Conjugated additions of amines and β-amino alcohols to trifluorocrotonic acid derivatives: synthesis of ψ[NHCH(CF3)]-retro-thiorphan

Article abstract of DOI:10.1016/j.tetlet.2006.11.124

The aza-Michael addition of amines and β-amino-alcohols to trifluorocrotonic acid derivatives is described. The reactions occurred in good to excellent yields, but low stereocontrol. The title reaction has been used as a key step for the synthesis of trifluoroethylamine mimics of retro-thiorphan, which showed low inhibitory activity of NEP (constant inhibition at the molar level).

Full text of DOI:10.1016/j.tetlet.2006.11.124

Tetrahedron Letters 48 (2007) 589–593
Conjugated additions of amines and b-amino alcohols
to trifluorocrotonic acid derivatives: synthesis of
w[NHCH(CF3)]-retro-thiorphan
Marco Molteni,a Alessandro Volonterio,a,* Giacomo Fossati,a
Paolo Lazzarib and Matteo Zandac,*
aDipartimento di Chimica, Materiali ed Ingegneria Chimica ‘G. Natta’ del Politecnico di Milano,
via Mancinelli 7, I-20131 Milano, Italy
bNeuroscienze PharmaNess S.c.a r.l., c/o Parco Scientifico e Tecnologico della Sardegna, Loc.Piscinamanna, I-09010 Pula (CA), Italy
cC.N.R., Istituto di Chimica del Riconoscimento Molecolare, via Mancinelli 7, I-20131 Milano, Italy
Received 2 August 2006; revised 17 November 2006; accepted 20 November 2006
Available online 8 December 2006
Abstract—The aza-Michael addition of amines and b-amino-alcohols to trifluorocrotonic acid derivatives is described. The reactions
occurred in good to excellent yields, but low stereocontrol. The title reaction has been used as a key step for the synthesis of tri-
fluoroethylamine mimics of retro-thiorphan, which showed low inhibitory activity of NEP (constant inhibition at the molar level).
Ó 2006 Elsevier Ltd. All rights reserved.
Incorporation of a trifluoromethyl group into peptide
and protein structures is emerging as a highly attractive
strategy for improving biological activity and modifying
structural properties of these molecules.1 Of particular
interest is the replacement of a peptide or amide bond
[CONH] with trifluoroethylamine units, to provide the
so called w[CH(CF3)NH] isosteres.2 This strategy, that
was originally proposed by our group,3 has recently
found important applications in the development of
highly potent and metabolically stable inhibitors
of cathepsin K for the therapy of osteoporosis,4 and of
cathepsin S for the therapy of several autoimmune-based
inflammatory diseases.5 Along with the replacement of
the native peptide bond,3c the trifluoroethylamine unit
was also proposed as a surrogate of the retro-peptide
bond [NHCO],6 thus giving rise to w[NHCH(CF3)] iso-
steres.3a,b To this end, we described the synthesis of
partially modified w[NHCH(CF3)]-retro-peptides by
aza-Michael addition of a-amino acid esters to chiral
N-trifluorocrotonoyl-oxazolidin-2-ones. However, the
use of different nucleophiles, such as amines and ami-
no-alcohols, has not been investigated yet, despite the
potential use of such strategy to achieve a straightfor-
ward entry in to highly valuable trifluoroethylamine ana-
logues of biologically active molecules.
One of the possible targets is the w[NHCH(CF3)] ana-
logue 1 of retro-thiorphan (2) (Fig. 1), which is a potent
and selective inhibitor of the metalloproteinase NEP
(neutral endopeptidase) sparing another zinc proteinase
ACE (angiotensin converting enzyme), which has a key
role in the control of blood pressure.7
The N-trifluorocrotonoyl-oxazolidin-2-one 3 (Scheme 1)
was prepared as described in the literature.3a,b First
experiments carried out by simply mixing the acceptor
3 and benzylamine 4a (2 equiv) in DCM at rt (the
standard procedure used for the aza-Michael additions
of a-amino-acid esters to acceptors like 3)3a,b were
surprisingly disappointing, since the main product,
obtained in modest yields along with several unidentified
Ph
HS
Ph
HS
CF3
O
CO2H
CO2H
N
H
N
H
*
*
*
1
Retro-thiorphan 2
*
Figure 1. Retro-thiorphan (2) and w[NHCH(CF3)]-retro-thiorphan
(1).
