2,5-disubstituted pyrrolidines: Versatile regioselective and diastereoselective synthesis by enamine reduction and subsequent alkylation

Article abstract of DOI:10.1039/b303789d

A direct and versatile route enabling the regio- and diastereoselective synthesis of 2,5-disubstituted pyrrolidines by reduction of enamines derived from pyroglutamic acid is reported.

Full text of DOI:10.1039/b303789d

View Online / Journal Homepage / Table of Contents for this issue
2,5-Disubstituted pyrrolidines: versatile regioselective and
diastereoselective synthesis by enamine reduction and subsequent
alkylation
Syed Raziullah Hussaini and Mark G. Moloney*
The Department of Chemistry, Dyson Perrins Laboratory, The University of Oxford,
South Parks Road, Oxford, OX1 3QY
Received 7th April 2003, Accepted 1st May 2003
First published as an Advance Article on the web 6th May 2003
A direct and versatile route enabling the regio- and
diastereoselective synthesis of 2,5-disubstituted pyrrol-
idines by reduction of enamines derived from pyroglutamic
acid is reported.
alcohols 2a,b (yields 50 and 100% respectively), resulting from
C-2 ester reduction and not enamine reduction (Scheme 1).32
Interest in the pyrrolidine ring nucleus has arisen not only
from its synthetic perspective but more recently due to its
unique ability to induce well-defined molecular architecture in
peptidomimetic compounds1–3 and for its wide-ranging bio-
logical activity. Examples of bioactive pyrrolidines include
the antibiotic lemonomycin,4 the novel glutamate receptor
antagonist kaitocephalin,5 the neuraminidase inhibitors
A-3156756,7 and A-192558,8 and the protease inhibitor aeru-
ginosin.9 Pyroglutamic acid is an obvious candidate for the
synthesis of modified pyrrolidines, but the close relationship
of functional groups and stereochemistry in this compound
make for unexpected complications in protecting group chem-
istry and chemical manipulations.10,11 We have been interested
in the development of reliable chemistry for the stereo-
controlled manipulation of any or all ring positions of pyro-
glutamic acid, and this has been achieved for positions 3 and
4, giving access to diversely substituted12–15 and structurally
well-defined16 pyroglutamates. Our focus has more recently
turned to manipulation of the lactam carbonyl as a route to
2,5-disubstituted pyrrolidines, compounds of considerable
interest.17 Literature chemistry for this transformation have
relied on four main routes: lactim ether formation followed by
condensation with a β-dicarbonyl compound;18,19 selective
reduction of the lactam carbonyl followed by reaction with
nucleophiles20 or Wittig reagents,21,22 or the reverse procedure
by direct addition of organoaluminium reagents followed by
reduction;23 opening of the lactam followed by reclosure;24,25
and Eschenmoser ring contraction.26 The latter reaction,
involving condensation of a thiolactam with an α-bromo-
ester, was of appeal since it has been extensively applied by
Rapoport and co-workers,27 and can occur under particularly
mild conditions, especially when the bromo component is di-
substituted with electron withdrawing groups, and should be
tolerant of other functional groups. Recent reports have begun
to expand the synthetic utility of these adducts.28,29 However,
literature reports indicate that reduction of the enamine
products from the Eschenmoser contraction is highly substrate
and reagent dependent, being particularly easy for β-enamino
esters, and that cis-stereocontrol with catalytic hydrogenation
and trans-stereocontrol with metal hydrides was possible.30
We report here methodology which extends the scope of
this approach for the direct synthesis of 2,5-disubstituted
pyrrolidines.
Scheme 1
Attempted reductions of ester 1a under acidic (NaBH4,
HOAc; Et3SiH, TFA) or neutral (CrCl2, THF, H2O) condi-
tions,33 or of alcohol 2b (Redal-H), returned unreacted starting
material. Protection of the nitrogen function of 1a,b as a BOC
derivative under forcing conditions gave 3a and 3b in 73 and
23% yields respectively, but this enamine function still proved to
be unreactive, since reduction of 3a with Redal-H or NaBH4
gave pyrrolidine 4a from ketone reduction (24%) or enamine 4b
from C-2 ester reduction and debenzoylation (11%) respectively.
