Highly stereocontrolled boron-mediated synthesis of β-hydroxy-α-amino acids and dipeptides. Part 2

Article abstract of DOI:10.1016/S0957-4166(99)00539-X

The chiral synthons 1(a-d) were submitted to boron-mediated asymmetric aldol condensation with acetaldehyde and benzaldehyde providing, in high diastereomeric excess (>95%), R,R-configured aldols 2(a-f) which are useful intermediates to enantiomerically pure β-hydroxy-α-amino acids. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0957-4166(99)00539-X

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 10 (1999) 4709–4714
Highly stereocontrolled boron-mediated synthesis of
β-hydroxy-α-amino acids and dipeptides. Part 2
Pina Di Felice, Marcella Maestri, Francesca Paradisi, Gianni Porzi and Sergio Sandri ∗
Dipartimento di Chimica ‘G. Ciamician’, Università degli Studi di Bologna, Via Selmi 2, 40126 Bologna, Italy
Received 18 October 1999; accepted 25 November 1999
Abstract
The chiral synthons 1(a–d) were submitted to boron-mediated asymmetric aldol condensation with acetaldehyde
and benzaldehyde providing, in high diastereomeric excess (>95%), R,R-configured aldols 2(a–f) which are useful
intermediates to enantiomerically pure β-hydroxy-α-amino acids. © 2000 Elsevier Science Ltd. All rights reserved.
1. Introduction
Recently, we described1 a new diastereoselective approach to β-hydroxy-α-amino acids and dipeptides
through aldol reactions using the lithium enolate of an appropriate chiral synthon. In the continuation of
our investigations, we focused our attention on improving the diastereoselection of aldol condensations
for the synthesis of enantiomerically pure common and uncommon α-amino acids and peptides. To this
aim, we performed the aldol addition reaction using the boron enolate of the chiral synthons 1(a–d)
synthesized as described in our preceding papers.1,2
Thus, in this paper very interesting and encouraging results are reported. The stereochemical outcome
of the aldol reaction which occurred with satisfactory chemical yields and with diastereomeric excesses
better than 95% (Table 1) providing aldols 2, which are intermediates for obtaining dipeptides and/or
β-hydroxy-α-amino acids, is explained.1
2. Results and discussion
The aldol reaction occurred with practically total conversion of the chiral synthon 1 with benzaldehyde,
while with acetaldehyde the chemical yield was not higher than 70% (Scheme 1).
∗
Corresponding author. E-mail: porzi@ciam.unibo.it
0957-4166/99/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00539-X
tetasy 3174

4710
P. Di Felice et al. / Tetrahedron: Asymmetry 10 (1999) 4709–4714
Table 1
Aldol condensation of boron enolates with acetaldehyde and benzaldehyde
Scheme 1. (i) Bu2BOTf/Et3N in CH2Cl2, rt; (ii) R00CHO in CH2Cl2, −78°C, then H2O2/NaOH at rt
The data reported in Table 1 show that in all the experiments performed, the reaction occurred
in a highly stereocontrolled fashion: actually, except for entries 4 and 5, total 1,4-cis induction and
anti stereoselection was achieved with only the isomer (3R,100R)-2 being obtained. The observed
stereochemical results are very different from those previously obtained by using the lithium enolates1
which, conversely, prevalently gave trans induction and anti selectivity. Furthermore, such unsatisfactory
diastereoselection has been confirmed by the results we obtained with the lithium enolates of 1(c,d).3
The high stereocontrol evidenced by the boron enolates (Table 1) showed that the Si-face of the
aldehyde approaches with a very large preference to the Re-face of the chiral boron enolate (Fig.
1) highly favoring relative topicity ul. Almost total formation of the (3R,100R)-2 diastereomer was
observed independent of both the substituent at N-1 (compare entries 1 and 5, 2 and 6) and its absolute
configuration (compare entries 1 and 3). Moreover, it appears evident that the stereocenter on the remote
C-6 influences the diastereoselectivity more than the chiral substituent on the adjacent N-1 (compare
entries 4 and 6).
In order to explain the stereochemical outcome of the condensation, a molecular modeling
investigation4 was carried out on the simplified model 3 of the boron enolate (Fig. 2). Energy
calculations have shown that in the minimum energy conformation the atoms (N-1)–(C-2)–(C-3)–(N-
4)–(C-5) are almost planar, while the C-6 is slightly out of the plane. The C-10 is not coplanar with
the heterocyclic ring with the dihedral angle (C-3)–(C-2)–(N-1)–(C-10) being 45°. In addition, the
boron enolate, independent of the C-10 stereocenter configuration, adopts a preferential conformation

