Diastereoselectivity in the alkylation of 4-fluoroproline methyl esters

Article abstract of DOI:10.1016/j.tetlet.2006.10.057

The reaction of alkylation of cis- and trans-4-fluoro-N-Boc-l-proline methyl esters has been examined by exposing their lithium enolates to a range of alkylating agents. The process showed a high degree of facial diastereoselectivity (except when methyl iodide was used as alkylating agent), invariably giving rise to products bearing the alkyl group in anti with respect to the fluorine atom. A tentative model to account for the observed stereoselectivity is also proposed.

Full text of DOI:10.1016/j.tetlet.2006.10.057

Tetrahedron Letters 47 (2006) 8929–8932
Diastereoselectivity in the alkylation of 4-fluoroproline
methyl esters
Rosanna Filosa,* Claude Holder and Yves P. Auberson
Novartis Institutes for BioMedical Research, Klybeckstrasse 141, 4002 Basel, Switzerland
Received 26 April 2006; revised 6 October 2006; accepted 9 October 2006
Available online 30 October 2006
Abstract—The reaction of alkylation of cis- and trans-4-fluoro-N-Boc-L-proline methyl esters has been examined by exposing their
lithium enolates to a range of alkylating agents. The process showed a high degree of facial diastereoselectivity (except when methyl
iodide was used as alkylating agent), invariably giving rise to products bearing the alkyl group in anti with respect to the fluorine
atom. A tentative model to account for the observed stereoselectivity is also proposed.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
In their hands, alkylation with benzylic iodide preferen-
tially gave products substituted at the b face (3b),
whereas allylic or aliphatic halides, furnished the corres-
ponding a-epimers (3a) as major products. They inter-
preted their stereochemical results in terms of a
balance between two opposite effects: (a) the stereoelec-
tronic interaction of N-lone pair with the forming C–C
bond orbital (anomeric effect), independent of the type
of alkylating agent, favouring substitution at the a-face,
and (b) the interaction of the N-lone pair with p-orbitals
of benzyl or allyl halide (n–p interaction), leading to the
preferential formation of b-epimers.
The stereo-selective alkylation of N-Boc-L-proline esters
is recognized as a synthetically important process in the
construction of a variety of natural products1 as well as
of conformationally restricted building blocks for the
synthesis of peptidomimetics.2 Over the last years, much
effort has been expended in attempting to rationalize the
stereochemical outcome of the alkylation of heterocyclic
enolates such as those produced by 4-silyloxy-N-Boc-L-
proline methyl esters.3
Sato et al.3c synthesized a series of 2-alkyl proline deri-
vatives with the aim of shedding more light on the
factors controlling the diastereoselectivity in this type
of reaction (Fig. 1). They report that the steric and/or
electronic factors of the N-Boc group would be more
crucial rather than that of the silyloxy or ester moieties
in the attack on the electrophilic carbon centre. Thus,
the preferential formation of 3a might be due to a shield-
ing of the syn-face of the enolate by the N-Boc group.
In light of the preceding results, we embarked on a pro-
gramme designed to readily examine the stereochemical
behaviour of the corresponding 4-fluoro-N-Boc-L-pro-
line methyl ester derivatives via classical alkylation con-
ditions. More particularly, we demonstrated the strong
influence of the fluorine atom in directing the reactivity
of these heterocyclic lithium enolates, leading to a par-
ticularly high degree of facial diastereoselectivity.
On the other hand, quite recently Kawahara and
co-workers3d found that the diastereoselectivity of the
alkylation of 4-silyloxy-N-Boc-L-proline methyl esters
is strongly dependent on the alkylating reagents.
In this letter, we describe the reaction of alkylation of
cis- and trans-4-fluoro-N-Boc-L-proline methyl esters,
by exposing their lithium enolates to a range of alkylat-
ing agents, leading to a series of alkylated N-Boc-L-pro-
lines (13–16).
Keywords: Diastereoselectivity; Proline derivatives; Fluorinated orga-
nic compounds; Alkylation.
Our synthesis (Scheme 1) commenced with the readily
available 2-allyl-4-silyloxy-N-Boc-L-proline methyl esters
5 and 6 which were converted to the corresponding
fluorinated L-proline derivatives 9 and 10 whose absolute
configuration at C-2 was formerly assigned.3c
*
Corresponding author at present address: Dipartimento di Scienze
`
Farmaceutiche, Universita degli Studi di Salerno, Via Ponte don
Melillo, 84084 Fisciano, SA, Italy. Tel.: +39 89962822; fax: +39
89962828; e-mail: rfilosa@unisa.it
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.10.057

