The diastereoselective and enantioselective substitution reactions of an isoindoline-borane complex

Article abstract of DOI:10.1039/a904788c

The alkylation of N-methylisoindoline-borane complex, using ⁿBuLi in THF is diastercoselective, the substitution occurring predominantly syn to the borane group. Use of the ⁿBuLi-sparteine reagent mixture in Et₂O changes t

Full text of DOI:10.1039/a904788c

View Article Online / Journal Homepage / Table of Contents for this issue
The diastereoselective and enantioselective substitution reactions of
an isoindoline–borane complex
Azhar Ariffin,a Alexander J. Blake,a Mark R. Ebden,a Wan-Sheung Li,a Nigel S. Simpkins*a
and David N. A. Foxb
a School of Chemistry, University of Nottingham, University Park, Nottingham, UK NG7 2RD
b Department of Discovery Chemistry, Pfizer Central Research, Sandwich, Kent, UK CT13 9NJ
Received (in Cambridge) 15th June 1999, Accepted 13th July 1999
The alkylation of N-methylisoindoline–borane complex, using nBuLi in THF is diastereoselective, the substitution
occurring predominantly syn to the borane group. Use of the sBuLi–sparteine reagent mixture in Et2O changes
the diastereoselectivity observed and enables the reaction to be conducted enantioselectively, giving the chiral
isoindoline–borane complexes in up to 89% ee. The relative and absolute configurations of the chiral products
were established by X-ray structure determinations and NMR studies. The new asymmetric process is shown to
be an enantioselective deprotonation reaction, and the intermediate organolithium is shown to be epimerisable.
Table 1 Diastereoselective alkylation of 3
Over the past few years several research groups have described
the use of Lewis acid activators, such as BF3 or BH3, to facili-
tate the α-metallation of a range of tertiary amines, including
benzylic, allylic and even saturated types, by either lithium
tetramethylpiperidide (LTMP) or sBuLi.1 We recently described
our own results in this area, in which the use of borane acti-
vation allowed interesting regiocontrol in the metallation and
electrophilic substitution of certain tertiary amines.2 Following
this first phase of our work, it became clear that the use of such
borane activation could also have implications for controlling
the stereochemistry of amine substitution reactions.
Firstly, a paper by Vedejs and Kendall showed that the
borane activation technique is applicable to aziridines, and that
the alkylation chemistry of such systems is highly diastereo-
controlled, both the metallation and electrophilic quench
occurring syn to the BH3 group, e.g. the conversion of amine–
borane complex 1 into substituted derivative 2, Scheme 1.3
Electrophile
Product
3→4–5 (%)
Dr 4:5
4–5→6 (%)
Bu3SnCl
Me3SiCl
AllylBr
BnBr
Me2CO
Ph2CO
ArCH2Bra
EtI
4–5a
4–5b
4–5c
4–5d
4–5e
4–5f
4–5g
4–5h
4–5i
95
77
84
99
—
—
66
93
75
5:1
4:1
6:1
3:1
—
—
—
87
83
62
78
94
71
63
—
Scheme 1
2:1
6:1
5:1
MeI
Our own preliminary report in this area described the stereo-
controlled alkylation of the amine–borane complex 3 derived
from N-methylisoindoline, for example the simple alkylation
reactions which furnished diastereomeric mixtures of products
in which 4 was the major component, Scheme 2.4
a Ar = 2-naphthyl.
(i) Diastereoselective substitution of 3 using BuLi in THF
Treatment of the readily accessible N-methylisoindoline–
borane complex 3 with nBuLi in THF, followed by addition of
an electrophile, gave the anticipated substituted complexes as
mixtures of diastereomers, Table 1. The results require some
detailed comment, but the overall observation is that alkylation
or addition to carbonyl compounds is possible in high yield and
that the reactions occur to give mainly diastereomer 4, in which
the new substituent is incorporated syn to the BH3 group on
nitrogen. Many of the product yields shown are somewhat
improved compared to the initial report, but it should be noted
that the very high diastereomeric ratio quoted in the original
communication for quenching with Bu3SnCl proved not to be
reproducible under these standard conditions. In the case of
silylated or stannylated intermediates 4–5a and 4–5b decom-
plexation was not possible without decomposition, whereas
in the case of carbonyl addition products 4–5f and 4–5g the
Scheme 2
An enantioselective variant of this reaction was also
described, which allowed the isolation of product complexes
in good levels of enantiomeric excess.4 Herein we describe
further details of this research, which has sought to probe the
scope of the reaction and the origins of the diastereo- and
enantiocontrol, which were not clear from our preliminary
findings.
J. Chem. Soc., Perkin Trans. 1, 1999, 2439–2447
2439

View Article Online
Table 2 Enantioselective alkylation of 3
Electrophile
Product
3→4–5 (%)
Dr 4:5
Ee (%)c
Bu3SnCl
Me3SiCl
AllylBr
BnBr
Ph2CO
ArCH2Bra
EtI
4–5a
4–5b
4–5c
4–5d
4–5f
4–5g
4–5h
4–5i
4–5j
95
84
65
1:20
1:14
1:3
1:5
—
1:5
10:1
1:2
89
85
87
83
—
—
75
83
56
53
(62)b
58
d
d
71
70
65
MeI
Fig. 1 X-Ray crystal structure of borane complex 4i.
nC5H11Br
20:1
a Ar = 2-naphthyl. b Yield of free amine. c Ee of major diastereomer in
each case. d Not determined.
In general the yields of products are very good and the levels
of enantiomeric excess are similar for most of the electrophiles,
being of the order of 85%. In the cases of iodoethane and
bromopentane as electrophiles, warming of the reaction mix-
ture was required in order to achieve satisfactory reaction,
which may explain the apparent erosion of enantiomeric excess.
However, perhaps the most striking aspect of the results is the
dramatic change in diastereoselectivity for most of the electro-
philes used, compared to the outcome of the corresponding
Fig. 2 Diagnostic NOE enhancements in borane complexes 4a and 5a.
borane complexes were unstable and only the free amine was
isolated.
The relative stereochemistry of the major product in each of
the reactions described in Table 1 was initially based entirely on
the X-ray structure determination carried out for the major
product, 4i, obtained from the methylation reaction, Fig. 1. The
X-ray structure of 4i reveals that the N2–B bond [1.629(2) Å] is
considerably longer than the N2–C2Ј bond [1.485(2) Å], allow-
ing the two nitrogen substituents to be clearly distinguished.
Moreover, the BH3 group may be slightly larger than the
N–CH3 group, judging from the B–H bond length [1.13(2) Å]
compared to the C–H (0.96 Å, idealised) bond length. It is
difficult to see how such minor differences could lead to the
observed levels of diastereocontrol, based on steric arguments,
and later results point to alternative controlling factors (vide
infra).
Additional NMR experiments were carried out on other
products in order to lend credence to our previous assumption
that all of the major products 4 belonged to the same diastereo-
meric series. In particular, detailed NOE experiments were
possible on either the stannylated (4–5a) or silylated (4–5b)
compounds, since in these cases the diastereomers could be
separated, Fig. 2.
