Synthesis and in vitro evaluation of novel small molecule inhibitors of bacterial arylamine N-acetyltransferases (NATs)

Article abstract of DOI:10.1016/S0960-894X(03)00484-0

The synthesis and inhibitory activity of a series of 5-substituted-(1,1-dioxo-2,3-dihydro-1H-1λ⁶-benzo[e] [1,2]thiazin-4-ylidene)-thiazolidine-2,4-dione derivatives as competitive inhibitors of recombinant bacterial arylamine-N-acetyltransferas

Full text of DOI:10.1016/S0960-894X(03)00484-0

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2527–2530
Synthesis and In Vitro Evaluation of Novel Small Molecule
Inhibitors of Bacterial Arylamine N-Acetyltransferases (NATs)
Edward W. Brooke,b Stephen G. Davies,a,* Andrew W. Mulvaney,a Minoru Okada,c
Frederique Pompeo,b Edith Sim,b Richard J. Vickersa and Isaac M. Westwooda
aThe Dyson Perrins Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QY, UK
bDepartment of Pharmacology, University of Oxford, Mansfield Road, Oxford OX1 3QT, UK
cChemistry Laboratories, Institute for Drug Discovery Research, Yamanouchi Pharmaceutical Company Ltd.,
21 Miyukigoaka, Tsukuba, Ibaraki, 305-8585, Japan
Received 10 February 2003; accepted 6 May 2003
Abstract—The synthesis and inhibitory activity of a series of 5-substituted-(1,1-dioxo-2,3-dihydro-1H-1l6-benzo[e][1,2]thiazin-4-
ylidene)-thiazolidine-2,4-dione derivatives as competitive inhibitors of recombinant bacterial arylamine-N-acetyltransferases
(NATs) are described. The most potent NAT inhibitors are those that contain planar hydrophobic substituents on the sultam
nitrogen.
# 2003 Elsevier Ltd. All rights reserved.
Arylamine N-acetyltransferases (NATs) encode for a
gene family of cytosolic 32-34kDa enzymes which
acetylate arylamines, N-aryl hydroxylamines1 and aryl
hydrazines2 using acetyl-CoA as the acetyl donor.
NATs are widely distributed in eukaryotes and pro-
karyotes, with over 20 NATs identified so far in the
latter.3 In Humans one of the two identified NAT iso-
enzymes, NAT2, inactivates the front-line anti-tuber-
cular drug isoniazid (INH).4 Interindividual variation
in rates of drug metabolism2,5 and susceptibility to
arylamine carcinogens6 have been linked to poly-
morphisms in these two Human isoenzymes.7 Homo-
logues of Human NAT are present in Mycobacterium
tuberculosis and Mycobacterium smegmatis (MSNAT)
where they have been shown to acetylate, and hence
inactivate, INH.8 Overexpression of NAT in myco-
bacteria has been shown to increase their resistance to
INH9 and a nat gene knockout of M. smegmatis
increased sensitivity to the drug.10 The detection of a
NAT polymorphism in drug-resistant isolates of M.
tuberculosis further suggests that NAT may be involved
in INH resistance in tuberculosis.8 The resurgence of
drug-resistance in tuberculosis, especially in relation to
the AIDS pandemic, is
problem.11
a major global health
The crystal structure of NAT from Salmonella typhi-
murium (STNAT) has been reported12 and reveals the
protein to consist of three domains such that a cysteine, a
histidine and an aspartate residue (Cys70, His110, Asp127
MSNAT nomenclature) are juxtaposed to form a cata-
lytic triad, reminiscent of the cysteine proteases papain
and cathepsin.13 The crystal structure of MSNAT has
been resolved to a resolution of 1.7A and shows a near
identical three-dimensional structure to STNAT.14
;
We have instigated a chemical genetics15 approach to
determine the endogenous role of mycobacterial NAT
and have identified a novel class of small molecule inhibi-
tors of the pure recombinant enzyme from M. smegmatis
(MSNAT). High throughput screening of a proprietary
small molecule library16 identified the N-(2-naphthyl)-
methyl substituted 1,1 - dioxo - 2,3 - dihydrobenzo[1,2]-
thiazine-4-ylidene thiazolidine-2,4-dione (TZD-sultam
adduct) 1 to be a weak inhibitor of MSNAT-catalysed
acetylation of INH, with half-maximal inhibitory con-
centration (IC50) of 78mM. We report herein the synth-
esis, in vitro evaluation and preliminary structure–
activity relationships (SAR) of a series of TZD-sultam
adducts as competitive inhibitors of bacterial NATs.
*Corresponding author: Tel.: +44-1865-275695; fax: +44-1865-
275633; e-mail: steve.davies@chem.ox.ac.uk
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00484-0

