The discovery of potent and selective kynurenine 3-monooxygenase inhibitors for the treatment of acute pancreatitis

Article abstract of DOI:10.1016/j.bmcl.2017.02.078

A series of potent, competitive and highly selective kynurenine monooxygenase inhibitors have been discovered via a substrate-based approach for the treatment of acute pancreatitis. The lead compound demonstrated good cellular potency and clear pharmacodynamic activity in vivo.

Full text of DOI:10.1016/j.bmcl.2017.02.078

Accepted Manuscript
The discovery of potent and selective kynurenine 3-monooxygenase inhibitors
for the treatment of acute pancreatitis
John Liddle, Benjamin Beaufils, Margaret Binnie, Anne Bouillot, Alexis A.
Denis, Michael M. Hann, Carl P. Haslam, Duncan S. Holmes, Jon P. Hutchinson,
Michael Kranz, Andrew McBride, Olivier Mirguet, Damian J. Mole,
Christopher G. Mowat, Sandeep Pal, Paul Rowland, Lionel Trottet, Iain J. Uings,
Ann L. Walker, Scott P. Webster
PII:
S0960-894X(17)30221-4
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.02.078
BMCL 24750
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
9 January 2017
27 February 2017
28 February 2017
Please cite this article as: Liddle, J., Beaufils, B., Binnie, M., Bouillot, A., Denis, A.A., Hann, M.M., Haslam, C.P.,
Holmes, D.S., Hutchinson, J.P., Kranz, M., McBride, A., Mirguet, O., Mole, D.J., Mowat, C.G., Pal, S., Rowland,
P., Trottet, L., Uings, I.J., Walker, A.L., Webster, S.P., The discovery of potent and selective kynurenine 3-
monooxygenase inhibitors for the treatment of acute pancreatitis, Bioorganic & Medicinal Chemistry Letters (2017),
doi: http://dx.doi.org/10.1016/j.bmcl.2017.02.078
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The discovery of potent and selective kynurenine 3-
monooxygenase inhibitors for the treatment of acute pancreatitis
John Liddle*a, Benjamin Beaufilsb, Margaret Binniec, Anne Bouillotb, Alexis A. Denisb,
Michael M. Hanna, Carl P. Haslama, Duncan S. Holmesa, Jon P. Hutchinsona, Michael
Kranza, Andrew McBridec, Olivier Mirguetb, Damian J. Moled, Christopher G. Mowate,
Sandeep Pala, Paul Rowlanda, Lionel Trottetb, Iain J. Uingsa, Ann L. Walkera, Scott P.
Websterc
a) GlaxoSmithKline, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, UK. b) GlaxoSmithKline, Les Ulis, Paris, France.
c) Centre for Cardiovascular Science, University of Edinburgh, UK. d) MRC Centre for Inflammation Research, University of Edinburgh,
U.K. e) EastChem School of Chemistry, University of Edinburgh, UK.
Acute pancreatitis (AP) is an acute inflammation of the pancreas usually triggered by gallstones or
excessive alcohol consumption. Aproximately 20% of patients will develop a systemic inflammatory
response progressing to multiple organ dysfunction syndrome. The mortality rate in this group of
patients exceeds 20% with most deaths occurring within the first 7 days of hospital admission.1 AP
remains a significant unmet medical need with the current standard of care being entirely supportive.
A recent prospective study of patients diagnosed with AP showed a strong association between
disease severity and the plasma concentrations of the cytotoxic tryptophan metabolite
3-hydroxykynurenine (3HK), which is formed exclusively by kynurenine 3-monooxygenase (KMO).2
KMO-deficient mice cannot generate 3HK and, compared to wild-type animals, appear resistant to
kidney and lung damage when subjected to a model of pancreatic injury driven secondary organ
dysfunction.3 This combination of clinical observation and preclinical rodent model data suggests that
the activation of this pathway and the formation of the cytotoxic metabolite 3HK make a significant
contribution to the observed pathology. We therefore sought to design potent and selective KMO
inhibitors for the treatment of AP with properties amenable to intravenous administration, a
requirement for critically ill patients. AP progresses relatively quickly and hence early treatment
following clinical diagnosis is important. Our objective therefore was to develop inhibitors with
excellent aqueous solubility that could achieve high levels of KMO inhibition rapidly from a single
bolus dose. Sustained KMO inhibition would be maintained from a continuous infusion.
KMO is a β-nicotinamide adenine dinucleotide 2′-phosphate (NADPH)-dependent flavin hydroxylase
that employs the redox co-factor flavin adenine dinucleotide (FAD) to catalyse the hydroxylation of
kynurenine (Kyn) to form 3-HK using molecular oxygen.4-5 This reaction occurs on the pathway of
tryptophan catabolism and leads ultimately to the de novo synthesis of the nicotinamide moiety of
NAD/NADP. The structure of human KMO has yet to be solved, but a recent paper disclosing the
crystal structure of Saccharomyces cerevisiae KMO suggests that: i) the phenyl ring of Kyn 1 (Figure
1) occupies a hydrophobic pocket, ii) the carbonyl forms a H-bond to a conserved Gln325 and, iii) the
carboxylate forms a salt bridge to Arg83 and H-bond to Tyr97.6 The aniline is thought to H-bond to

