Towards Gram-positive antivirulence drugs: New inhibitors of Streptococcus agalactiae Stk1

Article abstract of DOI:10.1016/j.bmcl.2010.05.004

A structure-activity relationship study from a screening hit and structure-based design strategy has led to the identification of bisarylureas as potent inhibitors of Streptococcus agalactiae Stk1. As this target has been directly linked to bacterial virulence, these inhibitors can be considered as a promising step towards antivirulence drugs.

Full text of DOI:10.1016/j.bmcl.2010.05.004

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3486–3490
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Towards Gram-positive antivirulence drugs: New inhibitors
of Streptococcus agalactiae Stk1
Mayalen Oxoby, François Moreau, Lionel Durant , Alexis Denis , Jean-Marie Genevard , Vanida Vongsouthi,
*
Sonia Escaich , Vincent Gerusz
MUTABILIS SA, 102 Avenue Gaston Roussel, 93230 Romainville, France
a r t i c l e i n f o
a b s t r a c t
Article history:
A structure–activity relationship study from a screening hit and structure-based design strategy has led
to the identification of bisarylureas as potent inhibitors of Streptococcus agalactiae Stk1. As this target has
been directly linked to bacterial virulence, these inhibitors can be considered as a promising step towards
antivirulence drugs.
Received 3 March 2010
Revised 3 May 2010
Accepted 3 May 2010
Available online 7 May 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Antivirulence drugs
Antibacterial
STK
Stk1
Streptococcus agalactiae
Gram-positive
Serine/threonine kinase
Bacterial resistance to antibiotics is a serious public health
problem which increasingly limits the effectiveness of current
drugs.1–3 Streptococcus agalactiae (Group B Streptococci, GBS), a
Gram-positive pathogen, represents the major cause of infection
during the neonatal period leading to pneumonia, sepsis and men-
ingitis. To fulfill the urgent medical and social need of effective
antibacterial therapies, our approach focuses on developing new
molecules with innovative mechanisms of action.
Kinase-mediated phosphorylation of intracellular proteins is a
crucial pathway by which signals are transmitted within the cell.
In bacteria, communication is achieved by two-component sys-
tems (TCS) as well as by eukaryotic-type serine/threonine kinases
(STK).4–6 In the past few years a conserved family of bacterial
STK, unique to Gram-positive species, has emerged as an attractive
potential antibacterial target.7–13 These kinases are proposed to be
regulators of metabolic processes, including transcription, growth,
differentiation, cell development, sporulation, biofilm formation
and interaction with host cells. Recently, direct relationships
between STKs and bacterial virulence have also been estab-
lished.7a,11–14 For instance, mutations or deletion of the Stk1 gene
of S. agalactiae exhibited pleiotropic effects on growth, cell segre-
gation and effectiveness in murine models of virulence.8,9,15 We
report herein our approach to the identification of S. agalactiae
Stk1 inhibitors and the preliminary structure–activity relationship
(SAR) generated from a family of 2-aryl-purines.
In S. agalactiae, Stk1 is a transmembrane protein harboring a
serine/threonine kinase domaine located in the cytoplasm of the
bacteria and 3 PASTA (for ‘penicillin-binding protein and serine/
threonine kinase associated’) domains outside the bacteria.8 The
soluble kinase domain has been expressed in Escherichia coli with
an N-Terminal 6-His tag.16 Stk1 can phosphorylate both serine
and threonine residues, and recognizes myelin basic protein
(MBP) as surrogate substrate which has been chosen in our
assays.8a,17 High-throughput screening of our in-house library of
molecules allowed identification of the 6-(3-aminopropanol)-
substituted purine18 compound 1b as an inhibitor of S. agalactiae
Stk1 with an IC50 of 66 lM. In order to better understand the po-
tential mode of binding of that molecule, a model of the ATP-bind-
ing site was built up with Easypred19 by homology modelling
based on the sequence and available structure of the Mycobacte-
rium tuberculosis STK enzyme PknB (pdb code: 1O6Y).20 M. tubercu-
losis H37rv STK shares 32/52% of identity/similarity with the S.
agalactiae NEM316 Stk1. Analogues were also synthesized as out-
lined in Scheme 1 to investigate the SAR of this series.
