Observation of a new dimeric amino acid derivative in the reaction of methyl N-BOC-(S)-(3-fluorophenyl)alanate with DIBAL-H and lithio ethyl propiolate

Article abstract of DOI:10.1016/S0957-4166(01)00054-4

A novel amino acid dimer 9a was isolated while trying to apply a known reduction/nucleophilic addition sequence. This dimer provided information both on the mechanism of the process and on the dependence of the desired reaction on the stoichiometry of rea

Full text of DOI:10.1016/S0957-4166(01)00054-4

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 12 (2001) 361–363
Observation of a new dimeric amino acid derivative in the
reaction of methyl N-BOC-(S)-(3-fluorophenyl)alanate with
DIBAL-H and lithio ethyl propiolate
V. John Jasys and Frank J. Urban*
Pfizer Global Research and Development, Chemical Research and Development, Groton, CT 06340, USA
Received 19 December 2000; accepted 17 January 2001
Abstract—A novel amino acid dimer 9a was isolated while trying to apply a known reduction/nucleophilic addition sequence. This
dimer provided information both on the mechanism of the process and on the dependence of the desired reaction on the
stoichiometry of reagents. © 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
Recently, we were examining the multikilogram prepa-
ration of (S,S)-lactone 1a. The des-fluorinated analogue
1b and various other homologues have been prepared
by a number of different syntheses.1 After reviewing
these, it was decided to apply a variation of the route
described by Kleinman involving addition of lithio
ethyl propiolate 4 to an amino acid aldehyde.2 The
possibility of racemisation of the isolated BOC-3-
fluorophenylalanine-derived aldehyde and the moderate
selectivities found in the addition of lithio ethyl propio-
late 4 to this aldehyde caused us to explore a one-pot
procedure shown in Scheme 1.3 In this reaction, methyl
BOC-(3-fluorophenyl)alanate 2a was reduced with
DIBAL-H to aluminoxy acetal 3a followed by addition
of lithio ethyl propiolate 4. In the course of this work,
a novel side reaction was found that was dependent on
the stoichiometry of the lithio ethyl propiolate and
resulted in the isolation of the novel diol dimer 9a
structurally related to building blocks of HIV protease
inhibitors.4
2. Results and discussion
In the one-pot reduction/organometallic addition pro-
cess described in the literature, a slight excess of
DIBAL-H was added to the ester. It was found that
adding DIBAL-H (1.2–1.5 equivalents) to ester 2a in
THF at −78°C, followed by quenching with acetic acid
always resulted in a mixture of aldehyde, starting ester,
and alcohol from over reduction. The use of 2.1 equiv-
alents of reductant completely removed the starting
ester, yet afforded substantial amounts of alcohol (up
to 20%). The optimal conversion was obtained by
reversing the order of addition; adding ester 2a to
DIBAL-H in THF at −78°C gave complete consump-
Scheme 1.
* Corresponding author. E-mail: frank j urban@groton.pfizer.com
– –
0957-4166/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(01)00054-4

362
V. J. Jasys, F. J. Urban / Tetrahedron: Asymmetry 12 (2001) 361–363
The one-pot, DIBAL-H/lithium propiolate procedure
had been performed at Pfizer with phenylalanine ana-
logue 2b and had produced alcohol 5b in moderate
yield with no mention of major side products. In our
hands, the same DIBAL-H/lithium propiolate proce-
dure was repeated with BOC-phenylalanine methyl
ester 2b, and the corresponding diol 9b was isolated by
chromatography of the crude reaction mixture. There-
fore, the problems of reproducing the process were
unrelated to the presence of a fluorine atom.
To optimise the desired process, the impact of the
stoichiometry of 4 was evaluated. One equivalent of
propiolate 4 gave little or no 5a or 9a. Two equivalents
of propiolate 4 gave a reaction mixture which displayed
no vinyl resonances in the 1H NMR spectrum, and
afforded an HPLC assay consistent with a 6.5:1 mixture
of (S,S)-5a and its (R)-hydroxyl diastereomer. The
reaction was repeated on an 18 mmol scale and the
crude 5a was hydrogenated and cyclised to lactone 1a.
After column chromatography, a mixture representing
a 38% yield of the desired (S,S)-1a lactone with 10% of
the (S,R)-diastereomer at the oxygenated centre was
isolated. This procedure was also carried out on >10
kilogram scale of 2a with identical results.
Figure 1. ChemDraw 3D drawing of the single crystal X-ray
structure of diol 9a.
tion of ester and <10% alcohol. During DIBAL-H
addition to the ester 2a, free aldehyde is released from
3a and undergoes further reduction to alcohol.
The reduction of 2a to aluminoxy acetal 3a was carried
out with DIBAL-H at −78°C and the cold solution was
combined with 1.5 equivalents of 4 at a temperature
below −60°C. While the reaction of lithium propiolate 4
with free aldehyde occurs at −78°C, addition of 4
affords product 5 only at >−10°C indicating the
absence of aldehyde.5 The crude reaction mixture
looked fairly clean by TLC, but was complex by NMR.
A crystalline solid having the same Rf on TLC as the
desired 5a was isolated in up to 20% yield by treating
the crude oil with isopropyl ether. This was found to be
the novel dimer 9a by single crystal X-ray analysis6
(Fig. 1). While the dimerisation of aldehydes or amino
acid derivatives in the presence of transition metals is
known, the absence of any transition metal and the
unprecedented unsaturation in one of the amino alco-
hol moieties made 9a an unusual product.
A mechanism for the formation of dimer 9a is sug-
gested in Scheme 2. The key point is that the initial
reduction product 3a was not fully deprotonated at the
BOC-NH, as described above. With 1.5 equivalents of
lithio propiolate 4, the deprotonation and formation of
a cyclic aluminate complex 6a is completed with 4 as a
base. Upon warming the reaction, methoxide is lost to
give the oxonium species 7a, and reaction with lithio
propiolate provides the desired alcohol 5a (pathway i).
If intermediate 7a is not trapped by a nucleophile such
as 4, a second deprotonation from the a-carbon centre
of 7a provides a nucleophilic species 8a that reacts with
oxonium 7a (pathway ii). Another possibility is that the
methoxy group is displaced by lithio propiolate directly
in pathway i or is eliminated as methoxide to generate
intermediate 8a. In literature examples of this
Scheme 2.