As can be seen from Fig. 2 the results for 4a and 5a support
our assignments, the NOE enhancement seen for the methine
CH on irradiating the NMe being ca. eight fold greater for one
isomer than the other. Analogous results were seen for 4b and
5b, and similar experiments were performed on the benzylated
and ethylated series—although these were done on the (in these
cases inseparable) mixtures of isomers.
At this stage our results seem to mirror quite closely the
findings of Vedejs and Kendall in terms of the preferred mode
of reaction being syn to the BH3 group.3 However, the vari-
ations in diastereomer ratios seen on changing the nature of the
electrophilic quench seemed to point to a stereochemically
labile intermediate organometallic, a contrasting situation
compared to that found with the aziridines.5
n
reactions using BuLi in THF. This aspect of the results was
not apparent when we submitted our initial communication
because only a few such asymmetric reactions had been
attempted. Now, with a broader range of results available com-
pared to our earlier report, and further NMR and X-ray data
(vide infra) which secure stereochemical assignments, our earlier
interpretation of the results requires slight reassessment.7
In the cases of series 4–5c, d, f and g were carried out decom-
plexation of the mixtures of diastereomeric borane complexes
to give the corresponding non-racemic amines 6 in good yields
(although we did not assay for the ee of these products). The
likely stereochemistry of these compounds was deduced follow-
ing further experiments described in the next section.
(iii) Origins of the diastereo- and enantioselectivity
The swap in diastereoselectivity from 4 to 5 as the major prod-
n
uct (for most electrophiles used) when changing from BuLi–
s
THF to BuLi–sparteine–ether could clearly be a result of the
change in alkyllithium or the solvent, or could be due to the
chiral diamine ligand. Several explanations could also be pro-
posed for the observed enantioselectivity, depending upon
whether the intermediate organolithium was stereochemically
labile, and whether the initial deprotonation event was the
principal determinant of the asymmetric induction. Further
experiments were designed in order to probe the origins of both
the diastereoselectivity and enantioselectivity of the substitu-
tion reactions.
In order to probe solvent effects we carried out the metal-
lation of 3 with sBuLi in either Et2O or THF, Scheme 3. As can
be seen, the stereochemical results for the two solvents are
highly complementary, with Et2O resulting in substitution anti
to the borane group to give 5, whilst the reaction in THF
furnishes 4, the product of substitution syn to boron.
This result indicated the powerful effect of the solvent on the
stereochemical outcome, although at this point it was not pos-
sible to reach a conclusion about which of the metallation or
electrophilic quenching steps was responsible for the results.
Certainly, the fact that rather variable ratios of 4:5 were
obtained under either set of conditions (Tables 1 and 2) pointed
to a stereochemically labile organolithium intermediate. How-
ever, an alternative explanation might be that the solvent
(ii) Enantioselective substitution of 3 using sBuLi–sparteine in
diethyl ether
An asymmetric variant of the substitution reactions of complex
3, presented in Table 1 was achieved simply by employing
sBuLi–(Ϫ)-sparteine as the base, which resulted in the form-
ation of the desired products in diastereo- and enantiocon-
trolled fashion, Table 2.6
2440
J. Chem. Soc., Perkin Trans. 1, 1999, 2439–2447

View Article Online
Table 3 Transmetallation–substitution of 5a
stereochemical “anchor”. This last feature suggested to us that
the observed asymmetric induction in the sBuLi–sparteine reac-
tions is due to an initial asymmetric deprotonation step, which
fixes the configuration at nitrogen (although not the stereo-
chemistry at carbon). One further reaction was carried out in
order to fully discount an alternative mechanism, which would
involve an asymmetric substitution mechanism, in which the
dominant feature responsible for the induction would be
association of the (racemic) organometallic with sparteine,
Scheme 4.
Electrophile
5a→4–5 (%)
Dr 4:5
Ee (%)
MeI
EtI
AllylBr
Me3SiCl3
83
80
70
81
6:1
5:1
9:1
78
80
79
71
10:1
Scheme 4
The borane complex 3 was metallated in Et2O in the usual
way, and then sparteine was added to the mixture at Ϫ78 ЊC,
followed by Me3SiCl. The expected product 5b was isolated as
a single diastereomer, but was found to be almost racemic
(ee ≤ 5%), this result being in accord with our proposed asym-
metric deprotonation mechanism.
In order to further clarify the mode of asymmetric induction
we re-examined one of the results in Table 2 in much more
detail. Firstly, for the product of electrophilic quenching with
Me3SiCl (4–5b), it was found that the enantiomers of both
diastereomers could be separated by HPLC. In fact, all four
possible stereoisomers could be analysed simultaneously,
making determination of dr and ee values very straightforward.
Furthermore, since both the enantiomerically enriched samples
were crystalline, we were able to determine the absolute and
relative stereochemistry for the major constituents of 4b
and 5b by recrystallisation and determination of the structures
by X-ray crystallography. The obtained structures are shown
in Figs. 3 and 4, and the overall picture for the asymmetric
deprotonation of 3 which emerges from the data is shown in
Scheme 5.
The figures shown for each component lead to the observed
(measured) dr and ee values, and the absolute configurations
can be assigned with certainty from the X-ray data. Therefore it
is possible to conclude that the principle origin of the asym-
metric induction is a deprotonation in which the chiral base
selects for the arrowed position to the extent of 92.6:7.4—i.e.
the initial ratio of A:B is about 9:1. This approximate ratio can
also be extracted from the HPLC data for other electrophiles,
although in these cases X-ray structural data on the intermedi-
ate borane complexes is lacking. It is also clear from the above
data that decomplexation of the two diastereomeric, enantio-
merically enriched, borane complexes should give amines of
opposite absolute configuration. This conclusion differs from
our initial assignment, which was made tentatively, based on
preliminary polarimetric results.9
Scheme 3
altered the diastereoselectivity in the initial metallation, giving
configurationally stable intermediates, which then reacted with
the different electrophiles with a variable degree of retention/
inversion. In order to further probe these possibilities we
carried out a range of transmetallation reactions employing
diastereomerically pure stannylated derivative 5a of 89% ee,
Table 3.
The reactions are seen to occur with overall inversion, i.e. the
tin substrate is pure 5, whilst the products are mainly diastereo-
mer 4, this result conforming to our expectation that the solvent
effect from THF would exert a powerful influence. This result,
taken in combination with the solvent effects in Scheme 3 pro-
vide compelling evidence that the intermediate organolithium is
stereochemically labile. Since in THF the reaction in Scheme 3
is highly selective for 4 then, if this was the result of a configur-
ationally stable intermediate, we would expect the transmetal-
lation sequence from 5a to give mainly product diastereomer 5.