8930
R. Filosa et al. / Tetrahedron Letters 47 (2006) 8929–8932
TBSO
TBSO
TBSO
TBSO
R1
OR
OLi
R1
LiHMDS, THF
-78 ˚C
Alkyl halide
CO2R
CO2R
N
N
N
CO2R
N
+
Boc
Boc
Boc
Boc
1
2
3a
3b
X
OMe
OLi
RO
N
Boc
OMe
F
X
N
Boc
OLi
A
B
Figure 1. Alkylation of 4-siloxy-N-Boc-L-proline methyl esters, and transition state’s hypothesis.
TBSO
TBSO
TBSO
i, ii
CO2Me
+
N
CO2Me
N
CO2Me
N
Boc
Boc
6
Boc
5
4
iii
iii
HO
HO
CO2Me
N
CO2Me
N
Boc
Boc
7
8
iv
iv
F
F
CO2Me
N
CO2Me
N
Boc
Boc
9
10
Scheme 1. Reagents and conditions: (i) LiHMDS, THF, ꢁ78 °C, 30 min; (ii) allyl bromide; (iii) TBAF, THF, rt; (iv) DAST, CH2Cl2, rt.
Each diastereoisomer (5 and 6) previously separated by
flash silica gel chromatography, was deprotected with
TBAF in THF to give the corresponding derivatives 7
and 8. Thus, exposure to diethylamino sulfur trifluoride
(DAST), as fluorinating agent, under mild conditions
(rt, under CH2Cl2, 8 h), of 7 and 8 yielded the desired
nuclear chemical shift correlation (ROESY) spectra that
show a crosspeak between CaH4 5.22, (m, 1H, H4) and
aH 2.55 (1H, CaH, br dd J = 14.3 and 7.5) of the meth-
ylene group of the allyl substituent.
Prompted by these reports, and by general focus on eval-
uating the experimental influence of fluorine in organic
chemistry, it appeared appropriate to also examine the
alkylation of also b-4-silyloxyproline enolates.
compounds
(2S,4S)-2-allyl-4-fluoro-N-Boc-L-proline
20
methyl ester 9 (½aꢀD +42 (c 0.55, CHCl3)) and (2R,4S)-
20
2-allyl-4-fluoro-N-Boc-L-proline methyl ester 10 (½aꢀD
ꢁ30 (c 0.7, CHCl3)), respectively. Their configuration
was assigned on the basis of mechanistic arguments,
namely considering that hydroxyl/fluorine substitution
promoted by DAST is a typical SN2 reaction,4 and con-
firmed for derivative 9 by two-dimensional (2D) homo-
Comparison with the results of the alkylation with allyl
bromide which gave 47% of (2S,4S)-2-allyl-product and
53% of (2R,4S)-2-allyl-product, confirms the lack of dia-
stereoselectivity with this substrate.

