Regiochemical observations on the lithiation of 1,2,4-trichlorobenzene and reaction with DMF and oxamide electrophiles in THF

Article abstract of DOI:10.1016/S0040-4039(03)01360-1

Unusual regiochemistry is observed in the products arising from the reaction of lithiated 1,2,4-trichlorobenzene with N,N-dimethylformamide and tetraalkyloxamides.

Full text of DOI:10.1016/S0040-4039(03)01360-1

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 5653–5656
Regiochemical observations on the lithiation of
1,2,4-trichlorobenzene and reaction with DMF and oxamide
electrophiles in THF
Andrew J. Burton,a,* Kevin S. Cardwell,a Matthew J. Fuchter,a Mika K. Lindvall,c Rajnikant Patel,b
Terry W. Packham,b Jeremy C. Prodger,a Mark B. Schillinga and Matthew D. Walkera
aGlaxoSmithKline R&D, Chemical Development, Medicines Research Centre, Stevenage, Hertfordshire SG1 2NY, UK
bGlaxoSmithKline R&D, Process Technologies, Temple Hill, Dartford, Kent DA1 5AH, UK
cGlaxoSmithKline R&D, Computational and Structural Sciences, Medicines Research Centre, Stevenage,
Hertfordshire SG1 2NY, UK
Received 23 April 2003; accepted 30 May 2003
This paper is dedicated to the late David Collier
Abstract—Unusual regiochemistry is observed in the products arising from the reaction of lithiated 1,2,4-trichlorobenzene with
N,N-dimethylformamide and tetraalkyloxamides.
© 2003 Elsevier Ltd. All rights reserved.
The a-keto amide 3 is a key intermediate in the synthe-
sis of the sodium channel blocker GW356194 4.1 The
synthesis of 3 has involved conversion of 2,3,5-
trichlorobenzaldehyde 2 into 3 via a four stage proce-
dure that uses both oxidising conditions and cyanide
chemistry (Scheme 1). We present here the results of
our investigations into the regiochemistry and mecha-
nism of the reaction of lithiated 1,2,4-trichlorobenzene
1 with N,N-dimethylformamide (DMF). In addition we
report a new, one-step synthesis of the a-keto amides
12, 13 and 14 for use as intermediates in the synthesis
of GW356194 4.
2). However, when a THF solution of anion 1 is reacted
with 5 equiv. DMF at −78°C and allowed to warm
slowly to room temperature the unexpected 2,3,5-
trichlorobenzaldehyde regioisomer 2 is the major reac-
tion product, accompanied by a small amount of the
2,4,5-trichlorobenzaldehyde
8 (conditions b). Con-
versely, if the reaction is quenched without aging then
the expected 2,3,6-trichlorobenzaldehyde 7 predomi-
nates (Scheme 2) and can be isolated in 50% yield
(conditions c).2 Detailed distributions of isomers (as
judged by capillary GC) as the reaction is warmed to
room temperature are reported in Table 1.
The chloro substituents in 1,2,4-trichlorobenzene com-
bine to make the anion in the 3-position the most stable
and this can be demonstrated by deuteration (Scheme
The reaction of tetramethyloxamide with organolithium
compounds has been described for the direct synthesis
a-keto amides.3 We wanted to see if lithiated 1,2,4-
Scheme 1. Reagents and conditions: (a) DMF (5 equiv.), THF, −78°C–rt, 76%; (b) (i) NaMnO4, NaOH, 85%; (ii) oxalyl chloride,
toluene; (iii) CuCN, CH3CN; (iv) 4 M HCl in 1,4-dioxane, 1 equiv. H2O, 55% (three stages).
Keywords: organolithium; 1,2,4-trichlorobenzene; 2,3,5-trichlorobenzaldehyde; a-keto amide.
* Corresponding author. Tel.: +44 1438 763667, fax: +44 1438 764869; e-mail: andrew.j.burton@gsk.com
0040-4039/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)01360-1

5654
A. J. Burton et al. / Tetrahedron Letters 44 (2003) 5653–5656
and 11 gave rise to the a-keto amides 12, 13 and 14,
respectively (Scheme 3, conditions a).4 However, we
were surprised to find that this reaction did not show
the same temperature dependence as the reaction of 1
with dimethylformamide and none of the 2,3,6-isomer
could be isolated as demonstrated for oxamide 9
(Scheme 3, conditions b). Also, none of the 2,4,5 iso-
mers were observed.
We envisaged that our observations for the reaction
with oxamides compared to DMF could be explained
by operation of a different reaction pathway. In the
formylation reaction, we propose that the adduct 17 is
formed in the 3-position and then rearranges. In order
to probe this mechanistic hypothesis we decided to
carry out labelling studies to see if the rearrangement of
the DMF adduct was inter- or intramolecular. There-
fore we looked for a way of generating the adduct 17
that ensured it would be the only reactive species in
Scheme 2. Reagents and conditions: (a) (i) nBuLi, THF,
−78°C, (ii) D2O; (b) (i) nBuLi, THF, (ii) DMF (5 equiv.),
−78°C–rt, (iii) H2O; (c) (i) nBuLi, THF, (ii) DMF (5 equiv.),
−78°C, (iii) H2O.
Table 1.
solution. Thus, 2,3,6-trichlorobenzaldehyde
7 was
Time (h)
T (°C)
Ratio of isomers
treated with lithium dimethylamide (LDMA) and the
composition of the resulting reaction mixture was deter-
mined by HPLC. On addition of LDMA we observed
rapid conversion to trichlorobenzene, indicating that
the aldehyde had decomposed to give lithiated
trichlorobenzene 1, presumably via adduct 17, and
DMF. The lithiated trichlorobenzene then gradually
reacted with the DMF over 18 h to give a mixture of
2,3,5 (2)
2,4,5 (8)
2,3,6 (7)
1.5
2.5
4
6
20
−65
−40
−20
15
0
42
88
92
92
0
4
7
7
7
100
50
5
0
0
20
2,3,5- and 2,4,5-trichlorobenzaldehyde
respectively.
2
and 8,
trichlorobenzene would react with oxamides to give
a-keto amides with the correct regiochemistry for use as
synthons in the synthesis of GW356194 4 and we were
delighted to find that reaction of 1 with oxamides 9, 10
Although not useful for adduct regeneration this result
suggests an intermolecular rearrangement mechanism in
which the adducts are in equilibrium with lithiated
Scheme 3. Reagents and conditions: (a) (i) nBuLi, THF, −78°C, (ii) R2N(CO)2NR2 (NR2=NMe2 9; N-pyrrolidyl 10; N-piperidyl
11), −78°C–0°C, (iii) NaHCO3 (aq.); (b) (i) nBuLi, −78°C, (ii) 9 −78°C/6 h, (iii) NaHCO3 (aq.) (−78°C).
Scheme 4.