Corresponding authors. Tel.: +39 02 23993084; fax: +39 02 23993080
(M.Z.); e-mail: matteo.zanda@polimi.it
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.11.124

590
M. Molteni et al. / Tetrahedron Letters 48 (2007) 589–593
R
R1
No H+
O
O
O
O
N
Bn
O
HN
*
DCM
rt
source
O
N
CF3
+
RR1NH
O
N
CF3
*
Bn
N
H
CF3
4
3 h
5
3
6
(38%)
O
O
BnNH2
O
N
CF3
H
N
MeO
NH2
OH
DCM, rt
NH2
NH2
Me
Me
3
Bn
4a
4b
4c
O
4d
O
HN
*
AcOH
NEt3
O
N
CF3
NH2
2 h
5a
(>98%)
OH
OH
H2N
(S)-4g
H2N
H2N
d.r. 1.3:1.0
4e
(S)-4f
(
S)-4h
Scheme 1. Set-up of the conditions for the aza-Michael reaction.
Scheme 2. The conjugate additions.
by-products, was amide 6. We next performed the reac-
tion of 3 with 4a under conditions of mild acid catalysis,
namely in the presence of a weak protic acid like tri-
ethylammonium acetate (2 equiv), generated in situ
from triethylamine and acetic acid (Table 1, entry 1).
Gratifyingly, 5a was obtained in 2 h at rt in nearly
quantitative yields, although with a very low diastereo-
control.8
tor 3. In order to improve the stereocontrol of the pro-
cess observed with 4a, the use of a chiral proton source
(Table 1, entry 11) and many different Lewis acids (en-
tries 12–15) was attempted, but we were unable to
achieve significantly better results in terms of diastereo-
selectivity, although the use of MgClO4 (entry 12) pro-
duced a reversal of diastereomeric ratio.
Chiral amine 4b (entry 2) also failed to provide good ste-
reocontrol. Dimethylamine 4c (entry 3) and cyclohexyl-
amine 4e (entry 5) followed the same trend, providing
synthetically useful yields and modest stereocontrol.
Only p-anisidine 4d (entry 4) failed to give rise to the
aza-Michael addition, owing to its lower nucleophilicity
with respect to aliphatic amines. The reactivity of b-ami-
no-alcohols was investigated next (entries 6–10). As in
the case of simple amines, (S)-phenylalaninol (4f, entry
6), (S)-valinol (4g, entry 8) and (S)-phenylglycinol (4h,
entry 10) gave rise to high-yielding but scarcely stereose-
lective reactions. It is worth noting that these reactions
Treatment of adduct 5a with benzylamine 4a gave rise to
exocyclic cleavage of the oxazolidin-2-one auxiliary,
affording amide 6 in 75% yield. However, the reaction
was very slow (48 h), suggesting that the observed for-
mation of 6 from 3 and 4a in the absence of acid catal-
ysis (Scheme 1) could take place through other pathways
as well.9
An array of aliphatic and aromatic amines, as well as b-
amino-alcohols 4 (Scheme 2) was used for exploring
scope and limits of the aza-Michael reaction with accep-
Table 1. The conjugate additions performed as in Scheme 3
Entry
Amine
Conditions
Diast. ratio
Yielda (%)
1
2
3
4
4a (2 equiv)
4b (2 equiv)
4c (2 equiv)
Triethylamine (2 equiv), acetic acid (2 equiv), 2 h
Triethylamine (2 equiv), acetic acid (2 equiv), 3 h
Triethylamine (2 equiv), HCl (2 equiv), 3 h
Triethylamine (2 equiv), acetic acid (2 equiv), 100 h
1.3/1.0
1.7/1.0
1.0:1.0
—
90
67
80
—
4d (2 equiv)
5
6
7
8
4e (2 equiv)
Triethylamine (2 equiv), acetic acid (2 equiv), 3.5 h
48 h
48 h
24 h
24 h
24 h
1.7/1.0
1.5/1.0
1.1/1.0
1.3/1.0
1.0/1.0
1.4/1.0
1.2/1.00
1.0/1.6
—
76
93
65
72
(S)-4f (2 equiv)
(R)-4f (2 equiv)
(S)-4g (2 equiv)
(R)-4g (2 equiv)
(S)-4h (2 equiv)
4a (2 equiv)
4a (1 equiv)
4a (1 equiv)
4a (1 equiv)
4a (1 equiv)
9
70
10
11
12
13b
14
15
95
Triethylamine (2 equiv), (S)-phenylpropionic acid (2 equiv), 3 h
