An approach to?the?stereo-controlled synthesis of?polycyclic derivatives of?l-4-thiazolidinecarboxylic acid active against?HIV-1?integrase

Article abstract of DOI:10.1016/j.ejmech.2006.03.032

Herein, we describe a new strategy for the preparation of thiazolothiazepine-based inhibitors of human immunodeficiency virus type-1 integrase (IN). The present method allows facile preparation of the title compounds in a single enantiomeric form starting

Full text of DOI:10.1016/j.ejmech.2006.03.032

European Journal of Medicinal Chemistry 41 (2006) 914–917
http://france.elsevier.com/direct/ejmech
Original article
An approach to the stereo-controlled synthesis of polycyclic derivatives
of L-4-thiazolidinecarboxylic acid active against HIV-1 integrase
F. Aiello a, A. Brizzi b, O. De Grazia a, A. Garofalo a,*, F. Grande a, M.S. Sinicropi a
R. Dayam c, N. Neamati c
a Dipartimento di Scienze Farmaceutiche, Università della Calabria, 87036 Arcavacata di Rende (Cs), Italy
b Dipartimento Farmaco Chimico Tecnologico, Università di Siena, Via A. Moro, 53100 Siena, Italy
c Department of Pharmaceutical Sciences, School of Pharmacy, University of Southern California, Los Angeles, CA 90089, USA
Received 18 October 2005; received in revised form 10 March 2006; accepted 20 March 2006
Available online 15 June 2006
Abstract
Herein, we describe a new strategy for the preparation of thiazolothiazepine-based inhibitors of human immunodeficiency virus type-1 inte-
grase (IN). The present method allows facile preparation of the title compounds in a single enantiomeric form starting from L-4-thiazolidinecar-
boxylic acid. This method could be easily extended to the synthesis of several analogs derived from optically active cyclic aminoacids. We also
present a putative model showing the interaction between L- and D-isomers of compound 1 in the IN active site. A sensibly lower IC50 value was
found for (–)-1 over racemic-1 in an anti-IN assay.
© 2006 Elsevier SAS. All rights reserved.
Keywords: HIV-1 integrase inhibitors; L-4-thiazolidinecarboxylic acid; Thiazolothiazepines
1. Introduction
recombinant IN and thus far two agents have been tested in the
clinic [5,6].
Thiazolothiazepines, previously prepared by us in a racemic
mixture, despite of their low potency, represent a class of com-
pounds with interesting characteristics, such as: 1) activity in
the presence of Mg2+ as a cofactor, 2) activity in cell-based
antiviral assays, 3) low cytotoxicity, and 4) good selectivity
against IN [7]. Such properties prompted us to optimize these
molecules by designing analogs bearing several structural
alterations, with the main aim at increasing potency [8]. The
structures of the two original leads 1 and 2 are shown in Fig. 3.
Currently, chemotherapeutic intervention for the treatment
of HIV infection is based on the administration of drugs target-
ing reverse transcriptase (RT) and protease (PR) [1]. Combina-
tions of RT and PR inhibitors have made a major impact in the
life of many AIDS patients. However, this approach does not
eradicate the virus. Because of persistent toxicity, difficulties in
patient adherence and development of drug-resistance, there is
an urgent need to develop novel therapeutics targeting other
viral proteins with better safety profiles [2].
HIV-1 integrase (IN) is such an attractive target and is
essential for viral replication. IN mediates the insertion of ret-
roviral DNA into host chromosomes by a complex process
consisting of two fundamental reactions known as 3′-proces-
sing and strand transfer [3,4]. The in vitro assaying of a large
number of compounds was made possible by the availability of
As a crucial step for the development of more potent inhi-
bitors, we wanted to know which enantiomers of compound 1
and 2 is responsible for activity. Previously, our most potent
thiazolothiazepine, 2, was obtained as an enantiomeric mixture
starting from ( )-thiazolidine-2-carboxylic acid. However, for
the preparation of compound 1, even though we started from
commercially available L-4-thiazolidinecarboxylic acid, the
optical activity was lost during the synthesis. The different
arrangement of the two enantiomers of compound 1 as well
* Corresponding author. Tel.: +39 0984 49 3118; fax: +39 0984 49 3298.
E-mail address: garofalo@unical.it (A. Garofalo).
0223-5234/$ - see front matter © 2006 Elsevier SAS. All rights reserved.
doi:10.1016/j.ejmech.2006.03.032

