'meso-Selective' functionalisation of N-benzyl-α-methylbenzylamine derivatives by α-lithiation and alkylation

Article abstract of DOI:10.1016/S0040-4039(02)00154-5

Lithiation and methylation of amide and carbamate derivatives of α-methylbenzylamine proceeds with high diastereoselectivity in favour of meso bis-α-methylbenzylamine derivatives. Carboxylation of the intermediate organolithium is also diastereoselective, and with N-Boc p-methoxy-α-methylbenzylamine as starting material, oxidative cleavage provides a new asymmetric route to phenylglycine. Other electrophiles give a range of stereochemical outcomes, apparently depending on the stereospecificity of their reactions with a pair of diastereoisomeric organolithiums of low to moderate configurational stability.

Full text of DOI:10.1016/S0040-4039(02)00154-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 1955–1959
‘meso-Selective’ functionalisation of
N-benzyl-a-methylbenzylamine derivatives by a-lithiation
and alkylation
Ryan A. Bragg, Jonathan Clayden* and Christel J. Menet
Department of Chemistry, University of Manchester, Oxford Road, Manchester M13 9PL, UK
Received 7 December 2001; revised 17 January 2002; accepted 18 January 2002
Abstract—Lithiation and methylation of amide and carbamate derivatives of a-methylbenzylamine proceeds with high diastereose-
lectivity in favour of meso bis-a-methylbenzylamine derivatives. Carboxylation of the intermediate organolithium is also
diastereoselective, and with N-Boc p-methoxy-a-methylbenzylamine as starting material, oxidative cleavage provides a new
asymmetric route to phenylglycine. Other electrophiles give a range of stereochemical outcomes, apparently depending on the
stereospecificity of their reactions with a pair of diastereoisomeric organolithiums of low to moderate configurational stability.
© 2002 Elsevier Science Ltd. All rights reserved.
(R,R)- and (S,S)-N,N-Bis(a-methylbenzyl)amine 1 and
its derivatives are important reagents for asymmetric
synthesis. Their lithium derivatives, for example, are
among the most widely used and generally successful of
the chiral lithium amide bases.1 Amine 1 is made
stereoselectively by reductive amination of the appro-
priate a-methylbenzylamine with acetophenone, a reac-
tion which proceeds with 8:1 selectivity in favour of
chiral stereoisomer 1 over the meso stereoisomer 2.2
Achiral 2 is by contrast much less useful, and rarely
required,3 and as a result there is no stereoselective
route to 2 and its derivatives.
with high stereoselectivity (97:3 for 5 and 90:10 for 6).7
Good yields of both meso amides were isolated (Scheme
1). The stereochemistry of the products was indicated
by their lack of optical rotation and by spectroscopic
comparison with authentic samples of amides derived
from 1.
Though ideal for our purposes, using a chiral com-
pound to make an achiral one, albeit stereoselectively,
is of limited appeal. We therefore investigated the
stereoselective synthesis of unsymmetrical benzylamine
derivatives by the lithiation and alkylation of the amide
7 and the carbamate 8 with a range of electrophiles.
Using a Carousel Reaction Station (Radleys Ltd.), sam-
ples of 7 and 8 were stirred with t-BuLi in THF at
−78°C for 30–120 min and quenched in parallel with
the electrophiles shown in Table 1 (Scheme 2). The
product ratios (determined by NMR and HPLC) and
the isolated yields are shown in Table 1.
Ph Ph
Ph Ph
N
H
N
H
1
2
As part of a study4 of stereospecificity in some cyclisa-
tion reactions,5 we required both diastereoisomers of
some amide derivatives of 1 and 2. We therefore made
amides 3 and 4 from a-methylbenzylamine and treated
them with t-BuLi in THF at −78°C. Quenching the
deep red organolithiums with methyl iodide after 2 h
gave amides 5 and 6.6 We had expected to obtain both
as a mixture of diastereoisomers, but surprisingly the
meso diastereoisomer of both compounds was formed
Stereoselectivity varied according to the electrophile
and was different for reactions of the amide 7 and the
* Corresponding author. Fax: +44 161 275 4939; e-mail: j.p.clayden@
man.ac.uk
Scheme 1. meso-Selective alkylation of amides.