MgClO4 (1 equiv), 2 he
>98c
>98c
BF3ÆOEt2 (1 equiv), 24 hd
—
—
TiCl4 (1 equiv), 4 hb
—
1.0/1.0
Me2AlCl (1 equiv), 4 hb
Not determinedf
a Isolated yields.
b The reaction was conducted at À78 °C.
c Measured by NMR.
d The reaction was conducted at À78 °C to rt.
e The reaction was conducted at 0 °C.
f The products were obtained in a mixture with about 20% of the starting material.

M. Molteni et al. / Tetrahedron Letters 48 (2007) 589–593
591
could be performed without acid catalysis, as a likely
consequence of the lower nucleophilicity of b-amino-
alcohols with respect to amines 4a–e.10 In order to
understand whether a match/mismatch recognition with
chiral acceptor 3 could be involved, the reactions of
(R)-enantiomers of 4f,g were also explored, but these
reactions afforded a nearly perfect 1:1 ratio of the corre-
sponding diastereomeric products.
ment of (S,S,S)-9 with KOH in degassed EtOH at reflux
was suitable for the preparation of the target diastereo-
pure retrothiorphan (S,S)-1 in moderate yields.12 How-
ever, the reaction suffered from somewhat low
reproducibility, occasionally providing large amounts
of the undesired disulfide dimer of 1 together with other
unidentified by-products. Indeed, the reaction outcome
was apparently strongly dependent on the degree of pur-
ity of solvents and reagents employed. We therefore
decided to develop an alternative procedure, by-passing
the critical oxazolidin-2-one cleavage. Due to the low
stereodirecting effect of the oxazolidin-2-one function,
we sorted out to drop this auxiliary and use commercial
ethyl trifluorocrotonate as starting Michael acceptor
(Scheme 4).
In order to see whether the low diastereocontrol of these
aza-Michael reactions could be due to interconversion
of the epimeric products 5, one of the pure diastereo-
mers of 5g (obtained by silica gel chromatography)
was submitted to the exact reaction conditions in the
presence of 4g for 30 h at rt. No epimerization was
observed (NMR control), thus confirming that the
aza-Michael reactions of 3 with amines and amino-
alcohols 4 occur under kinetic control.
Addition of (R)-4f to ethyl trifluorocrotonate took place
in good yields in EtOH at reflux, even though no stereo-
control was observed in this case too. Product 11 was
obtained in an equimolar mixture of epimers at the
CF3-substituted carbon, which was very hardly separa-
ble by flash chromatography. Epimers 11 were therefore
treated with MsCl according to the previously developed
conditions, affording chlorides 12, which were trans-
formed into the epimeric thiolacetates 13. The final
cleavage of both terminal ester and thiolester functions
was achieved by basic hydrolysis that provided the tar-
get retrothiorphan 1 having R-configuration at the benz-
ylic stereocentre in satisfactory yields. In this case, the
reaction was perfectly reproducible, although we were
unable to separate the two epimers.13
The diastereomeric adducts 5f were chosen as starting
materials for the synthesis of the trifluoroethylamine mi-
mic of retrothiorphan (Scheme 3). In order to transform
the hydroxyl into thiol function, we found that treat-
ment of diastereomerically pure (S,S,S)-5f with MsCl
(3 equiv) and triethylamine (3 equiv) afforded in nearly
quantitative yields the chloride (S,S,S)-8.11 We could
not isolate the intermediate mesylate that undergoes fast
SN2 reaction with the chloride counterion to give 8.
