Synthetic Optimization and Structural Limitations of the Nitrile Aldol Reaction

Article abstract of DOI:10.1021/jo9702148

In an effort to further optimize and to define the structural limitations of the nitrile aldol reaction, several studies were undertaken. Lithiated phenylacetonitrile 1 was demonstrated to exhibit diastereoselectivity for addition to benzaldehyde higher t

Full text of DOI:10.1021/jo9702148

6
316
J . Org. Chem. 1997, 62, 6316-6321
Syn th etic Op tim iza tion a n d Str u ctu r a l Lim ita tion s of th e Nitr ile
Ald ol Rea ction
Paul R. Carlier,* Kam Moon Lo, Michael M.-C. Lo, Priscilla C.-K. Lo, and Cedric W.-S. Lo
Department of Chemistry, Hong Kong University of Science and Technology,
Clear Water Bay, Kowloon, Hong Kong
Received February 6, 1997X
In an effort to further optimize and to define the structural limitations of the nitrile aldol reaction,
several studies were undertaken. Lithiated phenylacetonitrile 1 was demonstrated to exhibit
diastereoselectivity for addition to benzaldehyde higher than that of other metalated phenylac-
etonitriles. With a view toward practical lab-scale synthesis of aldols derived from lithiated
arylacetonitriles, the effects of reaction concentration and quenching method on isolated yields of
the anti-aldols were readdressed. Secondary aldehydes give reasonable recrystallized yields of anti-
aldols at reaction concentrations as high as 0.1 M; pivalaldehyde gives good results at concentrations
up to 0.3 M. Thirdly, the aldol reaction of lithiated 1 with para-substituted benzaldehydes 6b-g
was then studied; increasing electron-withdrawing power of the para substituent was found to
cause a marked decrease in diastereoselectivity. Finally, the aldol reactions of two aliphatic nitriles
(isovaleronitrile 4 and tert-butylacetonitrile 5) were studied. Suprisingly, aldol diastereoselectivity
of aliphatic nitriles is not uniformly anti-selective, and selectivities overall are lower than those
afforded by arylacetonitriles.
The aldol reaction of nitriles is potentially a very useful
as pivalaldehyde 9) typically give >20:1 anti:syn selectiv-
ity, and the minor syn-isomer normally cannot be de-
tected by H NMR. However, secondary aldehydes such
way to simultaneously construct two stereocenters and
install latent amino functionality. In 1987 Trost reported
that lithiated acetonitrile undergoes axial selective ad-
dition to conformationally biased cyclohexanones and
cyclohexenones.1 Addition of lithiated propionitrile to
cyclohexenones occurs with even greater axial selectivity
1
as benzaldehyde 6a often give undesirably low selectivity
(2.5:1 to 6:1). In an attempt to achieve higher levels of
anti selectivity, the use of different metal counterions was
explored. The various metalated phenylacetonitriles
were prepared by deprotonation with the appropriate
metal amide or organometallic (method A), or by trans-
metalation of lithiated 1 (method B). Kauffmann has
previously described transmetalation of lithiated nitriles
to the corresponding Fe(II), Cr(II), Mn(II), Co(II), Ni(II),
and Cu(I) derivatives.5 Transmetalation of lithiated
and with considerable stereoselectivity at the R-nitrile
carbon.2 On the other hand, initial reports of aldol
addition of nitriles to aldehydes were characterized by
little or no diastereoselectivity.3 We recently demon-
strated that lithiated arylacetonitriles undergo anti-
selective addition to aldehydes.4 In this report we
describe efforts to further optimize diastereoselectivities
and yields for the addition of arylacetonitriles to alde-
hydes. The effect of different metal counterions, reaction
concentration, and quenching methods are discussed. The
aldol reactions of substituted benzaldehydes, and two
aliphatic nitriles, were also studied in order to further
define the substrate domain within which useful diaste-
reoselectivities can be obtained.
1
,5,6
1
nitriles to the corresponding titanium,
zinc, and
1
aluminum derivatives has also been described. In the
course of our investigations more than 20 metal/oxidation
state/ligand combinations were examined. The actual
success or failure of the transmetalation in each case is
not known, but a change in diastereoselectivity and yield
was usually observed relative to lithiated 1. However,
in the end, no metal offered superior diastereoselectivity
relative to lithium; representative examples are given in
Table 1.
Previous optimization studies on the aldol reactions of
lithiated phenylacetonitrile 1 established that diastereo-
selectivity is highest in THF, at or below 0.025 M, and is
insensitive to the identity of amide base (LDA or LiH-
MDS) used for deprotonation.4 Under the conditions of
our standard protocol (THF, 0.025 M, -78 °C, 30 min),
nitrile aldol reaction diastereoselectivity was demon-
strated to be kinetically controlled. Yields are generally
high, and the selectivity obtained depends on the steric
bulk of the aldol substituent. Tertiary aldehydes (such
As can be seen, as the counterion changes from Li to
Na or K, a significant decrease in anti selectivity occurs
(entries 1-3, Table 1). These results match the expecta-
a
tion that, as M-O and M-N bonds lengthen, steric
differentiation of the competing diastereomeric transition
states should decrease, leading to poorer selectivity. The
use of boron enolates in aldol reactions has yielded
substantial improvements in selectivity relative to lithium,
ostensibly as a consequence of the extremely short B-O
bond length.7 However, despite extensive variation of the
boron alkyl moieties, deprotonation conditions, and work-
up method, poor yields resulted from use of boron in
X
Abstract published in Advance ACS Abstracts, August 1, 1997.
(
1) Trost, B. M; Florez, J .; J ebaratnam, D. J . J . Am. Chem. Soc.
987, 109, 613-615.
2) Trost, B. M.; Florez, J .; Haller, K. J . J . Org. Chem. 1988, 53,
396-2398.