Furthermore, application of conditions (nBuLi, CuI, Et2O) to
1a and 3a, recently reported to be successful for conjugate addi-
tions to lactam-derived enamine systems29 was not successful,
illustrating the lack of reactivity of this enamine system.
In view of these serious difficulties over chemoselectivity in
such diversely substituted substrates, reductions of diester 1b
were examined (Scheme 2). Sodium cyanoborohydride in
ethanol at pH 3–534 gave product 5a although this product
could not be separated from an unidentified impurity. Catalytic
hydrogenation (H2, Pt/C or Pd/C, 4.5 atm) also returned un-
reacted starting material, indicating a substantial lowering of
reactivity of β-enamino diesters relative to their β-enamino
ester counterparts, but application of more forcing conditions
(H2, PtO2.H2O,† TFA–HOAc (1 : 3), 4.5 atm, 48 h)31 gave a
quantitative yield of a 1 : 2 ratio of an inseparable cis : trans
product mixture 5a (this stereochemical assignment was estab-
lished subsequently on the benzoyl derivative 5b).35 This
approach provides a convenient solution to the problem of the
reduction of β-enamino diesters which has been noted in the
literature.31,34,36 Benzoylation of the amine nitrogen of 5a
(PhCOCl, py, 30–40 ЊC) proceeded without difficulty to give a
separable mixture of cis- and trans-5b in excellent yield but
We prepared lactams 1a,b according to the literature route31
and initially examined their reduction using conditions which
had been successfully applied to related enamine substrates
substituted with a single ester function.30 However, the enamine
function proved to be particularly resistant to reduction and
reaction of 1a with sodium borohydride for example gave
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 8 3 8 – 1 8 4 1
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
1838

View Online
Scheme 2
variable ratio, suggesting the possibility of equilibration. In the
1H NMR spectrum of these compounds (in CDCl3), all 3 ester
methylenes, H-2 and H-6 signals were co-incident for the cis-
product, but in the trans compound, one of the ester methylenes
was upfield from the others by 0.4 ppm and H-2 and H-6 were
downfield of the remaining two ester methylenes. Furthermore,
these signals exhibited extensive broadening for cis-5b, presum-
ably as a result of rotameric equilibration, but those of trans-5b
did not. Variable temperature (VT) analysis (373 K in d6-
DMSO) resulted in significant sharpening of all signals, and
enabled acquisition of NOE data (Fig. 1). For the cis- isomer,
both C(2)H and C(5)H, in addition to a weak mutual enhance-
ment, gave clear enhancements at C(3)H and C(4)H, but the
stereochemistry of trans-5b was indicated by C(2)H enhance-
ments at C(3)H and C(3)HЈ and C(5)H enhancements at
C(3)HЈ and C(4)HЈ, with no C(2)H–C(5)H enhancements.
Independent stereochemical assignment, which confirmed these
NOE results, was possible by single crystal X-ray analysis of
pyrrolidine cis-5b‡; this indicated a conformation in which
the C-2 substituent was pseudoaxial, the C-5 group pseudo-
equatorial, and C-3 out of plane of the remaining ring atoms
with a slightly pyramidalised nitrogen and the carbonyl group
directed towards the C-5 substituent. Presumably this arrange-
ment minimises eclipsing interactions in the ring which would
arise in the di(pseudoequatorial) conformer. Further investi-
gation of pyrrolidines 5b indicated that the pure cis isomer
could be converted (NaH–DMF, 0 ЊC RT) to a cis : trans
product mixture (1.7 : 1), and that treatment of the pure trans
isomer with LDA at low temperature also gave recovered start-
ing material in which the trans isomer predominated (5.4 : 1).