However, we found that increasing the ratio NEt3/MsCl
to 10:3, the main product became aziridine 10 that was
isolated in rather modest yields. Treatment of (S,S,S)-
8 with AcSK produced the desired thiol acetate
(S,S,S)-9 in excellent yields. Unfortunately, a number
of methods to achieve exocyclic oxazolidin-2-one cleav-
age (LiOOH, HCl, etc.) from 9 were unsuccessfully
tried, generally resulting in intractable mixtures of com-
pounds arising from oxidation of the thiol function, or
other side reactions. Eventually, we found that treat-
The same reaction sequence portrayed in Scheme 4 was
applied on enantiomer (S)-4f, thus leading to the epi-
meric retrothiorphan mimic (S,R/S)-1, obtained also in
this case as an equimolar mixture of epimers at the
CF3-stereocentre.
All the w[NHCH(CF3)]-retro-thiorphan diastereomers 1
were subjected to biological tests (fluorometric assay)
to evaluate their inhibitory capacity toward Neutral
Ph
OH
Ph
O
O
O
CF3
O
CF3
Cl
(95%)
MsCl (3 Eq.)
O
N
N
O
N
N
H
H
NEt3 (3 Eq.)
DCM, rt, 24 h
Ph
(S,S,S)-5f
CF3
(S,S,S)-8
(R)-4f
EtO2C
EtO2C
OH
CH3COSK
DMF, 0 °C
to rt, 24 h
CF3
N
H
*
EtOH
reflux, 18 h
(70%)
11
MsCl (3 eq)
Et3N (3 eq)
DCM
Ph
Ph
O
O
CF3
CF3
KOH, EtOH
SCOCH3
HO2C
SH
O
N
N
H
N
H
reflux, 30 h
Ph
Ph
(50-60%)
CF3
CF3
(S,S)-1
(S,S,S)-9
CH3COSK
DMF
(90%)
EtO2C
Cl
EtO2C
(89%)
SCOCH3
N
N
*
*
H
H
13
12
(94%)
NaOH 2M
EtOH, rt, 1 h
Ph
Ph
OH
O
O
CF3
Ph
O
O
N
CF3
CF3
MsCl (1.5 Eq.)
O
N
N
H
HO2C
SH
O
N
N
*
NEt3 (5 Eq.)
DCM, rt, 24 h
10
H
(S,S,S)-5f
(41%)
(R,R/S)-1 (60%)
Scheme 3. Synthesis of stereochemically pure w[NHCH(CF3)]-retro-
Scheme 4. Alternative route to w[NHCH(CF3)]-retro-thiorphan (mix-
thiorphan.
ture of epimers).

592
M. Molteni et al. / Tetrahedron Letters 48 (2007) 589–593
Endopeptidase 24.11 (NEP). The assay was carried out
by a method based on the procedure reported by Floren-
tin et al.14 All the w[NHCH(CF3)]-retro-thiorphan dia-
stereomers 1 showed IC50 values several orders of
magnitude higher than thiorphan, with Ki values over
4 M (for reference compound: IC50 = 5.06 nM, Ki =
2.53 nM). Moreover, the comparison of the results
obtained for the new w[NHCH(CF3)] isosteres 1 with
the data reported for (R) and (S)-retro-thiorphan
(Ki = 2.3 nM and 210 nM, respectively)15 confirmed
the loss of the NEP inhibition capacity. This dramatic
drop of inhibitory activity might be due to the fact that
the retropeptidic carbonyl group of retro-thiorphan is
known to be involved in critical interactions with the
active site of NEP as hydrogen bond acceptor.16 There-
fore its replacement with the trifluoroethylamine func-
tion, which is a very weak hydrogen bond acceptor,17
could be the underlying reason for the loss of potency.
