Heteroalicyclic carboxamidines as inhibitors of inducible nitric oxide synthase; The identification of (2R)-2-pyrrolidinecarboxamidine as a potent and selective haem-co-ordinating inhibitor

Article abstract of DOI:10.1016/j.bmcl.2011.03.038

Heteroalicyclic carboxamidines were synthesised and evaluated as inhibitors of nitric oxide synthases. (2R)-2-Pyrrolidinecarboxamidine, in particular, was shown to be a highly potent in vitro (IC₅₀ = 0.12 μM) and selective iNOS inhibitor (>100-fold vs both eNOS and nNOS), with probable binding to the key anchoring glutamate residue and co-ordination to the haem iron.

Full text of DOI:10.1016/j.bmcl.2011.03.038

Bioorganic & Medicinal Chemistry Letters 21 (2011) 3037–3040
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Heteroalicyclic carboxamidines as inhibitors of inducible nitric oxide
synthase; the identification of (2R)-2-pyrrolidinecarboxamidine
as a potent and selective haem-co-ordinating inhibitor
⇑
Robert J. Young , Wendy Alderton, Anthony D. R. Angell, Paul J. Beswick, David Brown, C. Lynn Chambers,
Miriam C. Crowe, John Dawson, Christopher C. F. Hamlett, Simon T. Hodgson, Savvas Kleanthous,
Richard G. Knowles, Linda J. Russell, Richard Stocker, James M. Woolven
GlaxoSmithKline R&D, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Heteroalicyclic carboxamidines were synthesised and evaluated as inhibitors of nitric oxide synthases.
Received 2 March 2011
Revised 9 March 2011
Accepted 9 March 2011
Available online 21 March 2011
(2R)-2-Pyrrolidinecarboxamidine, in particular, was shown to be a highly potent in vitro (IC50 = 0.12 lM)
and selective iNOS inhibitor (>100-fold vs both eNOS and nNOS), with probable binding to the key
anchoring glutamate residue and co-ordination to the haem iron.
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Nitric oxide synthase
Iron co-ordinating inhibitors
The simple radical gas nitric oxide (NO), which was shown to be
the endothelium derived relaxing factor,1 has been implicated in a
diverse range of physiological roles.2 It is produced by oxidation of
tween two nitrogens of the guanidine and the conserved glutamate
residue and a backbone tryptophan carbonyl.12 The third nitrogen
of the guanidine is thus held in close proximity to the haem iron to
facilitate its oxidation to NO.13,14 Evidence to date of compounds
exploiting both the glutamate–amidine salt bridge and forming a
classical type II co-ordination of the haem iron has been limited
to nNOS inhibitors with thioether substituted actetamidine mo-
tifs.15 In their detailed study, Martell et al. conclude that hydropho-
bic contacts around the haem facilitate a strong contribution to
stabilisation of the sulphur–iron binding. Thiocitrulline was pur-
L
-arginine by one of three isoforms of Nitric Oxide Synthase (NOS).3
Picomolar concentrations of NO have essential roles in vascular
homeostasis and neurotransmission; produced by the constitutive,
calcium-regulated, isoforms endothelial (eNOS) and neuronal
(nNOS) respectively.4 Nanomolar concentrations of NO are pro-
duced by the inducible isoform (iNOS), which is produced in re-
sponse to stimuli such as cytokines or bacterial endotoxins, to
play an essential role in host defence mechanisms.5 However, over-
production of NO can have pathophysiological consequences; the
upregulation of iNOS is implicated in septic shock6 and a number
of inflammatory conditions.7 The selective inhibition of iNOS (in
order to maintain the beneficial actions of the constitutive iso-
forms) has therefore been an attractive and well-pursued thera-
peutic target.8 However, to date, few compounds with genuine
potency and selectivity have been discovered;9 amongst this class,
the amino acid GW27415010 has progressed furthest in clinical
trials.11
O
O
Trp
HN
NH
O
Trp
HN
NH
O
(CH2)n
O
H
H
Fe
NH2
N+
X
Fe
N
N
H
H
B
N+
H
The majority of potent inhibitors feature a basic amidine or gua-
nidine motif, which satisfies the same binding interactions that
A
H
O
H
Arg
O
recognise guanidine in the natural substrate, L-arginine (Fig. 1A).
