Stereoselective γ-lactam synthesis via palladium-catalysed intramolecular allylation

Article abstract of DOI:10.1039/b504731e

A novel route to the synthesis of 3-(tolylsulfonyl)-4,5-cis-disubstituted γ-lactams using a diastereoselective palladium-catalysed intramolecular allylation of amino acid-derived allylic carbonates has been developed. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b504731e

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Stereoselective c-lactam synthesis via palladium-catalysed
intramolecular allylation{
Donald Craig,*a Christopher J. T. Hylanda and Simon E. Wardb
Received (in Cambridge, UK) 7th March 2005, Accepted 11th May 2005
First published as an Advance Article on the web 9th June 2005
DOI: 10.1039/b504731e
products where the cis diastereoisomer was favoured as a result
A novel route to the synthesis of 3-(tolylsulfonyl)-4,5-cis-
disubstituted c-lactams using a diastereoselective palladium-
catalysed intramolecular allylation of amino acid-derived allylic
carbonates has been developed.
of a metal-template effect.3 In light of these findings it was
proposed that palladium-catalysed cyclisation of amino acid-
derived allylic carbonates such as 2 might offer a diastereoselective
route to 4,5-cis-disubstituted c-lactams.4
During a programme directed towards the synthesis of alkaloids
related to morphine we required a stereoselective method for the
assembly of 3-(tolylsulfonyl)-4,5-cis-disubstituted c-lactams 1. A
search of the literature revealed few methods for the synthesis of
the thermodynamically disfavoured 4,5-cis-disubstituted isomers,
and it was felt that given their chemical and biological significance1
the development of a general route to these structures was
worthwhile.
Initial cyclisation studies were carried out on the L-serine
derived cyclisation substrate 2, which was synthesised as shown in
Scheme 1. N-Tosyl-L-serine5 was subjected to Grignard addition6
followed by reduction of the resulting ketone with Et3SiH–TFA to
give 3. Subsequent Swern oxidation and Wittig homologation was
followed by DIBAL-H-mediated reduction of the resulting ester.
The resultant alcohol was detosylated under dissolving metal
conditions and reductive amination carried out on the crude amine
product to yield 4. Finally, a DCC-assisted coupling with
tosylacetic acid and formation of the carbonate provided the
desired cyclisation substrate 2.
The use of p-allyl palladium chemistry for the construction of
c-lactams via formation of the C3–C4 bond has been described
extensively by Poli et al.2 Trost has reported the palladium-
catalysed cyclisation of allylic substrates to give cyclopentyl
For the proposed cyclisation Pd2(dba)3 was chosen as the Pd(0)
source due to its air stability and ease of handling, whilst the highly
basic tris(2,4,6-trimethoxyphenyl)phosphine (TTMPP)7 was cho-
sen as the ligand. The results of the investigation are summarised
in Table 1. It was found that the use of TTMPP in MeCN as
Table 1 Optimisation of palladium-catalysed cyclisation
Scheme 1 Reagents and conditions: (i) BrMgArOMe (8 equiv.), n-BuLi
(2 equiv.), rt, 37 h, 73%; (ii) Et3SiH (10 equiv.), TFA (20 equiv.), 45 uC,
24 h, 80%; (iii) (a) (COCl)2 (1.2 equiv.), DMSO (2.4 equiv.), CH2Cl2,
278 uC, 45 min; (b) Et3N (5 equiv.), rt, 10 min; (iv) Ph3PCHCO2Et
(4 equiv.), CH2Cl2, rt, 12 h, 86% over two steps; (v) DIBAL-H (3.6 equiv.),
CH2Cl2, 278 uC, 1 h, 78%; (vi) Na (5 equiv.), naphthalene (5.3 equiv.),
THF, rt, 1 h or NH3(l), Na (6 equiv.), 278 uC, 15 min; (vii) (a)
Entry
Conditionsa
cis : trans ratiob
1
2
3
4
5
6
7
8
9
a
TTMPP, PhMe, rt A 50 uC, 12 h
TTMPP, THF, rt A 50 uC, 12 h
TTMPP, MeCN, rt, 30 min
TTMPP, MeCN, 50 uC, 15 min
TIPP, MeCN, rt A 50 uC, 12 h
n-Bu3P, MeCN, rt A 50 uC, 12 h
P(Cy)3, MeCN, rt A 50 uC, 12 h
PPh3, MeCN, rt A 50 uC, 12 h
dppe, MeCN, rt A 50 uC, 12 h
No reaction
65 : 35
85 : 15c
78 : 22
67 : 33
67 : 33
63 : 37
63 : 37
50 : 50
˚
4-methylbenzaldehyde (4 equiv.), 4 A MS, AcOH, THF, rt, 1 h; (b)
NaCNBH3 (5 equiv.), 278 uC A rt, 10 h, 60% over two steps; (viii) DCC
(2.1 equiv.), TsCH2CO2H (2 equiv.), THF, rt, 12 h; (ix) methyl
chloroformate (2 equiv.), pyridine (2 equiv.), DMAP (cat.), CH2Cl2, rt,
2 h, 87% over two steps.
All reactions were carried out using 5 mol% [Pd2(dba)3] and
50 mol% phosphine ligand. Isomer ratios were determined by 1H
NMR analysis of crude products; reactions proceeded in . 90%
b
{ Electronic supplementary information (ESI) available: experimental
details and full spectroscopic data for all synthetic intermediates, allylic
carbonate substrates and lactam cyclisation products. See http://
www.rsc.org/suppdata/cc/b5/b504731e/index.sht
yield by 1H NMR. Product 1 + 5 was obtained in 90% isolated
yield.
c
*d.craig@imperial.ac.uk
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 3439–3441 | 3439

