Synthesis and evaluation of stereopure α-trifluoromethyl-malic hydroxamates as inhibitors of matrix metalloproteinases

Article abstract of DOI:10.1016/j.tetlet.2003.12.131

The total synthesis of trifluoromethyl (Tfm) analogs of known nanomolar matrix metalloproteinases (MMPs) inhibitors has been performed. The synthetic protocol is based on a moderately stereoselective aldol reaction of trifluoropyruvate with an N-acyl-oxaz

Full text of DOI:10.1016/j.tetlet.2003.12.131

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 1611–1615
Synthesis and evaluation of stereopure a-trifluoromethyl-malic
hydroxamates as inhibitors of matrix metalloproteinases
Monica Sani,a,b Dorina Belotti,c Raffaella Giavazzi,c Walter Panzeri,b
Alessandro Volonterioa,* and Matteo Zandab,*
aDipartimento di Chimica, Materiali ed Ingegneria Chimica ‘‘G.Natta’’ del Politecnico di Milano, via Mancinelli 7,
I-20131 Milan, Italy
bCNR––Istituto di Chimica del Riconoscimento Molecolare, sezione ‘‘A. Quilico’’, via Mancinelli 7, I-20131 Milan, Italy
cIstituto di Ricerche Farmacologiche Mario Negri, Laboratori Negri, via Gavazzeni 11, I-24125 Bergamo, Italy
Received 9 December 2003; revised 22 December 2003; accepted 23 December 2003
Abstract—The total synthesis of trifluoromethyl (Tfm) analogs of known nanomolar matrix metalloproteinases (MMPs) inhibitors
has been performed. The synthetic protocol is based on a moderately stereoselective aldol reaction of trifluoropyruvate with an
N-acyl-oxazolidin-2-thione for the construction of the core a-Tfm-malic unit. Both the diastereomeric forms of the target a-Tfm-
malic hydroxamates showed micromolar inhibitory potency toward MMP-2 and 9, with a substantial drop with respect to the
parent unfluorinated compounds.
Ó 2004 Elsevier Ltd. All rights reserved.
Matrix metalloproteinases (MMPs) are zinc (II)-depen-
dent proteolytic enzymes involved in the degradation of
the extracellular matrix.1 More than 25 human MMPs
have been identified so far. Loss in the regulation of
their activity can result in the pathological destruction
of connective tissue, a process associated with a number
of severe diseases, such as cancer and arthritis. The
inhibition of various MMPs has been envisaged as a
strategy for the therapeutic intervention against such
pathologies. To date, however, a number of drawbacks
have hampered the successful exploitation of MMPs as
pharmacological targets. In particular, the toxicity
demonstrated by many MMPsÕ inhibitors in clinical
trials has been ascribed to nonspecific inhibition.
tures a good mimicry with several naturally occurring
residues such as methyl, isopropyl, phenyl, etc.3
Recently, Jacobson et al. described a new family of
potent peptidomimetic hydroxamate inhibitors A (Fig.
1) of MMP-1, -3, and -9, bearing a quaternary a-methyl-
alcoholic moiety at P1 position, and several different R1
groups at P10.4 Interestingly, the other stereoisomers,
including the epimers at the quaternary carbinol func-
tion, showed much lower activity, as the authors dem-
onstrated that the hydroxamic binding function was
moved away from the catalytic Zn2þ center.
Within the frame of a project aimed at studying the
Ôfluorine effectÕ in peptides and identifying selective
Incorporation of fluorine into organic molecules is an
effective strategy for improving and modifying the bio-
logical activity.2 In particular, the trifluoromethyl group
is recognized in medicinal chemistry as a substituent of
distinctive qualities. It is in fact at the same time highly
hydrophobic, electron rich, sterically demanding,
moreover it can provide high in vivo stability, and fea-
P1
O
R2
HO
O
R
H
N
NHR3
P3'
HO
N
H
R1
O
P1'
R = CH3;
P2'
Keywords: Fluorine; Peptidomimetics; Matrix metalloproteinases.
* Corresponding authors. Tel.: +39-02-23993084; fax: +39-02-23993-
080; e-mail addresses: alessandro.volonterio@polimi.it; matteo.
zanda@polimi.it
A
1 R = CF3
Figure 1. The DuPont MerckÕs MMP inhibitors (A) and their CF3-
analogs (1).