This is clearly not the case and so the intermediate must be
stereochemically labile.8
Table 3 also shows that the transmetallation–substitution
process starting with material of 89% ee yields products which
are still highly enantiomerically enriched, although some
erosion of ee is evident depending upon the electrophile
employed. This shows that the intermediate organolithium is
reasonably configurationally stable with respect to racemisation,
despite the apparent ease of epimerisation at the lithiated
centre. This is because both of the possible diastereomeric
organolithiums resulting from 5a (or 3) have a fixed configur-
ation at the chiral nitrogen centre, which acts as a kind of
Although all of the data in Table 2 can be accommodated by
the above explanation, the switch in diastereoselectivity seen
n
for the electrophiles EtI and C5H11Br, compared to the other
examples is perplexing. In these two cases only, the alkylation
J. Chem. Soc., Perkin Trans. 1, 1999, 2439–2447
2441

View Article Online
Fig. 5 X-Ray crystal structure of ammonium salt 7.
Fig. 3 X-Ray crystal structure of borane complex 4b.
Fig. 4 X-Ray crystal structure of borane complex 5b.
reactions were sluggish enough to require slow warming, and
this may account for the different sense of diastereoselectivity
seen. In the case of reaction with EtI a more detailed analysis
can tentatively be made since we were able to establish the con-
figuration at the newly formed stereocentre in the major isomer
by obtaining an X-ray structure for the ammonium iodide 7,
Fig. 5.
This compound was obtained by treating the mixture of
optically active borane complexes 4–5h with 6 M HCl, followed
by MeI, and then recrystallising from hexane–CH2Cl2. If we
assume that the resulting structure originates from the major
enantiomer present in the intermediate non-racemic amine,
then the result can be seen to be in accord with HPLC data for
this reaction, Scheme 6.
Scheme 5
remains difficult to do more than speculate as to the detailed
origins of the diastereoselectivity and enantioselectivity. In the
former regard, the solvent plays a key role, but the striking
differences in diastereoselectivity seen on changing from diethyl
ether to THF could well be due to changes in aggregation state
of the intermediate organolithium. Likewise, it is impossible to
rationalise in detail the sense of asymmetric induction seen with
sparteine, simply because the usual steric arguments cannot be
applied with much confidence. In their study,3 Vedejs and
From the above analysis we can predict the stereochemistry
of the free amines 6 resulting from the asymmetric metallation
n
process. In the case of EtI and C5H11Br they should have the
(1S)-configuration analogous to the ammonium salt 7, whereas
for the other electrophiles the opposite (1R)-configuration
should predominate.
Even though the above study has uncovered the key aspects
of this novel type of asymmetric deprotonation reaction, it
2442
J. Chem. Soc., Perkin Trans. 1, 1999, 2439–2447

View Article Online
Conclusion
A new type of asymmetric deprotonation reaction has been
s
investigated, using the BuLi–sparteine reagent, which gives
interesting amine–borane complexes in synthetically useful
levels of enantiomeric excess. The mechanism of the
reaction has been demonstrated to involve an enantioselective
deprotonation to give a non-racemic organolithium which is
configurationally stable at nitrogen but stereochemically labile
with respect to the C–Li bond. Dramatic solvent effects on the
diastereoselectivity of electrophilic quenching are observed,
which are not easily rationalised.
The types of products prepared by this chemistry have previ-
ously been prepared in racemic and non-racemic form using
Meyers’ formamidine methodology.11 The initial adducts could
be further alkylated to give trans-1,3-dialkylated isoindolines,
which have potential applications as C2-symmetric chiral aux-
iliaries.12 Although we have not carried out further work on the
chiral borane complexes 4 and 5, the chemistry described herein
could provide a complementary approach to such auxiliaries.
Scheme 6
Experimental
General details
The general procedures and the experimental details for
preparation of complexes 3, 4i, 5i and 6i were as described
previously.2
Scheme 7
Kendall identified a distinct tendency for metallation syn to the
BH3 subunit, for example the transformation of 8 into 9,
Scheme 7. This result was attributed to a combination of steric
and electrostatic effects. Since the N–B bond is longer than the
corresponding exocyclic N–C bond it was considered that the
BH3 unit might be more tolerant of 1,2-eclipsing interactions
with an approaching alkyllithium than the N-methylene unit. In
addition there could be an electrostatic interaction between the
lithium ion of the base and the BH3 unit. Such interactions have
been observed in the solid state, and are illustrated by the X-ray
structure of TMEDA–LiBH4, which can be represented as
shown for 10.10
In the asymmetric deprotonation of our isoindoline system
such an attractive effect, involving the borane group and the
approaching lithium base, might be crucial to the asymmetric
induction by reducing the number of metallation pathways. For
example, it is possible that the key asymmetric step involves
chiral base selection between only one set of enantiotopic
hydrogens, which are oriented syn to the borane group, as high-
lighted in 11. It is even possible that such interactions might
play a part in determining the diastereoselectivity of the sub-
sequent electrophilic quench by offering some stabilisation to
the intermediate organometallic in which the lithium is syn with
respect to the borane, as suggested by structure 12.
Diastereoselective reactions in Table 1
n
Typical procedure: synthesis of 4a–5a using BuLi in THF. A
solution of nBuLi (0.84 ml of a 1.60 M solution in hexanes, 1.35
mmol) was added dropwise to a solution of complex 3 (100 mg,
0.68 mmol) in THF (10 ml) at Ϫ78 ЊC under an atmosphere of
nitrogen. The reaction mixture was then warmed to 0 ЊC and
stirred for 1 h at this temperature, before recooling to Ϫ78 ЊC.
Tributyltin chloride (0.55 ml, 2.0 mmol) was added in one por-
tion and the mixture stirred at Ϫ78 ЊC for 1 h before saturated
aqueous NaHCO3 (1 ml) was added with subsequent warming
to room temperature. After addition of EtOAc (10 ml), the
organic layer was separated, washed with saturated NaHCO3
(5 ml), brine (2 × 5 ml), dried (MgSO4) and the solvent was
removed under reduced pressure. The resulting light brown oil
was purified by flash column chromatography on silica gel (2%
EtOAc–light petroleum) to give firstly 4a as a colourless oil (230
mg, 78%), νmax (CHCl3)/cmϪ1 3155, 2956, 2923, 2871, 2363 (B–
H), 2320 (B–H) and 2253 (B–H); δH (400 MHz, CDCl3) 0.86
[9H, t, J 7, Sn(CH2)3CH3], 0.97 [6H, m, Sn(CH2)3CH3], 1.26
[6H, m, Sn(CH2)3CH3], 1.44 [6H, m, Sn(CH2)3CH3], 1.50–2.50
(3H, v br, BH3), 2.76 (3H, s, NCH3), 3.95 (1H, d, JAB 15,
NCHH), 4.26 (1H, s, NCHSn), 4.40 (1H, d, JAB 15, NCHH),
7.00 (1H, d, J 7, ArH), 7.11 (1H, dd, J 7 and 7, ArH) and 7.19
J. Chem. Soc., Perkin Trans. 1, 1999, 2439–2447
2443

View Article Online
(2H, m, ArH); δC (100 MHz, CDCl3) 12.1 (CH2), 13.7 (CH3),
27.5 (CH2), 29.0 (CH2), 53.3 (CH3), 68.1 (CH2), 73.3 (CH),
120.4 (ArCH), 122.9 (ArCH), 125.6 (ArCH), 127.8 (ArCH),
134.2 (ArC) and 143.5 (ArC); m/z (FAB) 436 [(M Ϫ H)ϩ, 24%]
(HRMS: found [M Ϫ H]ϩ, 436.2157. Requires [M Ϫ H],
436.2197) followed by 5a as a low melting point white solid (50
mg, 17%, 95% overall) (Found: C, 57.80; H, 9.31; N, 3.27.