This observation suggests important considerations for
a successful use of the trifluoroethylamine function as
a peptide/retropeptide bond mimic (Fig. 2). (1) The tri-
fluoromethyl group, contrarily to the carbonyl oxygen,
is a weak hydrogen-bond acceptor.17 The trifluoroethyl-
amine function can be therefore an effective peptide
bond replacement only if the carbonyl group of the ori-
ginal ligand’s amide/peptide-bond is not involved in
essential hydrogen-bonding with the receptor. (2) The
NH of the trifluoroethylamine unit is a good hydro-
gen-bond donor, due to the strong electronwithdrawing
effect exerted by the CF3 group, and could be always
considered a good mimic of a peptidic NH. (3) The
sp3 N atom of the trifluoroethylamine function is a
bad hydrogen bond acceptor and has very little Lewis
basicity, in close analogy with the peptide bond.4a
w[NHCH(CF3)] analogue 1 of retro-thiorphan. An
alternative route to 1, obtained in this case as an epi-
meric mixture at the CF3-substituted stereocentre, has
been developed as well from ethyl trifluorocrotonate.
Unfortunately, all the diastereomers of 1 showed low
inhibitory activity of NEP compared with the reference
compounds thiorphan and retro-thiorphan.
Acknowledgments
We thank MIUR (Cofin 2004 Project ‘Polipeptidi Bioat-
tivi e Nanostrutturati, and Project ‘Nuovi peptidomime-
`
tici ad attivita analgesica’, protocol No. 13378, 24/12/
2003), Politecnico di Milano, and C.N.R. for economic
support.
References and notes
1. (a) Zanda, M. New J. Chem. 2004, 28, 1401–1411; (b)
´
Begue, J.-P.; Bonnet-Delpon, D.; Crousse, B.; Legros, J.
Chem. Soc. Rev. 2005, 34, 562–572; (c) Gong, Y. F.; Kato,
K. Curr. Org. Chem. 2005, 8, 1659–1675.
´
2. Sani, M.; Molteni, M.; Bruche, L.; Volonterio, A.; Zanda,
M. In Fluorine-Containing Synthons; Soloshonok, V. A.,
Ed.; ACS Publications Division and Oxford University
Press: Washington, DC, 2005; pp 572–592.
3. (a) Volonterio, A.; Bravo, P.; Zanda, M. Org. Lett. 2000,
2, 1827–1830; (b) Volonterio, A.; Bellosta, S.; Bravin, F.;
´
Bellucci, M. C.; Bruche, L.; Colombo, G.; Malpezzi, L.;
Mazzini, S.; Meille, S. V.; Meli, M.; Ramirez de Arellano,
C.; Zanda, M. Chem. Eur. J. 2003, 9, 4510–4522; (c)
Molteni, M.; Volonterio, A.; Zanda, M. Org. Lett. 2003, 5,
3887–3890.
4. (a) Black, W. C.; Bayly, C. I.; Davis, D. E.; Desmarais, S.;
In summary, we have described the aza-Michael addi-
tion of amines and b-amino-alcohols to the chiral N-tri-
fluorocrotonoyl-oxazolidin-2-one 3. The reactions occur
in good to excellent yields, but low stereocontrol. One of
the aza-Michael adducts (5f) was used as the starting
material for the synthesis of the stereopure
´
´
Falgueyret, J.-P.; Leger, S.; Li, C. S.; Masse, F.; McKay,
D. J.; Palmer, J. T.; Percival, M. D.; Robichaud, J.; Tsou,
N.; Zamboni, R. Bioorg. Med. Chem. Lett. 2005, 15, 4741–
4744; (b) Li, C. S.; Deschenes, D.; Desmarais, S.;
´
Falgueyret, J.-P.; Gauthier, J. Y.; Kimmel, D. B.; Leger,
´
S.; Masse, F.; McGrath, M. E.; McKay, D. J.; Percival,
´
M. D.; Riendeau, D.; Rodan, S. B.; Therien, M.; Truong,
V.-L.; Wesolowski, G.; Zamboni, R.; Black, W. C. Bioorg.