3) (a) Hamana, H.; Sugasawa, T. Chem. Lett. 1982, 1401-1404.
b) Kasatkin, A. N.; Biktimirov, R. Kh.; Tolstikov, G. A.; Nikonenko,
A. G. J . Org. Chem. USSR 1990, 26, 1037-1045.
4) (a) Carlier, P. R.; Lo, K.-M. J . Org. Chem. 1994, 59, 4053-4055.
b) Carlier, P. R.; Lo, K.-M.; Lo, M. M.-C.; Williams, I. D. J . Org. Chem.
995, 60, 7511-7517.
1
(
2
(5) Kauffmann, T.; Kieper, H.; Pieper, H. Chem. Ber. 1992, 125,
899-905.
(
(
(6) Reetz, M. T.; Steinbach, R.; Wenderoth, B.; Westermann, J .
Chem. Ind. 1981, 541-542.
(
(7) Kim, B. M.; Williams, S. F.; Masamune, S. In Comprehensive
Organic Synthesis; Trost, B. M., Fleming, I., Heathcock, C. H., Eds.;
Pergamon Press: Oxford, 1991; Vol. 2, pp 239-275.
(
1
S0022-3263(97)00214-4 CCC: $14.00 © 1997 American Chemical Society

Nitrile Aldol Reaction
J . Org. Chem., Vol. 62, No. 18, 1997 6317
Ta ble 2. Isola ted Yield s of An ti Ald ols 10a -15
Ta ble 1. Effect of Meta l Cou n ter ion on Nitr ile Ald ol
Rea ction Dia ster eoselectivity
anti
aldol
[nitrile]
(M)
isolated
anti:syn (yield)a
entry
R
R′
yield,a %
entry
metal
base
MX
1
2
3
4
5
6
7
8
9
Li
Na
K
LiHMDS
NaHMDS
KHMDS
LiHMDS
EtMgBr
LDA
LDA
LDA
LDA
5.5:1 (86)
1.8:1 (83)
1
2
3
4
5
6
10a
11a
12
13
14
Ph
Ph
Ph
0.05
0.10
0.05
0.05
0.32
0.26
53
63
68
62
70
87
2-Naphth
4-MeOPh
Ph
2-Naphth
Ph
1.6:1 (95)
c-C6H11
1.1:1 (29)b
Mesb
B
(c-C6H11)2BCl
Mg
Al
Zn
Ti
Zr
1.2:1 (63)
1.2:1 (48)
1.1:1 (69)
0.83:1 (48)
1.1:1 (62)
t-Bu
t-Bu
AlCl3
ZnCl2
(i-PrO)3TiCl
Cp2ZrCl2
15
a
Isolated yield of pure anti aldol (anti:syn > 25:1). b Mes )
mesityl, 2,4,6-(Me)3C6H2.
a
1
Anti:syn ratios and yields determined by 400 MHz H NMR;
tallized yield. Thus for reactions with secondary aldehydes
we recommend concentrations of 0.1 M or less to ensure
reasonable isolated yields of the desired anti-aldol prod-
uct. On the other hand, with tertiary aldehydes such as
pivalaldehyde 9, higher reaction concentrations decrease
the anti:syn selectivity only slightly, and good to excellent
isolated yields of pure anti-aldols can be obtained at
concentrations as high as 0.3 M (entries 5-6).
Another important factor in achieving high yields of
the anti-aldol product is the quenching method employed.
In our experience, good yields of anti-aldols derived from
arylacetonitriles can be obtained if the protic quench
reported yields are weight recoveries of combined anti and syn
diastereomers. Reaction performed in Et2O; workup with pH 7
buffered H2O2.
b
reactions of lithiated 1 (entry 4, Table 1). The selectivity
obtained essentially matches that reported by Hamana
for the same reaction mediated by di(n-butyl)boron
3
a
triflate and EtN(i-Pr)
2
.
The use of the strongly chelat-
ing Mg, Al, and Zn counterions might be expected to
improve anti selectivity; however in practice, the opposite
result was obtained (entries 5-7, Table 1). Finally, given
the high aldol selectivities exhibited by titanium and
zirconium enolates, the use of these metals in the nitrile
aldol reaction was explored; representative examples are
given in entries 8 and 9 (Table 1). Again, poorer
(addition of saturated aqueous NH
4
Cl) is carried out at
8
-78 °C. If the reaction is allowed to warm up before the
quench, aldol yield and diastereoselectivity decrease,
possibly as a consequence of retro-aldol reaction. Silver-
man has recently reported that yields of â-hydroxy
nitriles derived from acetonitrile are greatly improved if
one uses a trimethylsilyl chloride trapping/in situ metha-
9
selectivity resulted relative to lithiated 1.
Despite the low diastereoselectivities obtained in react-
ing lithiated arylacetonitriles with secondary aldehydes,
we have found that reasonable isolated yields of the
diastereomerically pure anti-aldols can usually be ob-
tained, due to the high crystallinity of the major isomers.
The only drawback to using our published procedure to
prepare such materials on large scale is the large solvent
volumes engendered by our previously recommended
reaction concentration of 0.025 M. From an economic
standpoint, some decrease in selectivity could be tolerated
if it significantly reduced the amount of solvent required.
Although we have not undertaken a detailed study of the
effect of concentration on diastereoselectivity in each
case, we now report that somewhat higher reaction
concentrations may be used without significantly de-
creasing the isolated yields of anti-aldol obtained, par-
ticularly in the case of bulky aliphatic aldehydes (Table
).