C21H40BNSn requires C, 57.84; H, 9.25; N, 3.21%); νmax
(CHCl3)/cmϪ1 2957, 2924, 2854, 2416 (B–H), 2373 (B–H), 2314
(B–H) and 1460; δH (400 MHz, CDCl3) 0.90 [9H, t, J 7,
Sn(CH2)3CH3], 1.08 [6H, m, Sn(CH2)3CH3], 1.34 [6H, m,
Sn(CH2)3CH3], 1.49 [6H, m, Sn(CH2)3CH3], 1.50–2.50 (3H, v
br, BH3), 2.78 (3H, s, NCH3), 4.02 (1H, d, JAB 15, NCHH), 4.54
(1H, d, JAB 15, NCHH), 4.76 (1H, s, NCHSn), 7.01 (1H, d, J 7,
ArH) and 7.19 (3H, m, ArH); δC (100 MHz, CDCl3) 11.1
(CH2), 13.6 (CH3), 27.3 (CH2), 28.9 (CH2), 53.1 (CH3), 69.5
(CH2), 71.5 (CH), 120.9 (ArCH), 122.5 (ArCH), 126.0 (ArCH),
127.6 (ArCH), 135.6 (ArC) and 143.1 (ArC); m/z (FAB) 436
[(M Ϫ H)ϩ, 0.1%] (HRMS: found [M Ϫ H]ϩ, 436.2210. Requires
[M Ϫ H], 436.2200).
the resulting light brown oil was purified by flash column chro-
matography on silica gel (10% EtOAc–light petroleum) to give
an inseparable mixture of diastereomers (4d–5d) (159 mg, 99%)
in a ratio of 3:1. The diastereomeric mixture was dissolved in
EtOH (10 ml) and refluxed for 5 h. After cooling to room tem-
perature the solvent was removed under reduced pressure to
give 6d as a light brown oil (125 mg, 83%), νmax (CHCl3)/cmϪ1
3165, 2893, 2846 and 1359; δH (400 MHz, CDCl3) 2.45 (3H, s,
N-CH3), 3.02 (1H, dd, JAB 14 and 7, NCHCHH), 3.11 (1H, dd,
JAB 14 and 6, NCHCHH), 3.64 (1H, dd, JAB 13 and 2, NCHH),
3.99 (1H, m, NCHCH2), 4.26 (1H, dd, JAB 13 and 2, NCHH),
6.87 (1H, dd, J 7 and 1, ArH), 7.09 (1H, m, ArH), 7.15 (2H,
m, ArH), 7.20 (1H, m, ArH) and 7.28 (4H, m, ArH); δC (100
MHz, CDCl3) 41.2 (NCHCH2), 41.3 (N-CH3), 60.9 (NCH2),
71.4 (NCHCH2), {(122.1, 122.6, 126.2, 126.6, 126.9, 128.3,
129.6), ArCH}, 139.3 (ArC), 139.6 (ArC) and 143.6 (ArC);
m/z (EI) 221 [(M Ϫ 2H)ϩ, 24%] (HRMS: found [M Ϫ H]ϩ,
222.1250. Requires [M Ϫ H], 222.1283).
n
Synthesis of 6e using BuLi in THF. The above typical pro-
cedure was followed using complex 3 (100 mg, 0.68 mmol), and
the resulting dark brown oil was purified by Kugelrohr distil-
lation to give 6e as a colourless oil (80 mg, 62%), bp 70 ЊC/0.3
mbar; νmax (CHCl3)/cmϪ1 3459, 2973, 2946, 2853, 2794, 1687,
1458 and 1374; δH (400 MHz, CDCl3) 0.88 (3H, s, CH3), 1.32
(3H, s, CH3), 2.59 (3H, s, N-CH3), 3.68 (1H, d, J 15, NCHH),
3.76 (1H, s, NCHC), 4.43 (1H, d, J 15, NCHH) and 7.19–7.30
(4H, m, ArH); δC (100 MHz, CDCl3) 24.9 (CH3), 26.0 (CH3),
47.5 (N-CH3), 62.1 (NCH2), 73.5 [CH3C(OH)CH3], 82.4
(NCHC), 122.7 (ArCH), 124.1 (ArCH), 126.7 (ArCH), 127.4
(ArCH), 139.7 (ArC) and 140.6 (ArC) (HRMS: found Mϩ,
191.1309. Requires M, 191.1310).
n
Synthesis of 4b–5b using BuLi in THF. The above typical
procedure was followed using 3 (100 mg, 0.68 mmol), and the
resulting light brown oil was purified by flash column chrom-
atography on silica gel (3% Et2O–light petroleum) to give firstly
4b as a colourless oil (91 mg, 61%), νmax (CHCl3)/cmϪ1 2949,
2902, 2364, 2329, 2277, 1603 and 1461; δH (400 MHz, CDCl3)
0.23 (9H, s, SiMe3), 1.1–2.7 (3H, v br, BH3), 2.76 (3H, s, NCH3),
3.75 (1H, s, NCHSi), 3.90 (1H, d, JAB 15, NCHH), 4.50 (1H,
d, JAB 15, NCHH), 7.14 (1H, m, ArH) and 7.21–7.25 (3H, m,
ArH); δC (100 MHz, CDCl3) Ϫ0.5 [Si(CH3)3], 54.7 (NCH3),
68.5 (CH2), 73.2 (CH), 123.0 (ArCH), 123.4 (ArCH), 126.7
(ArCH), 127.4 (ArCH), 135.7 (ArC) and 141.0 (ArC); m/z
(FAB) 218 [(M Ϫ H)ϩ, 2%], 204 [(M Ϫ CH3)ϩ, 4] (HRMS:
found [M Ϫ H]ϩ, 218.1544. Requires [M Ϫ H], 218.1536) fol-
lowed by 5b as a colourless oil (24 mg, 16%, 77% overall), νmax
(CHCl3)/cmϪ1 2952, 2901, 2360, 2330, 2272 and 1460; δH (400
MHz, CDCl3) 0.28 (9H, s, SiMe3), 0.8–2.4 (3H, v br, BH3), 2.84
(3H, s, NCH3), 4.03 (1H, d, JAB 14, NCHH), 4.30 (1H, s,
NCHSi), 4.52 (1H, d, JAB 14, NCHH), 7.15 (1H, m, ArH)
and 7.21–7.25 (3H, m, ArH); δC (100 MHz, CDCl3) Ϫ0.3
[Si(CH3)3], 50.0 (NCH3), 70.0 (CH2), 70.7 (CH), 122.6 (ArCH),
122.9 (ArCH), 126.8 (ArCH), 127.6 (ArCH), 136.5 (ArC) and
140.8 (ArC); m/z (FAB) 218 [(M Ϫ H)ϩ, 8%], 204 [(M Ϫ CH3)ϩ,
22].
