Stereoelectronic control in addition of nucleophiles to an amidinium ion

Article abstract of DOI:10.1021/ja004240q

Nucleophilic addition to 1,3-dimethyl-5-phenyl-1,4,5,6-tetrahydropyrimidinium ion provides a quantitative measure of stereoelectronic control. This amidinium ion presents the nucleophile with two distinct paths for attack. Axial attack is favored by interaction between the orbital of the developing bond and antiperiplanar lone pairs on the nitrogens. Reaction of the amidinium salt with diverse nucleophiles (D^-, H₃C^-, n-Bu^-, PhCH₂^-, allyl^-, Ph^-, C₅F₆-, CH₂=CH^-, HC≡C^-, PhC≡C^-,CN^-) produces mixtures of cis and trans stereoisomers. Both kinetic and thermodynamic product distributions were measured by ¹H NMR, before and after acid-catalyzed equilibration. The values provide insight into the roles of steric and stereoelectronic forces at the transition state and in products. Stereoelectronic effects on reactivity are found to be weak (ca. 1 kcal/mol).

Full text of DOI:10.1021/ja004240q

J. Am. Chem. Soc. 2001, 123, 4451-4458
4451
Stereoelectronic Control in Addition of Nucleophiles to an
Amidinium Ion
Charles L. Perrin* and David B. Young
Contribution from the Department of Chemistry, UniVersity of California-San Diego,
La Jolla, California 92093-0358
ReceiVed December 11, 2000
Abstract: Nucleophilic addition to 1,3-dimethyl-5-phenyl-1,4,5,6-tetrahydropyrimidinium ion provides a
quantitative measure of stereoelectronic control. This amidinium ion presents the nucleophile with two distinct
paths for attack. Axial attack is favored by interaction between the orbital of the developing bond and
antiperiplanar lone pairs on the nitrogens. Reaction of the amidinium salt with diverse nucleophiles (D-, H3C-,
n-Bu-, PhCH2-, allyl-, Ph-, C5F6-, CH2dCH-, HCtC-, PhCtC-, CN-) produces mixtures of cis and trans
stereoisomers. Both kinetic and thermodynamic product distributions were measured by 1H NMR, before and
after acid-catalyzed equilibration. The values provide insight into the roles of steric and stereoelectronic forces
at the transition state and in products. Stereoelectronic effects on reactivity are found to be weak (ca. 1 kcal/
mol).
Introduction
intermediates by Deslongchamps, who hypothesized that cleav-
age is favored by two lone pairs antiperiplanar to the leaving
group.3
Antiperiplanar Lone Pairs. Stereoelectronic control, arising
from the positioning of lone pairs, is a topic of much current
interest.1 The aspect relevant to thermodynamic stabilities is
known as the anomeric effect.2 In connection with reactivity it
was first applied at the acetal level of oxidation, where an
antiperiplanar lone pair is expected to facilitate bond cleavage.3
This preference, often called the antiperiplanar lone-pair hy-
pothesis (ALPH) or the kinetic anomeric effect, is supported
by calculations.4 Yet experimental evidence is weak or elusive,5
except that an orthogonal lone pair is indisputably less effective
than a periplanar one.6 The ALPH was extended to tetrahedral
The role of antiperiplanar lone pairs is a fundamental aspect
of the dependence of chemical reactivity on structural features.
It is still an area of considerable uncertainty and controversy,7
with wide acceptance8 and only occasional skepticism.9
To what extent does an antiperiplanar lone pair facilitate
cleavage of a tetrahedral intermediate? Early evidence came
from hydrolysis of a cyclic hemiorthoester, which opens to
hydroxy ester, rather than lactone.10 Nevertheless, a serious
inconsistency is that the five-membered-ring hemiorthoester also
gives hydroxy ester, even though it ought to have cleaved to
(1) Kirby, A. J. The Anomeric and Related Stereoelectronic Effects at
Oxygen; Springer-Verlag: Berlin, 1983. Juaristi, E.; Cuevas, G. Tetrahedron
1992, 48, 5019. Thatcher, G. R. J., Ed.; The Anomeric Effect and Associated
Stereoelectronic Effects; ACS Symposium Series, American Chemical
Society: Washington, DC, 1993. Thibaudeau, C.; Chattopadhyaya, J.
Stereoelectronic Effects in Nucleosides and Nucleotides and their Structural
Implications; Uppsala University Press: 1999.
(5) Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Woerpel, K. A. J. Am.
Chem. Soc. 1999, 121, 12208. Dios, A.; Nativi, C.; Capozzi, G.; Franck,
R. W. Eur. J. Org. Chem. 1999, 1869. Zhu, J.; Bennet, A. J. J. Am. Chem.
Soc. 1998, 120, 3887. Moreau, C.; Lecomte, J.; Mseddi, S.; Zmimita, N. J.
Mol. Catal. A 1997, 125, 143. Bellucci, G.; Chiappe, C.; D’Andrea, F.; Lo
Moro, G. Tetrahedron 1997, 53, 3417.
(2) Alabugin, I. V. J. Org. Chem. 2000, 65, 3910. Box, V. G. S. J. Mol.
Struct. 2000, 522, 145. Carballeira, L.; Perez-Juste, I. J. Comput. Chem.
2000, 21, 462. Anderson, J. E. J. Org. Chem. 2000, 65, 748. Randell, K.
D.; Johnston, B. D.; Green, D. F.; Pinto, B. M. J. Org. Chem. 2000, 65,
220. Perrin, C. L.; Fabian, M. A.; Brunckova, J.; Ohta, B. K. J. Am. Chem.
Soc. 1999, 121, 6911. Uehara, F.; Sato, M.; Kaneko, C.; Kurihara, H. J.
Org. Chem. 1999, 64, 1436. Juaristi, E.; Cuevas, G. Tetrahedron 1999, 55,
359. Alber, F.; Folkers, G.; Carloni, P. J. Phys. Chem. B 1999, 103, 6121.
Mo, Y.; Zhang, Y.; Gao, J. J. Am. Chem. Soc. 1999, 121, 5737. Kirby, A.
J.; Komarov, I. V.; Wothers, P. D.; Feeder, N.; Jones, P. G. Pure Appl.
Chem. 1999, 71, 385. Verevkin, S. P.; Peng, W. H.; Beckhaus, H. D.;
Ru¨chardt, C. Eur. J. Org. Chem. 1998, 2323. Jones, P. G.; Kirby, A. J.;
Komarov, I. V.; Wothers, P. D. J. Chem. Soc., Chem. Commun. 1998, 1695.
Barrows, S. E.; Storer, J. W.; Cramer, C. J.; French, A. D.; Truhlar, D. G.
J. Comput. Chem. 1998, 19, 1111. Tvaroska, I.; Carver, J. P. Carbohydr.
Res. 1998, 309, 1. Lenz, R.; Ley, S. V.; Owen, D. R.; Warriner, S. L.
Tetrahedron: Asymmetry 1998, 9, 2471. Anderson, J. E.; Cai, J.; Davies,
A. G. J. Chem. Soc., Perkin Trans. 2 1997, 2633. Ganguly, B.; Fuchs, B.
J. Org. Chem. 1997, 62, 8892. Buckley, N.; Oppenheimer, N. J. J. Org.
Chem. 1996, 61, 8039.
(6) Briggs, A. J.; Evans, C. M.; Glenn, R.; Kirby, A. J. J. Chem. Soc.,
Perkin Trans. 2 1983, 1637.