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)00154-5

1956
R. A. Bragg et al. / Tetrahedron Letters 43 (2002) 1955–1959
Table 1.
Entry
E+
E
Yield 11 (%)
Ratio 11a:11b
Yield 12 (%)
Ratio 12a:12b
1
2
3
4
5
6
7
8
MeOD
MeI
BnBr
AllylBr
D
Me
Bn
Allyl
CO2H
CO2Me
88
80
84
85
78
90
–
–
–
75
97:3a
97:3b
\98:2a
93:7a
89:11b
98:2b
–
98
82c
68
70
–
\98:2a
96:4b
50:50
50:50
–
CO2
ClCO2Me
Cyclobutanone
Cyclohexanone
PhCHꢀNPh
Bu3SnCl
97d
63
40
87
67
\98:2b
75:25a
–
–
\98:2a
a,e
9
10
SnBu3
11:89a
36:64b
a Unconfirmed stereochemistry.
b Proven stereochemistry (see below).
c White solid, mp=64–66°C, [h]D25 −8 (c=1.21); lH (CDCl3, 300 MHz) 7.4–7.2 (7H, m, ArH), 6.91 (2H, d, J 8, ArH), 3.89 (3H, s, MeO), 1.2 (6H,
b, CH3), 1.12 (9H, s, Boc). lH (DMSO, 100°C, 300 MHz) 5.0 (1H, q, CHCH3), 4.84 (1H, q, CHCH3), 3.88 (3H, s, CH3O), 1.35 (3H, d, CH3),
1.33 (3H, d, CH3), 1.25 (9H, s, Boc). m/z (CI) 356 (M+H+, 15%), 300 (80%), 256 (20%), 125 (100%). Found M+H+ 356.2227; C22H30NO3
requires M, 356.2225. wmax/cm−1 (film) 1686, 1512.
d Colorless oil. [h]D23 +90.9 (c=0.44); lH (CDCl3, 300 MHz) 7.5–7.3 (7H, m, ArH), 6.9 (2H, d, J 9, ArH), 5.8 (1H, b, CHMe), 4.7 (1H, b,
CHCO2Me), 3.86 (3H, s, CH3O), 3.62 (3H, b, CH3OCO), 1.45 (9H, s, Boc), 1.3 (3H, d, J 7, CH3CH). m/z (CI), 400 (M+H+, 20%), 344 (55%),
300 (75%), 135 (100%). Found M+ 399.2044; C23H29NO5 requires M 399.20456. wmax/cm−1 (film) 1748, 1691.
e 3:1 mixture of 13a:13b.
O
O
PhN
N
PhN
N
OMe
OMe
Ph
Ph
Ha Hb
Ph
Ph
Ha Hb
13a
Jab = 4.7 Hz
13b
Jab = 8.7 Hz
that the reactions in Table 1 (entries 5 and 6) proceed
with the same stereochemical sense. We therefore assign
stereochemistry to 11a (E=CO2H and E=CO2Me) on
the assumption that 7 follows 9 in its reaction with
ClCO2Me. However, the stereoselectivity of deuteration
(entry 1), benzylation (entry 3), allylation (entry 4), and
reactions with ketones and imines (entries 7–9) have not
been confirmed and stereochemistry of the products is
assigned arbitrarily. Two imidazolinones 13a and 13b
were produced on reaction of 10 with N-phenylbenz-
aldimine by cyclisation of the newly-formed amide
anion onto the Boc group. Stereochemistry is assigned
on the basis of their Jab coupling constants, the signifi-
cant difference in the values of these coupling constants
suggesting that the compounds differ in stereochemistry
only at the CꢀN-derived stereogenic centre. The
stereoselectivity of stannylation (entry 10) is opposite to
that of deuteration, as explained below.
Scheme 2. Diastereoselective lithiation and electrophilic
quench of N-benzyl-N-a-methylbenzyl derivatives.
carbamate 8. However, methylation and carboxylation
were highly stereoselective with both starting materials,
and treatment of 12a (E=Me) with ceric ammonium
nitrate in aqueous acetonitrile gave the N-Boc-a-
methylbenzylamine 14 (E=Me), whose optical
rotation8 confirmed the overall sense of stereoselectivity
in the alkylation (Scheme 3 and Table 2).