0040-4039/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.12.131

1612
M. Sani et al. / Tetrahedron Letters 45 (2004) 1611–1615
fluorinated inhibitors of aspartic proteinases and
MMPs, we have recently described the synthesis, as well
as the conformational and biological properties of sev-
eral new types of fluorinated peptidomimetics.5 We
hypothesized that incorporation of fluoroalkyl substit-
uents on the P1 quaternary position of the inhibitors A
could be an effective strategy for optimizing and tuning
their binding properties, thus increasing the selectivity
toward the different MMPs. Now we describe the syn-
thesis of the Tfm-analogs 1 (Fig. 1) of A, and the effect
of the replacement of the a-CH3 group with a CF3 on
the inhibition of MMP-9.
S
O F3C
OH
CO2Et
O
N
S
O
4
Bn
O
N
Ph
Major
Ph
TiCl (2 equiv)
4
+
Bn
DIPEA (2 equiv)
2
S
+
O F3C
OH
CO2Et
CH Cl (dry)
2
2
CF3
O
N
0°C, then -70°C,
then rt
O
CO2Et
5
Bn
Minor
d.r. 2.7:1.0
yield 72% (hundreds
mgs scale), 90% (10
grams scale)
3
Ph
We decided to concentrate our efforts on the substrates 1
having R1 ¼ (CH2)4 Ph, whose analogs A were reported
to be very active. The a-Tfm-malic unit of 1 was recently
obtained by our group via titanium (IV) catalyzed aldol
reaction of trifluoropyruvic esters with enantiopure
N-acyl oxazolidin-2-ones.6 Although this reaction was
per se satisfactory (61–87% yields, dr up to 8:1
depending on the N-acyl group), the subsequent exocy-
clic cleavage of the oxazolidin-2-one auxiliary could not
be performed, despite intensive efforts. We therefore
decided to exploit the potential of oxazolidin-2-thiones,7
whose cleavage has been reported to occur much more
smoothly.8
Scheme 1. The aldol reaction to form the a-Tfm-malic framework.
partial (17%) a-epimerization to give ent-7. Even less
effectively, the same reaction performed on 5 gave 56%
yield of the diastereomeric Bn-ester 7, containing 33% of
the a-epimer ent-6. Although the unreacted starting
materials 4 and 5 could be recovered unchanged in good
yields, we felt that more efficient conditions were needed
in order to complete the synthesis. Disappointingly,
exploration of several different combinations of alco-
hols, solvents, and bases did not improve the situation
(Scheme 2).
The TiCl4 catalyzed reaction of the N-acyl-oxazolidin-2-
thione 2 (Scheme 1) with ethyl trifluoropyruvate 3
afforded the two diastereomeric adducts 4 and 5, out of
four possible, in low diastereomeric ratio. It is worth
noting that the reaction features a favorable scale-up
effect, affording ca. 70% yield on a 100-mg scale, and
90% on a 10-g scale (the reaction was repeated many
times on both scales). A number of alternative condi-
tions were explored, but neither significant improvement
nor switch of the diastereocontrol could be achieved.9
However, we were glad to find that the solid K2CO3 in
moist dioxane (rt, 10–12 h), was able to produce directly
the key carboxylic acid intermediates 8 and 9 (Scheme
3), from 4 and 5, respectively, in satisfactory yields and
with very low a-epimerization (2% for 8, 9% for 9).10
Coupling of the acid 8 with a-amino acid amides 10a–c
was achieved in good yields with the HOAt/HATU
system (Scheme 4).11 The resulting peptidomimetic
esters 11a–c were submitted to saponification, affording
the acids 12a–c in high yields.
Cleavage of the oxazolidin-2-thione was found to be
considerably more challenging than expected. In fact,
under the standard conditions reported in the literature
(BnOH, cat. DMAP, DCM, rt) the reaction on 4 was
very slow,8 affording modest conversion to the corre-
sponding Bn-ester 6 (60%) after one week at reflux, with
The subsequent coupling of 12a–c with O-Bn hydroxyl-
amine proved to be extremely challenging, owing to the
low reactivity and high steric hindrance of the carboxylic
F3C
F3C
S
O F3C
OH
CO2Et
OH
CO2Et
OH
CO2Et
BnO2C
BnO2C
BnOH (2 equiv)
O
N
+
DCM, reflux
1 week (60%)
6
ent-7
4
Bn
Ph
Ph
Major
Ph
4.8:1.0
F3C
S
N
F3C
O F3C
OH
CO2Et
OH
CO2Et
OH
CO2Et
BnOH (2 equiv)
DCM, reflux
BnO2C
BnO2C
O
+
7
5
ent-6
1 week
(56%)
Bn
Minor
Ph
Ph
2:1
Ph
Scheme 2. Attempted oxazolidin-2-thione cleavage with BnOH.