(7) Sinnott, M. L. AdV. Phys. Org. Chem. 1988, 24, 113. Sinnott, M. L.
In The Anomeric Effect and Associated Stereoelectronic Effects; Thatcher,
G. R. J., Ed.; ACS Symposium Series, American Chemical Society:
Washington, DC, 1993; Chapter 6. Thatcher, G. R. J.; Krol, E. S.; Cameron,
D. R. J. Chem. Soc., Perkin Trans. 2 1994, 683. Uchimaru, T.; Tsuzuki,
S.; Storer, J. W.; Tanabe, K.; Taira, K. J. Org. Chem. 1994, 59, 1835.
(8) Huber, R.; Vasella, A. Tetrahedron 1990, 46, 33. Urones, J. G.;
Marcos, I. S.; Basabe, P.; Sexmero, J.; Diez, D.; Garrido, N. M.; Prieto, J.
E. S. Tetrahedron 1990, 46, 2495. Messmer, A.; Hajo´s, G.; Tima´ri, G.
Tetrahedron 1992, 48, 8451. Brace, N. O. J. Org. Chem. 1993, 58, 1804.
Maligres, P. E.; Weissman, S. A.; Upadhyay, V.; Cianciosi, S. J.; Reamer,
R. A.; Purick, R. M.; Sager, J.; Rossen, K.; Eng, K. K.; Askin, D.; Volante,
R. P.; Reider, P. J. Tetrahedron 1996, 52, 3327. Loeppky, R. N.; Cui, W.
Tetrahedron Lett. 1998, 39, 1845. Berges, D. A.; Fan, J.; Devinck, S.;
Mower, K. J. Org. Chem. 2000, 65, 889.
(9) Perrin, C. L.; Engler, R. E.; Young, D. B. J. Am. Chem. Soc. 2000,
122, 4877. Graczyk, P. P.; Mikolajczyk, M. J. Org. Chem. 1996, 61, 2995.
Wipf, P.; Kim, Y. J. Am. Chem. Soc. 1994, 116, 11678. Brown, R. S.;
Bennet, A. J.; SÄlebocka-Tilk, H. Acc. Chem. Res. 1992, 25, 481. Brown,
R. S.; Bennet, A. J.; SÄlebocka-Tilk, H.; Jodhan, A. J. Am. Chem. Soc. 1992,
114, 3092. Caserio, M. C.; Shih, P.; Fisher, C. L. J. Org. Chem. 1991, 56,
5517. Agami, C.; Couty, F.; Prince, B.; Puchot, C. Tetrahedron 1991, 47,
4343. Bennet, A. J.; Slebocka-Tilk, H.; Brown, R. S.; Guthrie, J. P.; Jodhan,
A. J. Am. Chem. Soc. 1990, 112, 8497. Clennan, E. L.; L’Esperance, R. P.;
Lewis, K. K. J. Org. Chem. 1986, 51, 1440.
(3) Deslongchamps, P. Stereoelectronic Effects in Organic Chemistry;
Pergamon: Oxford, 1983. Deslongchamps, P. Tetrahedron 1975, 31, 2463.
(4) Lehn, J. M.; Wipff, G. J. Am. Chem. Soc. 1974, 96, 4048. Lehn,
J.-M.; Wipff, G. HelV. Chim. Acta 1978, 61, 1274. Faˇrcas¸iu, D.; Horsley,
J. A. J. Am. Chem. Soc. 1980, 102, 4906. Pullumbi, P.; Lemeune, S.; Barbe,
J.-M.; Trichet, A.; Guilard, R. THEOCHEM (J. Mol. Struct.) 1998, 432,
169.
10.1021/ja004240q CCC: $20.00 © 2001 American Chemical Society
Published on Web 04/20/2001

4452 J. Am. Chem. Soc., Vol. 123, No. 19, 2001
Perrin and Young
lactone. Therefore, it was proposed that the absence of lactones
could be attributed to their destabilization, independently of
ALPH.11
hindrance is the most common device for directing an incoming
nucleophile along a path chosen to create a chiral center
selectively. However, an alternative is to direct the nucleophile
to a position antiperiplanar to a lone pair.19 In some cases the
high stereoselectivity observed (>99:1) implies a strong influ-
ence (>3 kcal/mol). It is important to understand the circum-
stances under which an antiperiplanar lone pair governs
stereoselectivity.
Another System for Assessing ALPH. In the hydrolysis of
cyclic amidines there are two possible steps where stereoelec-
tronic control may operate. One is the cleavage of the hemi-
orthoamide intermediate. The other is the addition of hydroxide
to form that intermediate. If cleavage is favored by two lone
pairs antiperiplanar to the leaving group, then it follows from
the principle of microscopic reversibility that addition is favored
if the hydroxide enters antiperiplanar to the lone pairs. The
counter-ALPH results could be due to the involvement of syn
lone pairs in either step or in both. We have proposed that
stereoelectronic control is reduced in the cleavage step because
a lone pair on the oxyanion provides so strong a push that a
syn lone pair on the nitrogen is adequate.13 However, we would
like to separate the two steps.
To reduce ambiguity, hydrolysis of cyclic amidines was
studied. Reaction proceeds via a hemiorthoamide, with 1 as
initial conformer. After rotation about the exocyclic C-N bond,
two lone pairs are antiperiplanar to the endocyclic C-N, which
can cleave to the aminoamide (2). Cleavage of the exocyclic
C-N and formation of the lactam (3) could utilize an anti-
periplanar lone pair on O but would require the syn lone pair
on the ring N. Although ring inversion of 1 can produce
conformer 4, this too can cleave only the endocyclic C-N.
Conformers that could cleave to the lactam are inaccessible
because their formation requires nitrogen inversion, which is
slow. Thus, if ALPH holds, aminoamide 2 (n ) 5,6,7) is
predicted to be the kinetic product. Yet when leaving abilities
are balanced,12 five- and seven-membered rings produce con-
siderable lactam 3, and even in six-membered rings stereoelec-
tronic control is weak.13 Similar results were obtained in
hydrolysis of cyclic guanidines.14
As part of an ongoing effort to assess ALPH, we here examine
the stereochemistry of nucleophilic addition to an amidinium
ion, 1,3-dimethyl-5-phenyl-1,4,5,6-tetrahydropyrimidinium ion
(5). This study isolates the addition step and allows an
independent assessment of its stereoselectivity. Moreover, in
contrast to five- and seven-membered rings stereoelectronic
control is expected to be stronger in this six-membered ring.
One explanation for these results is the involvement of a syn
lone pair,13 which is supported by some computations,15 as well
as by several other counterexamples to ALPH.16 Although the
proponents of ALPH have accepted synperiplanar lone pairs in
some rigid acetals with obligatory eclipsing,17 they reject a more
general role for syn lone pairs in conformationally flexible
systems or in tetrahedral intermediates.18
Much of the interest in ALPH is for purposes of synthesis,
where it offers a novel method to control stereochemistry. Steric
This amidinium ion is expected to adopt conformation 5-eq,
since the conjugated N-C-N system constrains the ring to a
half-boat coplanar with both methyl groups, and since the phenyl
will be preferentially pseudoequatorial, owing to steric effects.
There are two possible addition paths, depending on whether
the nucleophile enters syn or anti to the developing lone pairs
on the nitrogens. According to ALPH, the latter path, leading
to cis product, with the nucleophile axial, is favored by
interactions between nitrogen lone pairs and the orbital of the
incipient bond. The two transition states are shown in Figure
1. The product ratio then provides a measure of the relative
stabilities of these two transition states, subject to the require-
ment that the products not interconvert under the reaction
conditions.