Deprotection of 12a (E=CO2Me) under the same con-
ditions provided a new synthesis of the (R)-N-Boc-
phenylglycine methyl ester 14 (E=CO2Me)9 and the
optical rotation10 confirmed the sense of stereoselectiv-
ity of the acylation with methyl chloroformate (Table
2). Mixtures of diastereoisomers of 12 (E=allyl) and 12
(E=benzyl) were deprotected in the same way to give
racemic samples of 14 (E=allyl, E=benzyl).
O
E
O
CAN, MeCN, H2O
OMe
t-BuO
N
t-BuO
NH
Ph
Ph
E
12
14
Methylation of the product of carboxylation of 7 with
CO2 gave material identical with that produced by
methoxycarbonylation of 7 with ClCO2Me, proving
Scheme 3. Oxidative deprotection of N-a-methyl-p-methoxy-
benzyl derivatives.

R. A. Bragg et al. / Tetrahedron Letters 43 (2002) 1955–1959
1957
Table 2.
As far as is known, deprotonation to form an organo-
lithium and reprotonation of the organolithium occur
with reliable retentive stereospecificity,13 so the lack of
stereospecificity in the deprotonation–reprotonation of
12a must be due to a lack of configurational stability in
10. Nonetheless, the lack of configurational stability in
10 cannot be complete, because methylation of 10
produced by lithiation of 8 gives a different stereochem-
ical outcome from methylation of d-10 produced by
lithiation of 12a (E=D). Organolithium 9, by contrast,
appears to have no configurational stability on the
reaction timescale, since methylation or protonation
(deuteration) of 9 formed by lithiation of either 7 or
11a (E=D) gives the same outcome (Scheme 4).14
Entry
E
Yield 11 (%)a
[h]D
−52b
[h]lDit.
1
2
3
4
Me
Bn
Allyl
CO2Me
80
79
47
41
−628
–
0
0
–
−112
−132.310
a From crude mixture of 12a+12b.
b 92% ee by HPLC on chiral stationary phase.
The stereoselectivity of the lithiation–alkylation
sequence, or lack of it, must originate from factors
which may include the diastereoselectivity of the depro-
tonation of 7 and 8, the rate at which the two
diastereoisomers of the organolithiums 9 and 10 inter-
convert (their configurational stability), the relative sta-
bility of the two diastereoisomers of 9 and 10, and the
rate and stereospecificity with which each reacts with
the electrophile. We carried out a series of experiments
aimed at disentangling these various factors.11
These results are consistent with a common mechanism
in which the stereoselectivity of the lithiation-quench of
7 and 8 is determined by kinetic and thermodynamic
factors combining to yield a single diastereoisomer of
organolithiums 9 and 10. For 9, equilibration to a
single diastereoisomer is fast, disguising the initial
stereoselectivity of the lithiation, but slower equilibra-
tion of 10 suggests that the initial kinetic stereoselectiv-
ity of the deprotonation is high. The single
diastereoisomer of 9 or 10 must then react with varying
stereospecificity to yield the observed product ratios.
To establish whether the intermediate organolithiums
are configurationally stable, the deuterated compounds
11a and 12a (E=D) were re-deprotonated with t-BuLi
at −78°C and quenched with MeI and, for 12a (E=D),
with NH4Cl. Mass spectrometry showed that >98% of
the product in each case remained deuterated, showing
On warming, 9 undergoes a cyclisation to give 15 in a
3:1 ratio of diastereoisomers.15 The cyclisation is effec-
tively an internal electrophilic quench, but it takes place
only at temperatures above −40°C, suggesting that at
higher temperatures the thermodynamic selectivity in
favour of 9a is decreased.
1
that only the H atom had been removed from 11a and
12a (E=D). This result is in accord with the high
values typical of primary kinetic isotope effects in low
temperature lithiations.12 However, while methylation
of lithiated 11a (E=D) returned d-11a (E=Me) as a
single diastereoisomer, methylation or protonation of
lithiated 12a (E=D) returned d-12a and d-12b (E=Me)
or 12a and 12b (E=D) as a 50:50 mixture (Scheme 4).