F. Aiello et al. / European Journal of Medicinal Chemistry 41 (2006) 914–917
915
Fig. 1. S and R enantiomers of compound 1 are shown side by side (a) and
superimposed (b).
Fig. 4. Retrosynthetic approaches for compound 1.
as their bound conformation inside the HIV-1 IN active site are
shown in Figs. 1 and 2.
The previous synthesis of 1 consisted of two subsequent
condensation steps between L-4-thiazolidinecarboxylic acid
and 3-mercapto-2-naphthoic acid (Fig. 4, path A) [7]. The for-
mation of the thiolactone bond by means of N,N′-carbonyldii-
midazole (CDI) required a fairly harsh condition (i.e.: pro-
longed reaction time and high temperature). Under this
condition a racemic mixture was obtained because of the rigid-
ity caused by the preformed amide moiety that hindered the
closure of the seven-membered ring. Similar results were
obtained during the preparation of other analogs using path A
[8]. Previously, when other precursors such as various L-4-oxa-
zolidinecarboxylic acids [9] and aromatic o-hydroxy carboxylic
acids were condensed, we always obtained racemic mixtures.
However, none of the compounds where heterocyclic sulfur
was replaced with oxygen(s) showed superior activity over
the parent compounds.
Therefore, a new synthetic route aimed at avoiding racemi-
zation was undertaken. First we wanted to form the thiophenol
ester bond in milder conditions so that the chiral center would
not be affected. In this case the seven-membered ring closure
reaction by the subsequent formation of the lactam moiety
would lead to the desired homochiral compounds (Fig. 4,
path B).
Fig. 2. The predicted bound conformation of S and R enantiomeric forms of
compound 1 inside the HIV-1 IN active site. The two isomers are shown as
stick models in green (S isomer), gray (R isomer). The yellow ribbon model
represents active site region of the HIV-1 IN. The prominent active site amino
acid residues are shown as stick models while magenta sphere represents Mg2+
.
Fig. 3. Chemical structures of the lead compounds.