View Article Online
solvent resulted in a rapid (, 30 min) cis-selective cyclisation
(85 : 15) producing c-lactams 1 and 5 in excellent yield (90%)
(Table 1, entry 3). The cis-relationship of the vinyl and
p-methoxybenzyl groups was assigned unambiguously from the
observation of NOE interactions between H-4 and H-5. With
TTMPP as the ligand, use of solvents other than MeCN resulted
in either lower selectivity or unviably slow reactions, even at
elevated temperatures. A range of other phosphines in conjunction
with MeCN were investigated for the cyclisation (entries 5–9),
but no increase in selectivity was observed. Interestingly, only
with TTMPP in MeCN did the reaction proceed at room
temperature.
Table 2 Palladium-catalysed cyclisation of 2 and 7a–d
Entry Substrate
cis : trans ratioa Isolated yield
R
1
2
3
4
5
a
2
CH2ArOMe 85 : 15
90%
7a
7b
7c
7db
Me
i-Pr
86 : 14
67 : 33
90 : 10
83 : 17
90%
78%
85%
79%
CH2i-Pr
n-Bu
Determined by 1H NMR analysis of the crude product and NOE
b
or NOESY experiments. Racemic substrate.
With the conditions optimised for the synthesis of 4,5-cis-
c-lactam 1 from substrate 2, we wished to investigate the generality
of the procedure for other amino acid-derived substrates 7a–d, and
in particular the effect of the steric bulk of R on the selectivity of
the cyclisation (Scheme 2). The required substrates were synthe-
sised from L-alanine, L-valine, L-leucine and DL-norleucine
respectively via an adaptation of the route used for 2. Thus, the
N-tosylamino acids were reduced with LiAlH4 and then subjected
to Swern oxidation and Wittig homologation. The resulting
unsaturated esters were reduced with DIBAL-H, desulfonylated
and the crude amines then subjected to reductive amination to give
secondary amines 6a–d. Amines 6a–d were coupled with tosylacetic
acid using PyBOP,8 and the resulting allylic alcohols carboxy-
methylated under standard conditions to give the desired
cyclisation substrates 7a–d. All steps proceeded in good to
excellent yield for all substrates.
Fig. 1 Proposed model for cis diastereoselectivity of the cyclisation.
preferred reactive conformation 9, where the palladium and its
associated ligands avoid unfavourable interactions with the R
group. As a linear relationship between the size of the R group and
the selectivity of the reaction was not observed, it is assumed that
there are other factors controlling the selectivity. For example, a
1,3-diaxial interaction between R and the enolate oxygen and/or
H1 might be considered in addition to the Pd–R interaction
(Fig. 1).
The allylic carbonates 7a–d were subjected to the conditions
optimised for substrate 2 (entry 3, Table 1; entry 1, Table 2).{ All
substrates underwent facile cis-selective cyclisation (Table 2). The
cis-diastereoselectivity of the cyclisation may be explained by the
In conclusion, we have developed a new method for the
efficient, diastereoselective synthesis of cis-4,5-disubstituted
c-lactams through the Pd(0)-catalysed cyclisation of a-tosyl-
substituted amides. These are the first examples of such
cyclisations to be carried out on acyclic substrates having this
level of complexity,2 and the substrates are readily synthesised as
single enantiomers using routine transformations starting from
a-amino acids. A model to explain the observed stereoselectivity
has been proposed. In addition, the highly functionalised nature of
c-lactams 1 and 8a–d render them amenable to further substitu-
tion. For example, 1 undergoes alkylation of the derived sulfone-
substituted enolate on the less hindered face with complete
stereoselectivity to provide 10 in good yield (Scheme 3).
The authors thank EPSRC and Knoll Pharmaceuticals
(CASE studentship to C. J. T. H.) and GlaxoSmithKline for
support.
Scheme 2 Reagents and conditions: (i) LiAlH4 (3.0 equiv.), THF, reflux,
2 h; (ii) (a) (COCl)2 (1.2 equiv.), DMSO (2.4 equiv.), CH2Cl2, 278 uC,
45 min; (b) Et3N (5 equiv.), rt, 10 min; (iii) Ph3PCHCO2Et (4 equiv.),
CH2Cl2, rt, 12 h; (iv) DIBAL-H (3.6 equiv.), CH2Cl2, 278 uC, 15 min,
then rt, 2 h; (v) NH3 (l), Na (6 equiv.), 278 uC; (vi) (a) 4-methylbenzalde-
˚
hyde (2 equiv.), MeOH, 4 A MS, 12 h, rt; (b) NaBH4 (2.4 equiv.), 0 uC–rt,
1 h; (vii) DCC (2.1 equiv.), HOBt (2.1 equiv.), TsCH2CO2H (2 equiv.),
CH2Cl2, rt, 12 h or PyBOP (2 equiv.), Hu¨nig’s base (5.5 equiv.),
TsCH2CO2H (2 equiv.), CH2Cl2, rt, 12 h; (viii) methyl chloroformate
(2 equiv.), pyridine (2 equiv.), DMAP (cat.), CH2Cl2, 1 h.
Scheme 3 Reagents and conditions: KH (1.1 equiv.), prenyl bromide
(10 equiv.), DMF, 0 uC, 30 min, 78%.
3440 | Chem. Commun., 2005, 3439–3441
This journal is ß The Royal Society of Chemistry 2005