M. Sani et al. / Tetrahedron Letters 45 (2004) 1611–1615
1613
Thus, we decided to prepare first the intermediate 15
(72%), which could be purified by short flash chroma-
tography (FC). The latter was reacted with free 10a,
affording the desired molecule 16 in high yields.14
S
O F3C
F3C
OH
CO2Et
OH
CO2Et
K CO (1.5 equiv)
HO2C
(71%)
2
3
O
O
N
moist dioxane
10 hours
4
8
Bn
Major
Ph
Ph
Saponification of the ester 16 occurred effectively, but
disappointingly
a
partial epimerization of the
[Ph(CH2)3] stereocenter occurred, affording a 3:1 mix-
ture of diastereomers 17 and 18 under optimized con-
ditions. Since their chromatographic separation proved
to be difficult, 17 and 18 were subjected together
to coupling with BnONH2 under the previously opti-
mized conditions. The resulting diastereomeric O-Bn
hydroxamates could be separated by FC, affording pure
19 (52%), that was hydrogenated to the target free
hydroxamate 20 in 83% yield.
S
N
F3C
O F3C
OH
CO2Et
OH
CO2Et
K CO (1.5 equiv)
HO2C
(69%)
2
3
moist dioxane
12 hours
9
5
Bn
Minor
Ph
Ph
Scheme 3. Cleavage of the oxazolidin-2-thione auxiliary with K2CO3
in moist dioxane.
The hydroxamates 14a–c and 20 where tested for their
ability in inhibiting MMP-2 and MMP-9 activity using
zymographic analysis. The four hydroxamates inhibited
in a dose-dependent manner the gelatinolytic bands at
92 and 72 kDa, corresponding, respectively, to pro-
MMP-9 and pro-MMP-2 released in the conditioned
medium by human melanoma cells WM983A. The IC50
values (lM) portrayed in Table 1 show that diastereo-
mers 14a–c displayed low inhibitory activity, in line with
the parent CH3 compounds. Disappointingly, 20
showed a much lower activity than the exact CH3 analog
A, that was reported to be a low nanomolar inhibitor of
MMP-9 (IC50 <1 nM toward MMP-9).4a
group bound the quaternary a-Tfm carbinolic center. A
number of ÔconventionalÕ coupling agents for peptides12
were tested (among them DCC/DMAP, EDC/HOBt,
DIC/HOBt, HATU/HOAt, PyBroP/DIPEA) but no
trace of the target O-Bn hydroxamates 13a–c could be
obtained. Finally, we found that freshly prepared
BrPO(OEt)2 was able to promote the coupling in rea-
sonable yields (32–61%).13 With 13a–c in hand we
addressed the final O-Bn cleavage by hydrogenolysis,
that provided the hydroxamates 14a–c in good yields.
Since 14a–c are the ÔwrongÕ diastereomers with respect
to A, we deemed necessary to synthesize at least one
analog having the correct stereochemistry, in order to
have a complete set of biological data on the effect of the
introduction of the Tfm group. However, a tailored
synthetic protocol had to be developed ex-novo, because
the minor diastereomer 9 (Scheme 5) featured a dra-
matically different reactivity in the key steps of the
synthesis. First of all, we noticed that the coupling of 9
and 10a with HATU/HOAt gave rise to relevant
amounts of the b-lactone 15, which had to be processed
separately, besides the expected coupling product 16.
It is also worth noting that 14a and 20 showed little
selectivity, whereas 14b and 14c showed a fairly better
affinity for MMP-9, in comparison with MMP-2. A
possible explanation for the drop of activity upon
replacement of the quaternary methyl with a CF3, is that
20, biased by the bulky CF3 group, is unable to assume
the crucial binding conformation of the CH3 analog A.