A. J. Burton et al. / Tetrahedron Letters 44 (2003) 5653–5656
5655
trichlorobenzene and DMF, with the mixture gradually
equilibrating to give the thermodynamically most
favourable adducts (Scheme 4). This suggestion was
further supported by the incorporation of a deuterium
label on addition of D7-DMF to the mixture and by a
double crossover experiment (Scheme 5).5 Three of the
four labelled components were observed directly by
selective HMBC experiments and the fourth by com-
parison with the coupled 13C NMR spectrum.
and with the product distribution itself at room
temperature.
Thus, while anion formation between the two chlorines
is favoured energetically, the corresponding product is
disfavoured, accounting for the opposite temperature
dependence trend between the 2,3,6 isomer and the
other two regioisomers. At low temperatures and short
reaction times, the relative stability of the aromatic
anions seems to dominate the product distribution.
Longer reaction time and warming to room tempera-
ture may allow extensive equilibration between the
regioisomeric adducts, and the relative energies of
adducts, modelled in THF, are broadly in line with the
relative ratio of 2,3,5, 2,4,5 and 2,3,6 isomers.
In reaction with the more sterically demanding oxam-
ides initial reaction at the 3-position appears not to
take place and we suggest that the anion 1 is in
equilibrium with 15 and reaction only takes place with
the less sterically hindered anion.6 This suggestion was
supported by our observation that both 1,3-
dichlorobenzene8 and 1,3,5-trichlorobenzene could be
formylated between the two chlorine atoms but showed
no reactivity towards oxamides. An analogous
crossover experiment using the a-keto amide 13, D9-13
and lithum pyrrolidyl amide was also carried out but
this led only to 2% deuterium crossover, as estimated
by high accuracy mass spectrometry.
In conclusion we have demonstrated that lithiated
1,2,4-trichlorobenzene is a valuable intermediate for the
one-step synthesis of 2,3,5 substituted a-keto-amides
when reacted with tetra-alkyloxamides and it is likely
that the reaction proceeds via a different mechanism to
the known synthesis of 2,3,5-trichlorobenzaldehyde.
In addition, the structures and energies of the regioiso-
mers of the aromatic anions, DMF adduct anions and
products were determined via a quantum chemical
approach.9 The energies of each regioisomeric series are
shown in Figure 1. The values are consistent both with
the temperature dependence of the product distribution
Acknowledgements
We would like to thank Peter Moore and Alec Simpson
for assistance in obtaining NMR and mass spectra for
the crossover experiments.
Scheme 5. Reagents and conditions: (a) (i) LDMA, −78°C, THF, 30 min, (ii) D7 DMF 1 equiv., (iii) −78°C–rt, (iv) H2O; (b) (i)
LDMA, −78°C, THF, 30 min; (c) (i) −78°C–rt over ca. 20 h, (ii) H2O.
Figure 1.

5656
A. J. Burton et al. / Tetrahedron Letters 44 (2003) 5653–5656
mer. This is supported by our observations.
References
7. Schlosser, M.; Marzi, E.; Cottet, F.; Bu¨ker, H. H.; Nib-
berling, N. M. M. Chem. Eur. J. 2001, 7, 3511.
1. PCT Int. Appl. (1999), WO 9932462 A1.
8. Kress, T. H.; Leanna, M. R. Synthesis 1988, 803.
9. All low energy conformers for all starting materials,
adduct anions and product isomers were located using
Low Mode Conformational Searching and the MMFF
force field in the gas phase, as implemented in Macro-
Model v. 7.0 (Schrodinger, Inc.). The resulting conform-
ers were further optimised at the B3LYP/6-31+G* level
using Gaussian98 A.7 (Gaussian, Inc., Pittsburgh, PA,
1998), and the energies of the adduct anions in THF
(using gas phase geometries) were calculated using
Tomasi’s PCM solvation model. The lowest energy con-
former of each isomer was considered in energy com-
parisons between isomers.
2. PCT Int. Appl. (1995), WO 9507877 A1.
3. Campaigne, E.; Skowronski, G.; Rogers, R. B. Synth.
Commun. 1973, 3, 325.
4. All new compounds gave concordant 1H, 13C NMR and
mass spectral analyses.
5. Deuterated trichlorobenzen2e, for the synthesis of 7-D/
D, was obtained from the corresponding Grignard
reagent, suggesting that these reagents do not display the
same anion equilibration.
6. Schlosser and co-workers have derived a set of coeffi-
cients to determine the acidity of chlorinated benzene
aromatics.7 Using their predictions, the 2,3,5-isomer
should be the next most stable anion after the 2,3,6-iso-