Specifically, there are three hydrogen-bonding interactions be-
O
O
Glu
Glu
Figure 1. Binding pattern of conserved Glu and Trp residues,12 which, in A, bind the
guanidine motif of the arginine substrate above the haem iron to facilitate oxidation
in all NOS isoforms. The envisaged heterocyclic carboxamidine interactions are
shown in B, illustrating the backbone recognition and proximity of the haem iron.
⇑
Corresponding author. Tel.: +44 1438 768372.
E-mail address: Rob.J.Young@gsk.com (R.J. Young).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.03.038

3038
R. J. Young et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3037–3040
ported to interact in this manner16 and some (aromatic) heterocy-
clic amidines have been made to probe for such an interaction, but
without apparent success.17
We report here on the synthesis and evaluation of heteroalicy-
clic carboxamidines (Fig. 1B) targeted from inspection of the NOS
crystal structures18 and data on pyrazolo19 and carbocyclic20 car-
boxamidines. Increased potency was sought by probing for addi-
tional interactions with the protein (type I interaction) or to
reach the co-ordination sphere of the haem iron to enable type II
interactions.
The synthetic strategy was to access carboxamidines (2–15) via
the requisite imidate (D) derived from available acid (A), amide (B)
or nitrile (C) precursors (Scheme 1). Specifically, acids (or N-Boc
protected amino acids) A were converted to amides B by sequential
treatment with carbonyl diimidazole then ammonia. Amides were
converted to the requisite imidates D by treatment with trim-
ethyloxonium tetrafluoroborate in dichloromethane. Alternatively,
the imidates D were formed directly by reaction of nitriles with a
saturated solution of hydrogen chloride in methanol. The reaction
of imidates D with ammonium chloride in methanol gave the re-
quired amidines as hydrogen chloride salts in good yield after puri-
fication by reverse phase (C-18) solid phase extraction. As
necessary, Boc protection was removed following amidination in
95:5 trifluoroacetic acid/water, affording the target compounds
as trifluoroacetate salts after similar C-18 purification.
In vitro screening21 against recombinant human iNOS22
indicated that some significant increases in potency were achieved
through our modifications (Table 1). Analysis of the SAR shows the
importance of ring size, presence/position of a heteroatom and
stereochemistry. Specifically, varying the ring size of carbocyclic
amidines showed that cyclobutyl carboxamidine gave the best
potency, closely followed by the smaller cyclopropyl analogue.
Expansion to cyclopentyl carboxamidine resulted in a substantial
loss of potency. The introduction of a heteroatom in the 2-position
could dramatically influence activity; in the case of oxygen 7 or
sulphur 8 there is little change compared with the carbocyclic 3,
OH
O
NH2
O
m
n
(a)
m
n
X
X
B
A
(b)
X = CH2, O, S, NBoc
OMe
.HCl
NH
NH2
NH
m
n
m
n
(d)
(e)
X
Y
D
1 to 11
Y = CH2, O, S, NH
(c)
N
m
X
n
C
Scheme 1. (a) CDI, NH3, DMF (b) Me3OBF4, DCM (c) HCl(g), MeOH (d) NH4Cl, MeOH; (e) TFA/H2O (95:5) if required. Compound 1 was purchased; compounds 3, 7 and 10–13
were made from available acids (A), 4, 5, 6, 9, 14, 15 from amides (B) and 2, 8 from nitriles (C). All compounds gave satisfactory analytical data (LSMS/1H NMR) supportive of
the proposed structures and at least >95% purity.
Table 1
Human iNOS IC50 values21 for a variety of ring systems, with GW27415010 as a comparison.
