Evaluation of N-(phenylmethyl)-4-[5-(phenylmethyl)-4,5,6,7-tetrahydro-1H-imidazo[4,5-c]pyridin-4-yl]benzamide inhibitors of Mycobacterium tuberculosis growth

Article abstract of DOI:10.1016/j.bmcl.2007.02.078

The biological evaluation of imidazopiperidines as FAS II inhibitors of Mycobacterium tuberculosis growth has been carried out with a view to assessment of potential as lead compounds for the development of a new TB drug. A summary of the hit evaluation a

Full text of DOI:10.1016/j.bmcl.2007.02.078

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2740–2744
Evaluation of N-(phenylmethyl)-4-[5-(phenylmethyl)-4,5,6,7-
tetrahydro-1H-imidazo[4,5-c]pyridin-4-yl]benzamide inhibitors
of Mycobacterium tuberculosis growth
Michael D. Wall,a,* Michael Oshin,a Gavin A. C. Chung,a Tony Parkhouse,c
Andrea Gore,a Esperanza Herreros,b Brian Cox,a Ken Duncan,a Brian Evans,a
Martin Everetta and Alfonso Mendozab
aGlaxoSmithKline, Medicines Research Centre, Gunnels Wood Road, Stevenage SG1 2NY, UK
bGlaxoSmithKline, Investigacion y Desarrollo S.L. C/Severo Ochoa 2. P.T.M. E-28760, Tres Cantos, Madrid, Spain
cGlaxoSmithKline, New Frontiers Science Park Third Avenue, Harlow, Essex, CM19 5AW, UK
Received 18 December 2006; revised 28 February 2007; accepted 28 February 2007
Available online 3 March 2007
Abstract—The biological evaluation of imidazopiperidines as FAS II inhibitors of Mycobacterium tuberculosis growth has been car-
ried out with a view to assessment of potential as lead compounds for the development of a new TB drug. A summary of the hit
evaluation and current challenges is described herein.
Ó 2007 Elsevier Ltd. All rights reserved.
The World Health Organisation (WHO) has prioritized
tuberculosis (TB) as one of the top three diseases requir-
ing attention by the international community.1 The
majority of TB cases occur in the developing world,
although TB is fast re-emerging as a killer in developed
world populations.2 Conservative WHO estimates cur-
rently attribute around two million deaths worldwide
per year to TB with the true figure probably masked
by the rising mortality of HIV/TB patients.3
OH OH
H
N
H
O
O
N
R
NH2
O
HO
OH
O
O
N
O
O
OH
O
Non-compliance with the six-month, multi-drug chemo-
therapeutic regimen4 is a major factor in the develop-
ment of multi-drug resistant (MDR) strains of
Mycobacterium tuberculosis (M.tb) despite more wide-
spread use of the WHO recommended DOTS (directly
observed therapy, short course) treatment. MDR-TB is
defined as strains of the M.tb resistant to at least two
of the front-line TB drugs (Fig. 1), namely isoniazid 1
(shown in its pro-drug form) and rifampicin 2.5
C
N
N
N
R =
H
1
2
Figure 1. Two of the DOTS regime recommended drugs; isoniazid 1
and rifampicin 2.
InhA, the biological target for 1, is an enoyl-ACP reduc-
tase, part of the FASII system, which is involved in fatty
acid chain elongation.7 These chains contribute to the
biosynthesis of mycolic acids which are key components
of the mycobacterial cell wall. It has been suggested that
1 is activated by a catalase-peroxidase enzyme, the prod-
uct of the katG gene, allowing radical formation of a
covalent bond to give isonicotinyl-NAD+, prior to inser-
tion into the protein structure.6e
We were interested in exploring inhibitors which exhibit
an alternative mechanism of inhibition to that of 1 thus
maintaining activity in MDR-TB strains.
Keywords: Mycobacterium tuberculosis; Imidazopiperidines; InhA.