This clearly indicates that equilibration of the two isomers
is possible; we presume that initial β-elimination, favoured by
a highly acidic H-6 and good N-benzoyl leaving group, is
Fig. 1
responsible for this outcome under these conditions.
Unexpected retro-Michael additions of cyclic β-aminoesters
have been observed previously.37
We anticipated that compound 5b would prove to be a
useful synthetic template by direct modification of the C-2
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 8 3 8 – 1 8 4 1
1839

View Online
ethoxycarbonyl or C-5 bis(ethoxycarbonyl)methyl substituents,
but observed some unexpected complications in its reactivity.
Attempted reduction of the C-2 methoxycarbonyl of 5b (which
had been straightforward for enamines 1a,b as noted above)
proved to be problematic, and formation of amide 6a (40%),
along with the inseparable by-products 6b (6%) and 6c (7%),
was observed. These results are also consistent with a facile
β-elimination of starting 5b, which in this case is followed by
α,β-double bond and ester reduction (once or twice) to give the
observed products.
manipulation, providing a versatile and diastereocontrolled
access to 2,5-pyrrolidines. Furthermore, an effective spectro-
scopic protocol involving 1H NOE analysis at high temperature
(373 K) has been identified which minimises conformational
effects and permits detailed stereochemical assignments to be
made; where possible, these have been confirmed independently
by crystallographic analysis.
Acknowledgements
SRH gratefully acknowledges support from The University of
Oxford (Lady Noon/OUP Fund), St Peter’s College, Oxford
for a Graduate Studentship Award, and Lancaster Synthesis
for additional support. We thank Steve Bell and Dr Andrew
Cowley for the crystallographic analysis. We acknowledge the
use of the EPSRC Chemical Database Service at Daresbury38
and the EPSRC National Mass Spectrometry Service Centre at
Swansea.
In an effort to extend the C-6 (bismethoxycarbonyl)methyl
substituent using a standard alkylation strategy, deprotonation
followed by electrophilic quench was examined. Treatment of
pure trans-5b with NaH then benzyl bromide in DMF gave a
51% yield of trans-7a. The stereochemistry of 7a was unequivo-
cally assigned by VT/NOE analysis in d6-DMSO at 373 K
(Fig. 1). Significantly, irradiation of the benzylic system gave
enhancements at C(2)H, C(3)H and C(4)H, suggesting a pre-
ferred conformation in which this group is folded under the
heterocyclic ring. However, the cis starting material 5b gave
none of the expected cis product 8a, but instead a 59% yield of
trans-7a. In each of these reactions, starting material (about
10%) was recovered as a mixture of cis and trans isomers. These
results are also consistent with the facile cis ↔ trans equilib-
ration as described above, followed by alkylation. Alternatively,
methylation of pure trans-5b with NaH/methyl iodide gave a
25% yield of trans-7b. The stereochemistry of 7b was again
assigned by VT/NOE analysis in d6-DMSO at 373 K (Fig. 1),
and confirmed by single crystal X-ray analysis;‡ this analysis
indicated a conformation in which both the C-2 and C-5 sub-
stituents were pseudoaxial, that C-3 was out of plane of the
remaining ring atoms, the nitrogen was slightly pyramidalised,
the benzoyl carbonyl group was directed towards the C-5 sub-
stituent, and C(6)Me was located under the heterocyclic ring.
The cis-starting material 5b also gave none of the expected cis
product 8b under these conditions, but instead a 25% yield of
trans-7b. The yield of trans-7b could be improved to 88% by
reacting trans-5b with methyl triflate. Again, in these reactions,
recovered starting material had been equilibrated.
References
† PtO2 obtained from BDH proved to be the most reliable; all other
material required longer reaction times.
‡ CCDC reference numbers 209273–209274. See http://www.rsc.org/
suppdata/ob/b3/b303789d/ for crystallographic data in .cif or other
electronic format.