For example, aldol reaction of 2-naphthylacetonitrile
with benzaldehyde 6a at 0.1 M gives an 63% isolated
yield of the anti-aldol after two recrystallizations (entry
, anti-11a ). This yield is quite reasonable in view of
the known anti:syn ratio of 2.3:1 at 0.025 M. Perform-
ing the reaction at 0.4 M however decreases the anti:
syn selectivity to 1.2:1 and drastically lowers the recrys-
10
nolysis protocol rather than a protic quench. However
independent reinvestigations of the Silverman quench
procedure for the synthesis of aldol anti-10a indicated
no improvement in yield or selectivity compared to our
11
published procedure. It is however possible that ap-
plication of the Silverman protocol to the other examples
in Table 2 might result in further yield improvements.
To date the diastereoselectivity of the nitrile aldol
reaction has been rationalized in terms of steric effects
in the competing transition structures. Clearly, elec-
tronic effects could play a major role in determining
diastereoselectivity and may limit the range of useful
aldehyde substrates. To determine the sensitivity of
nitrile aldol diastereoselectivity to electronic factors, the
reaction of lithiated 1 with various para-substituted
benzaldehydes 6b-f in THF at 0.025 M was studied
2
(Table 3).
2
As can be seen in Table 3, substitution of benzaldehyde
6
a with electron-donating groups resulted in little change
2
in the diastereoselectivity (entries 1-4). However, as the
electron electron-withdrawing power of the substituent
increases, selectivity decreases (entries 5-7). The effect
is quite significant: reaction of p-nitrobenzaldehyde 6g
4
b
(
8) Patterson, I. In Comprehensive Organic Synthesis; Trost, B. M.,
(10) Zhou, J . P. J .; Zhong, B.; Silverman, R. B. J . Org. Chem. 1995,
60, 2261-2262.
Fleming, I., Heathcock, C. H., Eds.; Pergamon Press: Oxford, 1991;
Vol. 2, pp 301-319.
(11) Silverman et al. previously reported that application of their
protocol to the reaction of phenylacetonitrile and benzaldehyde gave
a single aldol product of unspecified relative configuration. However
subsequent reexamination in our laboratory and in the Silverman
group revealed the product to be approximately a 4:1 mixture of anti-
and syn- diastereomers (private communication, R. B. Silverman to
P. R. Carlier).
(
9) Kasatkin (ref 3b) had previously reported an threo:erythro ) 54:
4
3
6 ratio for addition of (i-PrO) TiCl transmetalated lithiophenylaceto-
nitrile to benzaldehyde in THF. Taking into consideration the fact
that their assignment should be reversed (see ref 4b, footnote 5), the
actual ratio they obtained corresponds to 0.85:1, essentially what we
observe.

6
318 J . Org. Chem., Vol. 62, No. 18, 1997
Carlier et al.
Ta ble 3. Electr on ic P er tu r ba tion of Nitr ile Ald ol
Rea ction Dia ster eoselectivity
anti:syna
σp (X)
-
entry
aldehyde
X
1
2
3
4
5
6
7
6b
6c
6d
6a
6e
6f
Me2N
MeO
Me
H
Cl
5.6:1
4.7:1
5.6:1
5.5:1
3.4:1
1.5:1
0.9:1
-0.32
-0.12
-0.14
0
0.24
0.62
1.25
CF3
NO2
6g
a
Anti:syn ratios determined by 400 MHz 1H NMR. Yields of
the combined aldols range from 73 to 98% and are given in the
Experimental Section.
F igu r e 1. Plot of log(anti:syn) for aldols anti-10a -f versus
p
-
σ
(X).
is completely unselective (entry 7, Table 3). Studies on
the effect of benzaldehyde substituents on aldol thermo-
chemistry12 and diastereoselectivity have been reported
previously. A limited study (four substrates) of the
tris(dialkylamino)sulfonium enolate of cyclohexanone
demonstrated that diastereoselectivity with electron-
Ta ble 4. Com p a r ison of th e Ald ol Dia ster eoselectivity of
P h en yla ceton itr ile (1) w ith Isova ler on itr ile (4) a n d
ter t-Bu tyla ceton itr ile (5)
13
R
R′
(nitrile)
entry
aldol
(aldehyde)
anti:syna
1
2
3
4
5
6
10a
15
16a
17
18a
19
Ph (6a )
t-Bu (9)
Ph (6a )
t-Bu (9)
Ph (6a )
t-Bu (9)
Ph (1)
Ph (1)
i-Pr (4)
i-Pr (4)
t-Bu (5)
t-Bu (5)
5.5:1
60:1
0.26:1
4.4:1
1.3:1
>20:1
deficient benzaldehydes was slightly higher than that
with electron-rich aldehydes.1
lithium enolates of cyclohexanone
3a
Other studies on the
1
3b
and 2,6-dimethyl-
1
3d
cyclohexanone
and R-halogenated Sn(II) imide
enolates1 did not demonstate any meaningful trend. The
results obtained in our study are unique in that they
qualitatively match the expectation that decreasing
electron density in the aldehyde would render it more
reactive and consequently less selective (the so-called
reactivity-selectivity principle). Why similar trends
were not observed in the enolate studies described above
is not clear. However to gain further insight into our
results, correlation of the log(anti:syn) for reactions of
lithiated 1 with 6a -g versus various Hammett param-
eters was then performed. By far the best correlation
3c
a
Anti:syn ratios determined by 300 or 400 MHz 1H NMR, or
capillary GC for ratios higher than 20:1. Yields of the combined
aldols range from 63 to 84% and are given in the Experimental
Section.