*
Corresponding author. Tel.: +44 (0) 1438 768633; e-mail: michael.d.
wall@gsk.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.02.078

M. D. Wall et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2740–2744
2741
This has also been supported by crystallographic evi-
dence.6a Once within the protein, the modified co-factor
inhibits fatty acid chain saturation by preventing hy-
dride delivery to the enoyl-ACP fatty acid chains.7b
MDR-TB strains often contain mutations in the katG
gene,8 thus activation of 1 via the catalase-peroxidase
mechanism is prevented resulting in a significant reduc-
tion of its efficacy.9
give the linker resin 4.11 Treatment of 4 with selected
amines under reductive amination conditions gave the
secondary amines 5 which were then subjected to HATU
coupling conditions with 3 (R2 = OH) to give the Fmoc
protected imidazopiperidine loaded resins 6. Deprotec-
tion of the Fmoc group followed by a second round of
reductive aminations with a number of pre-selected alde-
hydes12 resulted in the desired imidazopiperidines
attached to the solid phase resin 7. Cleavage of the prod-
ucts from the resin was achieved using trifluoroacetic
acid to give the desired crude compounds in Table 1 in
good yields.
Our investigation focused on identifying non-covalent
inhibitors of InhA which are independent of upstream
biochemical processes, in an attempt to provide novel
chemotherapeutic agents for MDR-TB.
Imidazole replacement was carried out in solution using
a modification of the original route, Scheme 2. Phenyl-
ethylamine 10 was reacted under coupling conditions
with monomethyltetraphthalate to give 11 in good yield.
Treatment of 11 with POCl3 in the presence of a dehy-
drating agent at reflux gave the imine intermediate
which when subjected to hydride reduction afforded
the racemic scaffold 12 in 82% yield over two steps.
Structural diversity was attributed to the scaffold at this
point under the conditions of reductive amination using
pre-selected aldehydes to give the alkylated piperidine
analogues 13a–c in good yields. Treatment of the methyl
ester under mild conditions of hydrolysis afforded the
carboxylic acid which was coupled to a number of pre-
selected benzylamines to afford the final compounds
14a–c.
From a screen for inhibitors of InhA, we discovered the
imidazopiperidine series 3. Substructure searches led to
the discovery of analogues within this series, which sup-
ported our original hit. Confirmation of activity was
achieved from a resynthesised sample. We were inter-
ested in addressing the initial medicinal chemistry ques-
tions associated with the series, summarised in Figure 2.
It was recognized that the original scaffold 3 provided a
tractable hit which facilitated rapid expansions at both
the carbonyl and piperidine positions using solid phase
chemistry. In addition, it was thought that the imidazole
ring could be replaced in order to evaluate its capacity as
a hydrogen bond donor/acceptor. Since the original hits
were previously tested as racemates, it was necessary to
establish the activity of the individual enantiomers. With
these considerations, our initial synthesis programme
was prompted towards a suitable starting point from
which to evaluate the activity of this series.
The high-throughput screen was carried out using a 384-
well microplate assay in which reduction of 2-trans
dodecenoyl co-enzyme A and concomitant oxidation
of NADH by InhA was monitored indirectly by reacting
remaining NADH with the dye MTS (Promega), cou-
pled via the electron carrier PMS (Sigma), to produce
a coloured derivative14 (abs at 492 nm).
The 4-(4,5,6,7-tetrahydro-1H-imidazo[4,5-c]pyridin-4-
yl)benzoic acid (imidazopiperidine) scaffold was pre-
pared using literature procedures.10 Treatment of this
imidazopiperidine with activated Fmoc–succinimide in
aqueous 1,4-dioxane in the presence of sodium hydro-
gen carbonate gave the deprotected racemic imidazopi-
peridine key intermediate 3 (R1 = Fmoc, R2 = OH,
R3 = Fmoc) in 76% yield. The scaffold was used without
further purification for modification at the carboxy and
piperidine positions using solid phase chemistry11
described in Scheme 1.
Determination of whole cell activity was carried out by
measuring the minimum inhibitory concentration (MIC)
required to inhibit growth of broth cultures of M.tb
incubated in 96-well microtitre plates. Growth was mea-
sured indirectly via metabolic reduction of the fluores-
cent dye Alamar blue.
TentagelTM was allowed to react with the formyl meth-
oxybutanoic acid under HATU coupling conditions to
All prepared compounds were screened in the InhA bio-
chemical assay with the broad-spectrum antibacterial
triclosan (IC50 = 2.1 lM), a known InhA inhibitor, as
the standard and compounds displaying significant
activity were evaluated in the M.tb whole cell assay. A
selection of these results is presented in Table 1.