2528
E. W. Brooke et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2527–2530
with 4-bromobenzyl bromide, followed by Heck and
Suzuki couplings respectively with styrene and phe-
nylboronic acid. Deprotection and condensation with
TZD as before furnished the corresponding TZD-sul-
tams 19 and 2020 (Scheme 2).
Initial structure–activity studies based around
1
revealed that NAT inhibitory activity resided exclu-
sively in compounds incorporating the N-substituted
TZD-sultam motif; both thiazolidine-2,4-dione (TZD)
and 2, the ketone precursor to 1, were devoid of NAT
inhibitory activity. Similarly, N-unsubstituted TZD-sul-
tam adduct 3 was found to be inactive against MSNAT
while N-alkylation of the TZD moiety in the sultam
adducts was also poorly tolerated.17
Scheme 2. Synthesis of biaryl and stilbenoid derivatives. Reagents and
conditions: (a) NaH, DMFrt, 0.5 h then 4-BrC 6H4CH2Br, rt 12 h; (b)
styrene, Pd(OAc)2 (5.0 mol%), P(o-tolyl)3 (20.0 mol%), tetra n-butyl
ammonium bromide, DMF:Et3N (1:1), 110 ꢀC, 16 h; (c) PhB(OH)2,
Pd(PPh3)4 (5.0 mol%), NaOH, DMF, 100 ꢀC, 32 h.
The N-substituted TZD-sultam adducts were evaluated
for their in vitro inhibitory activity against recombinant
MSNAT-catalysed acetylation of INH by measuring the
rate of hydrolysis of acetyl CoA using 5,50-dithio-bis(2-
nitrobenzoic acid),16,21 with spectrophotometric deter-
mination of the product, thio-2-nitrobenzoic acid, lmax
412 nm. The IC50 values are summarized in Table 1.
A series of TZD-sultam adducts was synthesized in
which the substituent at the sultam nitrogen atom was
systematically varied in order to determine the optimal
substituent for binding to MSNAT. Previous studies to
determine the substrate specificity of MSNAT revealed
lipophilic arylamines to be preferred.16 Thus a series of
lipophilic substituents were introduced at the sultam
nitrogen according to Scheme 1.
Kinetic analysis of the inhibition was performed by
varying the concentration of inhibitor and isoniazid
substrate. Inhibition was shown to be competitive with
respect to the isoniazid substrate (Fig. 1) and non-linear
regression was used to determine the inhibition con-
stant, Ki, for the more potent members of the series
(Table 2).
N-diversification was achieved by alkylation of ketal-
protected sultam 4, with commercially available alkyl
and benzylic bromides. Ketal 4 was synthesised in four
steps from saccharin.18 Deketalisation with methanolic
hydrogen chloride gave N-subsituted ketone 5 which
was subsequently condensed with the boron enolate of
thiazolidine-2,4-dione to afford the final TZD-sultams
6–18.19 The (E)-stilbenylmethyl and biphenylmethyl-
substituted sultams were synthesized from 4-bromo-
benzyl intermediate 21, obtained by alkylation of ketal 4
As the crystal structure for MSNAT is now available, in
silico docking studies were performed to determine the
orientation of binding of the TZD-sultam adducts to
the protein. Simulated annealing was achieved using the
AutoDock suite of programs as previously described22
and ten docking solutions were produced for each com-
pound. All the compounds gave exothermic energies of
binding that did not necessarily correlate with the level
of inhibition observed. However, it was observed that
the lowest energy conformations of the most potent
Scheme 1. Synthesis of TZD-sultam adducts. Reagents and condi-
tions: (a) NaOH, H2O, rt, 1 h; (b) MeCOCH2Cl, cetyl trimethyl-
ammonium bromide, PhMe, reflux, 12–16 h; (c) NaOEt, EtOH, 55 ꢀC,
0.5 h then HCl(aq), rt, 0.5 h; (d) ethane-1,2-diol, p-TsOH, C6H6,
reflux, 2 h; (e) NaH, DMFrt, 0.5 h then RCH 2Br, rt 12–16 h; (f) HCl,
MeOH, THF, reflux, 1 h; (g) thiazolidine-2,4-dione, boron trifluoride
diethyl etherate, triethylamine, 1,4-dioxane, 0 ꢀC then rt, 48 h.
Figure 1. Inhibition of MSNAT by 20. Double-reciprocal plot of rate
against isoniazid concentration in the presence of 20 at 25 mM (open
circles), 12.5 mM (filled circles), 6.25 mM (open squares) and 0 mM
(filled squares). Non-linear regression (KyPlot) shows the inhibition to
be competitive with kinetic parameter Ki=14.4 mM.