The 3-aminopropanol was introduced on C6 position of the 2,6-
dichlopurine by refluxing in butanol to afford compound 1a with
72% yield. Alkylation led exclusively to substituted compounds
on position N9 (1b–d).21 Finally, palladium-mediated reactions of
* Corresponding author. Tel.: +33 (0)1 57 14 05 22; fax: +33 (0)1 57 14 05 24.
E-mail address: vincent.gerusz@mutabilis.fr (V. Gerusz).
Formerly at Mutabilis S.A.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.05.004

M. Oxoby et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3486–3490
3487
R1
N
HN
N
N
N
a
N
N
Cl
Cl
OH
b
Cl
N
N
H
2,6-dichloropurine
1a: R1 = H
N
c
N
1b: R1 = Me (64%)
1c: R1 = Et (71%)
1d: R1 = i-Pr (66%)
N
R2
N
N
OH
H
2a-f
Scheme 1. Reagents and conditions: (a) 3-aminopropan-1-ol, TEA, n-BuOH, 120 °C,
overnight, 72%; (b) NaH (60% in oil), alkyl halide, DMF, rt, overnight; (c) boronic
acid, K2CO3, Pd(PPh3)4, toluene, 100 °C, overnight, 19–43%; or aniline, t-BuONa,
Pd2dba3, R(+)-BINAP, toluene, 80 °C, overnight, 11%; or amine, neat, reflux, 2.5–5 h,
12–86%.
Figure 1. Overlay of 2a (green) with ATP in the active site of S. agalactiae Stk1 with
expanded key residues.
Table 2
S. agalactiae Stk1 inhibitory activities of compounds 3 and 4
the C2 chlorine with boronic acids or aniline provided derivatives
2a–c while amino-derivatives 2d–f were obtained by refluxing in
neat amines.
N
N
N
The methyl group at purine N9 position was rapidly determined
as optimal for the activity as seen in Table 1, since modifications by
larger alkyl substituents or in contrast no substitution resulted in a
significant decrease of activity. Exploration of the position C2 indi-
cated a hydrophobic aromatic bias. Indeed, a phenyl group in com-
pound 2a is the most promising substitution with a fivefold
improvement of activity compared to the initial hit 1b. In contrast,
the more hydrophilic pyridyl analogue 2b was less active than 2a.
The lipophilic aniline and benzylamine derivatives 2c and 2d were
more active than 1b, although 2–4 times less potent than 2a,
whereas the morpholino derivative 2e and the charged piperazine
2f were notably less active than 2a.
All these observations were in complete agreement with the
binding mode of 2a (Fig. 1) obtained by docking in the ATP site
of our model using QXP/Flo+.22 The 9-methyl purine ring system
fits well into the canonical ATP-binding site, the heterocycle being
sandwiched between the lipophilic residues Ala 40 and Leu 143.
Two crucial hydrogen bonding interactions occurring between
the nitrogen N7 of the purine and the nitrogen of the propanol-
R
N
N
OH
H
3a-h, 4a-h
a
a
#
R
IC50
(lM)
#
R
IC50 (lM)
H
H
N
N
3a Cl
3
4a
4b
4c
1
O
H
N
H
N
3b NO2
3
4
7.5
12
O
H
N
H2N
3c
CF3
O
H
H
N
N
3d Ph
4
4d
4e
4f
18
22
38
O
H
N
S
3e
3f
NH2
13
O
Table 1
O
S. agalactiae Stk1 inhibitory activities of compounds 1 and 2
O
N
Compound R1
IC50
(l
M) Compound R2
IC50
12
(lM)
a
a
NHSO2Me 15
N
H
O
1a
1b
H
>300
2a
N
N
3g NHAc
3h CO2H
18
39
4g
4h
174
Me 66
2b
34
26
O
1c
Et
175
2c
N
H
>300
(CH2)3NH
N
H
1d
iPr
>300
2d
48
66
a
Values are means of three experiments.