Entry
1
Compound
hiNOS IC50
(
lM)
Entry
Compound
hiNOS IC50
5.3
(lM)
NH2
NH2
NH
0.79
8
S
NH
NH2
NH2
2
0.38
9
0.7
N
N
NH
NH
NH2
NH2
3
4
5
6
4.4
>100
51
10
0.18
0.03
0.2
NH
NH2
NH
NH2
(R)-11
(S)-12
13
N
N
NH
N
NH
NH2
NH2
NH
N
NH
NH2
NH2
N
N
58
14
NH
NH
NH2
CO2H
NH2
S
O
7
8.9
GW 274150
1.4
HN
N
NH
H

R. J. Young et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3037–3040
3039
but nitrogen had a profound effect on increasing activity in four-
and five-membered rings compared with their carbocyclic ana-
logues. The physical basis of why nitrogen acted as a better ligand
for iron than oxygen or sulphur was not established, given struc-
tural similarities and availability of lone pairs.
Of most significance was the impact of stereochemistry on
activity—most notably in the big difference exhibited by the enan-
tiomeric pyrrolidine derivatives, (R)-14 and (S)-15 (Table 2). The R-
isomer was shown to be 100-fold more potent than the S, with an
IC50 value of 0.12 lM. These compounds were also highly selec-
tive—the more potent enantiomer being >100-fold selective over
both eNOS and nNOS. (Table 2)
Docking studies of these compounds within a model of the iNOS
active site were conducted using GOLD (Genetic Optimisation for
Ligand Docking).23 The model was built based on X-ray co-ordi-
indicated that the amidine pKa was 9.2 and that for the ring nitro-
gen 4.8; consistent with the ring nitrogen being essentially union-
ised at pH 7.4, thus presenting a lone pair to co-ordinate with iron.
The docking of (S)-15 in this manner indicated the requirement of a
conformational shift in the protein structure to accommodate the
ring.
Further evidence to support this hypothesis was provided by
the data on the azetidines, (R)-11 and (S)-12. These compounds
also showed variation in activity between the enantiomers, but
there was only 10-fold difference in potency. This is consistent
with binding in a similar fashion to that proposed for the pyrrol-
idines, but as the azetidines are smaller, any conformational shift
required for the S-isomer is less sterically demanding.
The most compelling evidence to support the postulated bind-
ing mode was secured through an examination of UV spectral
changes obtained in competitive binding experiments using the
nates of the
D114 monomer and published stereoscopic pictures
of the
D
65 dimer available at the time.24 The preferred docked
D
65-iNOS construct.25 Addition of both (R)-14 and (S)-15 to the
poses suggested a clear rationale for the difference between (R)-
14 and (S)-15. With the expected hydrogen bonds between
Glu377/Trp372 and the mono-protonated carboxamidine in place,
the neutral ring nitrogen of (R)-14 could be located in close prox-
imity to the haem iron (Fig. 2). This atom could therefore be acting
as the sixth ligand to the metal. Physical measurements on (R)-14
enzyme caused clear spectral shifts consistent with type II (Iron)
binding.26 The R-enantiomer clearly bound with much greater
affinity (binding Kd for (R)-14 was 0.63 0.16
lM) than the S-
(Kd >40 M for (S)-15), consistent with the modelling predictions.
l
However, the selectivity of the compound cannot be rationalised
by possible steric interactions around the binding site in the con-
stitutive isoforms, suggesting an alternative rationale; indeed the
selectivity of amino acid amidine derivatives9,10 is clearly not
based on steric factors given the high local sequence homology
and small size of the inhibitors.27 It might be that subtle differ-
ences in redox chemistry across the isoforms, which give rise to
selectivity in substrate based analogues,28 contribute to this.
In conclusion, (2R)-2-pyrrolidinecarboxamidine, (R)-14, is a
highly potent and highly selective inhibitor of iNOS. It is more po-
tent than cyclopropyl carboxamidine and has a level of selectivity
unprecedented in other simple amidines and guanidines. The in-
creases in potency were rationalised by an interaction with the
haem functionality, presenting an important improvement in
inhibitor design. It is interesting to note that (R)-14, with a c log P
of À0.93 and eight heavy atoms (HA) has a ligand efficiency
[ÀRT ln(IC50)/HA] of 1.22 and the Lipophilicity Ligand Efficiency
(pIC50 À c log P) is 7.9.