Substituent
2
O
R
tolerence
Preliminary observations for this series showed that sub-
stitution was required at both R1 and R2 to achieve
activity. The scope of SAR at R1 and R2 was evaluated
from an array of imidazopiperidine racemic analogues,
examples of which are presented in Table 1 (compounds
8e–l).
Chirality
Substituent
tolerence
1
R
N
N
N
Imidazole
Replacement
3
R
The most active compounds contained the electron
donating para-methoxybenzylamine group in the R2
position while complemented by the presence of either
the mono or dichloro benzyl analogue at the R1 position
3
Figure 2. The proposed medicinal chemistry issues requiring investi-
gation within the imidazopiperidine series 3.

2742
M. D. Wall et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2740–2744
OMe
OMe
OMe O
N
NH
H
N
H
O
H
N
(i)
H
N
(ii)
O
O
O
O
O
O
4
5
6
fmoc
N
N
N
fmoc
(iii)a,b
O
N
N
OMe
N
(iv)
H
N
O
O
O
N
N
H
N
N
N
H
8
7
Scheme 1. Reagents and conditions: (i) Benzylamine, CH3COOH, NH4HB(OAc)3, DCM, rt, 1 h; (ii) 3, HATU, DMF, rt, 1 h; (iii) a—piperidine,
DMF, rt, 1 h; b—benzaldehyde, CH3COOH, NH4HB(OAc)3, DCM, rt, 1 h; (iv) CF3COOH, DMF, rt, 0.5 h.
Table 1. InhA enzyme and whole cell activity (vs M.tb) for a selection of imidazopiperidine inhibitors
Compound
R1
R2
IC50/lMa
MIC/lMb
8e
8f
2,3-Di chlorobenzyl
3-Chloro benzyl
2,3-Di chlorobenzyl
3-Chloro benzyl
Naphthyl
4-Methoxy benzyl
4-Methoxy benzyl
Benzyl
0.24
0.41
1.39
3.02
3.5
32
63
32
32
32
32
63
16
63
63
32
63
8g
8h
8i
Benzyl
4-Methoxy benzyl
2-Trifluoro benzyl
4-Methoxy benzyl
Benzyl
8j
2,3-Di chlorobenzyl
2,4,5-Tri methoxy benzyl
Naphthyl
9.81
10.18
13.38
>10
8k
8l
9ac
9bd
9cc
9dd
2,3-Di chlorobenzyl
2,3-Di chlorobenzyl
3-Chloro benzyl
3-Chloro benzyl
4-Methoxy benzyl
4-Methoxy benzyl
4-Methoxy benzyl
4-Methoxy benzyl
0.08
>10
0.2
a Drug concentration at 50% reduction in enzymatic activity.
b Minimum drug concentration at 50% inhibition of cellular growth.
c Enantiomer 1 with retention time < 7 min.
d Enantiomer 2 with retention time > 7 min.
O
O
O
O
O
2
O
O
O
R
(i)
(ii)a,b
(iii)
(iv)a,b (v)
NH2
HN
1
1
HN
R
N
R
N
10
11
12
13
14
Scheme 2. Reagents and conditions: (i) C9H8O4, EDCI, TEA, HOBT, EtOAc/THF (1:1), 0 °C—rt, 18 h, 95%; (ii) a—POCl3,P2O5, reflux, 1.5 h, 95%;
b—NaBH4, MeOH/H2O, rt, 16 h, 82%; (iii) R1CHO CH3COOH, NaHB(OAc)3, DCM, rt, 16 h; (iv) a—NaOH, THF/H2O (80:20), rt, 70 h; b—HCl;
(v) R2CH2NH2, EDCI, TEA, Hobt, EtOAc/THF (1:1), 0 °C—rt, 20 h.
of the piperidine nitrogen (8e, IC50 = 0.24 lM). There is
a minor drop off in activity upon removal of one chlo-
ride group (8f, IC50 = 0.41 lM) however, in general,
where R2 was para-methoxybenzylamine a degree of
activity was preserved for example, in the case where
R1 was the naphthyl (8i, IC50 = 3.5 lM) or 2,4,5-tri-

M. D. Wall et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2740–2744
2743
Table 2. Whole cell activity (vs M.tb) and HEPG2 toxicity data for a
selection of imidazopiperidine inhibitors
methoxybenzyl (8k, IC50 = 10.2 lM) when compared to
the non-substituted benzyl group at R2 (8l,
MICa (M.tb)
Tox50(HEPG2)a(% Inh)
IC50 = 13.38 lM) indicating
requirement at this position.
a
potential lipophilic
Compound
8k
8l
63
16
32
63
>11.5 (20)
>13.2 (20)
>12.4 (25)
>12.8 (25)
Activity in the biochemical assay decreased significantly
when R1 was replaced with alkyl or cyclo-alkyl alterna-
tives13 concluding that the series displayed a narrow
SAR in the biochemical assay and essentially flat SAR
in the whole cell assay.