E. W. Brooke et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2527–2530
2529
Table 1. In vitro inhibition data for inhibition of recombinant
MSNAT-catalysed acetylation of INH for selected TZD-sultam NAT
inhibitors
Table 2. Kinetic parameters for competitive inhibition of MSNAT
with respect to isoniazid. Values were determined as in Figure 1
Compd
Ki (mM)
1
9
10
18
19
20
70Æ4
43Æ5
14.7Æ1
26.7Æ2
14.0Æ1
14.4Æ1
Compd
R substituent
IC50 (mM)a
78Æ5
1
TZD
—
—
H
Me
na
nab
>100
>100
92Æ5
2
3
6
7
8
>100
Figure 2. Simulated annealing of 20 to the MSNAT crystal structure.
Figure shows the molecular surface of the protein calculated by Swiss
PDB Viewer, including the side-chains mentioned in the text, and
the lowest energy docking solution of 20 displayed in standard CPK
colouring.
9
10
11
n-Octyl
n-Decyl
61Æ8
22Æ2
>100
12
13
>100c
inhibitors, listed in Table 2, all showed very similar
orientations of binding (shown for 20 in Fig. 2). These
docking solutions predict the sultam moiety to be
oriented nearest to the active-site cysteine hence block-
ing binding of the substrate. The variable R group on
nitrogen extends into a long groove in the protein, held
in place by a tryptophan residue (Trp97). Contact resi-
82Æ10
14
15
Bn
>100
>100
dues around the groove are Phe38, Phe130, Gln133
,
Val169, His203, Phe204, Asn220 and His229. The side-chain
of His203 appears to be involved in a p-stacking inter-
action with the thiazolidine-2,4-dione group of the
inhibitor, explaining the lack of activity with the ketal
and ketone derivatives.
16
17
18
19
>100
65Æ5
33Æ1
29Æ2
The hydrophobic residues here (Trp, Phe, Val) explain
the structure–activity relationship observed, with large
hydrophobic (long chain aliphatic and planar aromatic)
R-groups on the sultam-TZD adduct providing the
greatest inhibition due to non bonding, hydrophobic–
hydrophobic and p-stacking interactions between inhib-
itor and protein. However, there appears to be further
potential to improve the potency of this class of NAT
inhibitor through targeting of the interactions with the
His and Asn residues flanking the active site.
In conclusion we have shown that the TZD-sultam
adducts are competitive inhibitors of the acetylation of
the anti-tubercular drug isoniazid by mycobacterial
NATs. Ongoing work will determine whether these
compounds can be used to emulate the phenotype of a
genetic knockout in NATs from a range of sources, and
whether they affect the sensitivity of the mycobacteria to
isoniazid.
20
20Æ1
aIC50 value were determined from direct regression curve analysis.
bna, not active.
cIC50=205 mM.