Sch em e 1a
-
14
was obtained using the σ
). The optimum correlation of log (anti:syn) with σ
p
parameter (r ) 0.97, Figure
-
1
p
in
a
(
a) 1:1 LiAlH4/AlCl3, Et2O; (b) (Cl3CO)2CO, Et3N, CH2Cl2,
our case is interesting and may be indicative of the
nature of charge distribution in the transition state.15
However at this time it is only possible to point out that
synthetically useful diastereoselectivity cannot be achieved
with extremely electron deficient benzaldehydes such as
-78 f 0 °C.
to benzaldehyde with little selectivity (3:1 and 2:1
respectively, unassigned).5 To determine if increasing
steric bulk in the nitrile moiety would raise selectivities
to a useful level, aldol reactions of lithiated isovaleroni-
trile 4 and tert-butylacetonitrile 5 were studied in THF
at 0.025 M (Table 4).
6
f or 6g.
To date our studies of the nitrile aldol reaction have
been limited to arylacetonitriles. The utility of this
reaction would greatly increase if aliphatic nitriles could
be shown to react with reliable and high diastereoselec-
tivity. Kauffman reported that lithiated propionitrile and
the corresponding Fe(II) derivative underwent addition
From the standpoint of aldol reaction with aldehydes,
it was expected that lithiated isovaleronitrile 4 would be
approximately isosteric with lithiated phenylacetonitrile
1
. However, selectivities for the reactions of 4 are
(
12) Arnett, E. M.; Palmer, C. A. J . Am. Chem. Soc. 1990, 112, 7354-
markedly lower than those of 1 (cf. entries 1 and 3;
entries 2 and 4, Table 4). More significantly, the reaction
of 4 with 6a is actually syn-selective (entry 3, Table 4).
The assignment of the major diastereomer as syn was
made on the basis of the vicinal coupling constants and
was supported by transformation of the major aldol to
7
360.
(13) (a) Noyori, R.; Nishida, I.; Sakata, J . J . Am. Chem. Soc. 1983,
1
1
05, 1598-1608. (b) Majewski, M.; Gleave, D. M. Tetrahedron Lett.
989, 30, 5681-5684. (c)Pridgen, L. N.; Abdel-Magid, A. F.; Lantos,
I.; Shilcrat, S.; Eggleston, D. S. J . Org. Chem. 1993, 58, 5107. (d)
Mateos, A. F.; Coca, G. P.; Alonso, J . J . P.; Gonz a´ lez, R. R.; Hern a´ ndez,
C. T. Tetrahedron Lett. 1995, 36, 961-964.
4b
(14) Exner, O. In Correlation Analysis of Chemical Data; Plenum
the corresponding cyclic carbamate cis-20a (Scheme 1).
Press: New York, 1988; pp 61-62.
Observation that 3
J 5-6 of cis-20a is 4.1 Hz confirms
(15) J acobsen et al. have argued that the optimum correlation of
+
log(cis:trans) to σ (rather than σ ) for the epoxidation of cis-methyl
cinnamate derivatives indicates that a carbocationic intermediate is
not on the reaction path (J acobsen, E. N.; Deng, L.; Furukawa, Y.;
Martinez, L. E. Tetrahedron 1994, 50, 4323-4334.)
that the i-Pr and Ph substituents are cis; the fact that
3
J
4a-5 is greater than 7 Hz indicates that the i-Pr
substituent is equatorial, a result unexpected on the basis

Nitrile Aldol Reaction
J . Org. Chem., Vol. 62, No. 18, 1997 6319
of the relative A values of i-Pr and Ph.16 AM1 geometry
optimization calculations confirm however that the i-Pr-
and LDA (2.0 M, 18.2 mL, 36.4 mmol) were added via syringe,
and the mixture was stirred for 30 min. Pivalaldehyde 9 (4.0
mL, 3.17 g, 36.8 mmol) was added via syringe; after 30 min at
-
1
equatorial conformer is 1.3 kcal mol more stable than
the i-Pr-axial conformer.17 Reactions of tert-butylaceto-
-78 °C, the septum was removed and saturated NH
solution (10 mL) was added all at once. The cooling bath was
removed, and after 10 min 1 N HCl (100 mL) and Et O (100
4
Cl
nitrile 5 are of interest from the viewpoint that selectivity
with benzaldehyde 6a is remarkably low (anti:syn ) 1.3:
2
mL) were added. Separation of layers, extraction of the
1
, entry 5, Table 4). Useful stereoselectivity is achieved
aqueous phase (2 × 100 mL of Et O), drying of the combined
2
only when steric effects are pushed to the conceivable
limit (R′ ) R ) t-Bu, entry 6, Table 4). Clearly the aldol
selectivities exhibited by aliphatic nitriles 4 and 5 cannot
be accounted for by the transition state model previously
advanced for arylacetonitriles.4b It is possible that the
structures of lithiated aliphatic nitriles differ consider-
ably from their arylacetonitrile counterparts and that this
difference is also reflected in the corresponding nitrile
aldol transition structures. Solution structural studies
on lithiated 4 and 5 may shed light on these issues.
In conclusion, we have determined that lithium offers
the best diastereoselectivity in addition of metalated
arylacetonitriles to aldehydes. Secondly, a reinvestiga-
tion of the effect of concentration on aldol selectivity
reveals that for secondary aldehydes reasonable recrys-
tallized yields of the anti-aldol products can be obtained
at reaction concentrations up to 0.1 M, due to the high
crystallinity of these materials. In the case of pivalal-
dehyde, reaction concentrations up to 0.3 M provide good
to excellent isolated yields of the anti-aldol products.
Third, aldol diastereoselectivity with para-substituted
benzaldehydes decreases dramatically with increasing
electron-withdrawing power of the substituent, to such
an extent that reactions with p-(trifluoromethyl)benzal-
dehyde 6f and p-nitrobenzaldehyde 6g do not offer
synthetically useful selectivity. Finally we have estab-
lished that aliphatic nitriles, unlike arylacetonitriles, do
not offer reliable anti-selectivity in reactions with alde-
hydes.