Alternatively, one can hypothesize that the bulky and
highly electron-rich CF3 group is unable to fit the S1
pocket of the hitherto tested MMPs.
CF3
HO
EtO2C
CO2H
R2
CF3
O
R2
CF3
O
HO
HO
KOH 0.5 N
NHCH3
8
NHCH3
MeOH/H2O 7:3
HOAt/HATU EtO2C
N
H
HO2C
N
H
Ph
R2
O
O
(92-96%)
TMP, DMF
Ph
+
Ph
12a-c
11a R2 = t-Bu (71%)
11b R2 = p-MeO-Bn (65%)
NHCH3
11c R2 = (CH2)3NHBoc (68%)
H2N
O
10a-c
CF3
O
R2
HO
CF3
O
R2
HO HN
NHCH3
HO
BnONH2
BrPO(OEt)2
N
H2/Pd(OH)2
MeOH, rt
NHCH3
H
BnOHNOC
N
H
O
O
O
TMP, DMF
CH2Cl2
Ph
Ph
13a R2 = t-Bu (61%)
14a R2 = t-Bu (87%)
13b R2 = p-MeO-Bn (60%)
13c R2 = (CH2)3NHBoc (32%)
14b R2 = p-MeO-Bn (79%)
14c R2 = (CH2)3NHBoc (60%)
Scheme 4. Synthesis of the peptidomimetics 14 from the major diastereomer 8.

1614
M. Sani et al. / Tetrahedron Letters 45 (2004) 1611–1615
F3C
O
CF3 O
KOH 0.5 N
OH
CO2Et
HO
CF3
CO2Et
10a, TMP
HOAt/HATU
TMP, DMF
O
HO2C
MeOH/H2O 7:3
NHCH3
EtO2C
N
H
DMF
O
9
15
Ph
Ph
16
(89%)
(72%)
Ph
CF3
CF3
O
HO
NHCH3
HO2C
N
H
1. BnONH2, BrPO(OEt)2,
TMP, DMF, CH2Cl2
CF3
O
O
HO
Ph
H2/Pd(OH)2
MeOH, rt
17
NHCH3
+
BnOHNOC
N
H
2. Flash chromatogr.
O
Ph
19
(52%)
O
HO
NHCH3
HO2C
N
H
O
18
Ph
CF3
O
HO
O
3:1 mixture (90% overall)
HO HN
NHCH3
N
H
O
(83%)
20
Ph
Scheme 5. Synthesis of the target peptidomimetic 20 from the minor diastereomer 9.
3. Banks, R. E.; Tatlow, J. C.; Smart, B. E. Organofluorine
Chemistry: Principles and Commercial Applications;
Plenum Press: New York, 1994.
Table 1. IC50 (lM) of the target Tfm-hydroxamates
Compound
MMP-2
MMP-9
4. (a) Jacobson, I. C.; Reddy, P. G.; Wasserman, Z. R.;
Hardman, K. D.; Covington, M. B.; Arner, E. C.;
Copeland, R. A.; Decicco, C. P.; Magolda, R. L. Bioorg.
Med. Chem. Lett. 1998, 8, 837–842; (b) Jacobson, I. C.;
Reddy, G. P. Tetrahedron Lett. 1996, 37, 8263–8266.
5. See for example: (a) Molteni, M.; Volonterio, A.; Zanda,
M. Org. Lett. 2003, 5, 3887–3890; (b) Volonterio, A.;
14a
14b
14c
20
156
407
722
23
121
84
23
15
The preparation of other CF3 peptidomimetics 1 is
currently underway in order to have a more complete
picture of the effect of the CF3 introduction in this class
of peptidomimetics. Computational studies are also in
progress in order to understand the reasons for the
unfavorable effect of the CF3 group.15
ꢀ
Bellosta, S.; Bravin, F.; Bellucci, M. C.; Bruche, L.;
Colombo, G.; Malpezzi, L.; Mazzini, S.; Meille, S. V.;
ꢀ
Meli, M.; Ramırez de Arellano, C.; Zanda, M. Chem. Eur.
J. 2003, 9, 4510–4522; (c) Sani, M.; Bruche, L.; Chiva, G.;
ꢀ
Fustero, S.; Piera, J.; Volonterio, A.; Zanda, M. Angew.
