Definition of the heterocyclic pharmacophore of bacterial methionyl tRNA synthetase inhibitors: Potent antibacterially active non-quinolone analogues

Article abstract of DOI:10.1016/j.bmcl.2004.05.070

Potent inhibitors of bacterial methionyl tRNA synthetase (MRS) have previously been reported. Through SAR of the quinolone moiety, the right hand side pharmacophore for MRS inhibition has now been defined as an NH-C-NH functionality in the context of a bicyclic heteroaromatic system. Potent antibacterial fused-pyrimidone and fused-imidazole analogues have been obtained and enantioselective activity demonstrated. Compound 46 demonstrated very good antibacterial activity against panels of antibiotic-resistant staphylococci and enterococci.

Full text of DOI:10.1016/j.bmcl.2004.05.070

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3937–3941
Definition of the heterocyclic pharmacophore of bacterial methionyl
tRNA synthetase inhibitors: potent antibacterially active
non-quinolone analogues
Richard L. Jarvest,a,* Sula A. Armstrong,a John M. Berge,a Pamela Brown,a John S. Elder,a
Murray J. Brown,a Royston C. B. Copley,a Andrew K. Forrest,a Dieter W. Hamprecht,a
Peter J. O’Hanlon,a Darren J. Mitchell,a Stephen Rittenhouseb and David R. Wittya
aGlaxoSmithKline, New Frontiers Science Park, Third Avenue, Harlow, Essex CM19 5AW, UK
bGlaxoSmithKline, 1250 South Collegeville Road, Collegeville, PA 19426, USA
Received 1 March 2004; revised 24 May 2004; accepted 24 May 2004
Available online 19 June 2004
Abstract—Potent inhibitors of bacterial methionyl tRNA synthetase (MRS) have previously been reported. Through SAR of the
quinolone moiety, the right hand side pharmacophore for MRS inhibition has now been defined as an NH–C–NH functionality in
the context of a bicyclic heteroaromatic system. Potent antibacterial fused-pyrimidone and fused-imidazole analogues have been
obtained and enantioselective activity demonstrated. Compound 46 demonstrated very good antibacterial activity against panels of
antibiotic-resistant staphylococci and enterococci.
ꢀ 2004 Elsevier Ltd. All rights reserved.
O
The search for new antibacterial agents acting at novel
molecular targets is driven by the steadily increasing
OEt
incidence of bacterial resistance to established antibiotic
classes. In this context, inhibitors of bacterial aminoacyl
tRNA synthetases, essential enzymes in protein bio-
synthesis, have attracted interest. This enzyme class is
clinically validated by the topical antibiotic mupirocin
(marketed as Bactrobanꢁ), which acts by inhibition of
bacterial isoleucyl tRNA synthetase.
Br
N
H
N
H
N
H
Br
1
Analogues in which the exocyclic amine was replaced by
oxygen or sulfur were prepared as shown in Scheme 1.
Reaction of the p-methoxybenzyloxy-quinoline 23 with
an appropriately protected alcohol or thiol, followed
by standard elaboration, afforded 4 and 5. These com-
pounds were tested in the standard S. aureus MRS
acylation assay.1 They were found to have significantly
reduced potency with IC50 values of 38 and 680 nM,
respectively, compared to 16 nM for the direct NH
analogue. Compounds 4 and 5 also had very poor
antibacterial activity against S. aureus Oxford with MIC
values of 32 and >64 lg/mL. The exocyclic NH thus
plays an important role in the MRS inhibitors.
We have reported potent inhibitors of Staphylococcus
aureus methionyl tRNA synthetase (MRS) with excel-
lent antibacterial activity against staphylococci and
enterococci.1–3 This series of compounds, such as 1, has
been characterised by a 2-aminoquinolone right hand
side. One disadvantage of these molecules is their low
membrane permeability, thought to be due to the amino-
quinolone moiety. Here we describe the determination
of the essential right hand side pharmacophore and the
identification of potent non-quinolone inhibitors.