8i
8f
Whole cell activity (vs M.tb), HEPG2 toxicity data for a selection of
imidazopiperidine inhibitors.
a Concentration, lM; HEP, hepatic cell line; % Inh, percentage inhi-
bition at quoted concentration.
When the three most active compounds (8e,f,g) were
separated into their enantiomers by chiral HPLC, a gen-
eral gain in enantiospecific activity was observed when
compared to the racemic mixtures (compound 9). While
the absolute stereochemistry of the active enantiomer
was not determined, the specificity was highlighted by
the most active enantiomer 9b (IC50 = 80 nM vs
>10 lM for 9a) where the parent racemate 8e had previ-
ously displayed an IC50 = 240 nM.
was significantly less than the desired 10-fold require-
ment (typically 20–30% inhibition at ca. 12 lM) giving
a strong indication of dual off-target activity.
These data suggested that although potent compounds
could be identified, the strategy was not likely to
enhance M.tb cellular activity. Work is ongoing to in-
crease the therapeutic index by identification of the sec-
ondary mechanism of action.
The requirement of the imidazole group was investi-
gated by combining the features of activity for the active
imidazopiperidine compounds (8e,g) but replacing the
imidazole group with a phenyl group. This caused a
complete loss of activity in the biochemical assay and
no significant whole cell activity indicating that the imid-
azole group was indeed a necessary feature for activity
against the enzyme.
Acknowledgments
The author thank Mr. Eric Hortense for chiral chroma-
tography separations, Dr. Andy Merritt, Mr. Tony
Dean and Dr. Chun-Wa Chung for useful discussions
and reading this manuscript.
No significant broad-spectrum antibacterial activity was
seen for 8e, 8f, 8i, 8j, 8k, 9b, 9c and 9d when compared
with triclosan, amoxicillin and ciprofloxacin15 suggest-
ing that this series may be acting by a non-specific
mechanism.
References and notes
1. Current WHO initiatives to combat TB can be viewed on
the following websites; (a) http://www.stoptb.org/
stop_tb_initiative/; (b) http://www.theglobalfund.org/en/.
2. (a) Raviglione, M. C. Tuberculosis 2003, 83, 4; (b) Espinal,
M. A. Tuberculosis 2003, 83, 44.
3. Tuberculosis, HIV/Aids and Malaria, The status and impact
of the three diseases. WHO, Geneva (2005), report which is
available on the following website: <http://www.theglob-
alfund.org/en/about/publications/>.
4. Groups at Risk: WHO report on the tuberculosis epidemic.
WHO, Geneva (1996). This report is available on the
following website: <http://www.who.ch/programmes/gtb/
tbrep_96/tbreport.html/>.
In an attempt to address the disconnect between the
in vitro and cellular activity, it was decided to profile a
selection of these compounds through a high through-
put artificial membrane assay16 which relies on passive
diffusion for the compound to permeate through a
monolayer of phospho-lipid membrane. It was found
that these compounds displayed a range of permeabili-
ties with the separated enantiomers all displaying med-
ium to high permeabilities when compared with the
standard.
5. Gupta, R. Int. J. Tuberc. Lung. Dis. 2003, 7, 410.
6. (a) Dessen, A.; Quemard, A.; Blanchard, J. S.; Jacobs, W.
R.; Sacchettini, J. C. Science 1995, 267, 1638; (b) Quem-
ard, A.; Sacchettini, J. C.; Dessen, A.; Vilcheze, C.;
Bittman, R.; Jacobs, W. R.; Blanchard, J. S. Biochemistry
1995, 34, 8235; (c) Brennan, P. J.; Nikaido, H. Annu. Rev.
Biochem. 1995, 64, 29; (d) Lee, R. E. et al. Curr. Top.