P. Di Felice et al. / Tetrahedron: Asymmetry 10 (1999) 4709–4714
4711
Fig. 1.
with the chiral phenethyl group trans with respect to both the substituent at C-6 and the boron atom
which lies about 80° out of the plane and cis to the (C-6)-CH3 (Fig. 1). Conformational analysis
(using AM1 calculations) revealed that the isomer in which the boron atom and (C-6)-CH3 lie cis
is 3.5 and 4.4 kcal/mol more stable than the trans-isomer when the phenethyl group possesses R- or
S-configuration, respectively. Such a strongly preferred arrangement would favor the cis attack of the
aldehyde with respect to the (C-6)-CH3 (see Fig. 1): effectively, the experimental results are consistent
with the molecular modeling studies. These results suggest that the phenethyl moiety most probably
acts to transfer through space the chiral information of the C-6 stereocenter to the C-3, i.e. the point
of electrophilic attack. Thus, the (C-6)-CH3 would direct the adjacent phenethyl group into the anti
position which in turn directs the boron atom syn to the (C-6)-CH3. These findings are consistent with
the ‘chiral relay network’ already proposed.5
Fig. 2.
In order to confirm such a ‘chirality transfer’ we performed a conformational analysis of boron enolate
1d in which the chiral phenethyl group was changed to Ph-CH2−. Also for this substrate in the preferred
conformation, the boron is arranged cis with respect to the (C-6)-CH3 and trans with respect to the (N-
1)-CH2Ph. AM1 calculations4 showed that the preferred cis arrangement between B and CH3 is about
1.7 kcal/mol more stable than the trans. Thus, it can be asserted that the chiral substituent at N-1 does
not act as a chiral inductor, but more simply as a chirality transfer group, relaying the chirality from the
remote chiral group at C-6.
In this context, it is interesting to point out that the substrate 1c (entry 4), even if it gives total anti
selectivity, in addition to diastereomer 2d (80%) also furnished the diastereomer 2g (20%) due to an
attack of the Si-face of the chiral boron enolate of 1c on the Re-face of benzaldehyde. The observed
diastereoselection can be rationalized on the basis of the model used by d’Angelo and co-workers6 to
explain the diastereofacial differentiation in the alkylation of chiral enamine derived from (R)- or (S)-
phenethylamine. These authors demonstrated that the phenethyl C–H bond preferentially eclipses the
single C–C bond rather than the C_C, so that the enamine olefinic face syn to the phenyl ring is sterically

4712
P. Di Felice et al. / Tetrahedron: Asymmetry 10 (1999) 4709–4714
more hindered. A similar model, outlined in Fig. 3, could explain the observed facial stereoselection for
the substrate 1c. Indeed, because the phenethyl C–H bond preferentially eclipses the (C-5)–(C-6) bond,
the phenyl ring causes greater steric hindrance on the Si face of the enolate, and the boron atom is then
forced trans with respect to the aromatic ring favoring the approach of benzaldehyde from the Re face,
i.e. the less hindered side.
Fig. 3.
3. Stereochemical assignments
While the absolute configuration of the new introduced stereogenic centers of 2(a,b,c) was already
established,1 the stereochemistry of 2(d,f) and 2g was assigned by comparison of the specific rota-
tion value of the corresponding (2R,3R)- and (2S,3S)-methyl phenylserinate with that reported in the
literature7 obtained from the respective aldols by following the simple procedure previously reported.1
The configuration of 2e was assigned on the basis of 1H NMR spectroscopic data by using the approach
previously employed for analogous compounds1 and confirmed by the agreement of the measured
specific rotation value of the corresponding (2R,3R)-threonine, obtained from 2e,1 with that reported
in the literature.8
4. Experimental
4.1. General information
1H and 13C NMR spectra were recorded in CDCl3 at 300 MHz. Coupling constants (J) are in hertz.
Optical rotation values were measured at 25°C on a Perkin–Elmer 343 polarimeter. The reactions
involving organometallic reagents were carried out under an argon atmosphere in dry solvent. Bu2BOTf
was used as a commercially available 1 M solution in CH2Cl2. Et3N was distilled from CaH2 prior to
use. The aldehydes employed were distilled prior to use.
4.2. (10S)-5-Ethoxy-1-(10-phenethyl)-3,6-dihydro-1H-pyrazine-2-one 1c
The product was synthesized following the procedure we described for 1a2b by using the ethyl ester
of bromoacetic instead of bromopropionic acid. 1H NMR: δ 1.2 (t, 3H, J=7.1), 1.52 (d, 3H, J=7.1), 3.38
(ddd, 1H, J=2.6, 2.6, 17), 3.74 (ddd, 1H, J=2.6, 2.6, 17), 4.06 (m, 2H), 4.2 (dd, 1H, J=2.6, 2.6), 6.1 (q,
1H, J=7.1), 7.3 (m, 5ArH). 13C NMR: δ 13.9, 14.6, 40.2, 49.2, 50.6, 61.2, 127.1, 127.5, 128.4, 138.4,
157.6, 165.7. Anal. calcd for C14H18N2O2: C, 68.27; H, 7.37. Found: C, 68.46; H, 7.35.