The role of the functionality in the pyridone ring was
explored by comparing the quinolone 6 with analogues
locked in the keto 7 and enol 8 tautomers as well as the
* Corresponding author. Tel.: +44-(0)-1438-762074; fax: +44-(0)-1438-
763620; e-mail: richard.l.jarvest@gsk.com
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.05.070

3938
R. L. Jarvest et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3937–3941
OpMB
OpMB
cating that the endocyclic NH also plays a key role in
inhibitor binding to MRS.
i.
BocNH
Cl
S
N
N
Further analogues were prepared maintaining the cru-
cial NH–C–NH unit. Two thienopyridones 19 and 20
were prepared by reaction of N-(3,5-dibromobenzyl)-
1,3-diaminopropane with the chloropyridines 13 and 14,
followed by standard acid hydrolysis.4
3
2
ii., iii.
iv., ii.
O
Cl
Cl
N
H
X
N
H
O
OMe
OMe
4 X = O
5 X = S
Y
N
S
N
X
RS
N
H
N
H
S
Cl
Cl
N
N
Scheme 1. Reagents and conditions: (i) BocNH(CH2)3SH/NaH/THF;
(ii) TFA; (iii) 3,4-diCl-benzaldehyde/NaCNBH3/AcOH/MeOH; (iv)
3,4-diCl-benzyl-NH(CH2)3OH/NaH/THF/D.
14
15
13
16 X = Cl, Y = CH
17 X = MeS, Y = N
An array of more structurally diverse analogues was
also prepared. Suitable NH-containing heterocycles
bearing a leaving groupat the 2-position were either
reacted with N-(3,5-dibromobenzyl)-1,3-diaminoprop-
ane, or in some cases with 1,3-diaminopropane and the
product then reductively alkylated with 3,5-dibromo-
benzaldehyde. Pyrimidone-type analogues were pre-
pared from 2-alkylthio starting materials 15 and
imidazole-type analogues from 2-halo 16 or 2-alkylthio
17 starting materials. Either excess amine or diisopropyl-
ethylamine was used as base. The cyanoindole 28 was
prepared by displacement on a 2-chloroindole interme-
diate5 whilst the urea 29 was prepared by reaction with
phenyl isocyanate.
OMe
OMe
Cl
O
Cl
Cl
N
H
N
H
H2N
10
i.
12
iv., v.
ii., iii.
Cl
Cl
OH
OMe
O
N
H
N
H
BocNH
N
H
8
11
Cl
Scheme 2. Reagents and conditions: (i) BocNH(CH2)2CO2H/EDC/
HOBt/DMF; (ii) TFA/DCM; (iii) 2,3,5-tri-Cl-benzaldehyde/NaCN-
BH3/NaOMe/MeOH; (iv) AlCl3/LiAlH4/THF/)30 ꢂC; (v) BBr3/
dichloroethane/D.
Key analogues are detailed in Table 2, along with their
MRS IC50 values and antibacterial activity against S.
aureus and Enterococcus faecalis. Most of the fused
heteroaromatic analogues afforded potent enzyme inhi-
bition and very good antibacterial activity. A large
reduction in potency was observed when one ring was
saturated (24), when the second ring was pendant rather
than fused (25), or when the first ring was not present
(29).
deoxygenated compound 9. Compounds 7 and 9 were
prepared by the standard amine displacement reaction
with 2-methanesulfanylchromone or 2-chloroquinoline,
whilst 8 was synthesised via an amide coupling strategy
as shown in Scheme 2.
The enzyme inhibition of compounds 7–9 along with the
quinolone 6, are reported in Table 1. All the analogues
were substantially less potent than the quinolone, indi-
The poor inhibition of the thienopyrimidone 23 is in
notable contrast to that of the other bicyclic hetero-
aromatics. However, the low potency of 23 may be
attributable to a low concentration of the desired 1-NH
tautomer of this compound. Calculation of the energy of
the 1-H and 3-H tautomers of the 2-NHMe derivatives
of the heterocycles of 18–23 showed that the thieno-
pyrimidone 23 had a significantly higher preference for
the 3-H isomer compared to all the other structures.6
Thus an NH–C–NH functionality presented in the
context of a bicyclic heteroaromatic appears to define
the left hand side pharmacophore for bacterial MRS
inhibition.