4
organic layers (MgSO ), filtration, and concentration in vacuo
yielded the crude product, which was recrytallized from
toluene (9 mL)/hexane (60 mL) to afford 5.34 g (87%) of anti-
1
5 as white crystals. The material was greater than 95% pure
1
13
by H and C NMR (spectral and analytical data published
previously ).
4b
(
2RS,3RS)-3-Hyd r oxy-2,3-d ip h en ylp r op ion itr ile (a n ti-
1
0a ). Reaction of 1 and 6a was perfomed as above on a 15
mmol scale in 300 mL of THF; recrystallization from toluene/
hexane gave 1.81 g of anti-10a (53%). The material was
1
13
greater than 95% pure by H and C NMR (spectral data
published previously ).
1
8
(
2R S ,3R S )-3-H yd r oxy-2-(2′-n a p h t h yl)-3-p h e n ylp r o-
p ion itr ile (a n ti-11a ). Reaction of 2-naphthylacetonitrile 2
and 6a was performed as above on a 15 mmol scale in 150 mL
of THF; two successive recrystallizations from CH
2 2
Cl /hexane
gave 2.57 g of anti-11a (63%). The material was greater than
1
13
95% pure by H and C NMR (spectral and analytical data
4
b
published previously ).
(2R S ,3S R )-3-C yc lo h e x yl-3-h y d r o xy-2-(4′-m e t h o xy-
p h en yl)p r op ion itr ile (a n ti-12). Reaction of p-methoxyphe-
nylacetonitrile 3 and cyclohexanecarboxaldehyde 7 was per-
formed as above on a 20 mmol scale in 400 mL of THF; two
successive recrystallizations from CHCl
3
/hexane gave 3.51 g
of anti-12 (68%). The material was greater than 95% pure by
1
13
H and C NMR (spectral and analytical data published
4
b
previously ).
(2R S ,3R S )-3-H yd r oxy-2-p h e n yl-3-(2′,4′,6′-t r im e t h yl-
p h en yl)p r op ion itr ile (a n ti-13). Reaction of 1 and mesityl
aldehyde 8 was performed as above on a 8 mmol scale in 150
mL of THF; recrystallization from toluene/hexane gave 1.32 g
of anti-13 (62%). The material was greater than 95% pure by
1
13
H and C NMR (spectral and analytical data published
4
b
previously ).
Exp er im en ta l Section
(2RS,3RS)-3-Hyd r oxy-4,4-d im et h yl-2-p h en ylp en t a n e-
n itr ile (a n ti-14). Reaction of 2 and pivalaldehyde 9 was
performed as above on a 30 mmol scale in 90 mL of THF;
Gen er a l. All reactions were performed in oven-dried
glassware under a nitrogen atmosphere. THF was distilled
from Na/benzophenone immediately prior to use. Lithium
diisopropylamide (2.0 M in THF) and lithium hexamethyld-
isilazide (1.0 M in THF) were purchased from Aldrich and
titrated prior to use (2,2′-bipyridine as indicator). Sodium
hexamethyldisilazide (1.0 M in THF) and potassium hexam-
ethyldisilazide (0.5 M in toluene) were used as received from
Aldrich. tert-Butylacetonitrile (5) was prepared by dehydra-
tion of the corresponding amide; all other nitriles and all
aldehydes were purchased from Aldrich and distilled prior to
recrystallization from toluene/hexane gave 5.41 g of anti-14
1
(
70%). The material was greater than 95% pure by H and
13
4b
C NMR (spectral and analytical data published previously ).
(
2RS,3RS)-3-H yd r oxy-2-p h en yl-3-(N′,N′-d im et h yl-4′-
a m in op h en yl)p r op ion itr ile (a n ti-10b). The crude aldols
were obtained in 73% yield (anti:syn ) 5.6:1) from p-(dim-
ethylamino)benzaldehyde 6b and 1. Pure anti-10b was iso-
1
lated by recrystallization from dichloromethane/hexane.
H
NMR: 2.307 (d, J ) 2.9 Hz, 1H), 2.937 (s, 6H), 4.028 (d, J )
5
9
1
3
1
.9 Hz, 1H), 4.876 (dd, J ) 2.9 Hz, 5.9 Hz, 1H), 6.63-7.33 (m,
use. Unless otherwise noted, NMR spectra were recorded in
13
H). C NMR: 40.44, 47.27, 76.35, 112.18, 119.18, 126.99,
1
CDCl
3
, at 400 or 300 MHz for H, and 100.75 or 75.48 MHz
27.16, 128.29, 128.49, 128.77, 132.97, 150.77. IR (KBr):
1
3
for C. Elemental analysis was performed at the Shanghai
Institute of Organic Chemistry (Chinese Academy of Sciences,
PRC) and at Medac Ltd. (Brunel University, UK). Aldol
reactions were carried out according to the published
procedure, and unless otherwise noted, all reactions were
performed on a 1 mmol scale at a concentration of 0.025 M.
Rep r esen ta tive Gr a m -Sca le Ald ol Rea ction P r oce-
d u r e: (2RS,3RS)-3-Hyd r oxy-4,4-d im eth yl-2-(2′-n a p h th -
yl)p en ta n en itr ile (a n ti-15). An oven-dried 500 mL round-
bottomed flask was charged with THF (100 mL) and stirring
bar, capped with a septum, purged with nitrogen, and cooled
to -78 °C. Phenylacetonitrile 1 (3.5 mL, 3.55 g, 30.3 mmol)
+
432.0 (s), 2248.0 (s). MS (CI (NH
3
)): 266.7 (M + 1). Mp:
18 2
13.7-114.5 °C. Anal. Calcd for C17
H N : C, 76.67; H, 6.81;
N, 10.52. Found: C, 76.65; H, 6.79; N, 10.50.