N
2e
O
amine side-chain of the compounds with, respectively, the NH
and CO of Val 93 in the hinge region of the protein complete the
binding. In addition, similarly to the known CDK inhibitors such
as olomoucine,23 compound 2a made a close lipophilic contact
with its methyl at position N9 in the deeply buried hydrophobic
N
2f
110
HN
a
Values are means of three experiments.

3488
M. Oxoby et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3486–3490
N
N
a
N
N
Procedure A
N
R
N
H
H2N
HN
N
N
N
OH
N
N
OH
H
H
O
R = H, alkyl, aryl
3e
4a-d
5a-n
Procedure B
b
d
N
N
c
N
N
N
R
N
H
H2N
HN
N
N
N
OTBS
N
N
OTBS
H
H
O
6
7a-n R = aryl, heteroaryl
Scheme 2. Reagents and conditions: procedure A: (a) R = H: urea (4 equiv), water, AcOH (1.3 equiv), HCl, reflux, 4%; R = alkyl, aryl: RNCO (0.8–4.6 equiv), NaHCO3 (0–6 equiv),
toluene or DCE or THF, rt to reflux, 20–120 h, 4–27%; procedure B: (b) TEA (2 equiv), TBDMSCl (1.5 equiv), DMF, rt, 5 h, 85%; (c) 6, 4-nitrophenylchloroformate (1.2 equiv),
THF, rt, 45 min–2 h, then RNH2 (1.5 equiv), TEA (2 equiv), 50 °C, 0.5–5 h; (d) TBAF (3–5 equiv), THF or ammonium fluoride (13 equiv), MeOH, rt to reflux, 20 h, 40–46% over
two steps.
pocket lined by Met 90, Val 74 and the methyl of Thr 153. It ap-
pears also clearly that there is little room for bulkier substituents.
According to this model of docking we hypothesized that growing
the molecule from the meta position of the phenyl ring towards the
Table 3
S. agalactiae Stk1 inhibitory activities of compounds 5
N
N
N
H
H
N
N
R
N
N
OH
H
O
5a-n
a
a
#
R
IC50
(l
M)
#
R
IC50 (lM)
Cl
H2N
5ab
0.16
5hc
0.73
Cl
CH3CONH
N
5bb
5cb
5db
5eb
0.2
5ic
5jc
5kb
5lb
0.93
0.96
1.6
2
O
0.28
0.32
0.49
O
Cl
Cl
S
NC
Cl
O
O
S
5fc
0.64
0.63
5mb
5nb
16
Cl
5gb
59
MeO
Cl
a
b
c
Values are means of three experiments.
Synthesis according to Scheme 2, procedure A.
Synthesis according to Scheme 2, procedure B.

M. Oxoby et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3486–3490
3489
phosphate region of the ATP site could generate additional interac-
tions, especially with the hydrophilic residues Gln140 or Asp154
and the remote lipophilic loop between Gly 21 and Ala 24.
A series of substitution was therefore introduced at the position
meta of the aryl group by Suzuki couplings with the corresponding
boronic acids to provide compounds 3a–h (Table 2). With the
exception of substituents Cl, CF3, NO2 and Ph of 3a–d, no improve-
ment was achieved with short amino, sulfonylamido, amido or car-
boxylic acid substitutions (compounds 3e–h). The biochemical
characterization of 3c assessed that this compound worked by
competitive inhibition with regard to ATP, in agreement with our
model.24 At this stage, some counter-screening on the related
eukaryotic protein kinase A was performed on a few compounds.
The results indicated a potential for selectivity of this series with
respect to eukaryotic kinases.25
benzylurea 4b. Finally, we observed that longer and more flexible
substitutions yielded weaker activities as illustrated by com-
pounds 4f–g or complete inactivity in the case of 4h.