an oxygen atom within the FAD co-factor and the primary amine appears to be devoid of any direct
interactions.
Figure 1: Substrate derived KMO inhibitors
We sought to investigate a substrate-based approach, since Kyn 1 has attractive properties for
optimisation including a low molecular weight and low lipophilicity. Indeed, several groups have
generated potent KMO inhibitors via a substrate-based approach for the potential treatment of
Huntington’s disease, but all appeared to lack the physicochemical properties required for the clinical
development of an intravenous drug.7-8 Reports describing the SAR of the kynurenine phenyl ring
concluded that 3-chloro, 3,4-dichloro and 3-nitro substituents led to relatively potent KMO inhibitors,
exemplified by 2 (KMO IC50 900nM; pIC50 6.0) and UPF-648 3 (KMO IC50 20nM; pIC50 7.7).9-12
Subsequent generations of KMO inhibitors from Roche 4 and CHDI / Evotec 5 exploited a series of
heterocycles to mimic the H-bonding nature of the kynurenine ketone, thought to be important for
activity.13-14
Our approach focussed on removing the undesirable Kyn aniline and carbonyl moieties by cyclising
to form a series of bicyclic propanoic acids, illustrated by 6. Indeed, the simple indazole 7 (Table 1),
which maintains a hydrogen bond acceptor nitrogen in a position similar to the Kyn carbonyl oxygen,
had modest activity (pIC50 5.2) in the KMO biochemical assay.15 To confirm the appropriate position
for the chloro substituent, which had been reported to increase enzyme inhibition, both the 5-chloro 8
and 6-chloroindazole 9 analogues were prepared. The 6-chloro analogue 9 had a dramatic impact,
achieving excellent KMO potency (pIC50 7.3), an increase of approximately 100-fold. The 6-chloro
substituent was maintained to probe the SAR around the propanoic acid and various bicyclic
heterocycles. The propanoic acid chain was sensitive to modification. Introduction of a methyl
group, hydroxyl group or amino group at either position on the alkyl chain, exemplified with

analogues 10-13, was detrimental for potency. Extension of the propanoic chain, compound 14, was
also detrimental although the ethanoic acid 15 showed only a marginal decrease in KMO potency in
the indazole series.
Table 1: Optimisation of propyl chain
Compound
Structure
KMO enzyme15
Mean pIC50 (n)
5.2 (7)
Ligand
efficiency16
7
0.51
0.55
8
6.0 (4)
9
7.3 (7)
5.3 (3)
0.67
0.45
10
11
12
5.6 (4)
6.2 (4)
0.48
0.53
13
14
5.6 (7)
5.5 (4)
0.48
0.47
15
7.0 (6)
0.69
Compounds 10 -13 are racemic mixtures.
Data are given for parent acid, except for 13 which are given for the HCl salt.