2RS,3RS)-3-Hyd r oxy-3-(4′-m eth oxyp h en yl)-2-p h en yl-
p r op ion itr ile (a n ti-10c). The crude aldols were obtained in
(
4
b
8
1% yield (anti:syn ) 4.7:1) from p-methoxybenzaldehyde 6c
and 1. Pure anti-10c was isolated by recrystallization from
toluene. H NMR: 2.427 (d, J ) 2.4 Hz, 1H), 3.792 (s, 3H),
4
6
2
1
1
.023 (d, J ) 5.9 Hz, 1H), 4.931 (dd, J ) 3.4 Hz, 5.9 Hz, 1H),
13
.81-6.86 (m, 2H), 7.15-7.35 (m, 7H). C NMR: 47.38, 55-
9, 76.08, 113.90, 118.87, 127.50, 128-46, 128.88, 131.49,
32.55, 159.84. IR (KBr): 3460.0 (broad), 2238.0 (m). MS
+
(CI (NH
3
)): 252.9 (M + 1). Mp: 101.1-102.9 °C. Anal. Calcd
(
16) Eliel, E. L.; Wilen, S. H. In Stereochemistry of Organic
Molecules; J ohn Wiley & Sons: New York, 1994; pp 696-697.
17) Heats of formation for the i-Pr-equatorial and i-Pr-axial
for C16
H
15NO
2
: C, 75.87; H, 5.97; N, 5.53. Found: C, 75.59;
H, 5.95; N, 5.43.
(
conformers of cis-20 were determined by means of AM1 geometry
optimization calculations (Spartan 2.0 on a Silicon Graphics Personal
Iris 4D/35).
(18) Wade, P. A.; Bereznak, J . F. J . Org. Chem. 1987, 52, 2973-
2977.

6
320 J . Org. Chem., Vol. 62, No. 18, 1997
2RS,3RS)- a n d (2RS,3SR)-3-H yd r oxy-3-(4′-m et h yl-
Carlier et al.
(
7H), 8.143 (d, J ) 8.8 Hz, 2H). 13C NMR: 46.47, 75.05, 118.10,
123.46, 127.71, 128.62, 129.08, 129.15, 130.89, 145.66, 148.06.
IR (KBr): 3548.0 (s), 2244.0 (s), 1520.0 (vs), 1348.0 (vs). MS
p h en yl)-2-p h en ylp r op ion itr ile (a n ti- a n d syn -10d ). The
crude aldols were obtained in 77% yield (anti:syn ) 5.6:1) from
p-methylbenzaldehyde 6d and 1. Pure anti-10d was isolated
+
(CI (NH
3
)): 285.9 (M + NH
4
). Mp: 121.8-123.4 °C. Anal.
: C, 67.16; H, 4.51; N, 10.44. Found: C,
1
by recrystallization from toluene. H NMR: 2.327 (s, 3H),
Calcd for C15
H
12
N
2
0
3
2
.486 (d, J ) 3.4 Hz, 1H), 4.022 (d, J ) 5.9 Hz, 1H), 4.922 (dd,
67.77; H, 4.40; N, 10.40.
1
3
J ) 3.7 Hz, 5.7 Hz, 1H), 7.1-7.3 (m, 9H). C NMR: 21.14,
4
1
(2RS,3SR)- a n d (2RS,3RS)-3-Hyd r oxy-3-p h en yl-2-(2′-
p r op yl)p r op ion itr ile (syn - a n d a n ti-16a ). Aldol reaction
of isovaleronitrile 4 and 6a on a 20 mmol scale (crude anti:
syn ) 1:2.8), followed by flash chromatography (CH Cl )
7.29, 76.21, 118.78, 126.10, 128.42, 128.84, 129.19, 132.59,
36.45, 138.51. IR (KBr): 3418.0 (broad), 2252.0 (s). MS
+
(
C
EI ): 237.0 (M). Mp: 110.6-111.4 °C. Anal. Calcd for
2
2
16
H
15NO: C, 80.99; H, 6.37; N, 5.90. Found: C, 80.96; H,
afforded 1.87 g of the first eluting syn-16a , 0.37 g of anti-16a
(≈ 5% syn- impurity), and 0.53 g of a mixed fraction (anti:syn
≈ 1:1). Total isolated yield of aldols was 74%.
6
.37; N, 5.89.
To obtain syn-10d the mother liquor was chromatographed
1
(25% ethyl acetate in hexane). Although separation of the
syn -16a (R ) 0.39). H NMR: δ 1.062 (d, J ) 6.8 Hz, 6H),
f
diastereomers was not apparent on TLC, the last fraction
proved to be >95% syn-10d . H NMR: 2.266 (d, J ) 2.4 Hz,
1
Hz, IH), 7.1-7.3 (m, 9H).
1
1
2.174 (m, 1H), 2.679 (d, J ) 2.7 Hz, 1H), 2.768 (dd, J ) 4.6,
1
13
8.9 Hz, 1H), 4.687 (dd, J ) 2.7, 8.9 Hz, 1H), 7.35 (m, 5H).
C
H), 2.348 (s, 3H), 4.123 (d, J ) 6.8 Hz, 1H), 4.953 (d, J ) 6.8
NMR: δ 17.71, 21.50, 26.19, 47.90, 72.08, 118.36, 126.44,
1
3
-1
C NMR: 22.21, 46.63, 76.13,
128.63, 128.75, 140.82. IR (NaCl): 3446 (s), 2244 (m) cm .
+
18.76, 126.50, 128.60, 128.71, 128.89, 129.17, 132.11, 135.88,
MS (CI (CH )): 190 (M + 1). Mp: liquid. Anal. Calcd for
C H NO: C, 76.16; H, 7.99; N, 7.40. Found: C, 75.79; H,
4
38.81. IR(KBr): 3360.0 (broad,s), 2242.0 (m). MS
1
2
15
+
(CI (NH
3
)): 238.0 (M + 1). Mp: 101.0-103.0 °C. Anal. Calcd
7.94; N, 7.33.
a n ti-16a (R
for C16
H, 6.36; N, 5.89.
