Enantioselective synthesis of 2-substituted-N-Boc-Δ-4,5-piperidines

Article abstract of DOI:10.1016/S0957-4166(00)00462-6

The title compounds were prepared through a synthetic sequence involving: (i) a reaction of the condensation product between an enantiopure β-amino alcohol and an aldehyde with allylmagnesium chloride; (ii) an N-allylation of the resulting secondary amine; (iii) a chemoselective cleavage of the β-amino alcohol residue; and (iv) a protection of the secondary amine followed by a ring closing metathesis. The advantageous use of (1R,2S)-norephedrine was demonstrated in these syntheses.

Full text of DOI:10.1016/S0957-4166(00)00462-6

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 11 (2000) 4639–4643
Enantioselective synthesis of
2-substituted-N-Boc-D-4,5-piperidines
Claude Agami, Franc¸ois Couty* and Gwilherm Evano
Laboratoire de Synthe`se Asyme´trique (UMR CNRS 7611), Universite´ Pierre et Marie Curie, 4 place Jussieu,
75005 Paris, France
Received 24 October 2000; accepted 17 November 2000
Abstract
The title compounds were prepared through a synthetic sequence involving: (i) a reaction of the
condensation product between an enantiopure b-amino alcohol and an aldehyde with allylmagnesium
chloride; (ii) an N-allylation of the resulting secondary amine; (iii) a chemoselective cleavage of the
b-amino alcohol residue; and (iv) a protection of the secondary amine followed by a ring closing
metathesis. The advantageous use of (1R,2S)-norephedrine was demonstrated in these syntheses. © 2001
Elsevier Science Ltd. All rights reserved.
Piperidine-derived heterocycles are ubiquitous substructures in natural molecules and a huge
amount of work has been devoted towards the asymmetric synthesis of diversely substituted
piperidinic compounds.1 However, to our knowledge, a general and reliable method for the
synthesis of enantioenriched 2-substituted-N-Boc-D-4,5-piperidines such as 1 still remains to be
found.2
Our interest in this unsaturated piperidinic core especially lies in the possibility of further
functionalization of the C-4 or C-6 position of the heterocycle using Beak’s methodology.3 Our
synthetic plan to reach these target molecules is depicted in Scheme 1.
* Corresponding author. Fax: 33 1 44 27 26 20; e-mail: couty@ccr.jussieu.fr
0957-4166/00/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(00)00462-6

4640
C. Agami et al. / Tetrahedron: Asymmetry 11 (2000) 4639–4643
Scheme 1.
Reaction of an allyl Grignard reagent with the reaction product resulting from the condensa-
tion between an enantiopure b-amino alcohol 2 and an aldehyde accounts for the diastereoselec-
tive formation of the homoallylic amino alcohol 3. N-Allylation, followed by a chemoselective
N-dealkylation of the amino alcohol residue then gives diolefinic amine 4. The last steps of this
method consist in an N-protection by a Boc group followed by ring closing metathesis (RCM).
We report herein our preliminary results of this synthetic scheme and delineate the scope.
As regards the first step of this synthesis, addition of Grignard reagents,4 as well as lithium5
or cerium6 organometallics onto the reaction product7 resulting from the condensation between
an aldehyde and an enantiopure b-amino alcohol is a well-established methodology for the
preparation of enantioenriched a-substituted amines. This simple reaction of predictable stereo-
chemical outcome was therefore chosen to set the stereocenter in our target molecules. Two
enantiopure b-amino alcohols, namely (1R,2S)-norephedrine (2: R1=Me, R2=Ph) and (S)-
phenylglycinol (2: R1=Ph, R2=H) were used as chiral sources and were condensed with various
aldehydes. Allylation was effected on the crude condensation product using 3 equiv. of
allylmagnesium chloride in THF. The results are gathered in Table 1.
Table 1
Entry
Amino alcohol
R3
Product
Yielda
Deb
\80
20
76
\95
\95
54
1
2
3
4
5
6
7
8
9
(1R,2S)-Norephedrine
–
–
Ph
n-C9H20
n-C3H7
(E)-CHꢀCHPh
Ph
5
6
7
8
9
10
11
12
13
75
90
91
79
58
50
65
57
67
(S)-Phenylglycinol
–
–
–
–
n-C9H20
(E)-CHꢀCHPh
(E)-CHꢀCHC6H13
i-Pr
81
\95
67
a Yield of isolated product.
1
b Determined by H NMR. For all compounds, the diastereoisomeric excesses were ca. the same before and after
purification by chromatography. In the case of compound 5, the de was difficult to determine due to the presence of
several minor impurities (GC purity of 5: 95%).