(10) Deslongchamps, P.; Atlani, P.; Fre´hel, D.; Malaval, A.; Moreau,
C. Can. J. Chem. 1974, 52, 3651. Deslongchamps, P.; Cheˆnevert, R.;
Taillefer, R. J.; Moreau, C.; Saunders, J. K. Can. J. Chem. 1975, 53, 1601.
Deslongchamps, P.; Duhe´m S.; Lebreaux, C.; Patterson, D. R.; Taillefer,
R. J. Can. J. Chem. 1975, 53, 2791. Deslongchamps, P.; Lessard, J.; Nadeau,
Y. Can. J. Chem. 1985, 63, 2485.
(11) Perrin, C. L.; Arrhenius, G. M. L. J. Am. Chem. Soc. 1982, 104,
2839.
(12) Burdick, B. A.; Benkovic, P. A.; Benkovic, S. J. J. Am. Chem. Soc.
1977, 99, 5716. Perrin, C. L.; Nun˜ez, O. J. Am. Chem. Soc. 1987, 109,
522.
(13) Perrin, C. L.; Nun˜ez, O. J. Am. Chem. Soc. 1986, 108, 5997. Perrin,
C. L.; Thoburn, J. D. J. Am. Chem. Soc. 1993, 115, 3140.
(14) Perrin, C. L.; Young, D. B. J. Am. Chem. Soc. 2001, 123. 4446-
4450.
(15) Ratcliffe, A. J.; Mootoo, D. R.; Andrews, C. W.; Fraser-Reid, B. J.
Am. Chem. Soc. 1989, 111, 7661. Andrews, C. W.; Bowen, J. P.; Fraser-
Reid, B. J. Chem. Soc., Chem. Commun. 1989, 1913. Wilson, B. G.; Fraser-
Reid, B. J. Org. Chem. 1995, 60, 317.
A preference for cis product is counter-thermodynamic, since
the axial substituent suffers destabilizing steric interactions. An
independent experimental determination of the thermodynamics
is available through equilibration and measurement of the cis:
trans ratio.
(16) Somayaji, V.; Brown, R. S. J. Org. Chem. 1986, 51, 2676.
Konstantinidis, A.; Sinnott, M. L. Biochem. J. 1991, 279, 587. Perrin, C.
L.; Engler, R. E. J. Am. Chem. Soc. 1997, 119, 585.
(17) Deslongchamps, P.; Jones, P. G.; Li, S.; Kirby, A. J.; Kuusela, S.;
Ma, Y. J. Chem. Soc., Perkin Trans. 2 1997, 2621. Li, S.; Kirby, A. J.;
Deslongchamps, P. Tetrahedron Lett. 1993, 34, 7757. Li, S.; Deslong-
champs, P. Tetrahedron Lett. 1993, 34, 7759.
(18) Pothier, N.; Goldstein, S.; Deslongchamps, P. HelV. Chim. Acta
1992, 75, 604. Deslongchamps, P. Pure Appl. Chem. 1993, 65, 1161.
Deslongchamps, P. In The Anomeric Effect and Associated Stereoelectronic
Effects; Thatcher, G. R. J., Ed.; ACS Symposium Series; American Chemical
Society: Washington, DC, 1993; Chapter 3. Deslongchamps, P.; Dory, Y.
L.; Li, S. Heterocycles 1996, 42, 617. Dugas, H. Bioorganic Chemistry,
2nd ed.; Springer: New York, 1989; p 229.
Amidinium ion 5 can be reacted with a wide range of
nucleophiles, but there are experimental constraints. One is that
nucleophiles must be strong enough to form a covalent bond
and not leave the amidinium ion as an ion pair. Even if a
(19) Deslongchamps, P. Bull. Soc. Chim. France 1984, II-349. Stevens,
R. V. Acc. Chem. Res. 1984, 17, 289. Kahne, D.; Yang, D.; Lim, J. J.;
Miller, R.; Paguaga, E. J. Am. Chem. Soc. 1988, 110, 8716. Meyers, A. I.;
Bienz, S. J. Org. Chem. 1990, 55, 791. Durkin, K. A.; Liotta, D. J. Am.
Chem. Soc. 1990, 112, 8162.

Addition of Nucleophiles to an Amidinium Ion
J. Am. Chem. Soc., Vol. 123, No. 19, 2001 4453
Experimental Section
1
NMR Spectroscopy. All H and 13C NMR spectra were acquired
on a Varian Unity 500, Varian Mercury 400, or Varian Mercury 300
spectrometer. Spectra were taken in benzene-d6, chloroform-d, dimethyl
sulfoxide-d6, or DMF-d7. NOE and COSY experiments were performed
using the default sequence of the Varian software.
1
Integrations from H NMR spectra were used to evaluate product
ratios. If signals from cis and trans stereoisomers overlapped, other
well-resolved signals were integrated to determine individual contribu-
tions to the total.
Figure 1. Transition states for axial or equatorial addition to 5, anti
or syn to N lone pairs.
Synthesis. The synthesis of 5-phenyl-1,4,5,6-tetrahydropyrimidine
was adapted from a published procedure starting from phenylacetonitrile
and formamide.25 The resulting 4-amino-5-phenylpyrimidine (mp
154.5-155.5 °C) was hydrolyzed to 5-phenyl-4-pyrimidinone: mp
covalent compound is formed, it may be in equilibrium with
the ion pair, which permits interconversion of cis and trans and
masks the kinetic product. Another constraint is that additional
reactions must not intrude. Strong bases cannot be used, since
they form an adduct that is exceedingly fragile,20 or they
deprotonate 5 at C2, giving the carbene, which dimerizes.21 With
some nucleophiles, such as hydroxide, the initial products are
unstable tetrahedral intermediates that undergo rapid cleavage.22
Nucleophiles that we have successfully reacted with 5 are
NaBD4, LiAlD4, CH3Li, n-BuLi, PhCH2MgCl, CH2dCHCH2-
MgCl, PhMgBr, C6F5MgCl, CH2dCHMgBr, HCtCMgBr, and
PhCtCMgBr, leading to products 6-15.
Not only do these cyclic amidinium ions provide information
about ALPH but they and their adducts are also of independent
interest. Some 1,4,5,6-tetrahydropyrimidines show activity as
neuromuscular blockers or antidepressant agents.23 Imidazo-
linium ions are important precursors to carbenes.20 Tetrahydro-
pyrimidinium and imidazolinium salts have been treated with
nucleophiles to create a wide variety of saturated heterocycles.24
We now show that nucleophiles show a small but clear
preference for addition antiperiplanar to the lone pairs on the
two nitrogens of amidinium ion 5. This is consistent with ALPH,
whose influence though is weak.
1
176-177 °C (lit. mp 173-174 °C), H NMR (DMSO) δ 7.37-7.45
(m, 3H), 7.71-7.73 (m, 2H), 8.17 (s, 1H), 8.21 (s, 1H). This was
converted with POCl3 to 4-chloro-5-phenylpyrimidine: mp 73-74 °C
1
(lit. mp 71-72 °C), H NMR (DMSO) 7.53-7.60 (m, 5H), 8.87 (s,
1H), 9.08 (s, 1H). Hydrogenation over 5% Pd/C converted this to the
tetrahydropyrimidine hydrochloride: mp 165-168 °C, 1H NMR
(DMSO) δ 3.13 (sept, J ) 5 Hz, 1H), 3.44 (t, J ) 12 Hz, 2H), 3.52
(dd, J ) 12, 5 Hz, 2H), 7.32-7.41 (m, 5H), 8.29 (s, 1H), 10.20-
10.40 (bs, 1H).