Synthesis of 6f using nBuLi in THF. A solution of nBuLi (0.91
ml of a 1.49 M solution in hexanes, 1.35 mmol) was added
dropwise to a solution of complex 3 (100 mg, 0.68 mmol) in
THF (10 ml) at Ϫ78 ЊC under an atmosphere of nitrogen. The
solution was then warmed to 0 ЊC and stirred for 1 h, before
recooling to Ϫ78 ЊC. Benzophenone (0.250 g, 1.37 mmol) in
THF (1 ml) was added in one portion and the mixture stirred at
Ϫ78 ЊC for 1 h before MeOH (1 ml) was added with subsequent
warming to room temperature. Solvent was evaporated under
reduced pressure and the residue was dissolved in EtOH (10 ml)
and stirred at room temperature for 12 h. After evaporation of
the solvent, the residue was purified by flash column chrom-
atography on silica gel (30% EtOAc–light petroleum) to give 6f
as a colourless oil (167 mg, 78%), νmax (CHCl3)/cmϪ1 3382 (OH),
2953, 2856, 2796, 1461 and 1354; δH (400 MHz, CDCl3) 2.20
(3H, s, NCH3), 3.71 (1H, d, JAB 14, NCHH), 4.32 (1H, d, JAB
14, NCHH), 4.75 (1H, br, D2O exchange, OH), 5.05 (1H, s,
NCHC), 5.88 (1H, d, J 8, ArH), 6.84 (1H, m, ArH), 7.11 (2H,
m, ArH), 7.22–7.37 (6H, m, ArH), 7.54 (2H, m, ArH) and 7.76
(2H, m, ArH); δC (100 MHz, CDCl3) 44.3 (CH3), 62.2 (CH2),
78.7 (CH), 121.8 (ArCH), 125.0 (ArCH), {(126.3, 126.6, 126.8,
126.8, 127.5, 128.0, 128.2), ArCH}, {(139.0, 140.7, 144.2,
146.9), ArC}; m/z (FAB) 316 [(M ϩ H)ϩ, 23%] (HRMS: found
[M ϩ H]ϩ, 316.1695. Requires [M ϩ H], 316.1701).
n
Synthesis of 6c using BuLi in THF. The above typical pro-
cedure was followed using complex 3 (100 mg, 0.68 mmol), and
the resulting light brown oil was purified by flash column chro-
matography on silica gel (10% EtOAc–light petroleum) to give
an inseparable mixture of diastereomers (4c–5c) (106 mg, 84%).
The diastereomeric mixture was dissolved in EtOH (10 ml) and
refluxed for 16 h. After cooling to room temperature, the sol-
vent was removed under reduced pressure and the resulting oil
was purified by Kugelrohr distillation to give 6c as a colourless
oil (85 mg, 87%), bp 55 ЊC/0.5 mbar; νmax (CHCl3)/cmϪ1 3081,
2945, 2846, 2783, 1738, 1640, 1461, 1358; δH (400 MHz, CDCl3)
2.60 (3H, s, N-CH3), 2.65 (2H, m, NCHCH2), 3.66 (1H, dd, JAB
13 and 3, NCHH), 3.75 (1H, m, NCH), 4.28 (1H, dd, JAB 13
n
Synthesis of 6g using BuLi in THF. The above typical pro-
and 2, NCHH), 5.10 (1H, dd, J 10 and 2, CH᎐CHH), 5.19 (1H,
cedure was followed using complex 3 (100 mg, 0.68 mmol), and
the resulting light brown solid was purified by flash column
chromatography on silica gel (5% EtOAc–light petroleum) to
give an inseparable mixture of diastereomers (128 mg, 66%).
The diastereomeric mixture was dissolved in EtOH (10 ml) and
refluxed for 5 h. After cooling to room temperature the solvent
was removed under reduced pressure to give 6g as a white solid
(115 mg, 94%), mp 98 ЊC; νmax (CHCl3)/cmϪ1 2944, 2847, 2786,
1601, 1461 and 1358; δH (400 MHz, CDCl3) 2.47 (3H, s, NCH3),
3.20 (1H, dd, JAB 14 and 7, NCHCHH), 3.27 (1H, dd, JAB 14
and 7, NCHCHH), 3.68 (1H, dd, JAB 13 and 2, NCHH), 4.12
᎐
dd, J 17 and 2, CH᎐CHH), 5.90 (1H, ddt, J 17, 10 and 7,
᎐
CH᎐CHH) and 7.21–7.25 (4H, m, ArH); δ (100 MHz, CDCl )
᎐
C
3
37.5 (CH2), 40.8 (N-CH3), 60.8 (CH2), 69.5 (CH), 116.8
(CH᎐CH ), 122.1 (ArCH), 122.2 (ArCH), 126.6 (ArCH), 126.8
᎐
2
(ArCH), 135.5 (CH᎐CH ), 139.9 (ArC) and 143.4 (ArC); m/z
᎐
2
(FAB) 174 [(M ϩ H)ϩ, 21%] (HRMS: found [M ϩ H]ϩ,
174.1276. Requires [M ϩ H], 174.1283).
n
Synthesis of 6d using BuLi in THF. The above typical pro-
cedure was followed using complex 3 (100 mg, 0.68 mmol), and
2444
J. Chem. Soc., Perkin Trans. 1, 1999, 2439–2447

View Article Online
(1H, m, NCHCH2), 4.29 (1H, dd, JAB 13 and 2, NCHH), 6.89
(1H, d, J 7, ArH), 7.08 (1H, m, ArH), 7.17 (2H, m, ArH), 7.45
(3H, m, ArH) and 7.81 (4H, m, ArH); δC (100 MHz, CDCl3)
41.5 (NCHCH2, NCH3), 61.0 (NCH2), 71.3 (NCHCH2),
{(122.1, 122.6, 125.4, 126.0, 126.7, 127.0, 127.7, 127.7, 127.9,
128.1), ArCH}, {(132.3, 133.7, 137.0, 139.7, 143.7), ArC}; m/z
(EI) 272 [(M Ϫ H)ϩ, 2%] (HRMS: found [M Ϫ H]ϩ, 272.1452.
Requires [M Ϫ H], 272.1439).
ee, retention time 13.2 min (minor) and 14.5 min (major)] and
5b as the minor diastereomer [89% ee, retention time 12.8 min
(minor) and 20.6 min (major)].