The encouraging result obtained with 4a prompted us to further
explore this urea series. Indeed, the modelling and docking of 4a
into the ATP cavity indicated that the gain of potency could be ex-
plained by two new hydrogen bonds with the targeted residues
Gln140 or Asp154 as well as a favourable hydrophobic contact
with the Gly 21-Ala 24 loop. We also observed that the position
meta of the second phenyl group pointed into the direction of Ileu
157 and if properly substituted could therefore generate additional
lipophilic contact. New bis-aryl ureas have been synthesized either
via commercially available isocyanates (Scheme 2, procedure A) or
via activation of the aniline moiety of 3e as a 4-nitrophenylcarba-
mate (Scheme 2, procedure B). According to the model, we ob-
served that the position ortho (5m–n) was clearly less favourable
for substitution than the para or meta ones (Table 3). This was also
consistent with the requirement of planarity of the aryl-urea-aryl
system to interact with the abovementioned key residues and to
fit in the cavity below the loop Gly 21-Ala 24. In agreement with
our model, the most interesting compounds 5a–d were substituted
by lipophilic substituents at the meta position such as m,m0-di-
chloro, methyl, piperonyl and chloro ( Fig. 2) in contrast to the
para-chloro derivative 5k which was fivefold less active than its
meta counterpart 5d. The more polar groups such as amino, acet-
amido, cyano or methylsulfonyl (5e,f,h,i), while being more active
than 4a were less potent than 5a–d. As shown by 5j and 5l, the
replacement of the phenyl group by a heteroarene (pyridine or thi-
ophene) was not deleterious for the activity when compared to 4a.
Finally, having identified 5d as a potent inhibitor, we decided to
come back to the propanol moiety and explore its importance for
the activity. A more convergent pathway was set up (Scheme 3)
starting from 2,6-dichloropurine via a common methylsulfonyl
intermediate 10 that was treated with amines to afford the desired
compounds 11a–f in moderate yields.
The next step aimed at extending the compound. Reaction with
isocyanates afforded urea analogues 4a–d (Scheme 2, procedure
A), while longer sulfonamido and amido groups were obtained
by standard methods (sulfonyl chloride, 4-DMAP, pyridine and
amine, HATU, EDCI, DCM, respectively). A clear gain of activity
was obtained with the phenylurea 4a at 1 lM. Interestingly the
mono-aryl ureas 4b–d were significantly less potent, highlighting
the importance of the secondary aryl group for activity. The ben-
zylsulfonamide 4e turned out less active than the corresponding
In agreement with the proposed hydrogen-bond donation of the
propanolamine nitrogen with the protein, compound 11f was inac-
tive. Although we suspected from the docking model a potential
hydrogen bond of the hydroxy group of the propanol with Tyr
92, the activity of derivatives 11a–c clearly demonstrated that this
interaction did not occur. The propanolamine side-chain was most
likely making hydrophobic contact with the backbone of the pro-
tein as shown by the good activity of 11a–d compared to the
unsubstituted compound 11e (Table 4).
Figure 2. Compound 5d (green) in the active site of S. agalactiae STK with expanded
key residues.
N
HN
N
N
a-b
N
c-d
N
N
N
N
N
H
N
H
N
Cl
N
S
Cl
Cl
Cl
N
S
O
2,6-dichloropurine
8
9
e
N
N
f
N
N
N
N
H
N
H
N
H
H
R1
Cl
Cl
N
N
N
S
N
N
O O
O
O
R2
R1, R2 = H, alkyl
10
11a-f
Scheme 3. Reagents and conditions: (a) K2CO3 (1.1 equiv), MeI (1.4 equiv), DMSO, rt, 2 h, 61%; (b) NaSMe (1.2 equiv), DMSO, rt, 2 h, 51%; (c) 3-aminophenyl boronic acid
monohydrate (1.4 equiv), 2-cyclohexylphosphino-20,60-dimethoxybiphenyl (0.02 equiv), Pd(OAc)2 (4 ꢀ 0.04 equiv), CH3CN, reflux, 48 h, 21%; (d) 3-chlorophenyl isocyanate
(0.8 equiv), THF, rt, 2 h, 64%; (e) m-CPBA (8 equiv), THF, rt, 23 h, 94%; (f) HNR1R2 (5 equiv), THF, reflux, 5 h, 14–50%.