Conversion to 1,3-dimethyl-5-phenyl-1,4,5,6-tetrahydropyrimidinium
salts could be accomplished by using 2 equiv of NaH and slightly more
than 2 equiv of methylating agent in dry THF, either methyl iodide at
0 °C for 1 h, then at room temperature for 1 h or dimethyl sulfate at
40 °C for 1 h. The mixture was filtered, and the resulting solid was
taken up in dry CH2Cl2. The amidinium iodide 5a was purified by
recrystallization from dry CH2Cl2:THF (2:1). A colorless, hygroscopic
solid was obtained in 35% yield: 1H NMR (DMSO) δ 3.20 (s, 6H),
3.35 (m, 1H), 3.50 (m, 4H), 7.35-7.45 (m, 5H), 8.45 (s, 1H), 13C NMR
(DMSO) δ 152.8, 137.9, 128.9, 127.8, 127.7, 49.1, 41.1, 34.2. The
amidinium methosulfate 5b was obtained in 42% yield as an amorphous,
hygroscopic solid: 1H NMR (DMSO) δ 3.20 (s, 6H), 3.40 (s, 3H),
3.40-3.50 (m, 5H), 7.20-7.45 (m, 5H), 8.20 (s, 1H).
Addition of Nucleophiles. A sample of amidinium iodide 5a was
treated with excess 0.1 M NaOH. The reaction mixture was extracted
with methylene chloride, dried, and concentrated under vacuum,
yielding a colorless oil. The ring-opened aminoformamide (16) is the
only product, obtained in 62% yield. It exists as a 2:1 mixture of amide
stereoisomers: 1H NMR (DMSO) δ 2.22 and 2.25 (s, 3H), 2.65 (m,
2H), 2.65 and 2.71 (s, 3H), 3.08 (m, 1H), 3.2-3.6 (m, 2H), 7.2-7.4
(m, 5H), 7.7 and 7.9 (s, 1H), 13C NMR (DMSO) δ 163.8, 142.9, 129.9,
129.4, 128.1, 56.5, 54.0, 45.2, 38.0, 31.0 (major stereoisomer), δ 164.0,
143.1, 129.8, 129.6, 128.0, 56.9, 48.7, 44.4, 36.2 (minor stereoisomer).
(20) Winberg, H. E., (E. I. Du Pont de Nemours & Co.). U.S. Patent
No. 3,239,518, 1966; Chem. Abstr. 1966, 64, P15898c. Perrin, C. L.;
Armstrong, K. B.; Fabian, M. A. J. Am. Chem. Soc. 1994, 116, 715.
(21) Lemal, D. M.; Kawano, K. I. J. Am. Chem. Soc. 1962, 84, 1761.
Wanzlick, H. W. Angew. Chem., Int. Ed. Engl. 1962, 1, 75. Arduengo, A.
J.; Goerlich, J. R.; Marshall, W. J. J. Am. Chem. Soc. 1995, 117, 11027.
(22) Capon, B.; Ghosh, A. K.; Grieve, D. M. A. Acc. Chem. Res. 1981,
14, 306. McClelland, R. A.; Kanagasabapathy, V. M.; Steenken, S. Can. J.
Chem. 1990, 68, 375.
(23) Bramblecombe, R. W.; Hunt, R. R.; Rickard, R. L.; Taylor, J. V.
Br. J. Pharmacol. 1969, 37, 425. Weinhardt, K.; Wallach, M. B.; Marx,
M. J. Am. Chem. Soc. 1985, 28, 964. Dunbar, P. G.; Durant, G. J.; Fang,
Z.; Abuh, Y. F.; El-Assadi, A. A.; Ngur, D. O.; Periyasamy, S.; Hoss, W.
P.; Messer, W. S., Jr. J. Med. Chem. 1993, 36, 842. Messer, W. S., Jr.;
Abuh, Y. F.; Liu, Y.; Periyasamy, S.; Ngur, D. O.; Edgar, M. A. N.; El-
Assadi, A. A.; Sbeih, S.; Dunbar, P. G.; Roknich, S.; Rho, T.; Fang, A.;
Ojo, B.; Zhang, H.; Huzl, J. J., III; Nagy, P. I. J. Med. Chem. 1997, 40,
1230.
Reductions of amidinium salt 5a were carried out at 0 °C with 3
equiv of NaBD4 in methanol or with equimolar LiAlD4 in dry THF.
After 2 min the solution was shaken with saturated sodium carbonate
and methylene chloride, and the organic layer was washed twice with
water and dried. The solvent was removed under vacuum, and the
residue was taken up in an NMR solvent.
For addition of RMgBr or RLi two different procedures were utilized,
either to preserve kinetic product mixtures or to allow equilibration.
For preservation extreme care was taken to exclude trace acid or water.
Acid-free NMR solvents were prepared by adding 0.1 µg/mL 1,8-
diazabicyclo[5,4,0]undec-7-ene (DBU). All glassware was either flame-
dried or rinsed with this solvent. To a dry sample of the amidinium
ion 5 under dry N2 were added dry reaction solvent and excess
organometallic reagent. The mixture was stirred for a period of time
ranging from two minutes to several hours. Solvents were then removed
under vacuum. Acid-free NMR solvent was added to the reaction flask
(24) Dal Maso, M.; Orelli, L.; Perillo, I. A. J. Heterocycl. Chem. 1994,
31, 25. Salerno, A.; Ceriani, V.; Perillo, I. A. J. Heterocycl. Chem. 1992,
29, 1725. Salerno, A.; Ceriani, V.; Perillo, I. A. J. Heterocycl. Chem. 1997,
34, 709. Orelli, L. R.; Niemevz, F.; Garcia, M. B.; Perillo, I. A. J.
Heterocycl. Chem. 1999, 36, 105.
(25) Davies, W. H.; Piggott, H. A. J. Chem. Soc. 1945, 347.