Chem. Int. Ed. 2003, 42, 2060–2063.
6. Zucca, C.; Bravo, P.; Malpezzi, L.; Volonterio, A.; Zanda,
M. J. Fluorine Chem. 2002, 114/2, 215–223.
7. (a) Crimmins, M. T.; King, B. W.; Tabet, E. A.;
Chaudhary, K. J. Org. Chem. 2001, 66, 894–902; (b)
Crimmins, M. T.; McDougall, P. J. Org. Lett. 2003, 5,
591–594, and references cited therein; (c) Palomo, C.;
Acknowledgements
ꢀ
Oiarbide, M.; Garcıa, J. M. Chem. Eur. J. 2002, 8, 37–44.
We thank the European Commission (IHP Network
grant ‘‘FLUOR MMPI’’ HPRN-CT-2002-00181),
MIUR (Cofin 2002, Project ‘‘Peptidi Sintetici Bioat-
tivi’’), Politecnico di Milano, and CNR for economic
support.
8. Su, D.-W.; Wang, Y.-C.; Yan, T.-H. Tetrahedron Lett.
1999, 40, 4197–4198.
9. Among the conditions explored: TiCl4/())-sparteine (74%
yield, 4:5 ¼ 1.0/1.6); Sn(OTf)2/NEt3 (no reaction); Bu2
BOTf/NEt3 (no reaction); LDA (48% yield, 4:5 ¼ 2.6/1.0).
10. The stereochemistry of the a-Tfm-malic derivatives
described herein was unambiguously determined by
X-ray diffraction of a crystalline dipeptide derivative of
8, not shown in this paper. Full details will be reported in a
forthcoming full paper.
References and notes
1. (a) Coussens, L. M.; Fingleton, B.; Matrisian, L. M.
Science 2002, 295, 2387–2392; (b) Whittaker, M.; Floyd,
C. D.; Brown, P.; Gearing, A. J. H. Chem. Rev. 1999, 99,
2735–2776; (c) Bode, W.; Huber, R. Biochim. Biophys.
Acta. 2000, 1477, 241–252; (d) Giavazzi, R.; Taraboletti,
G. Crit. Rev. Oncol. Hematol. 2001, 37, 53–60.
2. Ojima, I.; McCarthy, J. R.; Welch, J. T. Biomedical
Frontiers of Fluorine Chemistry; ACS Books, American
Chemical Society: Washington, DC, 1996.
11. Carpino, L. A.; El Faham, A. J. Org. Chem. 1995, 60,
3561–3564, and references cited therein.
12. Humphrey, J. M.; Chamberlin, A. R. Chem. Rev. 1997, 97,
2243–2266.
13. (a) Gorecka, A.; Leplawy, M.; Zabrocki, J.; Zwierzak, A.
Synthesis 1978, 474–476, Contrary to the statement of
these authors, BrPO(OEt)2 undergoes rapid decomposi-
tion even at 4 °C, therefore it must be used within 1–2 h
from its preparation. Our results are in agreement with an

M. Sani et al. / Tetrahedron Letters 45 (2004) 1611–1615
1615
earlier report; (b) Goldwhite, H.; Saunders, B. C. J. Chem.
Soc. 1955, 3564.
14. For the addition of amine nucleophiles to b-lactones see:
Nelson, S. G.; Spencer, K. L.; Cheung, W. S.; Mamie, S.
J. Tetrahedron 2002, 58, 7081–7091.
15. Key to the abbreviations and acronyms used in this
paper: DCM ¼ dichloromethane; TMP ¼ sym-collidine
(2,4,6-trimethylpyridine); DCC¼ dicyclohexylcarbodiimide;
DMAP ¼ 4-(N,N-dimethylamino)pyridine;
EDC ¼ 1-(3-
dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride;
HOBt ¼ 1-hydroxybenzotriazole; DIC ¼ diisopropylcarbodi-
imide;
N0-tetramethyluronium hexafluorophosphate; HOAt ¼
HATU ¼ O-(7-azabenzotriazol-1-yl)-N,N,N0,
1-hydroxy-7-azabenzotriazole; PyBroP ¼ Bromotripyrrol-
idinophosphonium
hexafluorophosphate;
DIPEA ¼
diisopropylethylamine; DMF ¼ N,N-dimethylformamide.