R. Filosa et al. / Tetrahedron Letters 47 (2006) 8929–8932
F
8931
HO
F
F
iii, iv
i, ii
CO2Me
CO2Me
N
N
H
CO2Me
N
CO2Me
N
+
Boc
Boc
11
Boc
13b
13a
HO
F
F
F
iii, iv
i, ii
CO2Me
CO2Me
N
CO2Me
N
H
N
CO2Me
N
+
Boc
Boc
Boc
12
14a
14b
RX
Products
Yield
Ratio
Substrate
11
12
CH2=CHCH2Br
CH2=CHCH2Br
13a, 13b
14a, 14b
93%
95%
3:97
96:4
Scheme 2. Reagents and conditions: (i) Et3N, (Boc)2O; (ii) DAST, CH2Cl2, rt; (iii) LiHMDS, THF, ꢁ78 °C, 30 min; (iv) allyl bromide. Yields and
diastereomeric ratios observed in the allylation of 11 and 12.
F
In the second step of our investigation, the allylation of
(4R)-4-fluoro-N-Boc-L-proline methyl ester 11 and (4S)-
4-fluoro-N-Boc-L-proline methyl ester 12 was examined
(Scheme 2).
F
F
R
R
i, ii
CO2Me
CO2Me
CO2Me
N
N
N
+
Boc
Boc
Boc
11
15/16a
15/16b
Hence, treatment of 11 and 12 with allyl bromide and
1.25 M equiv of lithium bis-(trimethylsilyl)amide in
THF at ꢁ78 °C, directly gave the enantiopure alkylation
RX
Products
Yield
Ratio
Substrate
11
11
PhCH2Br 15a, 15b
CH3I 16a, 16b
97%
72%
98:2
20
20
products 13b (½aꢀD ꢁ46 (c 0.84, CHCl3)) and 14a (½aꢀD
+45 (c 0.865, CHCl3)) in high stereoselectivity ratio
(3:97 and 96:4), respectively (Scheme 2).
53:47
Scheme 3. Reagents and conditions: (i) LiHMDS, THF, ꢁ78 °C,
30 min; (ii) alkyl bromide. Yields and diastereomeric ratios observed in
the benzylation of 11.
The diastereomeric ratios were determined by HPLC
analysis and independently confirmed by analysis of
the relative intensities of appropriate signals in their
etc.) controlling the process, they are likely to be very
small in magnitude in this particular case.
1H NMR spectra, while the configuration of 14a and
20
13b was deduced by comparison of ½aꢀD values with
Taken together, our results indicate that the reaction of
enolates produced from 4-fluoroproline methyl ester
does not proceed in the same way as those coming from
O-protected 4-hydroxyprolines, as shown by Nagumo
et al.3d (Fig. 1).
those of an authentic sample of 9, synthesized starting
from D-proline (Scheme 1).
Intrigued by these results, we next examined the alkyl-
ation of compound 11 with benzyl and methyl halides.
It is apparent that the particularly relevant feature of
this reaction is its strong dependence on the final orien-
tation of the incoming alkyl group upon the initial posi-
tion (a or b) of the fluorine atom in the five-membered
Pro ring system. We wish to point out that this aspect
seems to be a somewhat distinctive attribute of 4-flu-
oro-Pro which does not appear to be shared by other
similar systems. Moreover, the influence of fluorine
atoms (positioned in 3 with respect to carbonyl group)
on the final outcome of enolate alkylation reactions
has very recently been examined, in light of the possibil-
ity that the fluorine may participate in the metal (lith-
ium) coordination, therefore shielding one face of the
enolate.5
The outcome of these experiments highlights, again the
substantially different behaviour of 4-fluoro-prolines,
under Ca alkylation conditions, as opposed to those
with a 4-hydroxy-substituent. In both aforementioned
cases, diastereomeric ratios were determined by HPLC
analysis and by measuring the relative intensities of
1
signals in their H NMR spectra (Scheme 3).
In practice, alkylation with benzyl halides proceeds
much along the same line as that with allyl halides dis-
playing a high yield and almost quantitative diastereo-
facial discrimination and basically furnishing compound
15a as the sole product. Conversely, the same reaction
with methyl halide led to a ca. 1:1 mixture of the two
diastereoisomers. Actually, the latter results were not
fully unexpected. In fact, whatever are the stereoelec-
tronic factors (steric hindrance, orbital interactions,
On the other hand, meaningful comparisons with other
Pro systems are not always straightforward, as—for