3490
M. Oxoby et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3486–3490
Table 4
S. agalactiae Stk1 inhibitory activities of compounds 11
N
N
N
H
N
H
N
Cl
R
N
O
a
a
#
R
IC50
(
lM)
#
R
IC50 (lM)
N
H
OMe
N
H
NH2
OH
11a
11b
0.28
0.3
11d
11e
0.48
1.3
N
H
NH2
OH
N
N
H
11c
0.32
11f
>300
CH3
a
Values are means of three experiments.
basic protein). The steady-state kinetic parameters of Stk1 were
KM(ATP) = 0.4 M at MBP = 2 M and KM(MBP) = 0.06 M at ATP=5 M. In the
luminescent assay, inhibitor in dimethylsulfoxide (DMSO) was pre-
incubated in a white polystyrene 96-well plate at room temperature for 30 min
with 27 L Stk1 in assay buffer AB (50 mM Hepes pH7.5, 0.5 mM MnCl2, 0.012%
Triton-X100, 1 mM DTT). 30 L of MBP/ATP mix in AB were then added to start
the reaction with final concentrations of 5 nM Stk1, 0.3 M myelin basic
protein (Sigma) and 0.3 M ATP (Sigma). After 90 min of incubation at room
temperature, 30 L of the revelation mix were added: 2 nM luciferase (Sigma),
30 -luciferin (Sigma), 100 N-acetylcysteamine (Aldrich).
In conclusion, we have identified new bisarylurea derivatives as
l
l
l
l
S. agalactiae Stk1 inhibitors in the course of a SAR study assisted by
molecular modelling. These small molecules represent encourag-
ing leads for the development of novel anti-infective drugs based
on the concept of antivirulence.
3 lL
l
l
l
l
l
Acknowledgment
lM
D
lM
Luminescence intensity was immediately measured on an Analyst-HT
(Molecular Devices) and converted into inhibition percentages. For IC50
determinations, the inhibitor was tested at 6–10 different concentrations,
and the related inhibitions were fitted to a classical Langmuir equilibrium
model using XLFIT (IDBS). In these conditions, the IC50 of the non specific
kinase inhibitor staurosporin was 21 nM. In the fluorescent assay, the
following components were pre-incubated for 30 min at room temperature
The authors wish to thank Oséo for financial support (Grant A
06 05 033Q).
References and notes
1. Croft, A. C.; D’Antoni, A. V.; Terzulli, S. L. Med. Sci. Monit. 2007, 13, RA 103.
2. Alanis, A. J. Arch. Med. Res. 2005, 36, 697.
3. Payne, D.; Tomasz, A. Curr. Opin. Microbiol. 2004, 7, 435.
4. Zhang, C. C. Mol. Microbiol. 1996, 20, 9.
in a black polystyrene 96-well plate: 5
45 L Stk1 in AB. 50 L of substrates-revelation mix in AB were then added in
each well leading to the following final composition: 10 nM Stk1, 2 M MBP,
0.3 M ATP, 5 u/mL Pyruvate Kinase (Sigma), 50 M phosphoenolpyruvate
(Sigma), 5 u/mL Lactate des-hydrogenase (Sigma) and 3 M NADH (Sigma).
lL inhibitor dissolved in DMSO and
l
l
l
l
l
5. Bakal, C. J.; Davies, E. J. Trends Cell Biol. 2000, 10, 32.
l
6. Ulrich, L. E.; Koonin, E. V.; Zhulin, I. B. Trends Microbiol. 2005, 13, 52.
7. (a) Fernandez, P.; Saint-Joanis, B.; Barilone, N.; Jackson, M.; Gicquel, B.; Cole, S.
T.; Alzari, P. M. J. Bacteriol. 2006, 188, 7778; (b) Székely, R.; Wáczek, F.;
Fluorescence intensity of NADH (kex = 360 nm, kem = 520 nm) was immediately
measured kinetically on a Fluostar Optima (BMG). Inhibition percentages were
derived from fitted initial velocities using XLFIT (IDBS).
}
Szabadkai, I.; Németh, G.; Hegymegi-Barakonyi, B.; Eros, D.; Szokol, B.; Pató, J.;
18. For other examples of kinase inhibitors based on purines, see: Gilbert, A. M.;
Nowak, P.; Brooijmans, N.; Bursavich, M. G.; Dehnhardt, C.; Santos, E. D.;
Feldberg, L. R.; Hollander, I.; Kim, S.; Lombardi, S.; Park, K.; Venkatesan, A. M.;
Mallon, R. Bioorg. Med. Chem. Lett. 2010, 20, 636.