Table 2
IC50 values21 and selectivity data for human NOS isoforms measured for single
enantiomers of 9 and 11, with GW27415010 as a comparison
Entry
R
IC50
(l
M)
Selectivitya
iNOS
0.03
eNOS
nNOS
0.26
i/e
i/n
NH2
(R)-11
(S)-12
0.25
8
9
N
N
NH
NH2
0.2
2.4
15
1.7
14
12
9
NH
NH2
(R)-14
0.12
125
117
N
NH
NH2
(S)-15
9.4
1.4
>100
466
>100
145
>10
333
>10
104
N
Acknowledgements
NH
GW 274150
—
We wish to thank Alan Hill for the pKa measurements, Mike
Hann for help with iNOS model construction and many in the
broader iNOS programme team for their contributions.
a
Selectivity expressed as the ratio of the IC50 of eNOS or nNOS to iNOS.
References and notes
1. Palmer, R. M. J.; Ferrige, A. G.; Moncada, S. Nature 1987, 327, 524.
2. Kroncke, K. D.; Fehsel, K.; Kolb-Bachofen, V. Clin. Exp. Immunol. 1998, 113, 147.
3. Groves, J. T.; Wang, C. C-Y. Curr. Opin. Chem. Biol. 2000, 4, 687.
4. Christopherson, K. S.; Bredt, D. S. J. Clin. Invest. 1997, 100, 2424.
5. Moncada, S.; Palmer, R. M.; Higgs, E. A. Pharmacol. Rev. 1991, 43, 109.
6. Wolfe, T. A.; Dasta, J. F. Annu. Pharmacother. 1995, 29, 36.
7. Kröncke, K-D.; Fehsel, K.; Klob-Bachofen, V. Clin. Exp. Immunol. 1998, 113, 147.
8. (a) Salerno, L.; Sorrenti, V.; DiGiacomo, C.; Romeo, G.; Siracusa, M. A. Curr.
Pharm. Des. 2002, 8, 177; (b) Maddaford, S.; Annedi, S. C.; Ramnauth, J.; Rakhit,
S. Annu. Rep. Med. Chem. 2009, 44, 27.
9. Alderton, W. K.; Cooper, C. E.; Knowles, R. G. Biochem. J. 2001, 357, 593.
10. Young, R. J.; Beams, R. M.; Carter, K.; Clark, H. A. R.; Coe, D. M.; Chambers, C. L.;
Davies, P. I.; Dawson, J.; Drysdale, M. J.; Franzman, K. W.; French, C.; Hodgson,
S. T.; Hodson, H. F.; Kleanthous, S.; Rider, P.; Sanders, D.; Sawyer, D. A.; Scott, K.
J.; Shearer, B. G.; Stocker, R.; Smith, S.; Tackley, M. C.; Knowles, R. G. Bioorg.
Med. Chem. Lett. 2000, 10, 597.
11. (a) Singh, D.; Richards, D.; Knowles, R. G.; Schwartz, S.; Woodcock, A.; Langley,
S.; O’Connor, B. J. Am. J. Respir. Crit. Care Med. 2007, 176, 988; (b) Hesslinger, C.;
Strub, A.; Boer, R.; Ulrich, W.-R.; Lehner, M. D.; Braun, C. Biochem. Soc. Trans.
2009, 37, 886.
12. These are identified as follows in the three human isoforms: hiNOS: Glu377/
Trp372; heNOS: Glu363/ Trp358; hnNOS: Glu592/ Trp587 and in murine iNOS
as: Glu371/ Trp366
Figure 2. Docked pose of (R)-14 in the active site model of iNOS (rendered using
Pymol from iNOS structural data).

3040
R. J. Young et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3037–3040
13. Raman, C. S.; Li, H.; Martásek, P.; Southan, G.; Masters, B. S. S.; Poulos, T. L.
Biochemistry 2001, 40, 13448.
Simon, M. I., Eds.; Academic Press: London, 1996; 268, pp 449–460. Nitric
Oxide, Part A.
14. Stuehr, D. L. Annu. Rev. Pharmacol. Toxicol. 1997, 37, 339.
15. Martell, J. D.; Li, H.; Doukov, T.; Martásek, P.; Roman, L. J.; Soltis, M.; Poulos, T.
L.; Silverman, R. B. J. Am. Chem. Soc. 2010, 132, 798.