4454 J. Am. Chem. Soc., Vol. 123, No. 19, 2001
Perrin and Young
Table 1. 1H NMR Data for cis and trans Hexahydropyrimidines 6-15
assignment
6c
7c
8c
9c
10c
H2
3.30 (bs)
1.84
3.42 (q J ) 6)
2.86
2.77 (t J ) 6)
2.90
3.53 (t J ) 6)
2.98
3.33 (t J ) 6)
2.85
H4,6c (t J ) 12 Hz)
H4,6t (dd J ) 12,4 Hz)
H5 (tt J ) 12,4 Hz)
N-CH3 (s)
2.74
2.58
2.71
2.64
2.59
3.20
1.97
7.2-7.4
3.29
2.32
7-7.2
1.09 (d)
3.31
2.50
7.2-7.4
1.7 (m),
1.3-1.5 (m),
0.94 (t)
3.27
2.37
7-7.4
2.84 (d)
3.24
2.37
7-7.2
b
Ph (m)
other
6t
7t
8t
9t
10t
H2
2.17 (bs)
2.74a
2.87 (q J ) 6)
2.94
2.42 (t J ) 6)
3.08
3.21 (t J ) 6)
2.92
2.81 (t J ) 6)
2.92
H4,6c (dd J ) 12,4 Hz)
H4,6t (t J ) 12 Hz)
H5 (tt J ) 12,4 Hz)
N-CH3 (s)
1.82
2.48
2.55
2.49
2.45
3.20a
1.96
7.2-7.4
3.29
2.10
7-7.2
1.17 (d)
3.29
2.32
7.2-7.4
1.62 (m),
1.3-1.5 (m),
0.94 (t)
3.27
2.21
7-7.4
2.84 (d)
3.28
2.18
7-7.2
2.32 (tt),
5.13 (m),
6.08 (m)
Ph (m)
other
assignment
11c
12cd
13c
14c
15c
H2
2.86 s
2.58 (d J ) 4)
2.82
4.23 (s)
4.51 (s)c
2.88
H4,6c (t J ) 12 Hz)
H4,6t (dd J ) 12,4 Hz)
H5 (tt J ) 12,4 Hz)
N-CH3 (s)
3.00 (dd J ) 12,1)
2.73
2.63
3.23
2.27
7-7.2
2.36
2.72
2.72
2.68 (m)
1.77
3.34
2.34
3.32
2.37
7-7.5
Ph (m)
other
7-7.6, 8.6
7-7.2
5.99b
11t
12t
13t
14td
15td
H2
2.92 (s)
3.04
3.72 s
3.25
2.33
3.34
2.14
2.66 (d J ) 4)
3.11
H4,6c (dd J ) 12,4 Hz)
H4,6t (t J ) 12 Hz)
H5 (tt J ) 12,4 Hz)
N-CH3 (s)
2.20
2.28
3.48
1.88
7-7.6, 8.6
3.34
2.25
7-7.2
5.87 (ddd J ) 12,8,4),
5.17 (dd J ) 12,2),
5.08 (dd J ) 8,2)
Ph (m)
other
7.2-7.4
c 1
a Indistinguishable from 6c. b Peak overlap precluded identification.
J
) 153 Hz. d Not observed.
CH
and stirred briefly. The resulting mixture was filtered through cotton
into an NMR tube. Each experiment was repeated several times, and
kinetic ratios were taken from samples with the highest proportion of
the less stable stereoisomer. Alternatively, the mixture was treated with
saturated aqueous NaCl solution and extracted with cyclohexane. The
organic extract was washed several times with water and then dried.
Solvent was removed under vacuum, and the residue was dissolved in
an NMR solvent.
Additions of lithium acetylide or lithium phenylacetylide (excess)
were also carried out in DMSO-d6 and dried over 4 Å molecular sieves.
Volatiles were removed under vacuum prior to NMR analysis. Kinetic
product mixtures were obtained with acid-free solvents and glassware.
Tetrabutylammonium cyanide (0.1-10 equiv) was added to ami-
dinium iodide 5a in an appropriate NMR solvent. The resulting mixture
was filtered through cotton and transferred to an NMR tube.
Equilibration of Hexahydropyrimidines. Conversion of kinetic
product mixtures to thermodynamic mixtures was carried out with acetic
acid or phenol and monitored by 1H NMR. Samples that contained an
unknown excess of DBU were first titrated with small quantities of
acid until the start of stereoisomerization was detected, and then they
were allowed to equilibrate for days.
Nucleophile Exchange. A sample of the 2-pentafluorophenyl
hexahydropyrimidine (12) in benzene-d6 was stirred briefly with excess
lithium phenylacetylide. Solvent was removed under vacuum prior to
NMR analysis. Decomposition of 2-pentafluorophenyl-2-propanol27 (δ
1.73, t, J ) 2 Hz) in DMSO-d6 containing NaOD was followed by 1H
NMR at room temperature.
Molecular-Mechanics Calculations. Energies of the stereoisomers
+
of PhCHdNHCH3 were calculated with MM2 parametrization (CS
Chem3D Pro 4.0).28
Results
Unsuccessful Reactions. tert-Butyl- and isopropyl-magne-
sium bromides produced complex mixtures in low yield. It is
likely that these bulky reagents act as bases rather than
nucleophiles. Reductions with milder reducing agents sodium
cyanoborohydride and diisobutylaluminum hydride, even at high
temperature, left amidinium salts unreacted. The addition of
excess weak nucleophiles sodium azide or tetrabutylammonium
fluoride to 5 in various organic solvents had no appreciable
1
effect on the H NMR spectrum.
Reductive Trapping of Iminium Ion Intermediates. Approxi-
mately 50 mg of hexahydropyrimidine was dissolved in 4 mL of
acetonitrile/acetic acid (1:1) and stirred overnight with 3 equiv of
NaBH3CN at 50 °C.26 The resulting solution was neutralized with Na2-
CO3, extracted into CHCl3, washed with saturated NaCl, and concen-
trated under vacuum.
Signal Assignments. Tables 1-2 list the 1H and 13C chemical
shifts for cis and trans products 6-15. Peak assignments for
each stereoisomer were made on the basis of H-H couplings
(27) Marek, E. M.; Kolenko, I. P.; Ryabinin, N. A.; Agishev, Yu. N. J.
Gen. Chem. USSR (Engl. Transl.) 1969, 39, 2257.
(26) Hiemstra, H. C.; Biera¨ugel, H.; Wijnberg, M.; Pandit, U. K.
Tetrahedron 1983, 39, 3981.
(28) Allinger, N. L.; Burkert, U. Molecular Mechanics; ACS Monograph
17; American Chemical Society: Washington, DC, 1982.

Addition of Nucleophiles to an Amidinium Ion
J. Am. Chem. Soc., Vol. 123, No. 19, 2001 4455
Table 2. 13C NMR Data for cis and trans Hexahydropyrimidines 6-15
assignment
6c
7c
8c
9c
10c
C2
C4,6
C5
N-CH3
Ph
77.5 t
59.8
37.8
a
79.5
80.1
51.6
34.8
41.4
127.8, 127.2,
126.9, 126.0,
125.3, 124.4
140.8, 139.5
27.1
80.0
53.2
42.5
37.2
128.3, 127.5,
126.3
52.0
37.0
41.9
a
52.7
35.5
42.2
a
42.2
128.1,
127.2,
126.2
141.8
a
ipso
other
a
a
a
142.5
136.6, 115.0,
30.1, 26.8,
22.9, 14.2
35.8
6tb
7t
8t
9t
10t
C2
C4,6
C5
N-CH3
Ph
78.6c
62.1
36.9
39.1
128.1, 127.5,
126.8
83.0
62.5
36.9
38.9
128.2, 127.1,
126.9
82.4
61.5
36.2
38.1
127.6, 127.2,
127.0, 126.0,
125.3, 124.8
140.2, 137.6
35.3
82.5
62.9
42.3
38.8
127.9, 127.9,
126.4
ipso
other
141.2
18.7
141.5
29.9, 26.9,
142.1
135.6, 115.2,
23.1, 14.1
35.4
assignment
11c
13c
14c
15c
C2
C4,6
C5
N-CH3
Ph
ipso
other
84.0
53.0
40.1
44.1
a
82.0
53.1
38.4
40.9
73.6
54.3
40.6
41.7
73.9
54.0
40.1
41.9
127.0, 126.0, 125.2
127.8, 127.6, 127.2
142.0
73.8, 76.7
131.6, 128.2, 128.0, 127.9, 127.3, 126.4
141.4
89.9, 81.0
a
a
119.1a
11t
13t
14td
15td
C2
C4,6
C5
91.6
62.0
41.0
88.5
60.3
39.2
N-CH3
Ph
42.8
41.2
127.1, 126.3,
125.4
128.3,
128.2,
128.1,
128.0,
127.1,
127.1
141.4,
140.6
ipso
140.3
other
137.5, 118.8
c 1
a Peak overlap precluded identification. b Indistinguishable from 6c.