19. Lambert, C.; Leonard, N.; De Bolle, X.; Depiereux, E. Bioinformatics 2002, 18,
1250.
}
Hafenbradl, D.; Satchell, J.; Saint-Joanis, B.; Cole, S. T.; Orfi, L.; Klebl, B. M.; Kéri,
G. Immunol. Lett. 2008, 116, 225.
8. (a) Rajagopal, L.; Clancy, A.; Rubens, C. E. J. Biol. Chem. 2003, 278, 14429; (b)
Rajagopal, L.; Vo, A.; Silvestroni, A.; Rubens, C. E. Mol. Microbiol. 2005, 56(5),
1329.
9. Echenique, J.; Kadioglu, A.; Romao, S.; Andrew, P. W.; Trombe, M. C. Infect.
Immun. 2004, 72, 2434.
20. Ortiz-Lombardia, M.; Pompeo, F.; Boitel, B.; Alzari, P. M. J. Biol. Chem. 2003, 278,
13094.
10. Hong Jin, H.; Pancholi, V. J. Mol. Biol. 2006, 357, 1351.
11. Donat, S.; Streker, K.; Schirmeister, T.; Rakette, S.; Stehle, T.; Liebeke, M.; Lalk,
M.; Ohlsen, K. J. Bacteriol. 2009, 191, 4056.
21. Schow, S. R.; Mackman, R. L.; Blum, C. L.; Brooks, E.; Horsma, A.; Joly, A.;
Kerwar, S. S.; Lee, G.; Shiffman, D.; Nelson, M. G.; Wang, X.; Wick, M. M.; Zhang,
X.; Lum, R. T. Bioorg. Med. Chem. Lett. 1997, 7, 2697.
12. (a) Débarbouillé, M.; Dramsi, S.; Dussurget, O.; Nahori, M. A.; Vaganay, E.;
Jouvion, G.; Cozzone, A.; Msadek, T.; Duclos, B. J. Bacteriol. 2009, 191, 4070; (b)
Cozzone, A. J. J. Mol. Microbiol. Biotechnol. 2005, 9, 198.
13. Beltramini, A. M.; Mukhopadhyay, C. D.; Pancholi, V. Infect. Immun. 2009, 77,
1406.
14. Kristich, C. J.; Wells, C. L.; Dunny, G. M. Proc. Natl. Acad. Sci. U.S.A. 2007, 104,
3508.
15. Moreau, F.; Vongsouthi, V.; Prouvensier, L.; Genevard, J. M.; Durant, L.; Bonvin,
Y.; Gerusz, V.; Denis, A.; Escaich, S. Abstracts No F2-495, 46th Interscience
Conference on Antimicrobial Agents and Chemotherapy (ICAAC), September
26–30, 2006; San Francisco, CA.
22. Flo+ is
a molecular modeling software developed by Colin McMartin of
Thistlesoft: McMartin, C.; Bohacek, R. J. Comput. Aided Mol. Des. 1997, 11,
333.
23. Otyepka, M.; Krystof, V.; Havlícek, L.; Siglerová, V.; Strnad, M.; Koca, J. J. Med.
Chem. 2000, 43, 2506.
24. Compound 3c (0–300
note20) in the presence of 2
at 5 M. Initial velocities were fitted with XLFIT (IDBS) leading to intersecting
lines on y-axis at 1/Vmax in the double reciprocal plot (1/initial velocity versus
1/[ATP]). The Ki of 3c was 2.9 M.
l
M) was tested on Stk in the fluorescent assay (see
l
M MBP and ATP (0.08–10 M). NADH was raised
l
l
l
25. Human protein kinase A (PKA) shares 99/99% of identity/similarity with its
16. In S. agalactiae GBS NEM316 strain, the relevant gene is gbs0307 with
NP_734776 as the corresponding protein accession number.
17. Stk1 biochemical assays were based either on luminescent ATP detection, or on
fluorescent ADP detection. They used the surrogate substrate MBP (myelin
bovine counterpart. Both PKAs share 24/48% of identity/similarity with the S.
agalactiae NEM316 Stk1. Screening 2a and 3c on the bovine PKA led to IC50
s
>100 M.
l