Recrystallisation of 4b from hexane gave colourless crystals
suitable for X-ray analysis. Storage of 5b in a freezer for two
days caused the oil to solidify, and subsequent recrystallisation
from hexane gave colourless crystals (mp 59 ЊC) suitable for
X-ray analysis.
n
s
Synthesis of 6h using BuLi in THF. The above typical pro-
Synthesis of (1R)-6c mediated by BuLi–(؊)-sparteine. The
cedure was followed using complex 3 (100 mg, 0.68 mmol), and
the resulting light brown oil was purified by flash column
chromatography on silica gel (10% EtOAc–light petroleum) to
give 4h–5h as an inseparable mixture of diastereomers (110 mg,
93%). The diastereomeric mixture (90 mg, 0.51 mmol) was dis-
solved in EtOH (10 ml) and refluxed for 16 h. After cooling to
room temperature the solvent was removed under reduced pres-
sure to give 6h as a brown oil (64 mg, 71%), νmax (CHCl3)/cmϪ1
2877, 2844, 2780, 1484, 1460 and 1360; δH (400 MHz, CDCl3)
0.93 (3H, t, J 7, CH2CH3), 1.61 (2H, m, CH2CH3), 2.54 (3H, s,
NCH3), 3.64 [2H, m, (NCHH ϩ NCHCH2CH3)], 4.25 (1H, d,
J 11, NCHH) and 7.19 (4H, m, ArH); δC (100 MHz, CDCl3) 9.2
(CH3), 24.9 (CH2), 41.0 (CH3), 61.0 (CH2), 71.1 (CH), 122.0
(ArCH), 122.1 (ArCH), 126.6 (ArCH), 126.7 (ArCH), 140.2
(ArC) and 143.5 (ArC); m/z (CI) 162 [(M ϩ H)ϩ, 15%] (HRMS:
found [M ϩ H]ϩ, 162.11287. Requires [M ϩ H], 162.1283).
above typical procedure was followed using 3 (100 mg, 0.68
mmol) and allyl bromide (0.29 ml, 3.38 mmol), and the result-
ing brown residue was purified by flash column chromatog-
raphy on silica gel (5% EtOAc–light petroleum) to give an
inseparable mixture of diastereomers (83 mg, 65%). HPLC
analysis using a Chiralcel OD column (1.5% IPA in hexane as
eluent) indicated that the product mixture contained 5c as the
major diastereomer [87% ee, retention time 11.0 min (minor)
and 12.3 min (major)] and 4c as the minor diastereomer [88%
ee, retention time 16.6 min (minor) and 18.7 min (major)].
The diastereomeric mixture (83 mg, 0.44 mmol) was then
dissolved in EtOH (5 ml) and heated at reflux for 12 h. After
removal of the solvent, the resulting oil was purified by Kugel-
rohr distillation to give 6c as a colourless oil (60 mg, 79%), [α]D20
Ϫ115 (c 0.03 in CHCl3). Spectroscopic details were in accord-
ance with those for the racemic sample described above.
s
Enantioselective reactions in Table 2
Synthesis of (1R)-6d mediated by BuLi–(؊)-sparteine. The
above typical procedure was followed using 3 (100 mg, 0.68
mmol) and benzyl bromide (0.24 ml, 2.0 mmol), and the result-
ing residue was purified by flash column chromatography on
silica gel (5% EtOAc–light petroleum) to give an inseparable
mixture of diastereomers (85 mg, 53%). HPLC analysis using a
Chiralcel OD column (2% IPA in hexane as eluent) indicated
that the product mixture contained 5d as the major diastereo-
mer [83% ee, retention time 12.9 min (major) and 14.7 min
(minor)] and 4d as an undetermined mixture of enantiomers.
The diastereomeric mixture (85 mg, 0.36 mmol) was then
dissolved in EtOH (5 ml) and refluxed for 12 h. After cooling to
room temperature the solvent was evaporated under reduced
pressure to give 6d as a yellow oil (70 mg, 87%), [α]D20 Ϫ34 (c 0.75
in CHCl3). Spectroscopic details were in accordance with those
for the racemic sample described above.
Typical procedure: synthesis of (1S,2S)-4a and (1R,2S)-5a
mediated by BuLi–(؊)-sparteine. A solution of (Ϫ)-sparteine
s
(3.4 ml, 14.87 mmol) and sBuLi (10.4 ml of a 1.3 M solution in
cyclohexane, 13.51 mmol) in Et2O (20 ml) was stirred at Ϫ20 ЊC
for 20 min prior to cannula addition to a suspension of 3 (1.00
g, 6.76 mmol) in Et2O (50 ml) at <Ϫ78 ЊC. The suspension was
stirred at this temperature for 1 h and tributyltin chloride (9.0
ml, 33.8 mmol) was added in one portion. Stirring was main-
tained at Ϫ78 ЊC for a further 1 h before saturated aqueous
NaHCO3 (10 ml) was added with subsequent warming to room
temperature. After addition of EtOAc (40 ml), the organic layer
was separated, washed with saturated brine (2 × 20 ml) and
dried (MgSO4). The solvent was evaporated under reduced
pressure. The resulting yellow residue was subjected to short
column chromatography (neat light petroleum then 5% EtOAc–
light petroleum) to remove tributyltin chloride. Further purifi-
cation by flash column chromatography on silica gel (1%
EtOAc–light petroleum) gave 4a as a colourless oil (0.15 g, 5%),
[α]D20 Ϫ32 (c 1.24 in CHCl3) followed by a low melting point solid
5a (2.66 g, 90%), [α]D20 Ϫ9 (c 1.0 in CHCl3). Spectroscopic details
for both complexes were in accordance with those for the
racemic samples described above. HPLC analysis using a
Chiralcel OD column [0.1% isopropyl alcohol (IPA) in hexane
as eluent] indicated that the product mixture contained 5a as
the major diastereomer [89% ee, retention time 10.3 min
(minor) and 11.6 min (major)] and 4a as the minor diastereo-
mer [45% ee, retention time 6.5 min (major) and 8.3 min
(minor)].
s
Synthesis of (1R)-6f mediated by BuLi–(؊)-sparteine. The
above typical procedure was followed using 3 (100 mg, 0.68
mmol) and benzophenone (0.250 g, 1.35 mmol). The solvent
was evaporated from the crude product under reduced pressure
and (Ϫ)-sparteine was removed by flash column chrom-
atography on silica gel (20% EtOAc–light petroleum). The
crude sample was dissolved in EtOH (5 ml) and stirred at room
temperature for 12 h. After removal of the solvent, the residue
was purified by flash column chromatography on silica gel (20%
EtOAc–light petroleum) to give 6f as a colourless oil (132 mg,
62%), [α]D20 ϩ78 (c 0.61 in CHCl3). Spectroscopic details were in
accordance with those for the racemic sample described above.
s
s
Synthesis of (1S,2S)-4b and (1R,2S)-5b mediated by BuLi–
Synthesis of (1R)-6g mediated by BuLi–(؊)-sparteine. The
(؊)-sparteine. The above typical procedure was followed using
3 (500 mg, 3.40 mmol) and chlorotrimethylsilane (2.16 ml, 17.0
mmol), and the resulting yellow residue was purified by flash
column chromatography on silica gel (4% EtOAc–light petrol-
eum) to give 4b as a white solid (36 mg, 5%), [α]D21 32 (c 0.17 in
CHCl3) followed by 5b as a colourless oil (586 mg, 79%, 84%
overall), [α]D21 0.9 (c 0.71 in CHCl3). Spectroscopic details for
both complexes were in accordance with those for the racemic
samples described above. HPLC analysis using a Chiralcel OD
column (0.5% IPA in hexane as eluent) indicated that the
product mixture contained 4b as the major diastereomer [85%
above typical procedure was followed using 3 (100 mg, 0.68
mmol) and 2-(bromomethyl)naphthalene (2-ArCH2Br) (0.30 g,
1.35 mmol), to give a crude product, which was purified by flash
column chromatography on silica gel (5% EtOAc–light petrol-
eum) to give an inseparable mixture of diastereomers (113 mg,
58%). The diastereomeric mixture (113 mg, 0.39 mmol) was
then dissolved in EtOH (5 ml) and refluxed for 12 h. After
cooling to room temperature, the solvent was removed under
reduced pressure to give 6g as a white solid (100 mg, 93%), [α]D20
Ϫ53 (c 0.59 in CHCl3). Spectroscopic details were in accord-
ance with those for the racemic sample described above.