Microbiol. Immunol. 1996, 215, 1; (e) Rozwarski, D. A.;
Grant, G. A.; Barton, D. H. R.; Jacobs, W. R.; Sacchet-
tini, J. C. Science 1998, 279, 98; (f) Parikh, S. L.; Xiao, G.;
Tonge, P. J. Biochemistry 2000, 39, 26, 7645.
7. (a) Vilcheze, C.; Morbidoni, H. R.; Weisbrod, T. R.;
Iwamoto, H.; Sacchettini, J. C.; Jacobs, W. R. J. Bacteriol.
2000, 182, 4059; (b) Schroeder, E. K.; de Souza, O. N.;
Santos, D. S.; Blanchard, J. S.; Basso, L. A. Curr. Pharm.
Biotechnol. 2002, 3, 197.
Potential aggregation or promiscuous inhibition within
the series was assessed using an in-house assay.17 Com-
pounds screened were found not to have significant
activities. This was reaffirmed with the enantiomers:
9a, b, c and d where there was clear demonstration of
enantiospecificity for one enantiomer over the physio-
chemically similar other. With the level of potency ob-
served for specific enantiomers it was concluded that
this series is highly unlikely to be acting in an aggrega-
tive fashion.
However, upon further examination of this series within
alternative cell lines it was found that a degree of toxic-
ity was prevalent. When incubated against liver
(HEPG2) cells Table 2 at concentrations 10-fold less
than the highest meaningful MIC, selectivity (vs MIC)
8. (a) Sacchettini, J. C.; Blanchard, J. S. Res. Microbiol.
1996, 147, 36; (b) Quemard, A.; Dessen, A.; Sugantino,
M.; Jacobs, W. R.; Sacchettini, J. C.; Blanchard, J. S.

2744
M. D. Wall et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2740–2744
J. Am. Chem. Soc. 1996, 118, 1561; (c) Stoeckle, M. Y.
for 3 h, filtered and washed with DCM and Methanol
(20 mL). The filtrate was concentrated to give 8 and
analysed by HPLC-MS. All compounds which showed
greater than 90% purity and a weight greater than 1 mg
were selected for testing.
et al. J. Infect. Dis. 1993, 168, 1063; (d) Heym, B. et al.
Lancet 1994, 344, 293; (e) Heym, B. et al. Mol. Microbiol.
1995, 15, 235; (f) Cockerill, F. R. et al. J. Infect. Dis. 1995,
171, 240; (g) Kapur, V. et al. Arch. Pathol. Lab. Med.
1995, 119, 131; (h) Rouse, D. A. et al. Antimicrob. Agents
Chemother. 1995, 39, 2472; (i) Musser, J. M. et al.
J. Infect. Dis. 1996, 173, 196; (j) Telenti, A. et al. J. Clin.
Microbiol. 1997, 35, 719.
Separation of enantiomers from the parent racemates 8e, f
andg was achieved using preparative chiral HPLC. On a
Chiralpak AD column (2 cm) with
a loading of
15 mg mLÀ1, a flowrate of 15 mL minÀ1 and detection at
k = 215 nm, fractions were collected using an ethanol/
heptane gradient up to 30% ethanol (over 15 min) to give
the separation of the enantiomers 9a–d.
9. (a) Heym, B.; Cole, S. T. Res. Microbiol. 1992, 143, 721;
(b) Cockerill, F. R.; Uhl, J. R.; Temesgen, Z.; Zhang, Y.;
Stockman, L.; Roberts, G. D.; Williams, D. L.; Kline, B.
C. J. Infect. Dis. 1995, 171, 240; (c) Haym, B.; Alzari, P.
M.; Honore, N.; Cole, S. T. Mol. Microbiol. 1995, 15, 235;
(d) Musser, J. M.; Kapur, V.; Williams, D. L.; Krieswirth,
B. N.; Soolingen, D. V.; van Embden, J. D. A. J. Infec.
Dis. 1996, 173, 196.
10. (a) Stocker, F. B.; Fordice, M. W.; Larson, J. K.;
Thorstenson, J. H. J. Org. Chem. 1966, 31, 7, 238; (b)
Stocker, F. B.; Evans, A. J. J. Org. Chem. 1990, 55, 3370.