J
) 136 Hz. d Not detected.
CH
The data in Tables 1-2 provide insight into the conformations
of 2-substituted hexahydropyrimidines. Signals for substituents
antiperiplanar to a nitrogen lone pair are shifted upfield. This
is clearly the case with CHD isotopomers 6c and 6t. Moreover,
H4,6t signals in trans isomers are always upfield of the
corresponding H4,6c signals in the cis. This indicates that
nitrogen lone pairs are less often axial in cis isomers, where
the N-methyl can avoid gauche interactions by being axial.
Similarly, the N-methyl is generally downfield in the cis isomers,
because it is axial, which also generally leaves the N-methyl
13C signal upfield in the trans isomers, owing to a γ-gauche
interaction with the 2-substituent. These results are consistent
with previous ones on some conformationally flexible hexahy-
dropyrimidines.31
Figure 2. Observed NOE enhancements.
and by analogy to other hexahydropyrimidines in this study.
Labels H4,6c and H4,6t designate cis or trans to the substituent
at C2.
Stereoisomer assignments were made with NOE experiments.
Observed enhancements for a typical set of hexahydropyrim-
idines are shown in Figure 2. Only with trans hexahydropyri-
midines 6-13t are NOEs observed between H2 and H4,6. The
acetylenyl-substituted hexahydropyrimidine stereoisomers 14c,15c
were assigned from a 1-2 Hz W-coupling between H2 and
H4,6t,29 detected by 2D COSY. The cis stereochemistry for 15c
is consistent with 1JCH ) 153 Hz, higher than the 136 Hz in 7t,
where the CH is antiperiplanar to the lone pairs.30
A further observation concerns the 2-phenyl-substituted cis
isomer (11c), which shows only a 1-Hz coupling between H5
and H4,6c, not the usual 12 Hz. Apparently this exists chiefly
(31) Jones, R. A. Y.; Katritzky, A. R.; Snarey, M. J. Chem. Soc. B 1970,
131. Ferguson, I. J.; Katritzky, A. R.; Reed, D. M. J. Chem. Soc., Perkin
Trans. 2 1977, 818. Katritzky, A. R.; Baker, V. J.; Ferguson, I. J.; Patel, R.
C. J. Chem. Soc., Perkin Trans. 2 1979, 143.
(29) Eberstadt, M.; Gemmecker, M.; Mierke, D. F.; Kessler, H. Angew.
Chem., Int. Ed. Engl. 1995, 34, 1671.
(30) Perlin, A. S.; Casu, B. Tetrahedron Lett. 1969, 2921.

4456 J. Am. Chem. Soc., Vol. 123, No. 19, 2001
Perrin and Young
Table 3. Product Distributions from Addition of Nucleophiles to
5a
products could be maintained for days in organic solvents with
an excess of DBU. Stereoisomerization of most samples
occurred cleanly in the presence of trace acid. No other products
were observed, except with the vinyl and allyl derivatives, which
decomposed over several hours.
nucleophile
NaBD4
LiAlD4
solvent
% cis (kinetic)
% cis (equil)
CH3OH
THF
88
90
50
50
CH3Li
n-BuLi
n-BuLi
n-BuLi
n-BuLi
n-BuLi
PhCH2MgCl
PhCH2MgCl
PhCH2MgCl
CH2dCHCH2MgCl
PhMgBr
C6F5MgCl
CH2dCHMgBr
HCtCMgBr
PhCtCMgBr
PhCtCMgBr
PhCtCMgBr
PhCtCMgBr
DME
65
75
75
75
75
75
65
65
65
60
80
<2
15
15
17
17
17
17
17
35
35
Rates of product stereoisomerization depend on the substituent
at C2. In the presence of acetic acid kinetic mixtures of 6c and
6t established equilibrium after several days. In contrast,
conversion of 7-11c to 7-11t occurred under these conditions
before the sample could be introduced into the NMR spectrom-
eter.
Rates of phenol-catalyzed stereoisomerization were assayed
by NMR in an attempt to assess the effect of 2-substitution.
Exact rate constants could not be obtained because the unknown
amount of DBU neutralizes the added acid and renders its
concentration uncertain. To restrict these stereoisomerization
experiments to a narrow range of acidities, the samples were
titrated with small aliquots of acid until stereoisomerization
commenced. The time required for equilibration of any particular
hexahydropyrimidine was then reproducible within a factor of
2. Alkyl and aryl hexahydropyrimidines 8-11 equilibrated in
approximately 1 h, the same for all four species. In contrast,
equilibration of 6 required >1 month.
DME
DMEa
benzene
THF
DMEa
DME
DMEb
DMEa
DME
35
25
DMEa,b
DME
<2
<2
15
>98
>98
>98
>98
>98
DME
DMSO
DMSO
DME
>98
>98
>98
>98
>98
DMSOa
DMSOc
a 5b. b With 0.5 equiv of tetramethylethylenediamine (TMEDA). c 5c.
as the conformer where the C5 phenyl is axial, allowing the
C2 phenyl to be equatorial. This is the opposite sense of ring
inversion from the other 5-phenylhexahydropyrimidines because
repulsions between the 5-phenyl and H4,6 or lone pairs are less
severe than between the 2-phenyl and the N-methyls.
Tetrabutylammonium Cyanide Additions. Amidinium salt
5a is soluble in polar solvents DMSO, acetonitrile, and DMF,
slightly soluble in CH2Cl2 and CHCl3, and insoluble in pyridine
or THF. It readily dissolves in all of these solvents upon addition
of tetrabutylammonium cyanide. We attribute this increased
solubility in nonpolar solvents to a replacement of iodide by
cyanide, which can form a covalent adduct.
The acid-catalyzed stereoisomerization is likely to occur by
reversible ring-opening to an iminium ion (17 in Figure 3). This
intermediate could be trapped in mild acid with NaBH3CN, as
1
with some similar heterocycles.26 The H NMR data for the
products (18, R ) PhCH2 or CCPh) from reduction of benzyl-
and phenylacetylenyl-substituted hexahydropyrimidines (9,15)
are given in Table 4.
The resulting species 5c is a mixture of ionic and covalent
forms. The existence of an equilibrium between them was
detected by H NMR in a number of solvents. The chemical
Figure 3. Stereoisomerization of hexahydropyrimidines and trapping
of the intermediate (17).
1
shift of H2 responds to the change in hybridization at C2. The
most drastic change was observed in DMF-d7. Upon addition
of 1 equiv of cyanide, the signal shifts from δ 8.6 to δ 6.8 and
broadens. Addition of excess cyanide narrows the signal
somewhat, with a further upfield shift to δ 5.5, aproaching the
δ 4.2 or 4.5 of covalent acetylene adducts 14-15. The spectral
data for this sample are: 1H NMR (DMF) δ 7.2-7.4 (m, 5H),
5.5 (bs, 1H, 1JCH ) 164 Hz), 3.15 (tt, J ) 12, 4 Hz, 1H), 3.00
(dd, J ) 12, 4 Hz, 2H), 2.58 (t, J ) 12 Hz, 2H), 2.54 (s, 6H).
The resolution of H4,6c from H4,6t is like that of other covalent
adducts in Tables 1-2, but the chemical shifts do not permit
assignment as cis or trans.