8932
R. Filosa et al. / Tetrahedron Letters 47 (2006) 8929–8932
On the other hand, in the case of derivative 12, the fluo-
rine is in a pseudo-axial orientation, and we can argue
that fluorine may chelate lithium in a sort of bicycle
[3.3.0] TS, as shown in the model (Fig. 2) in analogy
with what was recently found in acyclic systems.5
In this chelate species the nitrogen becomes a tertiary
centre and its steric influence is clearly significant and
will certainly contribute to the efficiency of diastereo-
selectivity of the reactions.
In conclusion, we found that alkylation of 4-fluoro-N-
Boc-L-Pro methyl esters proceeds in high yield and dia-
stereoselectivity. The influence of the fluorine atom on
the final outcome of these reactions appears to be quite
relevant. Except when MeI is used, our data suggest that
the major product is a compound in which the incoming
alkyl group is invariably oriented from the other side of
the fluorine atom. Very likely, the factors governing
such process depend on the possibility to form a chelate
in which there is an active participation of fluorine.
Acknowledgments
We thank Professor Giuseppe Bifulco, Pharmaceutical
Science Department of Salerno, for helpful discussion.
Figure 2. (a) Representation of the postulated transition states formed
in the course of alkylation of O-protected 4-hydroxy-N-Boc-proline
ester enolates. (b) Stereo view of the enolate highlighting the statement
of the shielding of the syn face by the Boc.
References and notes
1. (a) Coppola, G. M.; Schuster, H. F. Asymmetric Synthesis;
Wiley: New York, 1987; (b) Confalone, P. N.; Huie, E. M.;
Cole, G. M. J. Org. Chem. 1988, 53, 482; (c) Ikeda, M.;
Matsubayashi, K.; Imoto, T.; Kitao, K.; Ishibashi, H.;
Sato, T. Heterocycles 1994, 38, 1237.
2. (a) Thairivongs, D. T.; Pals, J. A.; Lawson, S. R.; Harris,
D. W. J. Med. Chem. 1987, 30, 536; (b) Khalil, E. M.;
Ojala, W. H.; Pradhan, A.; Nair, V. D.; Gleason, W. B.;
Mishara, R. K.; Johnson, R. L. J. Med. Chem. 1999, 42,
626.
3. (a) Seebach, D.; Naef, R. Helv. Chim. Acta 1981, 64, 2704;
(b) Noe, R.; Knollmuller, H.; Vollenkle, N. Pharmazie
1996, 51, 800; (c) Sato, T.; Kugo, Y.; Nakaumi, E.;
Ishibashi, H.; Ikeda, M. J. Chem. Soc., Perkin Trans. 1
1995, 1801; (d) Nagumo, S.; Mizukami, M.; Akutsu, N.;
Nishida, A.; Kawahara, N. Tetrahedron Lett. 1999, 40,
3209.
4. Doi, M.; Nishi, Y.; Kiritoshi, N.; Iwata, T.; Nago, M.;
Nakano, H.; Uchiyama, S.; Nakazawa, T.; Wakamiya, T.;
Kobayashi, Y. Tetrahedron 2002, 58, 8453.
instance—the literature data refer to reactions effected
only on a a-oriented 4-OH group. For this reason, in
the attempt of rationalizing the outcome of this series
of 4-fluoro-N-Boc-L-Pro alkylation reaction, we prefer
to discuss separately from each other the case of the
two 4-fluoro epimers 11 and 12. A preliminary consider-
ation is that, since the presence of the fluorine atom at
C4 is likely to induce in compounds 11 and 12 a decrease
in the pKa of the Ha, the corresponding enolates should
be considered more stabilized species than other similar
heterocyclic enolates.
Thus, in the case of 4a-fluoro-substituted 11, the transi-
tion states (TS) may benefit of further stabilization by
interaction of the n-lone pair orbital of N-Boc moiety
with the p-system of the incoming alkylating reagent,
and we can assume that the n–p interaction becomes
the predominant factor determining the diastereofacial
differentiation, leading to a lower energy TS for the
alkylation of the b-face affording 13b.
5. Yamazaki, T.; Kitazume, M. T.; Kubota, T.; Omura, M.
Org. Lett. 1999, 1, 905.