V. J. Jasys, F. J. Urban / Tetrahedron: Asymmetry 12 (2001) 361–363
363
reaction, a large excess of the organometallic reagent,
typically 3 equivalents, were used. The cyclic aluminate
complex provides a rationale for the relationship
between the stoichiometry of lithio propiolate 4 and the
improved diastereomeric ratio of alcohol 5a over its
diastereomer as compared with direct addition to the
free aldehyde. This also explains the higher reaction
temperature and the origin of the double bond in 9a.
Pinacol 9a was isolated as a single enantiomer.
Chem. 1993, 36, 320–330; (c) Kammermeier, B.; Beck, G.;
Holla, W.; Jacobi, D.; Napierski, B.; Jendralla, H. Chem.
Eur. J. 1996, 2, 307–315.
5. When isolated aldehyde was treated with excess lithio ethyl
propiolate 4 at −78°C, reaction occurred without warming
and the desired 5a and its diastereomer were isolated in a
3:1 ratio and 60% yield.
6. trans-2-(5S)-Bis-(tert-butoxycarbonylamino)-1,6-bis-(3-
fluorophenyl)-(3S,4R)-dihydroxy-hex-1-ene 9a: Methyl N-
BOC-(S)-(3-fluorophenyl)alaninate 2a (10.8 g, 36.3 mmol)
in dry toluene (50 mL) was cooled to −78°C under nitro-
gen and treated with DIBAL-H (1.5 M in toluene, 49.6
mL, 74.4 mmol) dropwise over 40 min and the mixture
was stirred at low temperature for 0.5 h. In a separate
flask, ethyl propiolate (5.5 mL, 54.3 mmol) was added
dropwise at −78°C to lithium hexamethyldisilazane in tet-
rahydrofuran (1 M solution, 54.3 mL, 54.3 mmol) over 20
min. The toluene solution was transferred to the lithio
propiolate solution by means of a cannula over 2 min and
the reaction was allowed to warm to room temperature
and stirred overnight. The reaction was recooled to −10°C
and acetic acid (13 mL) was added dropwise over 20 min.
The reaction was poured into a mixture of an equal
volume (100 mL each) of ethyl acetate and aqueous citric
acid (17 g) and this was stirred for 45 min to solubilise the
aluminium salts. The organic layer was separated and
washed with water (two washes) and brine. The solvent
was evaporated in vacuo and the crude oil was stirred with
isopropyl ether to provide a first crop of diol 9a (0.9 g).
Further material was recovered after chromatography over
silica gel with 10% ethyl acetate in methylene chloride. Mp
191–193°C (dec.). [h]D −40.6 (c=0.88, MeOH). NMR
(CDCl3) l 7.27–6.85 (m, 8), 6.85 (s, 1), 5.97 (bs, 1), 4.90 (d,
1), 4.43 (d, 1), 4.15 (m, 1), 3.93 (m, 1), 3.65 (bs, 1), 3.32 (m,
1), 2.95 (d, 2), 1.43 (s, 9), 1.32 (s, 9). Mass spectrum: m/e
534 (M+). Anal. calcd for C28H36F2N2O6: C, 62.91; H,
6.79; F, 7.11; N, 5.24. Found: C, 62.70; H, 6.72; F, 7.28;
N, 5.16%.
3. Conclusion
Altering the stoichiometry of the reagents in a known
synthetic procedure resulted in the formation of a novel
dimer 9a and provided some insight into the mechanism
of the transformation.7
Acknowledgements
We would like to thank Professor David Collum for
helpful discussions and Dr. Jon Bordner for providing
the single crystal X-ray of 9a.
References
1. Yokomatsu, T.; Yuasa, Y.; Shibuya, S. Heterocycles 1992,
33, 1051–1078.
2. Fray, A. H.; Kaye, R. L.; Kleinman, E. F. J. Org. Chem.
1986, 56, 4828–4833.
3. Kano, S.; Yuasa, Y.; Yokomatsu, T.; Shibuya, S. Chem.
Pharm. Bull. 1989, 37, 2867–2869.
4. (a) Kempf, D. J.; Sowin, T. J.; Doherty, E. M.; Hannick,
S. M.; Codavoci, L.; Henry, R. F.; Green, B. E.; Spanton,
S. G.; Norbeck, D. W. J. Org. Chem. 1992, 57, 5692–5700;
(b) Kempf, D. J.; Codacovi, L.; Wang, X. C.; Kohlbren-
ner, W. E.; Wideburg, N. E.; Saldivar, A.; Vasavanonda,
S.; Marsh., K. C.; Bryant, P.; Sham, H. L.; Green, B. E.;
Betebenner, D. A.; Erickson, J.; Norbeck, D. W. J. Med.
7. For an additional discussion of this type of process see:
Polt, R.; Peterson, M. A.; DeYoung, L. J. Org. Chem.
1992, 57, 5469–5480 and references cited therein.
.