H15NO: C, 80.99; H, 6.37; N, 5.90. Found: C, 80.67;
) 0.31). 1H NMR (CDCl
f
3
+ D O): δ 1.064 (d,
2
J ) 6.8 Hz, 3H), 1.115 (d, J ) 6.8 Hz, 3H), 1.66-1.75 (m, 1H),
2.728 (dd, J ) 7.3, 5.1 Hz, 1H), 4.852 (d, J ) 7.3 Hz, 1H), 7.36-
(
2RS,3RS)-3-Hyd r oxy-2-p h en yl-3-(4′-ch lor op h en yl)p r o-
1
3
p ion itr ile (a n ti-10e). The crude aldols were obtained in 76%
yield (anti:syn ) 3.4:1) from p-chlorobenzaldehyde 6e and 1.
TLC (25% ethyl acetate in hexane) did not exhibit visible
separation of anti- and syn-10e, but flash chromatography and
7.41 (m, 5H). C NMR: δ 18.64, 21.39, 27.27, 48.51, 72.26,
119.18, 126.12, 128.62, 128.77, 140.58. IR (KBr): 3248 (s),
-
1
+
2256 (m) cm . MS (CI (CH
4
)): 190 (M + 1). Mp: 68.8-69.6
°C. Anal. Calcd for C12 15NO: C, 76.16; H, 7.99; N, 7.40.
H
concentration of the first few aldol fractions afforded pure anti-
Found: C, 76.20; H, 8.12; N, 7.40.
1
1
0e. H NMR: 2.522 (d, J ) 3.4 Hz, 1H), 4.023 (d, J ) 5.9
(2RS,3RS)-3-H yd r oxy-4,4-d im et h yl-2-(2′-p r op yl)p en -
ta n en itr ile (a n ti-17). The crude aldols were obtained in 63%
yield (anti:syn ) 4.4:1) from 4 and pivalaldehyde 9. An
analytically pure sample of anti-17 was prepared by recrys-
tallization from hexane. 1H NMR: δ 1.000 (s, 9H), 1.074 (d, J
) 6.4 Hz, 3H), 1.168 (d, J ) 6.8 Hz, 3H), 1.827 (d, J ) 7.8 Hz,
1H), 2.008 (d, J ) 6.8 Hz, 1H), 2.578 (dd, J ) 1.0 Hz, 6.8 Hz,
1H), 3.351 (d, J ) 7.3 Hz, 1H). 13C NMR: δ 20.12, 21.01, 25.79,
Hz, 1H), 4.974 (dd, J ) 3.7 Hz, 5.6 Hz, 1H), 7.18-7.35 (m,
9
1
1
3
H). C NMR: 47.33, 75.69, 118.43, 127.63, 128.42, 128.73,
29.06, 132.04, 134.58, 137.75. IR (KBr): 3484.0 (s), 2244.0
+
(s). MS (CI (NH
3
)): 275.1 (M + NH
12NOCl: C, 69.90; H, 4.69; N, 5.43.
Found: C, 69.89; H, 4.65; N, 5.41.
2RS,3RS)- a n d (2RS,3SR)-3-Hyd r oxy-2-p h en yl-3-(4′-
tr iflu or om eth yl)p h en yl)p r op ion itr ile (a n ti- a n d syn -
0f). The crude aldols were obtained in 98% yield (anti:syn
1.5:1) from p-(trifluoromethyl)benzaldehyde 6f and 1. Flash
4
). Mp: 75.3-76.5 °C.
Anal. Calcd for C15
H
(
(
1
30.71, 35.81, 41.28, 77.49, 119.27. IR (KBr): 3476.0 (s), 2248.0
+
(m). MS (CI (NH
3
)): 187.0 (M + NH
19NO: C, 70.96; H, 11.31; N, 8.27.
Found: C, 70.86; H, 11.57; N, 8.29.
4
). Mp: 76.9-77.9 °C.
)
Anal. Calcd for C10H
chromatography (25% ethyl acetate/hexane) afforded separa-
tion of the aldol diastereomers.
(2R S ,3R S )-3-H y d r o x y -3-p h e n y l-2-t -b u t y lp r o p i o -
n itr ile (a n ti-18a ). The crude aldols were obtained in 71%
yield (anti:syn ) 1.3:1) from tert-butylacetonitrile 5 and 6a .
a n ti-10f (R
.9 Hz, 1H), 5.033 (d, J ) 4.9 Hz, 1H), 7.20-7.58 (m, 9H).
f
) 0.19). 1H NMR: 2.893 (s, 1H), 4.041 (d, J )
13
5
C
1
NMR: 47.27, 75.60, 118.36, 123.91 (q, J CF ) 274 Hz), 125.46
2
Pure anti-18a was obtained by recrystallization from Et O/
3
2
1
(
q, J CF ) 4 Hz), 126.71, 128.40, 128.86, 129.13, 130.84 (q, J CF
hexane. H NMR: δ 1.201 (s, 9H), 2.569 (d, J ) 2.1 Hz, 1H),
13
)
31 Hz), 131.89, 143.21. IR (KBr): 3482.0 (s), 2244.0 (s),
5.038 (d, J ) 2.1 Hz), 7.31-7.44 (m, 5H). C NMR: δ 28.32,
+
1
°
330.0 (s). MS (CI (NH
C. Anal. Calcd for C16
Found: C, 66.00; H, 4.08; N, 4.81.
syn -10f (R
.137 (d, J ) 6.8 Hz, 1H), 5.005 (dd, J )3.4 Hz, 6.4 Hz, 1H),
3
)): 308.8 (M + NH
4
). Mp: 94.7-96.6
33.79, 53.14, 70.57, 118.30, 125.44, 128.33, 128.76, 142.51. IR
-
1
+
H
12NOF
3
: C, 65.98; H, 4.18; N, 4.81.