Table 4. NMR Data for Products from Reduction of Iminium Ions
16
assignment
CH2Nu
18, R ) PhCH2
18, R ) CCPh
3.42 (d J ) 16 Hz),
3.58 (d J ) 16 Hz)
2.29, 2.40
2.52, 2.74
2.82, 2.91
3.56 (bs)
N-CH3
2.38, 2.39
2.98
2.76
CH2N (m)
CH2NH (m)
CH (m)
3.24
3.11
The cis pentafluorophenyl adduct could never be detected.
With careful workup the trans adduct 12t was obtained, but
this was hydrolyzed to 16 within several hours, even in the
presence of drying agents Na2SO4 or MgSO4. This reactivity is
surprising. It is quite similar to that of the ionic amidinium iodide
5a or the weakly covalent cyanide 5c, each of which undergoes
exchange with nucleophiles such as cyanide or hydroxide. It is
most unlikely that hydroxide or water is capable of SN2
displacement of pentafluorophenyl anion at a tertiary center. It
is more likely that hydrolysis occurs by reversion of 12 to 5,
the common cationic intermediate in all of these substitutions.
The reversion must be reversible, allowing for the rapid
stereoisomerization of 12c to 12t (Figure 4). When 12 was
allowed to react with lithium phenylacetylide, adduct 15 was
produced, with a 1H NMR spectrum matching that of 15 created
directly from 5.
The greater solubility of amidinium cyanide 5c permits its
application to addition reactions with nucleophiles, in the same
manner as iodide or methosulfate salts (5a,b). Products and
product distributions were identical in all cases tested. The
exclusion of water from solvents containing 5c made it easier
to avoid hydrolysis to aminoamide (16).
1
Product Ratios. The results from H NMR integrations are
listed in Table 3. For most of these nucleophiles the kinetic
product mixture contained mostly cis isomer, whereas thermo-
dynamic mixtures were mostly trans. Exceptions are pentafluo-
rophenyl- and vinyl-magnesium bromide and lithium acetylide
and phenylacetylide, each of which produced a constant product
mixture that did not show stereoisomerization.
Product Stability. Adducts 6-15 are generally stable to
neutral or basic conditions, including aqueous alkali. Kinetic

Addition of Nucleophiles to an Amidinium Ion
J. Am. Chem. Soc., Vol. 123, No. 19, 2001 4457
the stereoselectivity in reactions of cyclohexanones, where axial
approach of bulky nucleophiles is hindered.33 However, addition
to ketones is not a good model, since approach is at an angle
of ca. 110° to the C-O,34 passing near H3,5 (19), whereas
approach to an amidinium ion is more external (20). This is a
distinction between inter-exo-trig and inter-endo-trig, analogous
to Baldwin’s rules for ring-closure.35 Therefore, steric effects
on the kinetic product distributions in Table 3 are minimal.
Figure 4. Trapping of amidinium ion 5.
Discussion
Rates of Stereoisomerization. The rates of hexahydropyri-
midine stereoisomerization are interesting in their own right,
since two mechanisms operate, depending on substituent. The
variation of ring-opening rates can be attributed to the ability
of substituents R to stabilize the positive charge in 17. The
greater reactivity of substituted hexahydropyrimidines 7-11,
relative to the parent 6, is analogous to substituent effects on
rates of hydrolysis of 1,3-dioxolanes and on ring-chain equilibria
in some tetrahydropyrimidines.36 The rapid stereoisomerization
of 13-15, even in the absence of added acid, then must be due
to further stabilization of 17 by conjugation with CHdCH2, Ct
CH, or CtCPh. Yet the rate of ring-opening stereoisomerization
for 11 is comparable to that of alkyl derivatives (8-10) and
thus low enough to permit isolation of the kinetic mixture, even
though the phenyl substituent ought to stabilize 17 as vinyl and
acetylenyl do. The reason it does not is that conjugation between
phenyl and CdN+ induces steric crowding between an ortho
hydrogen and the cis methyl or alkyl. Indeed, a molecular
mechanics calculation indicates that the trans isomer of PhCHd
Product Distributions. The data in Table 3 show different
kinetic and thermodynamic product mixtures for nucleophiles
D-, Me-, n-Bu-, PhCH2-, allyl-, and Ph-. Kinetics favors
formation of the less stable cis isomer, which is converted to
trans upon equilibration. These results are consistent with
ALPH. However, there are cases where kinetic mixtures cannot
be distinguished from thermodynamic. For some nucleophiles
the major product is always cis (HCC-, PhCC-) or always trans
(C6F5-), or an unchanging mixture of the two (H2CdCH-), or
else a product of indeterminate stereochemistry (CN-). Such
behavior could arise if the kinetic mixture is identical with the
thermodynamic or if the kinetic product quickly stereoisomerizes
to the thermodynamic. If the latter, the observed products do
not reflect stereoelectronic control.
This ambiguity can be resolved by considering two mecha-
nisms for stereoisomerization, reversible dissociation of the
nucleophile and reversible ring opening. The amidinium ion (5)
forms covalent adducts with CN- and C6F5-, but both nucleo-
philes can then be replaced by others (HO-, PhCC-). Thus,
there must be a rapid equilibrium between covalent and ionic
forms, as is also consistent with the variation of chemical shift
of 5c with the concentration of cyanide. This interconversion
then provides a mechanism for rapid stereoisomerization, and
+
NHCH3 is 5 kcal/mol more stable than the cis. Stabilization
by phenyl is variable, since it does accelerate stereoisomerization
ca. 103-fold, relative to D, whereas no such effect was observed
with 1,3-imidazolidines.37
As for the other mechanism, it is not surprising that CN-
binds only weakly to the amidinium ion, since it is the least
-
the exclusively trans product 12t from C6F5 must be the
equilibrium mixture.
basic nucleophile studied (pKaHCN ) 9.2, pKaC F H ) 25, pKaRCCH
) 24 or 25.2).38 When a solution of 5c is washed with water,
it reverts to the ion and is hydrolyzed. The pentafluorophenyl
adduct is covalent, as expected and as demonstrated by its
stability to aqueous workup, but it does slowly undergo
dissociation to an ion pair, allowing stereoisomerization or
capture by water or PhCCLi. Yet the acetylenyl adducts do not
hydrolyze in base, even though the leaving abilities are
6
5
Acid-catalyzed stereoisomerization of hexahydropyrimidines
6-11 does convert kinetic product mixtures into thermodynamic
ones, whereas these are stable in base. These are in equilibrium
with open-chain iminium ions (17),26,32 which could be trapped
with NaBH3CN. Reversal of the ring opening then allows for
stereoisomerization. Since iminium ion 17 (R ) CCPh) could
also be trapped, this mechanism, catalyzed by unknown acid,
is probably responsible for stereoisomerizations of 13-15,
whose observed product distributions are equilibrium values.
The reason this is so rapid is addressed below.
In contrast to the nucleophiles where equilibration is rapid,
with D-, Me-, n-Bu-, PhCH2-, allyl-, and Ph- stereoisomer-
ization could be made slow enough that initial products are
kinetically determined. The predominance of the cis isomer in
the kinetic products in Table 3 is consistent with ALPH.
The aggregation state of the nucleophile, as moderated by
solvent or added TMEDA, has no appreciable effect on product
distribution. Nor does the counterion of the amidinium salt,
whether iodide or methosulfate. There is little variation with
the size of the nucleophile, although the smallest nucleophile,
D-, gives the largest percent of axial substitution. This parallels
-
comparable. It is likely that the leaving ability of C6F5 is
enhanced by relief of steric repulsions, whereas that of the
acetylides is reduced by the strength of the sp-sp3 C-C bond.