We tried to observe the stereoselectivity of the deproto-
1
nation of 10 directly by H NMR spectroscopy, carry-
O
O
BuLi
11a + 11b (E = SnBu3) 11:89
12a + 12b (E = SnBu3) 36:64
minor stannane 12a
regenerates major
organolithium 10a
R
N
Ar
R
N
Ar
Ph
+
Bu3Sn
Ph
Bu3Sn
minor
major
9a or 10a
O
d-9a or d-10a
O
O
R
Bu3SnCl
MeOD
R
R
N
Ar
Ph
N
Ph
Ph
R
N
Ph
Ph
O
Li
D
D
Li
warm to
20 ºC
MeI or H2O
t-BuLi
11a or 12a
(E = D)
fast for 9
slow for 10
fast for d-9
slow for d-10
R
N
Ar
Ph
O
O
7 or 8
t-BuLi
N
Ar
R
N
Ph
Ph
Li
Ph
Li
D
9b or 10b
d-9b or d-10b
O
O
E
+
O
O
H
H
R
N
Ph
Ph
R
N
Ph
Ph
Me
Me
Ph
+
N
N
D
E
D
Ph
MeO
MeO
H
H
d-11a (E = Me) 97:3 d-11b (E = Me)
Ph
Ph
50:50
50:50
d-12a (E = Me)
d-12a (E = H)
d-12b (E = Me)
d-12b (E = H)
15a
75:25 ratio
15b
Scheme 4. Investigating the origin of the stereoselectivity.

1958
R. A. Bragg et al. / Tetrahedron Letters 43 (2002) 1955–1959
ing out the lithiation in d8-THF in an NMR tube. At
−78°C, the starting material 8 was clearly a mixture of
N–CO rotamers, but lithiation gave a 85:15 mixture of
two species, presumably the two diastereoisomers of 10
formed under kinetic control. Under the same conditions,
d-12a gave the same two species, but with a considerably
higherproportion(30–40%)oftheminordiastereoisomer.
The latter result, and therefore possibly also the former,
must represent an already partly epimerised mixture of
organolithiums 10.
CropScience SA (Lyon) for support, and to Dr. O.
Ichihara and Dr. D. J. Mansfield for helpful discussions.
References
1. O’Brien, P. J. Chem. Soc., Perkin Trans. 1 1998, 1439.
2. (a) Marshall, J. A.; Lebreton, J. J. Am. Chem. Soc. 1988,
110, 2925; (b) Overberger, C. G.; Marullo, N. P.; Hiskey,
R. G. J. Am. Chem. Soc. 1961, 83, 1374.
3. The meso amine 2 has principally been used for NMR
studies of amide conformation and stereodynamics. See
(a) Langga˚rd, M.; Sandstro¨m, J. J. Chem. Soc., Perkin
Trans. 2 1996, 435; (b) Bragg, R. A.; Clayden, J. Org.
Lett. 2000, 2, 3351; (c) Bragg, R. A.; Clayden, J.; Morris,
G. A.; Pink, J. H. Chem. Eur. J., in press.
4. (a) Bragg, R. A.; Clayden, J. Tetrahedron Lett. 1999, 40,
8323; (b) Bragg, R. A.; Clayden, J. Tetrahedron Lett.
1999, 40, 8327.
5. (a) Ahmed, A.; Clayden, J.; Rowley, M. Chem. Commun.
1998, 297; (b) Ahmed, A.; Clayden, J.; Rowley, M.
Tetrahedron Lett. 1998, 39, 6103; (c) Ahmed, A.;
Clayden, J.; Yasin, S. A. Chem. Commun. 1999, 231; (d)
Ahmed, A.; Clayden, J.; Rowley, M. Synlett 1999, 1954;
(e) Clayden, J.; Tchabanenko, K. Chem. Commun. 2000,
317; (f) Clayden, J.; Menet, C. J.; Mansfield, D. J. Org.