KMO enzyme activity was also very sensitive to modification of the core heterocycle, in particular
around the 3-position of the indazole (Table 2). Methyl substitution on the 3-position led to a
significant loss in potency. The indole 17 maintained comparable potency confirming that the
indazole N, postulated to mimic the Kyn carbonyl in H-bonding to Gln 325, was not critical for
inhibition.
Table 2: Optimisation of heterocyclic core
Compound
KMO
enzyme15
mean pIC50
(n)
Ligand
efficiency16
R=
9
7.3 (7)
0.67
0.49
16
5.7 (4)
17
18
19
6.9 (6)
6.3 (4)
0.63
0.47
NA
< 5.0 (3)
20
< 5.2 (10)
0.45
21
22
7.9 (6)
7.6 (3)
0.68
0.69
23
7.1 (6)
0.57
Data are given for parent acid, except for 22 which are given for the 2-amino-(2-
hydroxymethyl)-propane-1,3-diol salt

The benzotriazole 18, benzimidazole 19 and benzimidazolone 20 all lost considerable potency.
Interestingly, the introduction of an oxygen atom at the 3-position, as exemplified by the
benzoxazolone 21, benzisoxazole 22 and benzoxazinone 23, restored good activity.
Before initiating wider exploration of the SAR of the phenyl ring, it was important to confirm that the
chloro substituent was optimal. Indeed, chlorine appeared to be the optimal substituent for the 5-
position of the benzoxazolone with only a lipophilic bromine atom retaining good enzyme inhibition
in the indazole series (Table 3). It is noteworthy that the trifluoromethyl and methyl compounds 26
and 28 lost potency by approximately 100-fold.
Table 3: Optimisation of phenyl substituent
Compound
R
KMO enzyme15
mean pIC50 (n)
Compound
R
KMO enzyme15
mean pIC50 (n)
21
9
Cl
CN
7.9 (6)
6.2 (4)
6.0 (2)
5.3 (4)
Cl
Br
7.3 (7)
7.4 (5)
5.5 (6)
24
27
28
25
CH3O
CF3
CH3
26
Data are given for parent acid
Although the chloroindazole analogue 9 showed potent KMO inhibition in the biochemical assay,
there was no significant inhibition in a cell-based assay using human embryonic kidney (HEK) cells
engineered to overexpress human KMO (Table 4).3 The corresponding benzoxazolone 21 and
benzisoxazole 22 analogues, which were marginally more potent in the biochemical assay, showed
modest KMO inhibition in the cell assay. Because of high ligand efficiency, encouraging aqueous
solubility, photostability and relatively low binding to human serum albumin (HSA), the
benzoxazolone core 21 was chosen for further exploration of the SAR with the aim of increasing
cellular potency. Further substitution on the phenyl ring was tolerated only at the 6-position.
Substitution at the 4 (data not shown) or 7 position was detrimental to enzyme inhibition. The 7-
methyl analogue 29 lost approximately 3-fold in potency. The 6-methyl compound 30 (GSK428)
maintained good enzyme potency (pIC50 7.9) and the increase in lipophilicity (logD 1.3) was
accompanied by a significant increase in cellular potency (pIC50 6.2), likely due to improved cell