Neither of these features is reflected in the pKa’s.
The ease of dissociation of C6F5- from a neutral molecule is
surprising, but not unreasonable in light of the electron-
withdrawing fluorines. Carbanions are known to dissociate from
(33) Dauben, W. G.; Fonken, G. J.; Noyce, D. S. J. Am. Chem. Soc.
1956, 78, 2579.
(34) Bu¨rgi, H. B.; Dunitz, J. D. Acc. Chem. Res. 1983, 16, 153.
(35) Baldwin, J. J. Chem. Soc., Chem. Commun. 1976, 734.
(36) Fife, T. H.; Hagopian, L. J. Org. Chem. 1966, 31, 1772. Piotrowska,
H.; Sas, W.; Urbanski, T. Tetrahedron 1981, 33, 1979.
(37) Lambert, J. B.; Wang, G.; Huseland, D. E.; Takiff, L. C. J. Org.
Chem. 1987, 52, 68.
(32) Chapuis, C.; Gauvreau, A.; Klaebe, A.; Lattes, A.; Perie, J. J. Bull.
Soc. Chim. Fr. 1973, 977. Fife, T. H.; Hutchins, J. E. C.; Pellino, A. M. J.
Am. Chem. Soc. 1978, 100, 6455. Fife, T. H.; Pellino, A. M. J. Am. Chem.
Soc. 1980, 102, 3062. Lambert, J. B.; Majchrzak, M. W. J. Am. Chem.
Soc. 1980, 102, 3588.
(38) Perrin, D. D. Ionization Constants of Inorganic Acids and Bases in
Aqueous Solution; Pergamon Press: Oxford, 1982. Streitweiser, A. J.;
Scannon, P. J.; Noemeyer, H. M. J. Am. Chem. Soc. 1972, 94, 7936.
Wooding, N. S.; Higginson, W. E. C. J. Chem. Soc. 1952, 774. Streitweiser,
A. J.; Reuben, D. M. E. J. Am. Chem. Soc. 1971, 93, 1794.

4458 J. Am. Chem. Soc., Vol. 123, No. 19, 2001
Perrin and Young
highly substituted tertiary alkoxides.39 As evidence that C6F5
free-energy difference of ca. 1 kcal/mol. Even in the most
selective case, reaction of LiAlD4 or NaBD4, the preference is
only 1.3 kcal/mol.
-
can dissociate even from a simple alkoxide, we find that base
converts 2-pentafluorophenyl-2-propanol to acetone partially
within 10 min and completely within 24 h.
Thermodynamic Product Ratios. The product distributions,
after equilibration, are listed in Table 3. The 50:50 mixture for
6 is evidence that the CHD group has equilibrated. Impurities
precluded accurate integration that might have indicated a greater
proportion of equatorial D.40
For the other groups the more stable product stereoisomer is
the trans, except for the acetylenes. This preference is quite
large for phenyl and pentafluorophenyl, as expected from the
steric repulsions with an axial aryl. These are so large that the
phenyl-substituted cis isomer (11c) exists chiefly as the ring-
inverted conformer, with the C5 phenyl axial. For alkyl groups
the equatorial preference is small, with a ratio ranging from
2:1 to 6:1, appreciably less than in cyclohexanes.41 This is
because of repulsion between an equatorial alkyl and one
N-methyl that is predominantly axial.31
It is remarkable that the acetylenes strongly favor the cis
isomer, with CtCR axial, contrary to the behavior in cyclo-
hexanes. This is likely to be an anomeric effect, analogous to
the 2 kcal/mol preference in 2-cyanopiperidine.42 It also requires
that the 5-phenyl greatly favors equatorial.
Contribution of Stereoelectronic Control. With nucleo-
philes D-, Me-, n-Bu-, PhCH2-, allyl-, and Ph- products are
stable enough that the kinetic mixtures provide a measure of
stereoelectronic control. The excess of cis product is good
evidence that the nucleophile prefers to add antiperiplanar to
the lone pairs on the two nitrogens. This excess is opposite to
that seen in the equilibrium mixture, which is governed by steric
effects. However, in the transition state steric effects are
minimal, owing to a trajectory (20) that avoids axial hydrogens.
Yet the influence of stereoelectronic control is also minimal.
The anti preference, varying from 60 to 90%, corresponds to a
An alternative measure of stereoelectronic control might be
the excess of the kinetic preference for cis over the thermody-
namic, since stereoelectronic control must overcome the prefer-
ence for trans. This excess varies from 0.75 to 1.6 kcal/mol,
except for R ) Ph (11), where it is >3 kcal/mol. This exception
shows that such a correction is not warranted, since the
contribution of stereoelectronic control ought not increase with
the steric bulk of the nucleophile. Steric interactions are not
comparable in the products, because the conformations are
different, one of the N-methyls being predominantly axial. In
the transition state both are equatorial to allow the nucleophile
to add antiperiplanar to the lone pairs, and those might even
have been expected to hinder equatorial approach. Nevertheless,
axial approach is not strongly favored.
We had proposed that stereoelectronic control is reduced in
the cleavage step of amidine hydrolysis, rather than in the
addition of hydroxide.13 Now that the two steps can be separated,
it becomes clear that addition does not show strong stereoelec-
tronic control. The nucleophiles here may not be comparable
to hydroxide in the amidine reactions, and it is unfortunate that
alkoxide addition could not be studied. The weakness of
stereoelectronic control in the addition step can be attributed to
a transition state that is early along the reaction coordinate, with
little interaction between the lone pair and the bond that is
forming, whose stereochemistry then does not matter. This
reverses the argument from the earlier proposal, where the
transition state was assumed to be early for the cleavage. Such
an assumption can always be made, but it weakens ALPH by
reducing its predictive power.
Conclusions
Kinetic products from reaction of nucleophiles D-, Me-,
n-Bu-, PhCH2-, allyl-, and Ph- with amidinium ion 5 show a
clear preference for addition antiperiplanar to the lone pairs on
the two nitrogens. This is counter to the thermodynamic
preference and is good evidence for stereoelectronic control.
However, its influence is weak, contributing ca. 1 kcal/mol.
(39) Zook, H. D.; March, J.; Smith, D. F. J. Am. Chem. Soc. 1959, 81,
1616. Cram, D. J.; Mateos, J. L.; Hauck, F.; Langemann, A.; Kopecky, K.
R.; Nielsen, W. D.; Allinger, J. J. Am. Chem. Soc. 1959, 81, 5774. Benkeser,
R. A.; Broxterman, W. E. J. Am. Chem. Soc. 1969, 91, 5162. Bartlett, P.
D.; Steadman, T. R.; Tidwell, T. T.; Weber, W. P. Tetrahedron Lett. 1970,
2915.
(40) Forsyth, D. A.; Hanley, J. A. J. Am. Chem. Soc. 1987, 109, 7930.
(41) Bushweller, C. H. In Conformational BehaVior of Six-Membered
Rings: Analysis, Dynamics and Stereoelectronic Effects; Juaristi, E., Ed.;
VCH: New York, 1995; p 25.
Acknowledgment. This research was supported by National
Science Foundation Grant CHE94-20739. Purchase of the NMR
spectrometers was made possible by grants from NIH and NSF.
(42) Booth, H.; Dixon, J. M.; Khedair, K. A. Tetrahedron 1982, 48, 6161.
JA004240Q