J. Chem. Soc., Perkin Trans. 1, 1999, 2439–2447
2445

View Article Online
s
Synthesis of (1S,2S)-4h and (1R,2S)-5h mediated by BuLi–
Transmetallation-substitution of 5a in THF solution. In a typ-
ical transmetallation reaction: a solution of BuLi (87 µl of a
n
(؊)-sparteine. The above typical procedure was followed using
3 (100 mg, 0.68 mmol) and iodoethane (0.27 ml, 3.4 mmol),
except that alkylation was facilitated by warming slowly (ca.
12 h) to room temperature. Work-up as before, followed by flash
column chromatography on silica gel (5% EtOAc–light petrol-
eum) gave an inseparable mixture of diastereomers (84 mg,
71%). HPLC analysis using a Chiralcel OD column (5% IPA
in hexane as eluent) indicated that the product mixture
contained 4h as the major diastereomer [75% ee, retention time
11.4 min (minor) and 12.8 min (major)] and 5h as the minor
diastereomer [65% ee, retention time 8.4 min (minor) and 10.0
min (major)].
1.57 M solution in hexanes, 0.14 mmol) was added dropwise to
a solution of 5a (50 mg, 0.11 mmol) in THF (5 ml) at Ϫ78 ЊC
under an atmosphere of nitrogen. The reaction mixture was
stirred at this temperature for 1 h before the electrophile (0.57
mmol) was added in one portion. Stirring was maintained at
this temperature for a further 1 h before saturated aqueous
NaHCO3 (3 ml) was added with subsequent warming to room
temperature. After addition of EtOAc (5 ml), the organic layer
was separated, washed with brine (2 × 5 ml) and dried
(MgSO4). The solvent was evaporated under reduced pressure
and the resulting residue was purified by flash column chrom-
atography on silica gel (2–5% EtOAc–light petroleum) to give a
1
mixture of diastereomers. The products were identified by H
s
Synthesis of (1S,2S)-4i and (1R,2S)-5i mediated by BuLi–
NMR, and were identical to the authentic samples described
above. HPLC analysis using a Chiralcel OD (0.5–5% IPA in
hexane as eluent) indicated that the product mixture contained
products 4 and 5 as indicated in Table 3.
(؊)-sparteine. The above typical procedure was followed using
3 (74 mg, 0.5 mmol) and iodomethane (0.16 ml, 2.5 mmol), and
the residue purified by flash column chromatography on silica
gel (3% then 5% EtOAc–light petroleum) gave 4i (21 mg, 26%),
[α]D25 ϩ15.1 (c 1.07 in CHCl3), followed by 5i (35.4 mg, 44%, 70%
overall), [α]D25 ϩ8.5 (c 1.77 in CHCl3) spectroscopic details for
both compounds were in accordance with those for the racemic
samples described above. HPLC analysis using a Chiralcel
OD column (10% IPA in hexane as eluent) indicated that the
product mixture contained 5i as the major diastereomer [83%
ee, retention time 7.4 min (minor) and 10.1 min (major)] and 4i
as an undetermined mixture of enantiomers.
Data for crystal structure determinations
Crystal structure determination of 4i. A crystal was mounted
on a glass fibre and transferred into the cold stream of the low
temperature device of the diffractometer.
Crystal data. C10H16BN, M = 161.05, monoclinic, a =
10.061(3), b = 9.145(3), c = 11.597(5) Å, β = 114.76(3)Њ,
U = 968.9(4) Å3, T = 150(2) K, space group P21/n (No. 14),
Z = 4, Dc = 1.104 g cmϪ3, µ(Mo-Kα) = 0.062 mmϪ1, 2457 reflec-
tions measured, 1699 unique (Rint 0.011), 1698 used in all calcu-
lations. Final R1 [1425F > 4σ(F)] = 0.0425 and wR(all F2) was
0.104.
s
Synthesis of (1S,2S)-4j and (1R,2S)-5j mediated by BuLi–
(؊)-sparteine. The above typical procedure was followed using
3 (74 mg, 0.5 mmol) and 1-bromopentane (0.31 ml, 2.5 mmol),
except that alkylation was facilitated by warming slowly (ca.
12 h) to room temperature. Work-up as before, followed by
flash column chromatography on silica gel (5% EtOAc–light
petroleum) to give a trace amount of 5j followed by 4j (70 mg,
65%), [α]D21 ϩ52 (c 3.09 in CHCl3); νmax (CHCl3)/cmϪ1 2959, 2861,
2384, 2277 and 1463; δH (400 MHz, CDCl3) 0.89 (3H, t, J 7,
CH2CH3), 1.0–2.2 (3H, v br, BH3), 1.63 (1H, m), 2.74 (3H, s,
NCH3), 3.90 (1H, dd, J 10 and 4, NCH), 3.96 (1H, d, J 15,
ArCHH), 4.53 (1H, d, J 15, ArCHH) and 7.24–7.32 (4H, m,
ArH); δC (68 MHz, CDCl3) 13.9 (CH3), 22.4 (CH2), 25.8 (CH2),
31.5 (CH2), 31.6 (CH2), 52.9 (NCH3), 66.8 (NCH2), 76.7 (CH),
123.0 (ArCH), 123.8 (ArCH), 127.5 (ArCH), 127.9 (ArCH),
135.8 (ArC) and 141.1 (ArC); m/z (EI) 216 (Mϩ Ϫ 1, 2%) and
144 (11) (HRMS: found Mϩ Ϫ 1, 216.1927. Requires M Ϫ 1,
216.1924). HPLC analysis using a Chiralcel OD column (2%
IPA in hexane as eluent) indicated that the product mixture
contained 4j as the major diastereomer [56% ee, retention time
9.3 min (minor) and 10.6 min (major)] and 5j as an
undetermined mixture of enantiomers.
Crystal structure determination of 4b. A crystal was mounted
on a multi-filament glass fibre and transferred into the cold
stream of the low temperature device of the diffractometer.