Med. Chem. Lett. 2006, 16, 1985–1989.
5. Link, J. O.; Zipfel, S. Curr. Opin. Drug Discovery Dev.
2006, 9, 471–482.
Receptor
H
Peptide bond
O
6. (a) Fletcher, M. D.; Campbell, M. M. Chem. Rev. 1998,
98, 763–795; (b) Goodman, M.; Chorev, M. Acc. Chem.
Res. 1979, 12, 1–7.
N
H
7. Roques, B. P.; Lucas-Soroca, E.; Chaillet, P.; Costentin,
X
Receptor
C=O good H-bond acceptor
N bad H-bond acceptor
NH good H-bond donor
´
J.; Fournie-Zaluski, M. C. Proc. Natl. Acad. Sci. U.S.A.
1983, 80, 3178–3182.
8. Synthesis of 5a. TEA (2 equiv) and AcOH (2 equiv) were
added at rt to a solution of 3 (1 equiv) in DCM (0.2 M
solution). After 5 min 4a was added at rt and the reaction
monitored by TLC. The mixture was diluted with 1 N
aqueous HCl solution, extracted with DCM, the collected
organic layers dried (Na2SO4), filtered, concentrated under
reduced pressure, and the crude purified by flash chroma-
tography. 5a (one diastereoisomer): 1H (400 MHz,
CDCl3): d 7.29 (m, 5H), 4.39 (m, 1H), 4.21 (m, 2H), 4.03
(d, J = 13.2 Hz, 1H), 3.85 (d, J = 13.2 Hz, 1H), 3.79 (m,
1H), 3.43 (dd, J = 16.2 and 9.4 Hz, 1H), 3.11 (dd, J = 16.2
and 4.1 Hz, 1H), 2.39 (m, 1H), 0.92 (d, J = 7.1 Hz, 3H),
0.84 (d, J = 6.8 Hz, 3H).
Receptor
H
Trifluoroethylamine
function
CF3
N
H
X
Receptor
CH-CF3 bad H-bond acceptor
N bad H-bond acceptor
NH good H-bond donor
X = O, N, etc.
Figure 2. Comparison between the peptide (amide) bond and the
trifluoroethylamine function.

M. Molteni et al. / Tetrahedron Letters 48 (2007) 589–593
593
9. Another possible pathway to 6 could be the fragmentation
into ketene 7 and the corresponding deacylated 2-oxazol-
idinone, followed by reaction with benzylamine.
4.41 (m, 1H), 4.23 (m, 2H), 3.86 (m, 1H), 3.49–3.32 (m,
4H), 3.04 (dd, J = 16.6 and 3.9 Hz, 1H), 2.85 (dd, J = 13.5
and 7.3 Hz, 1H), 2.74 (dd, J = 13.5 and 6.2 Hz, 1H), 2.40
(m, 1H), 1.61 (br s, 1H), 0.91 (d, J = 6.9 Hz, 3H), 0.86 (d,
J = 6.9 Hz, 3H).
Bn
12. (a) The stereochemistry was assigned by X-ray diffraction
of a single crystal of (S,S)-1. The crystallographic data will
be reported in a forthcoming full letter. (b) Synthesis of (S,
S)-1. Solid KOH (8 equiv) was added to a solution of
(S,S,S)-9 (1 equiv) in degassed abs. EtOH (0.1 M solu-
tion). The mixture was refluxed for 30 h., cooled to rt and
the solvent evaporated. The crude was dissolved in a 1 N
aqueous HCl solution and extracted several times with
AcOEt. The collected organic layers were dried (Na2SO4),
filtered, concentrated under reduced pressure, and the
HN
O
C
CF3
*
7
The protic acids successfully used in Table 1, entries 1–
5,11 could therefore play the role of rapidly quenching the
intermediate a-carbonyl carbanion formed by aza-Michael
addition, thus suppressing the fragmentation via ketene 7.
