of the o-F compound alters the ratio in the
direction of the thermodynamically favored product 12d ,
Th e syn th esis of (S,R)-r-[(1-p h en yleth yl)a m in o]-r-(2-
br om op h en yl)a ceton itr ile h yd r och lor id e (6b) is repre-
sentative of the procedure used for the synthesis of 5a ,b,d ,e-
,g,h ,k and 6b. (S)-R-Methylbenzylamine hydrochloride (20.25
g, 128.46 mmol) and sodium cyanide (6.60 g, 134.67 mmol)
were dissolved in H2O (130 mL). MeOH (130 mL) and 2-bro-
mobenzaldehyde (15.0 mL, 128.49 mmol) were added, and the
mixture was stirred for ca. 16 h. The reaction mixture was
diluted with H2O (ca. 500 mL), and the resulting solid was
collected via filtration. It was dissolved in a minimum amount
of warm PhCH3 and concentrated in vacuo to give a yellow
oil. Trituration of this material with hexanes gave a white solid
which was collected via filtration. 1H NMR (300 MHz, CDCl3)
of this material shows it to be the free base of 6b. The filtrate
2
as observed, regardless of the actual value of ∆∆Hf .
Con clu sion
If the o-substituent is large (i.e., iodo or bromo) and/or
generates a negative electrostatic potential, an unfavor-
able interaction may occur in the transition state of the
step that controls stereoselectivity, namely, the one
leading from 18 to 19/20. The steric and electronic
repulsion associated with the 18 f 20 step involves the
chlorine atom of the catalytic HCl molecule as it extracts
the sp3 hydrogen from the carbon bearing the ortho-
substituted aromatic ring. Lack of this transition-state
repulsion favors the 18 f 19 reaction, which leads to the
observed atropisomer. With smaller groups and those
possessing a significant positive partial charge, this
1
was then concentrated in vacuo to give a yellow oil. H NMR
(300 MHz, CDCl3) shows this material to be a 5:1 mixture of
6b and 5b. The oil was taken up in Et2O (ca. 100 mL). The
addition of methanolic HCl (10 mL, ca. 2 M) gave 6b as a white
crystalline solid (6.614 g, 15%).
Th e syn th esis of (a S,S)-6-(2-r-iodoph en yl)-3,5-dich lor o-
1-(1-p h en yleth yl)-2(1H)-p yr a zin on e (9a ) is representative
of the cyclization process using chlorobenzene as the solvent.

100 J . Org. Chem., Vol. 64, No. 1, 1999
Tulinsky et al.
Aminonitrile 5a (10.145 g, 28.01 mmol) was added to a solution
of oxalyl chloride (7.33 mL, 84.02 mmol) in chlorobenzene (140
mL), and the resulting solution was heated at 70 °C for 14 h.
The reaction was cooled to room temperature and concentrated
in vacuo, and the resulting residue was purified by silica gel
flash chromatography (CHCl3) to afford 7.515 g (57%) of pure
9a . An alternate procedure which offered improved yields
follows. A heavy-walled glass tube was charged with the
aminonitrile hydrochloride 5a (402 mg, 1.01 mmol), (COCl)2,
(0.30 mL, 3.4 mmol), and PhCH3 (5 mL). The tube was sealed
and the reaction mixture was heated (bath temp 70-75 °C)
for 23 h. Upon cooling to ambient temperature, the reaction
mixture was concentrated in vacuo, and the residue was
purified via flash chromatography on silica (eluant CHCl3).
Recrystrallization (EtOAc/hexanes) gave 9a as a white crystal-
line solid (271 mg, 60%).
carried out by scanning P in 15- increments while allowing
all other angles in the molecule to relax. In locating stationary
points on the potential-energy surface of each reaction, all
degrees of freedom were relaxed until the gradient norm fell
below 3.0. Transition-state complexes were established by
diagonalizing the calculated Hessian matrix and checking for
a single negative eigenvalue.20
Su p p or tin g In for m a tion Ava ila ble: X-ray crystallogr-
phic structures of 4, 11a , and 12a (including tables of X-ray
crystallographic data; spectroscopic and elemental analyses
of the compounds synthesized in the course of the study; and
simulation of the reaction kinetics in the system 15b + HCl
f ... f 19b using data from the AM1 calculations and absolute
reaction rate theory to estimate rate constants (41 pages). This
material is contained in libraries on microfiche, immediately
follows this article in the microfilm version of the journal, and
can be ordered from the ACS; see any current masthead page
for ordering information.
Com p u ta tion a l Meth od s. AM1 calculations were per-
formed at the Hartree-Fock level using the MOPAC 6.0
program.19 Computation of the potential-energy function, V(P),
corresponding to rotation of the substituted phenyl ring, was
J O9811930
(20) McIver, J . W.; Kormornicki, A. J . Am. Chem. Soc. 1972, 94,
2980.
(19) Stewart, J . J . P. J . Comput. Aided Mol. Des. 1990, 4, 1.