Table 1. MRS inhibition of non-N–H containing quinolone analogues
Cl
Cl
Het
N
H
N
H
Cl
No.
Het
X
Y
IC50 (nM)
S. aureus MRS
O
Further elaboration was focused on the active hetero-
cycles that were more different from the quinolone,
namely the benzopyrimidone (BP) of 21, the thieno-
pyrimidone (TP) of 22, the benzimidazole (BI) of 26 and
the azabenzimidazole (AB) of 27. Left hand side ana-
logues were prepared by the standard reductive amina-
tion methodology1;2 with aldehydes (Table 3) or cyclic
ketones (Table 4).
6
7
NH
O
––
––
<3
1000
X
Y
8
9
CH
N
OH
H
1800
100
X

R. L. Jarvest et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3937–3941
3939
Table 2. MRS inhibition and antibacterial activity of N–H containing right hand side analogues
Br
Het
N
H
N
H
Br
Het
IC50 (nM)
MIC (lg/mL)
E. faecalis 1
S. aureus MRS
S. aureus Oxford
O
18
19
20
21
22
23
24
25
<3
0.25
6 0.06
6 0.06
6 0.06
6 0.06
6 0.06
32
N
H
O
S
S
4
6.2
8.1
3.9
150
580
0.13
0.25
0.25
0.5
N
H
O
N
H
O
N
N
N
N
N
N
H
O
S
N
H
O
>64
>64
>64
S
N
H
O
>64
N
H
O
>1000
>64
N
H
Ph
N
26
27
29
1
0.13
0.25
N
H
N
N
5.0
0.5
N
H
CN
28
29
54
0.5
0.25
8
N
H
O
330
>64
N
H
In the benzyl series, the 2,3,5-substituted analogues had
potent MRS inhibition and good antibacterial activity
(Table 3). The benzimidazole analogues, compounds 32,
35, 38 and 41 were clearly not as good as the pyrim-
idones. From the tetrahydroquinoline series there were
also a number of analogues derived from the four

3940
R. L. Jarvest et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3937–3941
Table 3. MRS inhibition and antibacterial activity of 2,3,5-trisubstituted phenyl analogues
OEt
Het
N
H
N
H
R
R
Het
IC50 (nM)
MIC (lg/mL)
E. faecalis 1
S. aureus MRS
S. aureus Oxford
30
31
32
3,5-diCl
3,5-diCl
3,5-diCl
BP
TP
BI
<3
<3
<3
0.13
0.25
2
6 0.06
0.13
0.5
33
34
35
3-Br,5-Cl
3-Br,5-Cl
3-Br,5-Cl
BP
TP
BI
<3
<3
3.8
0.13
0.13
1
6 0.06
0.13
0.5
36
37
38
3,5-diBr
3,5-diBr
3,5-diBr
BP
TP
BI
9.7
4.9
17
0.13
0.13
1
6 0.06
6 0.06
0.25
39
40
41
3-Br,5-OMe
3-Br,5-OMe
3-Br,5-OMe
BP
TP
BI
7.0
3.3
17
0.13
0.25
1
6 0.06
6 0.06
0.13
Table 4. MRS inhibition and antibacterial activity of tetrahydroquinoline analogues
HN
Het
3
N
H
N
H
R
5
R
Het
IC50 (nM)
MIC (lg/mL)
S. aureus MRS
S. aureus Oxford
E. faecalis 1
42
43
3-Cl,5-Br
3-Cl,5-Br
BP
TP
8.2
6 0.06
6 0.06
6 0.06
6 0.06
<3
44
45
46
3,5-diBr
3,5-diBr
3,5-diBr
TP
BI
13
13
11
6 0.06
0.25
6 0.06
6 0.06
6 0.06
AB
6 0.06
47
48
49
50
3-I,5-Et
3-I,5-Et
3-I,5-Et
3-I,5-Et
BP
TP
BI
17
<3
16
6 0.06
6 0.06
0.5
0.25
0.5
1
AB
18
0.25
1
selected heterocycles, which showed excellent antibac-
terial activity against S. aureus and E. faecalis (Table 4).