The formation of ketenes by fragmentation of N-acyl-
oxazolidinones has been observed previously: Bravo, P.;
Fustero, S.; Guidetti, M.; Volonterio, A.; Zanda, M. J.
Org. Chem. 1999, 64, 8731–8735 and references cited
therein. Attempts to isolate these hypothetic ketene
intermediates by adding external nucleophiles, such as
dialkylamines, in the reaction mixture were unsuccessful.
10. b-Amino-alcohols are known to have a nucleophilic
reactivity in Michael-type reactions which is intermediate
between that of amines and a-amino-acids. See for
example: Um, I.-H.; Lee, E.-J.; Seok, J.-A.; Kim, K.-H.
J. Org. Chem. 2005, 70, 7530–7536. However, a referee
suggested that the difference in reactivity observed
between amines and b-amino-alcohols could lie in the fact
that the latter can immediately protonate the carbanion
formed after aza-Michael addition of the amino moiety to
3. In our opinion, this hypothesis does not explain why a-
amino-acid esters, which lack this proton, have the same
reactivity of b-amino-alcohols, as shown in Ref. 3a,b.
11. Synthesis of 8. Dry TEA (3 equiv) followed by freshly
distilled MsCl (3 equiv) was added at rt under N2 to a
solution of (S,S,S)-5f (1 equiv) in dry DCM (0.1 M
solution). After 24 h the mixture was diluted with water
and extracted with DCM. The collected organic layers
were dried (Na2SO4), filtered, concentrated under reduced
pressure, and the crude was purified by flash chromato-
graphy. 8: 1H (400 MHz, CDCl3): d 7.33–7.18 (m, 5H),
1
crude was purified by flash chromatography. (S,S)-1: H
(400 MHz, CD3OD): d 7.29–7.18 (m, 5H), 3.71 (m, 1H),
3.16 (m, 1H), 2.78 (m, 2H), 2.67 (dd, J = 16.0 and 4.2 Hz,
1H), 2.63 (dd, J = 14.0 and 4.7 Hz, 1H), 2.44 (dd, J = 16.0
and 9.4 Hz, 1H), 2.39 (dd, J = 13.6 and 4.7 Hz); 13C
(100 MHz, CD3OD): d 173.7, 139.9, 130.4, 129.5, 129.3,
127.9 (q, J = 283.7 Hz), 126.5, 60.2, 55.8 (q, J = 28.4 Hz),
40.7, 36.2, 28.4 ; [a]D À23.3 (c 0.6, methanol).
13. Compounds 1 are indefinitely stable upon storage at 4 °C
(neat). However, the stability is much lower when diaste-
reomers 1 are stored in solution at room temperature, and
definitely low when heated in the presence of bases, and in
general at pH > 7. This precluded, in our hands, the
possibility to isolate diastereopure 1 by crystallization with
chiral bases.
14. Florentin, D.; Sassi, A.; Roques, B. P. Anal. Biochem.
1984, 141, 62–69.
´
15. Roques, B. P.; Noble, F.; Dauge, V.; Founie-Zaluski,
M. C.; Beaumont, A. Pharmacol. Rev. 1993, 45, 87–146.
16. (a) Gomez-Monterrey, I.; Beaumont, A.; Nemecek, P.;
Roques, B. P.; Fournie-Zaluski, M.-C. J. Med. Chem.
1994, 37, 1865–1873; (b) Coric, P.; Turcaud, S.; Meudal,
H.; Roques, B. P.; Fournie-Zaluski, M. C. J. Med. Chem.
1996, 39, 1210–1219, and references cited therein.
17. Dunitz, J. D.; Taylor, R. Chem. Eur. J. 1997, 3, 89–
98.