916
F. Aiello et al. / European Journal of Medicinal Chemistry 41 (2006) 914–917
Scheme 1. Reagents and conditions: a) BOP/DIEA/DCM/rt/2 h (42%); b) TFA/
anisole (excess)/rt; c) PyBOP/DIEA/DMF/1 h (35%).
Scheme 2. Reagents and conditions: a) 3/BOP/TEA/DCM/rt/3 h (65%); b)
TFA/anisole (excess)/rt; c) PyBOP/DIEA/DMF/1 h (75%).
However, several synthetic challenges have to be resolved
in this case. For example, appropriate protections for the sec-
ondary amino group of the aminoacid as well as the aromatic
carboxylic group were required prior to the thiophenol ester
bond formation. After such consideration, the synthetic path-
way depicted in Scheme 1 was pursued.
ditions, to give essentially pure L-2-[(1,3-thiazolidin-4-ylcarbo-
nyl)oxy]benzoic acid 10. The final cyclization step was carried
out with PyBOP/DIEA in DMF to give optically pure L-(–)-7
in a 75% yield; [α]20D = –111 (CHCl3, c = 1); ee = > 97%. The
present procedure could therefore be used for the synthesis of
homochiral compounds, when the starting cyclic amino acid is
available in a single enantiomeric form.
2. Chemistry
N-Boc-L-4-thiazolidinecarboxylic acid 3 and the tert-butyl
3-mercapto-2-naphthoate 4 both bearing a protective group
easily removable in mild hydrolytic conditions were selected
as ideal precursors. Their condensation to form the correspond-
ing N-Boc-D-4-{[2-(tert-butoxycarbonyl)naphthalen-3-ylthio]
carbonyl}-1,3-thiazolidine 5 was accomplished with BOP/
DIEA in DCM, under nitrogen and in dark condition, in a
42% yield [10]. The fully protected intermediate was purified
by column chromatography on silica gel and then amino and
carboxylic groups were deprotected simultaneously by trifluor-
oacetic acid and excess anisole at room temperature. The thio-
phenol ester bond was not affected under this condition.
Because of easy removal of excess reagents no further purifi-
cation of D-3-[(1,3-thiazolidin-4-ylcarbonyl)thio]-2-naphthoic
acid 6 was needed. Compound (–)-1 was obtained by a cycli-
zation reaction attempted under different experimental condi-
tions. The best result was achieved by using PyBOP/DIEA as
an intramolecular coupling reagent in DMF [11]. Optical activ-
ity and purity measurements on the final compound confirmed
the stereo-controlled synthesis: [α]20D = –91 (CHCl3, c = 0.8);
ee = > 93%. Enantiomeric excesses were determined by 1H
NMR spectroscopy using europium tris[3-(heptafluoropropyl-
hydroxymethylene)-(+)-camphorate].
3. Results and discussion
Racemic 1 and pure (–)-1 have been singly subjected to
preliminary anti-IN evaluation in order to detect any possible
difference of potency. The extent of 3′-processing and strand
transfer was essentially determined as described in Ref. [8].
In earlier experiments, ( )-1 showed IC50 (μM) = 92 and
100 for the 3′-processing and strand transfer, respectively, in
the inhibition of HIV-1 IN catalytic activity [7]. In the same
conditions, pure (–)-1 showed IC50 (μM) = 65 and 88, respec-
tively.
The comparison of such values proves that both enantio-
mers possess inhibitory activity, though a greater potency has
to be ascribed to pure (–)-1. Thus (–)-1 may be considered the
eutomer. By merely theoretical conjectures, an eudismic
ratio = 2.4 (–/+) could be then calculated in the inhibition of
the 3′-processing step, assuming that both stereoisomers
would act as inhibitors at the active site, each being indepen-
dent from the other. More definitive conclusions could be
drawn after the synthesis and testing of the remaining stereoi-
somer (+)-1.
To investigate the generality of this method, we prepared a
related benzoxazepine. In particular the synthesis of L-(–)-
In conclusion, these preliminary results would prove that
chirality is a crucial feature to be taken into consideration for
optimizing interaction of thiazolothiazepine-based IN-inhibi-
tors inside the IN active site. Along with other structural mod-
ifications, the development of stereoselective synthetic proce-
dures therefore represents an additional opportunity for the
production of even more active compounds.
1,11a-dihydro-5H,11H-[1,3]thiazolo[4,3-c]
[1,4]benzoxaze-
pine-5,11-dione 7 was undertaken by only slight modifications
in the above reaction sequence (Scheme 2). Condensation
between N-Boc-L-4-thiazolidinecarboxylic 3 and tert-butyl sal-
icilate 8, using the BOP/TEA system in DCM [12], gave the
corresponding ester 9, which was deprotected in the above con-