In view of the encouraging properties of 46, the indi-
vidual enantiomers were targeted, as the tetrahydro-
quinolines had only been prepared as racemates. The
synthesis of the enantiomers is shown in Scheme 3. The
diastereoisomers of amide 54 were separated chro-
matographically and the absolute configuration was
assigned via an X-ray crystal structure of isomer 56.7
The separate diastereomers were then deprotected to
give the enantiomeric amines 57 and 58. A reverse
coupling strategy was used, whereby reductive alkyl-
ation of the tetrahydroquinoline amines was effected
with the aldehyde 59. The enantiomeric excess (ee) of the
final products was determined by chiral capillary zone
electrophoresis (cze) to be 98.4% for the (R)-isomer 60
and 99.0% for the (S)-isomer 61.8
Inhibitors containing the new right hand side heterocy-
cles all retained selectivity for the bacterial enzyme, with
none of the compounds giving significant inhibition of
mammalian (rat liver) MRS upto the highest concen-
tration tested (either 1 or 10 lM).
Compound 46 was tested against a wider range of
clinical isolates of S. aureus, Staphylococcus epidermidis,
E. faecalis and Enterococcus faecium to determine
MIC90 values (the concentration required to inhibit
90% of the organisms). The panels of isolates included a
large proportion of organisms resistant to various clin-
ical antibiotics.1 Very good activity was seen against all
the organisms, with all MIC90 values at 6 1 lg/mL
(MIC90’s: S. aureus, 1 lg/mL; S. epidermidis, 0.5 lg/mL;
E. faecalis, 0.06 lg/mL; and E. faecium 0.03 lg/mL).
The enantiomers 60 and 61 were assayed in the usual
way (Table 5). The (R)-enantiomer 60 was found to be
the more active isomer with a lower IC50 value and

R. L. Jarvest et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3937–3941
3941
O
active configuration of the tetrahydroquinoline series
has been identified as possessing (R)-stereochemistry.
Ph
OMe
X
NH2
HN
Br
Br
Br
iii.
ii.
N
H
N
H
N
H
Br
Br
Br
54 (RS)
55 (R)
56 (S)
51 X = O
Acknowledgements
i.
53
iv.
52 X = NHOMe
v.
We thank Dr. J. Zukowski for performing the chiral cze
analysis and Dr. C. S. V. Frydrych for participation in
array synthesis.
NH2
H
vi.
vii.
N
N
17
Br
O
N
H
N
H
N
H
59
Br
viii.
N
57 (R)
58 (S)
HN
N
Br
References and notes
N
N
H
H
N
H
1. Jarvest, R. L.; Berge, J. M.; Berry, V.; Boyd, H. F.; Brown,
M. J.; Elder, J. S.; Forrest, A. K.; Fosberry, A. P.; Gentry,
D. R.; Hibbs, M. J.; Jaworski, D. D.; O’Hanlon, P. J.;
Pope, A. J.; Rittenhouse, S.; Sheppard, R. J.; Slater-
Radosti, C.; Worby, A. J. Med. Chem. 2002, 45, 1959.
2. Jarvest, R. L.; Berge, J. M.; Brown, M. J.; Brown, P.; Elder,
J. S.; Forrest, A. K.; Houge-Frydrych, C. S. V.; O’Hanlon,
P. J.; McNair, D. J.; Rittenhouse, S.; Sheppard, R. J.
Bioorg. Med. Chem. Lett. 2003, 13, 665.
Br
60 (R)
61 (S)
Scheme 3. Reagents and conditions: (i) MeONH2ÆHCl/NaOAc/EtOH/
H2O/D; (ii) ZrCl4/LiBH4/THF; (iii) (S)-PhCH(OMe)CO2H/EDC/
HOAt/N-Me-morpholine/DMF; (iv) silica gel chromatography/pet.
ether/EtOAc; (v) 8 M HCl/dioxan/D; (vi) (MeO)2CH(CH)2NH2/D; (vii)
1 M HCl/D; (viii) NaCNBH3/NaOAc/AcOH/MeOH.
3. Jarvest, R. L.; Berge, J. M.; Brown, P.; Houge-Frydrych, C.
S. V.; O’Hanlon, P. J.; McNair, D. J.; Pope, A. J.;
Rittenhouse, S. Bioorg. Med. Chem. Lett. 2003, 13, 1265.