C. Agami et al. / Tetrahedron: Asymmetry 11 (2000) 4639–4643
4641
The above results deserve some comments. First, the diastereoisomeric excesses obtained in
these reactions dramatically depend upon the nature of the starting aldehyde. No minor
stereoisomer, except in the case of the previously reported compound 11,8 could be detected in
the reaction product when enals were used (entries 4, 7 and 8). Furthermore, in these cases, no
competitive 1,4- versus 1,2-addition was detectable. Very good de’s were also reached when
benzaldehyde was used, but in the case of compound 5 (entry 1) accurate determination of the
de was difficult (see Table 1). On the other hand, the diastereoselectivity dropped considerably
when aliphatic aldehydes, and especially long chain aldehydes, were used (entries 2, 3, 6 and 9).
As regards the starting b-amino alcohol, the use of norephedrine instead of phenylglycinol
was beneficial when cinnamaldehyde was used as the starting aldehyde. Indeed, compound 8 was
obtained with better yield and de than 11 (compare entries 4 and 7).
Having in hand some allylic amino alcohols, the next step of the synthesis was studied. It
consisted of an N-allylation of the secondary a,a%-disubstituted amine. The best conditions to
achieve this reaction were reacting the amine with 4 equiv. of allyl bromide in refluxing
acetonitrile for 12–24 h in the presence of K2CO3 and of a stoichiometric amount of tetra-
butylammonium iodide (Scheme 2).
Scheme 2.
The above results clearly show that the success of this reaction depends on the steric crowding
around the amine. Much better yields were obtained with norephedrine-derived amino alcohols
(R1=Me) than with phenylglycinol-derived substrates (R1=Ph): compare the yields for 14 and
18 or 17 and 20. This again highlights the advantageous use of norephedrine rather than
phenylglycinol in this synthesis.
Chemoselective N-dealkylation was then effected on substrates 14, 16, 17 (derived from
norephedrine) and 21 (derived from phenylglycinol). This procedure9 consists of treatment of the
b-amino alcohol with thionyl chloride in THF, followed by reaction with an excess of KCN in
a mixture of DMSO/THF for 72 h. Following this two-step dealkylation procedure, diolefinic
amines 23–26 were obtained with modest to fair yields, as shown below; these amines were then
protected in quantitative yields as an N-Boc group by treatment with (Boc)2O in refluxing ethyl
acetate to give compounds 27–30 (Fig. 1).

4642
C. Agami et al. / Tetrahedron: Asymmetry 11 (2000) 4639–4643
Figure 1.
Finally, these N-Boc amines were subjected to ring closing metathesis in refluxing
dichloromethane (1 h) in the presence of the Grubbs catalyst (3% molar ratio). Under these
conditions, compounds 27 and 28 gave the expected cyclized piperidines 31 and 32 in respective
yields of 98 and 73%. On the other hand, using substrates 29 and 30, a problem of chemoselec-
tivity arises from the presence of an olefinic moiety in R3. As a matter of fact, piperidine 34 was
isolated pure from minor by-products in 57% yield when 30 was used in this reaction, but
piperidine 33 resulting from the metathesis of 29 could not be obtained free of pyrrolidine 35,10
resulting from a metathesis reaction between the styrenic substituent and the N-allyl group
(ratio 33/35: 88/12, yield 82%).
Since the N-dealkylation procedure used above was proved not to alter the stereochemical
integrity of the released amine,11 the enantiomeric excesses of the target compounds correspond
to the diastereoisomeric ratio of their allylated precursor and are 80,12 >95 and >95% for 32, 33,
and 34, respectively. Since some imprecision persisted in the determination of the enantiomeric
excess of 31 (vide supra), racemic 31 was prepared as depicted in Scheme 3.
Scheme 3.
Chiral GC analysis (Supelco b-DEX-120) of N-Boc deprotected (TFA, CH2Cl2, 0°C, 1 h)
racemic and enantioenriched 31 then allowed accurate determination of its ee that was found to
be 87%.
In conclusion, this work delineates the scope of an enantioselective synthesis of 2-substituted-
N-Boc-D-4,5-piperidines in which the advantageous use of norephedrine as chiral inductor was
demonstrated. Further studies towards the subsequent functionalization of these unsaturated
piperidines are in progress.
Acknowledgements
Marie Roger and Minh-Thu Dinh are thanked for their technical assistance.