F. Aiello et al. / European Journal of Medicinal Chemistry 41 (2006) 914–917
917
4. Experimental section
A solution of the hydroxyacid 10 (0.10 g, 0.39 mmol) in dry
DMF (2 ml) is slowly added to a solution of PyBOP (0.20 g,
0.39 mmol) and DIEA (0.05 g, 0.39 mmol) in dry DMF (5 ml).
The mixture is stirred for 1 h and then the solvent is evaporated
under vacuum. The residue obtained is partitioned between
water and ethyl acetate and the organic layer is separated and
washed with 1 N hydrochloric acid and water. After evapora-
tion of the solvent, the oily residue is purified by column chro-
matography (ethyl acetate/hexanes, 1:6) to give pure com-
pound (–)-7 (70 mg, 75% yield) which shows identical data
as reported in Ref. [8], but [α]20D = –111 (CHCl3, c = 1),
ee = > 97%.
4.1. Chemistry
For general experimental informations see Ref. [8]. Ana-
lyses indicated by the symbols of the elements were within
0.4% of the theoretical values.
4.2. Synthetic procedure
A typical procedure is described for the synthesis of com-
pound (–)-7. N-Boc-L-4-thiazolidinecarboxylic 3 (1.40 g,
6 mmol), tert-butyl salicilate 8 (1.17 g, 6 mmol), triethylamine
(1.21 g, 12 mmol) and BOP reagent (2.65 g, 6 mmol) are dis-
solved successively in dichloromethane (45 ml). The solution
is kept at room temperature for 3 h. Saturated aqueous sodium
chloride solution (100 ml) is added and the resultant mixture is
extracted with ethyl acetate. The organic phase is separated,
washed successively with 1 N hydrochloric acid, saturated
sodium hydrogen carbonate solution and sodium chloride solu-
tion, dried (MgSO4) and evaporated in vacuo. Pure N-Boc-L-[2-
References
[1] E. De Clercq, Nat. Rev. Drug Discov. 1 (2002) 13–25.
[2] D.D. Richman, Nature 410 (2001) 995–1001.
[3] A.M.E. Skalka, Adv. Virus Res. 52 (1999) 271–475.
[4] P.O. Brown, Integration, Cold Spring Harbor Press, Cold Spring Harbor,
1999.
[5] R. Dayam, N. Neamati, Curr. Pharm. Des. 9 (2003) 1789–1802.
(tert-butoxycarbonyl)phenyl]-1,3-thiazolidine-4-carboxylate
was obtained as a white solid after column chromatography
purification on silica gel (diethyl ether/light petroleum, 1:1 as
9
[6] D.J. Hazuda, S.D. Young, J.P. Guare, N.J. Anthony, R.P. Gomez, J.S.
Wai, J.P. Vacca, L. Handt, S.L. Motzel, H.J. Klein, G. Dornadula, R.M.
Danovich, M.V. Witmer, K.A. Wilson, L. Tussey, W.A. Schleif, L.S.
Gabryelski, L. Jin, M.D. Miller, D.R. Casimiro, E.A. Emini, J.W. Shiver,
Science 305 (2004) 528–532.
1
eluent); 1.60 g, 65% yield. M.p. 89.5 °C. H NMR (300 MHz,
CDCl3, δ ppm) 1.51 (s, 9 H), 1.62 (s, 9 H), 3.48 (m, 1 H), 3.75
(m, 1 H), 4.55 (t, 1 H, J = 1.7), 4.70 (dd, 1 H, J = 6.5, 1.7),
5.05 (d, 1 H, J = 6.5), 7.2–7.9 (m, 4 H). Anal. C20H27NO6S (C,
H, N).
[7] N. Neamati, J.A. Turpin, H.E. Winslow, J.L. Christensen, K. William-
son, A. Orr, W.G. Rice, Y. Pommier, A. Garofalo, A. Brizzi, G. Cam-
piani, I. Fiorini, V. Nacci, J. Med. Chem. 42 (1999) 3334–3341.
[8] F. Aiello, A. Brizzi, A. Garofalo, F. Grande, G. Ragno, R. Dayam, N.
Compound 9 (0.20 g, 0.49 mmol) is dissolved in a mixture
of trifluoroacetic acid/anisole, 1:1 (20 ml) at 0 °C. The mixture
is then stirred at room temperature for 50 min. The volatiles are
evaporated under vacuum and the residue is azeotropized sev-
eral times with toluene to give essentially pure 2-[(1,3-thiazo-
lidin-4-ylcarbonyl)oxy]benzoic acid 10 which is subjected to
cyclization without further purification.
Neamati, Bioorg. Med. Chem. 12 (2004) 4459–4466.
[9] M. Falorni, S. Conti, G. Giacomelli, S. Cossu, F. Soccolini, Tetrahedron
Asymmetry 6 (1995) 287–294.
[10] Y. Sun, G. Lu, J.P. Tam, Org. Lett. 3 (2001) 1681–1684.
[11] T. Rückle, P. de Lavallaz, M. Keller, P. Dumy, M. Mutter, Tetrahedron
55 (1999) 11281–11288.
[12] B. Castro, G. Evin, C. Selve, R. Seyer, Synthesis 6 (1977) 413.