1 R. Millet, J. Domarkas, P. Rombaux, B. Rigo, R. Houssin and
J.-P. Henichart, Tetrahedron Lett., 2002, 34, 5087.
2 M. Angiolini, S. Araneo, L. Belvisi, E. Cesarotti, A. Checchia,
L. Crippa, L. Manzoni and C. Scolastico, Eur. J. Org. Chem., 2000,
2571.
3 S. Hanessian, G. McNaughton-Smith, H.-G. Lombart and
W. D. Lubell, Tetrahedron, 1997, 53, 12789.
4 H. He, B. Shen and G. T. Carter, Tetrahedron Lett., 2000, 41, 2067.
5 H. Watanabe, M. Okue, H. Kobayashi and T. Kitahara, Tetrahedron
Lett., 2002, 43, 861.
6 D. A. DeGoey, H.-J. Chen, W. J. Flosi, D. J. Grampovnik,
C. M. Yeung, L. L. Klein and D. L. Kempf, J. Org. Chem., 2002, 67,
5445.
7 S. Hanessian, M. Bayrakdarian and X. Luo, J. Am. Chem. Soc.,
2002, 124, 4716.
Having compounds 7a,b in hand gave the opportunity to
confirm our earlier hypothesis concerning eliminations in
derivatives 5b, since the acidic malonyl H-6 proton was now
blocked; when 7b was treated with base followed by benzyl
bromide, alkylation at C-2 occurred to give a mixture of cis and
trans adducts 9, but only in very low yield (27%) along with
unreacted starting material (41%); presumably the low yield is
a result of steric hinderance at the C-2 position due to the
proximity of the benzyl function.
8 G. T. Wang, Y. W. Chen, S. Wang, R. Gentles, T. Sowin, W. Kati,
S. Muchmore, V. Giranda, K. Stewart, H. Sham, D. Kempf and
W. G. Laver, J. Med. Chem., 2001, 44, 1192.
9 N. Valls, M. Vallribera, S. Carmelli and J. Bonjoch, Org. Lett., 2003,
5, 447.
10 C. Najera and M. Yus, Tetrahedron: Asymmetry, 1999, 10, 2245.
11 F. J. Sardina and H. Rapoport, Chem. Rev., 1996, 96, 1825.
12 J. H. Bailey, D. Cherry, J. Dyer, M. G. Moloney, M. J. Bamford,
S. Keeling and R. B. Lamont, J. Chem. Soc., Perkin Trans. 1, 2000,
2783.
13 J. Dyer, S. Keeling, A. King and M. G. Moloney, J. Chem. Soc.,
Perkin Trans. 1, 2000, 2793.
14 P. W. H. Chan, I. F. Cottrell and M. G. Moloney, J. Chem Soc.,
Perkin Trans. 1, 2001, 2997.
15 P. W. H. Chan, I. F. Cottrell and M. G. Moloney, J. Chem Soc.,
Perkin Trans. 1, 2001, 3007.
16 R. Goswami and M. G. Moloney, Chem. Commun., 1999, 2333.
17 M. Pichon and B. Figadere, Tetrahedron, 1996, 7, 927.
18 D. Fasseur, B. Rigo, C. Leduc, P. Cauliez and D. Couterier,
J. Heterocycl. Chem., 1992, 29, 1285.
19 T. Nagasaka, A. Tsukada and F. Hamaguchi, Heterocycles, 1986, 24,
2015.