Lett. 2000, 2, 4229; (g) Clayden, J.; Tchabanenko, K.;
Yasin, S. A.; Turnbull, M. D. Synlett 2001, 302; (h)
Ahmed, A.; Bragg, R. A.; Clayden, J.; Tchabanenko, K.
Tetrahedron Lett. 2001, 42, 3407.
6. At higher temperatures, or in the presence of HMPA or
DMPU, the benzyllithium intermediates are unstable and
undergo cyclisation reactions: see Ref. 5. Similar alkyla-
tions: (a) Snieckus, V.; Rogers-Evans, M.; Beak, P.; Lee,
W. K.; Yum, E. K.; Freskos, J. Tetrahedron Lett. 1994,
35, 4067; (b) Fraser, R. R.; Boussard, G.; Potescu, I. D.;
Whiting, J. J.; Wigfield, Y. Y. Can. J. Chem. 1973, 51,
1109.
Similar conclusions about configurational stability can be
drawn from reactions of the stannylated compounds 11
(E=SnBu3) and 12 (E=SnBu3). The 89:11 mixture of 11b
and 11a (E=SnBu3) was transmetallated with BuLi and
the product organolithium was methylated. 11a (E=Me)
was obtained with 97:3 stereoselectivity: the stereochem-
istry of the starting stannanes has no bearing on the
stereochemistry of the products, which arises by rapid
equilibration to a single organolithium. However,
transmetallation of the 64:36 mixture of 12b and 12a
behaved differently. Transmetallation was incomplete
even after 90 min at −78°C, and the minor diastereoiso-
mer, 12a (E=SnBu3), transmetallated faster than the
major diastereoisomer, leaving a 73:27 mixture of 12b and
12a (E=SnBu3) after this time. Methylation of the
organolithium formed by transmetallation gave, in 15%
yield, an 88:12 mixture of 12a and 12b (E=Me). Tin–
lithium exchange of a-heterosubstituted stannanes occurs
with reliable retention of stereochemistry,16 so formation
of the same diastereoisomer of 12 (E=Me) by lithiation
of 9 and by transmetallation of the minor diastereoisomer
of the stannane (E=SnBu3) suggests that both organo-
lithiums are the same and therefore that their stannylation
proceeds principally with inversion of stereochemistry.17
The widely varying stereospecificity of electrophilic sub-
stitution reactions of a-heterosubstituted benzyl-
lithiums18 makes us wary of proposing definitive retentive
or invertive pathways for other electrophiles. We know
for certain the stereochemistry of the final products 12a
(E=Me) and 12a (E=CO2Me), and MeI and ClCO2Me
must therefore react with the same stereospecificity
(whether retention or inversion) as each other. For
simplicity we have inferred that both substitute reten-
tively, but alkylation is commonly invertive,17 and this
assignment must be taken purely tentatively. Mixtures of
diastereoisomers formed by allylation and benzylation
may arise from the intervention of radical intermedi-
ates.19,20
7. We have found (Ref. 4) that stereospecificity in the
cyclisation of
6 and related compounds is due to
diastereoselective ortholithiation in which the stereogenic
Ar–CO axis plays a role. Although axial chirality may
play a role in the stereoselective methylation of 4, benz-
amide 3 lacks a potentially stereogenic axis and its
stereoselective alkylation must be governed by other
factors.
8. Albanese, D.; Gibson (ne´e Thomas), S. E.; Rahimian, E.
Chem. Commun. 1998, 2571.
9. For an example of the use of an N-bonded auxiliary to
direct amide lithiation and alkylation, see: Le´autey, M.;
Castelot-Deliencourt, G.; Jubault, P.; Pannecoucke, X.;
Quirion, J.-C. J. Org. Chem. 2001, 65, 5566.
10. Merino, P.; Castillo, E.; Franco, S.; Merchan, F. L.;
Tejero, T. J. Org. Chem. 1998, 63, 2371.
11. See Beak, P.; Basu, A.; Gallagher, D. J.; Park, Y. S.;
Thayumanavan, S. Acc. Chem. Res. 1996, 29, 552.