4. The methoxy-thienopyridines were prepared from the 2,4-
dichloro precursors by treatment with sodium methoxide.
The thieno[3,2-b]pyridine isomer gave exclusively the
unwanted 2-methoxy isomer unless 15-crown-5 was added,
which resulted in an almost complete reversal of specificity
(81% 4-methoxy, 3% 2-methoxy). See Ref. 2 for a discus-
sion of a related but less-pronounced case.
Table 5. MRS inhibition and antibacterial activity of enantiomeric
tetrahydroquinolines
HN
N
N
Br
N
H
N
H
N
H
Br
5. 2-Chloro-3-cyano-1-(2-trimethylsilylethoxymethyl)-indole
was subjected to the following reaction sequence: i.
BocNH(CH2)3NH2/DMSO/D; ii. TFA/anisole then AcOH;
iii. 3,5-diBr-benzyl bromide/K2CO3/THF; iv. LiBF4/
MeCN/TFA/CH2(CH2SH)2.
Stereochemistry
IC50 (nM)
MIC (lg/mL)
S. aureus
S. aureus
E. faecalis 1
MRS
Oxford
46
60
61
RS
R
11
6.3
48
6 0.06
6 0.06
16
6 0.06
6 0.06
2
6. Based on STO-3G ab initio calculations of gas phase
energies using AM1 optimised geometries.
S
7. Colourless needle, 0.38 · 0.05 · 0.04 mm, orthorhombic,
ꢀ
space group P212121 (#19), T ¼ 150 K, a ¼ 4:8880ð4Þ A,
3
ꢀ
ꢀ
ꢀ
b ¼ 13:7295ð12Þ A, c ¼ 26:441ð2Þ A, V ¼ 1774:4ð3Þ A ,
Z ¼ 4, Dcalcd ¼ 1:700 Mg/m3, F ð000Þ ¼ 904, l(CuKa, k ¼
potent antibacterial activity. The compression of IC50
values <10 nM due to the limit of the enzyme concen-
tration in the assay (3 nM)1 makes it hard to calculate
the enantiomeric inhibitory ratio. However, the ratio of
the antibacterial activity of the two isomers suggests a
high degree of enantioselectivity, of the order of at least
two orders of magnitude.
1:54178 A) ¼ 5.902 mmÀ1, Bruker SMART 6000 diffractom-
ꢀ
eter, 11,989 reflections collected ð6:68ꢂ 6 2h 6 145:52ꢂÞ,
3430 unique reflections ðRint ¼ 0:0530Þ, Gaussian absorp-
tion correction (transmission ¼ 0.37514–0.80284), full-ma-
trix least-squares refinement (on F 2) of 226 variables,
R1 ¼ 0:0316 (wR2 ¼ 0:0789) for 3284 observed data with
I P 2rðIÞ, R1 ¼ 0:0327 (wR2 ¼ 0:0805) for all data, S ¼
2
1:037,
w ¼ 1=½r2ðFo2Þ þ
ð0:0604PÞ ꢀ
where
P ¼
½MaxðFo2; 0Þ þ 2Fc2ꢀ=3, residual electron density between
In conclusion, the key right hand side pharmacophore
for bacterial MRS inhibition has been defined as an
NH–C–NH unit in the context of a bicyclic heteroaro-
matic system. Potent non-quinolone analogues have
been obtained with excellent antibacterial activity
against staphylococci and enterococci, including anti-
biotic resistant isolates. In addition, the biologically
À3
ꢀ
)0.471 and 0.942 e A
, absolute structure parame-
ter ¼ )0.046(19). Crystallographic data for the structure
has been deposited with the Cambridge Crystallographic
Data Centre as supplementary publication number CCDC
232517.
8. Fused silica 50 cm · 50 lm i.d., 20 kV, 100 mM sodium
phosphate buffer pH 2.5 containing 40 mM a-cyclodextrin.