P. Di Felice et al. / Tetrahedron: Asymmetry 10 (1999) 4709–4714
4713
4.3. General procedure for the aldol reaction of 1(a–d) with aldehydes
To 10 mmol of the lactime 1(a–d) dissolved in 20 ml of dry CH2Cl2 under an argon atmosphere, 2.8
ml (20 mmol) of dry Et3N and 18 ml (18 mmol) of Bu2BOTf were added. The reaction was stirred 1
h at rt, then the bath was cooled to −78°C and 20 mmol of aldehyde were slowly added. After 30 min
the reaction was quenched with phosphate buffer (pH=7) and extracted with ethyl acetate. The organic
solution was evaporated in vacuo and 10 ml of 1N NaOH were added to the residue dissolved in 40 ml
of methanol, then 10 ml of 30% H2O2 were added dropwise to the solution. After 1 h water was added,
the methanol removed in vacuo and the aqueous layer was extracted with ethyl acetate. The organic layer
was washed with water and dried. After evaporation of the organic solvent, the residue was subjected to
1H NMR analysis.
4.4. (3R,6R,10S,100R)-5-Ethoxy-3-(100-hydroxyethyl)-6-methyl-1-(10-phenethyl)-3,6-dihydro-1H-
pyrazine-2-one 2a
See the literature for details.1
4.5. (3R,6R,10S,100R)-5-Ethoxy-3-(100-hydroxybenzyl)-6-methyl-1-(10-phenethyl)-3,6-dihydro-1H-
pyrazine-2-one 2b
See the literature for details.1
4.6. (3R,6R,10R,100R)-5-Ethoxy-3-(100-hydroxyethyl)-6-methyl-1-(10-phenethyl)-3,6-dihydro-1H-
pyrazine-2-one 2c
See the literature for details.1
4.7. (3R,10S,100R)-5-Ethoxy-3-(100-hydroxybenzyl)-1-(10-phenethyl)-3,6-dihydro-1H-pyrazine-2-one 2d
1H NMR: δ 1.2 (t, 3H, J=7.1), 1.42 (d, 3H, J=7.1), 2.65 (dd, 1H, J=2.4, 17.1), 3.47 (dd, 1H, J=2.7,
17.1), 3.97 (d, 1H, J=7.3), 4.12 (m, 2H), 4.59 (m, 1H), 5.3 (dd, 1H, J=5.5, 7.3), 5.95 (q, 1H, J=7.1), 6.95
(m, 2ArH), 7.25 (m, 8ArH). 13C NMR: δ 13.6, 14.4, 39.6, 48.8, 60.9, 63.6, 74, 126.4, 127, 127.3, 127.6,
128, 128.5, 137, 138.8, 159, 166. [α]D=−90.2 (c 2.54, CHCl3). Anal. calcd for C21H24N2O3: C, 71.57;
H, 6.86. Found: C, 71.8; H, 6.84.
4.8. (3R,6R,10R)-1-Benzyl-5-ethoxy-3-(10-hydroxyethyl)-6-methyl-3,6-dihydro-1H-pyrazine-2-one 2e
1H NMR: δ 1.24 (t, 3H, J=7.1), 1.31 (d, 3H, J=6.1), 1.43 (d, 3H, J=6.8), 3.82 (dq, 1H, J=1.9, 6.8), 4
(d, 1H, J=15), 4.15 (m, 4H), 5.34 (d, 1H, J=15), 7.3 (m, 5ArH). 13C NMR: δ 14, 19.2, 20, 45.9, 51.2,
61.5, 64.1, 70.1, 127.7, 128, 128.6, 131.5, 160.4, 168.1. [α]D=22.6 (c 0.61, CHCl3). Anal. calcd for
C16H22N2O3: C, 66.18; H, 7.64. Found: C, 66.32; H, 7.65.
4.9. (3R,6R,10R)-1-Benzyl-5-ethoxy-3-(10-hydroxybenzyl)-6-methyl-3,6-dihydro-1H-pyrazine-2-one 2f
1H NMR: δ 0.5 (d, 3H, J=7), 1.2 (t, 3H, J=7.1), 3.6 (dq, 1H, J=2.1, 7), 3.78 (d, 1H, J=14.9), 3.82 (d,
1H, J=7.3), 4.1 (q, 2H, J=7.1), 4.64 (dd, 1H, J=2.1, 5.4), 5.3 (d, 1H, J=14.9), 5.3 (dd, 1H, J=5.4, 7.3),