2530
E. W. Brooke et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2527–2530
10. Payton, M.; Gifford, C.; Schartau, P.; Hagemeier, C.;
Acknowledgements
Mushtaq, A.; Lucas, S.; Pinter, K.; Sim, E. Microbiology 2001,
147, 3295.
11. Dye, C.; Scheele, S.; Dolin, P.; Pathania, V.; Raviglione,
M. C. JAMA 1999, 282, 677.
12. Sinclair, J. C.; Sandy, J.; Delgoda, R.; Sim, E.; Noble,
M. E. Nature Structural Biology 2000, 7, 560.
13. Otto, H.-H.; Schrimeister, T. Chem. Rev. 1997, 97, 133.
14. Sandy, J.; Mushtaq, A.; Kawamura, A.; Sinclair, J.; Sim,
E.; Noble, M. E. J. Mol. Biol. 2002, 318, 1071.
The authors wish to thank the MRC for a studentship
(E.W.B.), the MRC and EPSRC for a Discipline-Hop-
per Award (R.J.V.), the Blaschko Foundation for a
Fellowship (F.P.) and the Wellcome Trust for financial
support (E.S.).
15. (a) Specht, K. M.; Shokat, K. M. Curr. Opin. in Cell Biol.
2002, 14, 155. (b) Smith, A. Nat. Rev. Drug Disc. 2002, 1, 95.
(c) Stockwell, B. R. Nat. Rev. Genetics 2000, 1, 116.
16. Brooke, E. W.; Davies, S. G.; Mulvaney, A. W.; Pompeo,
F.; Sim, E. Bioorg. Med. Chem. 2003, 11, 1227.
References and Notes
1. Hein, D. W.; Doll, M. A.; Rustan, T. D.; Gray, K.; Feng,
Y.; Ferguson, R. J.; Grant, D. M. Carcinogenesis 1993, 1,
1633.
17. Data not shown.
2. Weber, W. W.; Hein, D. W. Pharmacol. Rev. 1985, 37, 25.
3. Payton, M.; Mushtaq, A.; Yu, T.-W.; Wu, L. J.; Sinclair,
J.; Sim, E. Microbiology 2001, 147, 1137.
4. Evans, D. A. P.; Manley, K. A.; McKuisick, V. A. Brit.
Med. J. 1960, 2, 485.
5. (a) Peters, J. H.; Miller, K. S.; Brown, P. J. Pharmocol.
Exp. Ther. 1965, 150, 298. (b) Jenne, J. W. J. Clin. Invest.
1965, 44, 1992.
6. Hein, D. W.; Doll, M. A.; Fretland Xiao, G. H.; Devana-
boyina, U. S.; Nangju, N. A.; Feng, Y. Cancer Epidemiology,
Biomarkers and Prevention 2000, 9, 29.
7. Upton, A.; Johnson, N.; Sandy, J.; Sim, E. Trends in
Pharmacological Sciences 2001, 151, 846.
8. Upton, A.; Mushtaq, A.; Victor, T. C.; Sampson, S. L.;
Sandy, J.; Smith, D. M.; van Helden, P. V.; Sim, E. Mol.
Microbiol. 2001, 42, 309.
18. (a) Zinnes, H.; Comes, R. A.; Zuleski, F. R.; Caro, A. N.;
Shavel, J., Jr. J. Org. Chem. 1965, 30, 2241. (b) Zinnes, H.;
Comes, R. A.; Shavel, J., Jr. J. Org. Chem. 1966, 31, 162.
19. All compounds gave satisfactory spectroscopic data.
Complete synthetic protocols and spectroscopic data will be
reported elsewhere in a full paper.
20. Spectroscopic data for 20: mp 175–177 ꢀC; IR (KBr)
cmÀ1: 1743 (C¼O), 1708 (C¼O), 1340 (SO2sym), 1167 (SO2asym);
1H NMR (CDCl3) 400 MHz d, ppm 4.14 (s, 2H, endocyclic
CH2), 4.95 (s, 2H, ArCH2N), 7.28–8.02 (m, 13H, unresolved
Ar); 13C NMR (CDCl3) 100 MHz, d 52.3, 53.9, 125.2, 127.0,
127.5, 128.2, 128.8, 130.0, 131.1, 131.1, 132.5, 132.7, 133.0,
138.0, 138.6, 140.4, 164.7, 166.1; HRMS (ESI) (MÀH)À
C24H17N2O4S2 requires 461.0637, observed 461.0630.
21. Riddles, P. W.; Blakeley, R. L.; Zerner, B. Methods
Enzymol. 1983, 91, 49.
9. Payton, M.; Auty, R.; Delgoda, R.; Everitt, M.; Sim, E. J.
Bacteriol. 1999, 181, 1343.
22. Mushtaq, A.; Payton, M.; Sim, E. J. Biol. Chem. 2002,
277, 12175.