16. Frey, C.; Krishnaswamy, N.; McMillan, K.; Spack, L.; Gross, S. S.; Masters, B. S.;
Griffith, O. W. J. Biol. Chem. 1994, 269, 26083.
17. Moormann, A. E.; Metz, S.; Toth, M. V.; Moore, W. M.; Jerome, G.; Kornmeier, C.;
Manning, P.; Hansen, D. W.; Pitzele, B. S.; Webber, R. K. Bioorg. Med. Chem. Lett.
2001, 11, 2651.
23. Dockings into a model of the crystal structure of iNOS were performed with
GOLD (Genetic Optimisation for Ligand Docking), Cambridge Crystallographic
Data Centre, Cambridge, UK (http://www.ccdc.cam.ac.uk); see Jones, G.;
Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R. J. Mol. Biol. 1997, 267, 727.
24. (a) Crane, B. R.; Arvai, A. S.; Ghosh, D. K.; Wu, C.; Getzoff, E. D.; Stuehr, D. J.;
Tainer, J. A. Science 1998, 279, 2121; (b) Raman, C. S.; Li, H.; Martsek, P.; Kral, V.;
Masters, B. S.; Poulos, T. L. Cell 1998, 95, 939.
25. McMillan, K.; Masters, B. S. S. Biochemistry 1993, 32, 9875.
18. Poulos, T. L.; Raman, C. S.; Li, H. Structure 1998, 6, 255.
19. (a) Lee, Y.; Martásek, P.; Roman, L. J.; Silverman, R. B. Bioorg. Med. Chem. Lett.
2000, 10, 2771; (b) Southan, G. J.; Gauld, D.; Lubeskie, A.; Zingarelli, B.;
Cuzzocrea, S.; Salzman, A. L.; Szabo, C.; Wolff, D. Biochem. Pharmacol. 1997, 54,
409.
20. Moore, W. M.; Webber, R. K.; Fok, K. F.; Jerome, G. M.; Kornmeier, C. M.; Tjoeng,
F. S.; Currie, M. G. Bioorg. Med. Chem. 1996, 9, 1559.
21. Time dependant kinetics were observed on iNOS and inhibitory values are
quoted over the 15–30 min timeframe using the oxyhaemaglobin protocol
26. The measured Kd for (R)-14 was 0.63 0.16
lM, with saturation of spectral
changes at 10.5 M. 40 M of (S)-15 showed only a minor shift in the spectrum
l
l
from type I to type II and higher concentrations caused precipitation of the
protein.
27. (a) Fischmann, T. O.; Hruza, A.; Niu, X. D.; Fossetta, J. D.; Lunn, C. A.; Dolphin, E.;
Prongay, A. J.; Reichert, P.; Lundell, D. J.; Narula, S. K.; Weber, P. C. Nat. Struct.
Biol. 1999, 6, 233; (b) Ji, H.; Li, H.; Martásek, P.; Roman, L. J.; Poulos, T. L.;
Silverman, R. B. J. Med Chem. 2009, 52, 779; (c) Garcin, E. D.; Arvai, A. S.;
Rosenfeld, R. J.; Kroeger, M. D.; Crane, B. R.; Andersson, G.; Andrews, G.;
Hamley, P. J.; Mallinder, P. R.; Nicholls, D. J.; St-Gallay, S. A.; Tinker, A. C.;
Gensmantel, N. P.; Mete, A.; Cheshire, D. R.; Connolly, S.; Stuehr, D. J.; Aberg, A.;
Wallace, A. V.; Tainer, J. A.; Getzoff, E. D. Nat. Chem. Biol. 2008, 4, 700.
28. Bryk, R.; Wolff, D. J. Biochemistry 1998, 37, 4844.
with 30 lM of arginine: Dawson, J.; Knowles, R. G. In Meth. Mol. Biology;
Titheradge, M. A., Ed.; Humana Press: Totowa NJ, 1998; 100, pp 237–242.
22. Charles, I. G.; Scorer, C. A.; Moro, M. A.; Fernandez, C.; Chubb, A.; Dawson, J.;
Foxwell, N.; Knowles, R. G.; Baylis, S. A. In Methods in Enzymology; Abelson, J. N.,