12. (a) Hoppe, D.; Paetow, M.; Hintze, F. Angew. Chem.,
Int. Ed. 1993, 32, 394; (b) Anderson, D. R.; Faibish, N.
C.; Beak, P. J. Am. Chem. Soc. 1999, 121, 7553; (c)
Clayden, J.; Pink, J. H.; Westlund, N.; Wilson, F. X.
Tetrahedron Lett. 1998, 39, 8377.
In summary, the lithiation and alkylation of a-methyl-
N,N-dibenzylamine derivatives provides the best route to
derivatives of meso bis-(a-methylbenzyl)amine 2 and an
alternative route to derivatives of phenylglycine. N-Acyl-
a-methyl-N,N-dibenzylamines also provide a convenient
platform from which to investigate the stereochemistry
in the reactions of amino-substituted benzyllithiums.
Acknowledgements
We are grateful to the EPSRC for studentships (to R.A.B.
and to C.J.M.), to Oxford Asymmetry and to Aventis

R. A. Bragg et al. / Tetrahedron Letters 43 (2002) 1955–1959
1959
13. Earlier reports (Carstens, A.; Hoppe, D. Tetrahedron
1994, 50, 6097) that deuteration using deuterated car-
boxylic acids or their ammonium salts proceed with
inversion have been corrected (Hammerschmidt, F.; Han-
ninger, A. Chem. Ber. 1995, 128, 1069; Derwing, C.;
Hoppe, D. Synthesis 1996, 149) but there is still some
uncertainty in this area, see: Norsikian, S.; Marek, I.;
Klein, S.; Poisson, J. F.; Normant, J. F. Chem. Eur. J.
1999, 5, 2055.
14. Configurational stability of related benzylic organolith-
iums a to nitrogen varies with conditions or additives.
See: Faibish, N. C.; Park, Y. S.; Lee, S.; Beak, P. J. Am.
Chem. Soc. 1997, 119, 11561 and Ref. 15.
15. We used a phenylglycinol-derived auxiliary to control the
stereoselectivity of a similar cyclisation: Bragg, R. A.;
Clayden, J.; Bladon, M.; Ichihara, O. Tetrahedron Lett.
2001, 42, 3411. We recently found that chiral lithium
amides initiate asymmetric dearomatising cyclisations by
enantioselective benzylic deprotonation: Clayden, J.;
Menet, C. J.; Mansfield, D. J. Chem. Commun. 2002, 38,
in press.
16. The only reported example of a non-stereospecific tin–
lithium exchange involved non-a-heterosubstituted
a
organolithium. For a discussion, see: Clayden, J.; Helli-
well, M.; Pink, J. H.; Westlund, N. J. Am. Chem. Soc.
2001, 123, 12449.
17. See Refs. 13, 14 and 16: (a) Hoppe, D.; Carstens, A.;
Kra¨mer, T. Angew. Chem., Int. Ed. Engl. 1990, 29, 1424;
(b) Derwing, C.; Frank, H.; Hoppe, D. Eur. J. Org.
Chem. 1999, 3519; (c) Hammerschmidt, F.; Hanninger,
A.; Simov, B. P.; Vo¨llenkle, H.; Werner, A. Eur. J. Org.
Chem. 1999, 3511; (d) Hammerschmidt, F.; Hanninger,
A.; Vo¨llenkle, H. Chem. Eur. J. 1997, 3, 1728; (e) Basu,
A.; Beak, P. J. Am. Chem. Soc. 1996, 118, 1575; (f)
Thayumanavan, S.; Lee, S.; Liu, C.; Beak, P. J. Am.
Chem. Soc. 1994, 116, 9755; (g) Clayden, J.; Pink, J. H.
Tetrahedron Lett. 1997, 38, 2565; (h) Clayden, J.; Pink, J.
H. Tetrahedron Lett. 1997, 39, 2561.
18. Gawley, R. E. Tetrahedron Lett. 1999, 40, 4297.
19. Gawley, R. E.; Low, E.; Zhang, Q.; Harrris, R. J. Am.
Chem. Soc. 2000, 122, 3344.
20. Gawley, R. E.; Zhang, Q. J. Org. Chem. 1995, 60, 5763.