View Article Online
Donald Craig,*a Christopher J. T. Hylanda and Simon E. Wardb
aDepartment of Chemistry, Imperial College London, South Kensington
Campus, London, UK SW7 2AZ. E-mail: d.craig@imperial.ac.uk;
Fax: +44 (0)20 7594 5868; Tel: +44 (0)20 7594 5771
bGlaxoSmithKline Research Ltd, New Frontiers Science Park North,
Third Avenue, Harlow, Essex, UK CM19 5AW.
M. R. Reese, L. M. Reeves, E. J. Stam and S. A. Yocum, J. Med. Chem.,
2000, 43, 2293.
2 (a) G. Giambastiani, B. Pacini, M. Porcelloni and G. Poli, J. Org. Chem.,
1998, 63, 804; (b) G. Poli and G. Giambastiani, J. Org. Chem., 2002, 67,
9456; (c) S. Thorimbert, G. Giambastiani, C. Commandeur, M. Vitale,
G. Poli and M. Malacria, Eur. J. Org. Chem., 2003, 2702; (d) S. Lemaire,
G. Giambastiani, G. Prestat and G. Poli, Eur. J. Org. Chem., 2004,
2840–2847.
E-mail: simon.e.ward@gsk.com; Fax: +44(0)1279 622895;
Tel: +44 (0)1279 627728
3 B. M. Trost and P. H. Lee, J. Am. Chem. Soc., 1991, 113, 5076.
4 For other examples of Pd(0)-catalysed cyclisation of a-tosyl esters,
see: Y.-G. Suh, J.-K. Jung, B.-C. Suh, Y.-C. Lee and S.-A. Kim,
Tetrahedron Lett., 1998, 39, 5377; J. A. Marshall, R. C.
Andrews and L. Lebioda, J. Org. Chem., 1987, 52, 2378; A. S.
Kende, I. Kaldor and R. Aslanian, J. Org. Chem., 1988, 53,
6265.
5 J. J. Caldwell, D. Craig and S. P. East, Synlett, 2001, 1602.
6 T. F. Buckley and H. Rapoport, J. Am. Chem. Soc., 1981, 103,
6157.
7 TTMPP 5 tris(2,4,6-trimethoxyphenyl)phosphine. See: M. Wada and
S. Higashizaki, J. Chem. Soc., Chem. Commun., 1984, 482.
8 PyBOP 5 (benzotriazol-1-yloxy)tripyrrolidinophosphonium hexafluoro-
phosphate; B. Castro, J. R. Dormoy, G. Evin and C. Selve, Tetrahedron
Lett., 1975, 16, 1219.
Notes and references
{ Typical experimental procedure for cyclisation reactions of carbonates 2,
7a–d: A solution of the carbonate (0.103 mmol, 1.0 equiv.) in MeCN (1 ml)
was added to a flask charged with Pd2(dba)3 (5.0 mol%) and TTMPP
(0.052 mmol, 0.5 equiv.) at rt. After the carbonate had been consumed as
indicated by tlc, the mixture was concentrated under reduced pressure.
Chromatography (30% EtOAc–petrol) then gave the c-lactams as an
inseparable cis : trans mixture.
1 For example, potent MMP-13 inhibitors based on a c-lactam scaffold
have recently been reported: R. P. Robinson, E. R. Laird, J. F. Blake,
J. Bordner, K. M. Donahue, L. L. Lopresti-Morrow, P. G. Mitchell,
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 3439–3441 | 3441