4714
P. Di Felice et al. / Tetrahedron: Asymmetry 10 (1999) 4709–4714
7.25 (m, 10ArH). 13C NMR: δ 14.1, 17.6, 45.6, 50.9, 61.5, 64.6, 74.9, 126.9, 127.6, 127.7, 128.2, 128.6,
135.4, 140, 160.8, 166.4. [α]D=−34 (c 1.95, CHCl3). Anal. calcd for C21H24N2O3: C, 71.57; H, 6.86.
Found: C, 71.45; H, 6.88.
4.10. (3S,10S,100S)-5-Ethoxy-3-(100-hydroxybenzyl)-1-(10-phenethyl)-3,6-dihydro-1H-pyrazine-2-one
2g
1H NMR: δ 1.2 (d, 3H, J=7.1), 1.22 (t, 3H, J=7.1), 2.7 (dd, 1H, J=2.3, 16.9), 3.04 (dd, 1H, J=2.2,
16.9), 3.82 (d, 1H, J=7.1), 4.1 (m, 2H), 4.6 (m, 1H), 5.34 (dd, 1H, J=5.4, 7.1), 5.97 (q, 1H, J=7.1), 7.3
(m, 10ArH). 13C NMR: δ 14.1, 14.4, 39.8, 49.3, 61.8, 64.6, 75.1, 126.9, 127.3, 127.6, 127.8, 138.4,
139.8, 159.2, 165.9. [α]D=−57.6 (c 1, CHCl3). Anal. calcd for C21H24N2O3: C, 71.57; H, 6.86. Found:
C, 71.45; H, 6.84.
Acknowledgements
Thanks are due to the University of Bologna (F.R.I., ex 60%) and MURST (COFIN 1998–2000) for
financial support.
References
1. Di Felice, P.; Porzi, G.; Sandri, S. Tetrahedron: Asymmetry 1999, 10, 2191.
2. (a) Porzi, G.; Sandri, S.; Verrocchio, P. Tetrahedron: Asymmetry 1998, 9, 119; (b) Favero, V.; Porzi, G.; Sandri, S.
Tetrahedron: Asymmetry 1997, 8, 599.
3. The following diastereomeric distribution was observed:
4. Energy calculations were performed by the AM1 method (HyperChem Program, 1994) by using the ‘Polak-Ribiere’
algorithm (RMS gradient 0.01 kcal).
5. Bull, S. D.; Davies, S. G.; Epstein, S. W.; Ouzman, J. V. A. Chem. Commun. 1998, 659.
6. d’Angelo, J.; Desmaele, D. Tetrahedron Lett. 1989, 30, 2645, and references cited therein.
7. Hoenig, H.; Seufer-Wasserthal, P.; Weber, H. Tetrahedron 1990, 46, 3841: [α]D=−33.8 (c 2, CH2Cl2).
8. Beil. 4(3), 1629: [α]D=−9 (c 2, H2O).