(KBr): 3434 (s), 2252 (m) cm . MS (CI (NH
3
)): 221 (M +
NH ). Mp: 114.3-116.3 °C. Anal. Calcd for C13
4
H
17NO: C,
) 0.13). 1H NMR: 2.500 (d, J ) 3.4 Hz, 1H),
76.81; H, 8.43; N, 6.89. Found: C, 76.84; H, 8.47; N, 7.03.
(2R S ,3R S )-3-H yd r oxy-4,4-d im e t h yl-2-t er t -b u t ylp e n -
ta n en itr ile (a n ti-19). anti-19 was obtained in 84% yield from
f
4
7
1
3
.19-7.60 (m, 9H). C NMR: 46.56, 75.51, 118-27, 123.90
1
3
(
1
q, J CF ) 272 Hz), 125.38 (q, J CF ) 4 Hz), 127.10, 128.69,
5 and 9. An analytically pure sample was prepared by
recrystallization from n-heptane. H NMR (CDCl
2
1
28.97, 129.11, 131.27 (q, J CF ) 31 Hz), 131.33, 142.57. IR
3
2
+ D O):
+
13
(
(
C
4
KBr): 3452.0 (s), 2242.0 (s), 1336.0 (s). MS (CI (NH
NH ). Mp: 102.0-103.4 °C. Anal. Calcd for
: C 65.98; H, 4.18; N, 4.81. Found: C, 65.69; H,
.19; N, 4.72.
2RS,3RS)- a n d (2RS,3SR)-3-Hyd r oxy-2-p h en yl-3-(4′-
3
)): 308.8
0.984 (s, 9H), 1.109 (s, 9H), 2.56 (s, 1H), 3.447 (s, 1H).
NMR (CDCl + D O): 25.62, 27.83, 33.55, 35.91, 45.34, 119.14.
IR (KBr): 3492 (s), 2236 (m) cm . MS (CI (NH
NH
C
M
+
4
3
2
-
1
+
16
H
12NOF
3
3
)): 201 (M +
). Mp: 129.5-129.7 °C. Anal. Calcd for C11H21NO: C,
2.08; H, 11.55; N, 7.64. Found: C, 71.85; H, 11.83; N, 7.57.
4
7
(
n itr op h en yl)p r op ion itr ile (a n ti- a n d syn -10g). The crude
aldols were obtained in 88% yield (anti:syn ) 0.9:1) from
p-nitrobenzaldehyde 6g and 1. Flash chromatography (25%
ethyl acetate/hexane) afforded separation of the aldol diaster-
eomers.
(5RS,6SR)-6-P h en yl-5-isop r op yltetr a h yd r o-1,3-oxa zin -
2-on e (cis-20a ). The conversion of syn-16a to the carbamate
cis-20a was carried out according to our previously published
two-step procedure.4b H-NMR: δ 0.837 (d, J ) 6.8 Hz, 1H),
1.026 (d, J ) 6.8 Hz, 1H), 1.3-1.4 (m, 1H), 2.0-2.1 (m, 1H),
3.163 (dd, J ) 9.8, 11.7 Hz, 1H), 3.4-3.5 (m, 1H), 5.550 (d, J
1
) 0.46). 1H NMR: 3.001 (s, broad, 1H), 4.081
a n ti-10g (R
f
1
3
)
3.9 Hz, 1H), 7.10 (br, 1H), 7.3-7.4 (m, 5H). C NMR: δ
(
d, J ) 5.9 Hz, 1H), 5.123 (d, J ) 5.4 Hz, 1H), 7.17-7.45 (m,
13
19.67, 21.81, 25.89, 39.75, 41.59, 80.90, 126.49, 128.15, 128.46,
7
1
9
H), 8.160 (d, J ) 8.8 Hz, 2H). C NMR: 47.18, 75.27, 118.09,
23.65, 127.32, 128.40, 129-06, 129.26, 131.47, 146.25, 147-
-
1
1
37.02, 154.93. IR (KBr): 3232 (w, br), 1711 (vs) cm . MS
+
(CI (CH
4
)): 220.2 (M + H). Mp: 109.7-113.8 °C. Anal. Calcd
9. IR (KBr): 3464.0 (s), 2244.0 (s), 1514.0 (vs), 1350.0 (vs).
+
for C13
H
17NO
2
: C, 71.21; H, 7.81; N, 6.39. Found: C, 71.18;
MS (CI (NH
3
)): 285.9 (M + 2NH
C, 67.16; H, 4.51; N, 10.44.
Found: C, 67.08; H, 4.49; N, 10.45.
syn -10g (R
) 0.40). 1H NMR: 2.919 (s, broad, 1H), 4.187
d, J ) 6.4 Hz, 1H), 5.140 (d, J ) 6.4 Hz, 1H), 7.16-7.53 (m,
3
). Mp: 143.5-144.3 °C.
H, 7.86; N, 6.45.
Anal. Calcd for C15
12 2 3
H N O :
Ack n ow led gm en t. We thank Mr. Chi Yeung Kwok
for performing initial aldol experiments with nitrile 5.
f
(

Nitrile Aldol Reaction
J . Org. Chem., Vol. 62, No. 18, 1997 6321
We also thank the Research Grants Council (H.K.) for
financial support of this work (Competitive Earmarked
Grants HKUST 203/93E and 588/95P).
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.
Su p p or tin g In for m a tion Ava ila ble: 13C and 1H NMR
spectra of anti-16a and anti-18a (4 pages). This material is
J O9702148