Crystal data. C12H22BNSi, M = 219.21, orthorhombic, a =
8.201(2), b = 9.792(3), c = 16.905(6) Å, U = 1357.5(7) Å3,
T = 150(2) K, space group P212121 (No. 19), Z = 4, Dc = 1.073 g
cmϪ3, µ(Mo-Kα) = 0.144 mmϪ1, 5373 reflections measured
(including Friedel opposites), 2630 unique (Rint 0.029), 2630
used in all calculations. Final R1 [2464F > 4σ(F)] = 0.0361 and
wR(all F2) was 0.0860. The Flack enantiopole parameter
refined to 0.00(14).13
Crystal structure determination of 5b. A crystal was mounted
on a dual-stage glass fibre and transferred into the cold stream
of the low temperature device of the diffractometer.
Crystal data. C12H22BNSi, M = 219.21, orthorhombic,
a = 9.877(4), b = 10.076(5), c = 13.669(9) Å, U = 1360.3(12) Å3,
T = 150(2) K, space group P212121 (No. 19), Z = 4, Dc = 1.070 g
cmϪ3, µ(Mo-Kα) = 0.143 mmϪ1, 3156 reflections measured
(including Friedel opposites), 2647 unique (Rint 0.021), 2647
used in all calculations. Final R1 [2362F > 4σ(F)] = 0.0563 and
wR(all F2) was 0.143. The Flack enantiopole parameter refined
to 0.0(2).13
Solvent effect: metallation of 3 using sBuLi in different solvent
s
systems. A solution of BuLi (1.04 ml of a 1.3 M solution in
hexanes, 1.35 mmol) was added dropwise to a solution of com-
plex 3 (100 mg, 0.68 mmol) in solvent (either Et2O or THF, 10
ml) at Ϫ78 ЊC under an atmosphere of nitrogen. The reaction
mixture was stirred for 1 h at Ϫ78 ЊC. Chlorotrimethylsilane
(0.43 ml, 3.4 mmol) was added in one portion and the mixture
stirred at Ϫ78 ЊC for 1 h before saturated aqueous NaHCO3 (1
ml) was added with subsequent warming to room temperature.
After addition of EtOAc (10 ml), the organic layer was separ-
ated, washed with saturated aqueous NaHCO3 (5 ml), brine
(2 × 5 ml), dried (MgSO4) and the solvent removed under
reduced pressure. The resulting light brown oil was purified
by flash column chromatography on silica gel (3% Et2O–light
petroleum) to give either 4b or 5b as the major product depend-
ing on the solvent used (see text). The product was identified by
1H NMR, which was identical to the authentic sample.
Crystal structure determination of 7. A crystal was mounted
on a dual-stage glass fibre and transferred into the cold stream
of the low temperature device of the diffractometer.
Crystal data. C12H18IN, M = 303.17, monoclinic, a =
8.639(3), b = 7.619(4), c = 9.731(4) Å, β = 97.22(3)Њ, U =
635.5(3) Å3, T = 150(2) K, space group P21 (No. 4), Z = 2,
Dc = 1.585 g cmϪ3, µ(Mo-Kα) = 2.486 mmϪ1, 3969 absorption-
corrected reflections measured (including Friedel opposites),
2227 unique (Rint 0.014) used in all calculations. Final R1
[2207F > 4σ(F)] = 0.0125 and wR(all F2) was 0.0318. The Flack
enantiopole parameter refined to 0.15(3).13
CCDC reference number (for 4i, 4b, 5b and 7) 207/352. See
http://www.rsc.org/suppdata/p1/1999/2439 for crystallographic
files in .cif format.
2446
J. Chem. Soc., Perkin Trans. 1, 1999, 2439–2447

View Article Online
Tetrahedron, 1994, 50, 6077; (c) T. R. Elworthy and A. I. Meyers,
Tetrahedron, 1994, 50, 6089.
Acknowledgements
We are grateful to Pfizer Central Research, Sandwich, UK, and
to the Royal Society of Chemistry for a Hickinbottom Fellow-
ship, which provided the funding to support M. R. E. We are
also grateful to the University of Malaya (Lembah Pantai,
59100 Kuala Lumpur, Malaysia) for financial support of A. A.
through the SLAB scheme. We also acknowledge the Engin-
eering and Physical Sciences Research Council (EPSRC) for
funding for a diffractometer and for a postdoctoral award to
W.-S. L.
6 Sparteine is a very important ligand for asymmetric metallation, see
for reviews (a) P. Beak, A. Basu, D. J. Gallagher, Y. S. Park and
S. Thayumanavan, Acc. Chem. Res., 1996, 29, 552; (b) D. Hoppe
and T. Hense, Angew. Chem., Int. Ed. Engl., 1997, 36, 2282.
7 In our original communication diastereomer 4 was mistakenly
reported as the major product from quenching with Me3SiCl.
8 This argument relies on the transmetallation reaction of 5a
proceeding with initial retention of configuration, which can usually
be assumed, see for example P. Beak and A. I. Meyers, Acc. Chem.
Res., 1986, 19, 356.
9 The very small quantities of the minor isomer 4b available in non-
racemic form, combined with the small values of the apparent
specific rotations resulted in the misassignment.
10 (a) D. R. Armstrong, W. Clegg, H. M. Colquhoun, J. A. Daniels,
R. E. Mulvey, I. R. Stephenson and K. Wade, J. Chem. Soc., Chem.
Commun., 1987, 630; (b) see also H. Nöth, S. Thomas and
M. Schmidt, Chem. Ber., 1996, 129, 451.
References
1 For reviews, see (a) S. V. Kessar and P. Singh, Chem. Rev., 1997,
97, 721; (b) B. Carboni and L. Monnier, Tetrahedron, 1999, 55,
1197.
2 M. R. Ebden, N. S. Simpkins and D. N. A. Fox, Tetrahedron, 1998,
54, 12923.
3 E. Vedejs and J. T. Kendall, J. Am. Chem. Soc., 1997, 119, 6941.
4 A. J. Blake, M. R. Ebden, D. N. A. Fox, W.-S. Li and N. S. Simpkins,
Synlett, 1998, 189.
5 Lithiated N-formamidine tetrahydroisoquinoline derivatives are
similarly configurationally unstable, see for example (a) A. I.
Meyers, J. Guiles, J. S. Warmus and M. A. Gonzalez, Tetrahedron
Lett., 1991, 32, 5505; see also (b) R. E. Gawley and Q. Zhang,
11 (a) L. J. Beeley and C. J. M. Rockell, Tetrahedron Lett., 1990, 31,
417; (b) A. I. Meyers and B. Santiago, Tetrahedron Lett., 1995,
36, 5877; (c) R. E. Gawley, S. R. Chemburkar, A. L. Smith and
T. V. Anklekar, J. Org. Chem., 1988, 53, 5381.
12 J. K. Whitesell, Chem. Rev., 1989, 89, 1581.
13 H. D. Flack, Acta Crystallogr., Sect. A, 1983, 39, 876.
Paper 9/04788C
J. Chem. Soc., Perkin Trans. 1, 1999, 2439–2447
2447