C. Agami et al. / Tetrahedron: Asymmetry 11 (2000) 4639–4643
4643
References
1. For a review, see: (a) Bailey, P. D.; Millwood, P. A.; Smith, P. D. Chem. Commun. 1998, 633–640 and references
cited therein; (b) Couty, F. Amino Acids 1999, 16, 297–320.
2. For some examples of synthesis of D-4,5-piperidines, see: (a) Rutjes, F. P. J. T.; Shoemaker, H. E. Tetrahedron
Lett. 1997, 38, 677–680. (b) Abell, A. D.; Garbiner, J.; Phillips, A. J.; Robinson, W. T. Tetrahedron Lett. 1998,
39, 9563–9566. (c) Huwe, C. H.; Velder, J.; Blechert, S. Angew. Chem., Int. Ed. Engl. 1996, 35, 2376–78, (d)
Sauers, A. J.; Ellman, J. A. J. Org. Chem. 2000, 65, 1222–1224.
3. Beak, P.; Lee, W. K. J. Org. Chem. 1993, 58, 1109–1117.
4. Takahashi, H.; Chida, Y.; Yashi, T.; Suzuki, T.; Yanaura, S. Chem. Pharm. Bull. 1986, 34, 2071–2077.
5. Agami, C.; Comesse, S.; Kadouri-Puchot, C.; Lusinchi, M. Synlett 1999, 1094–1096.
6. Pridgen, L. N.; Mokhallalati, M. K.; McGuire, M. A. Tetrahedron Lett. 1997, 38, 1275–1278.
7. This product is exclusively the acyclic eneimine when enal are used. On the other hand, oxazolidines are formed
when starting with saturated aldehydes.
8. Allin, S. M.; Button, M. A. C.; Baird, R. D. Synlett 1998, 1117–1120. In this paper, the authors report a de of
96% for 11 prepared by the same reaction, but run in ether instead of THF.
9. Agami, C.; Couty, F.; Evano, G. Tetrahedron Lett. 1999, 40, 3709–3712.
10. The nature of the minor compound 35 was determined by GC–MS analysis. All new compounds gave satisfactory
analytical data. Selected data: Compound 34: Rf (ether/petroleum ether: 5/95): 0.35; [h]2D0 −25 (c 0.7, CHCl3); oil;
1H NMR: 0.87 (t, J=7 Hz, 3H), 1.25–1.28 (m, 8H), 1.47 (bs, 9H), 1.9–2.1 (m, 3H), 2.45 (dd, J=2.2 and 17 Hz,
1H), 3.5 (d, J=18 Hz, 1H), 4.11 (d, J=18 Hz, 1H), 4.88 (very broad singlet, 1H), 5.41 (dd, J=16 and 5.8 Hz,
1H), 5.49–5.72 (m, 3H); 13C NMR: 14.5, 23.0, 28.9, 29.1, 29.6, 32.1, 32.7, 40.8, 49.9, 79.8, 123.1, 124.1, 128.5,
132.4, 155.5; anal. calcd for C18H31NO2: C, 73.67; H, 10.64; N, 4.77%. Found: C, 73.65; H, 10.61; N, 4.69%.
11. Agami, C.; Couty, F.; Lam, H.; Mathieu, H. Tetrahedron 1998, 54, 8783–8796.
12. Compound 32 was hydrogenated (Pd/C, EtOH, 95% yield) to give (R)-N-Boc-coniine. The specific rotation of
this compound ([h]2D0 −25 (c 0.7, CHCl3)) matched with an ee of 80%. For (S)-N-Boc-coniine, a specific rotation
of: [h]2D0 +33.5 (c 0.43, CHCl3) is reported. See: Jo, E.; Na, Y.; Chang, S. Tetrahedron Lett. 1999, 40, 5581–5582.
.