11. General procedure for solid phase synthesis of imidazopi-
12. A range of aldehydes were chosen from work carried out
on docking within the InhA active site, pharmacophore
generation and best fit models.
13. Data for these analogues have been omitted.
14. Assays were performed using an appropriate dilution of
enzyme with 0.25 mM DDCoA and 0.2 mM NADH in
30 mM PIPES buffer (pH 6.8). The final reaction
volume was 50 lL. The reaction was stopped and the
signal developed by the addition of 25 lL of 0.4
mg mLÀ1 MTS, 4 lg mLÀ1 PMS and 0.3 M urea in
PBS buffer.
peridine analogues. The modified resin
4 (200 mg,
0.094 mmol) was washed with dichloromethane and then
allowed to stir in DMF for 15 min. The amine was added
and the suspension allowed to stir for a further 1 h prior to
the addition of sodium triacetoxy borohydride (10 equiv,
0.94 mmol, 247 mg). The suspension was allowed to stir at
room temperature for 18 h. after which it was filtered and
washed with DCM (50 ml) DMF (50 ml), ethanolamine
(10%, in DMF, 50 ml) DMF and finally methanol (50 mL)
to give 5. The resin was dried under vacuum at 40 °C.
The acid 3 (323 mg, 5 equiv), HATU (5 equiv, 178.6 mg)
andPEA (163 lL) were added to DMF (5 mL) and stirred
for 1 h. The modified resin 5 as a suspension in DMF was
added and the suspension allowed to stir at room
temperature for 18 h.
The resin was filtered, and washed with DMF, DCM,
methanol and DCM (20 mL,respectively) to give 6 which
was used immediately without drying. The resin 6 was
added to a solution of piperidine in DMF (20% solution)
and the suspension allowed to stir at room temperature for
2 h prior to filtration and washing with DCM, Methanol
and DCM (20 mL). The loaded resin, aldehyde and acetic
acid were stirred at room temperature for 2 h prior to the
addition of the sodium triacetoxy borohydride in one
portion. The suspension was allowed to stir at room
temperature for 56 h prior to filtration and washing with
DCM, MeOH and DCM (20 mL) to give 7, which was
used without further modification for the next step.
To a solution of TFA/H2O (80:20) was added 7 in one
portion. The suspension was stirred at room temperature
15. The following strains of bacteria using the standards
amoxicillin, ciprofloxacin and triclosan were used to
profile the compounds cited vide supra: Staphylococcus
aureus ATCC29213, Streptococcus pneumoniae 1629,
S. pneumoniae Ery2, Klebsiela pneumoniae KP1, K. pneu-
moniae KP3, Escherichia coli EC2, E. coli ATCC25922,
Proteus mirabilis PM1, P. mirabilis PM5, Pseudomonas
aeruginosa PA1, P. aeruginosa ATCC 27853, Stenotroph-
omonas maltophilia T68214, Haemophilus influenzae
20001H, H. influenzae 07001H. Activity was only seen
with 8K at 2 lg mLÀ1 for S. aureus ATCC 29213.
16. Veber, D. F.; Johnson, S. R.; Cheng, H. Y.; Smith, B. R.;
Ward, K. W.; Kopple, K. D. J. Med. Chem. 2002, 45,
2615, Permeability is classified as High > 100 nm sÀ1
,
medium > 10 < 100 nm sÀ1 and low < 10 nm sÀ1. Triclo-
san = 240 nm sÀ1
,
9a and b = 80–180 nm sÀ1
, 9c and
.
d = 210–240 nm sÀ1
17. We use an in-house assay for determination of compounds
which are likely to behave in an non-specific fashion, the
results of which complement the work of (a) McGovern, S.
L.; Caselli, E.; Grigorieff, N.; Shoichet, B. K. J. Med.
Chem. 2002, 45(8), 1712; (b) McGovern, S. L.; Helfand, B.
T.; Feng, B.; Shoichet, B. K. J. Med. Chem. 2003; (c)
Seidler, J.; McGovern, S. L.; Thompson, D. N.; Shoichet,
B. K. J. Med. Chem. 2003, 4477; (d) Feng, Y. B.; Shoichet,
K. B. J. Med. Chem. 2006, 49, 2151; (e) Ryan, A. J.; Gray,
N. M.; Lowe, P. N.; Chung, C-W. J. Med. Chem. 2003, 46,
3448.