Alternatively, regioselective C-2 modification was also pos-
sible: double deprotonation of trans-5b with excess LDA and
alkylation (MeI) gave monomethyl adduct 10 in good overall
yield (56%). In this compound, the room temperature 1H NMR
spectrum exhibited very broad signals, but at 373 K (d6-DMSO)
a highly resolved spectrum was observed; NOESY analysis
(Fig 1) indicated the expected enhancements around the ring,
but no C-2(Me)/H-5 enhancements, suggesting the trans-
relative stereochemistry. This tentative assignment is corro-
1
borated by the observation that in the H NMR spectrum, the
20 I. Collado, J. Ezquerra and C. Pedregal, J. Org. Chem., 1995, 60,
OCH2CH3 signals were not coincident, and that H-6 did not lie
under these signals; this arrangement has also been observed
for other trans isomers as described above. When this procedure
was applied to cis-5b, a low yield (25%) of ent-10 was obtained,
since the material from this reaction exhibited opposite optical
rotation to product 10 from the reaction with trans-5b sug-
gesting that the trans-C(2)Me–C(5)H stereochemistry is the
thermodynamically more stable arrangement.
We have therefore established reliable conditions suitable
for the reduction of enamines derived from pyroglutamic acid
substituted with ester functions, and this has generated syn-
thetic intermediates which permit regioselective C-2 or C-6
5011.
21 J. Mulzer, F. Schulzchen and J.-W. Bats, Tetrahedron, 2000, 56, 4289.
22 I. Collado, J. Ezquerra, J. J. Vaquero and C. Pedregal, Tetrahedron
Lett., 1994, 35, 8037.
23 H. Li, T. Sakamoto and Y. Kikugawa, Tetrahedron Lett., 1997, 38,
6677.
24 A. J. Mota and N. Langlois, Tetrahedron Lett., 2003, 44, 1141.
25 J. Ezquerra, A. Rubio, C. Pedregal, G. Sanz, J. H. Rodriguez and
J. L. Garcia Ruano, Tetrahedron Lett., 1993, 34, 4989.
26 K. Shiosaki, in Comprehensive Organic Synthesis, ed. B. M. Trost,
vol. 2, Oxford, 1991.
27 R. Lin, J. Castells and H. Rapoport, J. Org. Chem., 1998, 63, 4069.
28 M. C. Elliott and M. S. Long, Tetrahedron Lett., 2002, 43, 9191.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 8 3 8 – 1 8 4 1
1840

View Online
29 T. Honda and M. Kimura, Org. Lett., 2000, 2, 3925.
30 G. V. Thanh, J.-P. Celerier, A. Fleurant, C. Grandjean, S. Rosset and
G. Lhommet, Heterocycles, 1996, 43, 1381.
31 M. D. Bachi, R. Breiman and H. Meshulam, J. Org. Chem., 1983,
48, 1439.
32 All new compounds gave satisfactory spectroscopic and high reso-
lution mass spectrometric or analytical data.
genation apparatus was flushed 3 times with hydrogen, and finally
set to a pressure of 4.5 atm and the mixture stirred for 48 h. Careful
filtration through Celite® using EtOAc, followed by solvent
removal, gave a residue which was redissolved in DCM (50 ml) and
washed with dilute aq. ammonia (3.2%, 50 ml), water and brine. The
solution was dried over MgSO4 and the solvent removed to give the
product as a clear oil.
33 A. B. Smith, P. A. levenberg and J. Z. Suits, Synthesis, 1986,
184.
34 J. S. Petersen, G. Fels and H. Rapoport, J. Am. Chem. Soc., 1984,
106, 4539.
35 The enamine (1 g, 3.3 mol) was dissolved in HOAc (7.5 ml) to which
was added Adam’s catalyst (0.164 g) in TFA (2.5 ml). The hydro-
36 A. V. R. Rao, M. K. Gurjar and J. Vasudevan, J. Chem. Soc.,
Chem. Commun., 1995, 1369.
37 E. Vazquez, A. Galindo, D. Gnecco, S. Bernes, J. L. Teran and
R. G. Enriquez, Tetrahedron: Asymmetry, 2001, 12, 3209.
38 D. A. Fletcher, R. F. McMeeking and D. Parkin, J. Chem. Inf. Com-
put. Sci., 1996, 36, 746.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 8 3 8 – 1 8 4 1
1841