permeability. 30 also maintained a good free fraction in HSA. The introduction of an additional
chlorine atom 31 (GSK180) led to excellent enzyme potency although the cell potency was
surprisingly disappointing.
Table 4: Optimisation of phenyl substituents
Compound
Structure
KMO
Enzyme15
mean
KMO Cell3
mean
MW /
HT Sol
HSA
(µM)18
ChromlogD
pH 7.417
Binding19
(%)
pIC50 (n)
R=
pIC50 (n)
9
7.3 (7)
7.9 (6)
<5.0 (4)
5.4 (3)
225 / 1.1
242 / 0.8
>= 207
>= 518
93.6
90.8
21
22
29
7.6 (3)
7.2 (5)
5.3 (3)
ND
256 / 1.0
256 / 1.3
>= 414
>= 354
97.3
94.7
30
7.9 (13)
8.2 (6)
6.2 (3)
5.7 (4)
255 / 1.3
276 / 1.6
337
95.0
96.3
GSK428
31
>= 421
GSK180
Data are given for the 2-amino-(2-hydroxymethyl)-propane-1,3-diol salt except for 9, 21 and 31 which are given for the parent acids
We recently reported a co-crystal structure of 31 bound to P. fluorescens KMO.3 Subsequent work
led to a significantly higher resolution (1.8Å) structure being solved (Figure 2).20 All the residues that
surround the catalytic site in the bacterial enzyme are conserved in human KMO, by sequence
alignment, with the exception of His320, which is a phenylalanine in the human protein. Despite this
high degree of conservation, 31 demonstrated significantly lower potency against P. fluorescens
KMO compared to human KMO (IC50 = 500 nM versus 6 nM, respectively). Nonetheless, significant
interactions within the catalytic site were revealed: the carboxylate forms a salt bridge with Arg84 and
forms a hydrogen bonds with Tyr98 and Asn369; while the oxazolidinone carbonyl forms a hydrogen
bond with the side chain of the C-terminal domain residue Tyr404.

Figure 2: Co-crystal structure of P. fluorescens
KMO in complex with 31. Enzyme residues (grey)
surrounding the bound inhibitor (orange) are
shown in stick representation, with hydrogen
bonds shown as dashed lines (red). Heteroatoms
are coloured according to atom type: nitrogen
(blue), oxygen (red), sulfur (yellow) and chlorine
(green). FAD, flavin adenine dinucleotide.
Ile106
Ala56
Tyr98
FAD
Tyr404
Pro318
His320
Arg84
Asn369
Phe238
Phe319
Met373
The 5-chlorine substituent, critical for good activity within this series and several other series of KMO
inhibitors (exemplified in Figure 1), forms two types of interactions (Figure 3). Firstly, an interaction
(coloured cyan) from a Phe319 aromatic ortho hydrogen (δ +ve) to the “side” of the chorine atom (δ -
ve), and secondly, an interaction (coloured purple) from the π cloud of the Phe 238 aromatic ring (δ -
ve) and the sigma hole “end” of the chlorine atom (δ +ve). Both these types of interactions are
observed for chlorine atoms interacting with electron-rich and electron-poor centres in crystal
structures in the Cambridge Structural Database, and are supported by high level ab initio quantum
mechanics studies.21-22 The nature of these interactions reveal why only a bromine atom (which also
has a δ+ve sigma hole in this position) is able to deliver similar activity, and why methyl substitution
is considerably less potent.
Figure 3: Interaction of 5-chlorine atom substituent with Phe238 and Phe319

The co-crystal structure indicated that further exploration was likely to be limited to substitution on
the 6-position of the benzoxazolone occupying a lipophilic pocket close to the FAD co-factor.
Indeed, a range of groups were tolerated although substitution was sensitive to electronic effects and
lipophilicity. A polar nitrile substituent was detrimental for enzyme potency and approximately 10-
fold weaker than the corresponding methyl compound 30 (Table 5). The lipophilic CF3 group was
also weaker than 30, and this decrease in activity may reflect the impact of the electron-withdrawing
nature of the CF3 group on the H-bonding interaction between the adjacent chlorine atom and the CH
of Phe319. Ethyl substitution maintained good levels of enzyme and cellular inhibition (pIC50 6.5),
albeit with an increase in HSA binding. The isopropyl compound 35 was significantly weaker
confirming the steric restrictions in this pocket.
Substitution at the 6-position with alkoxy groups followed a similar trend in enzyme activity.
Methoxy and ethoxy compounds 36 and 37 were tolerated, achieving excellent levels of enzyme
inhibition. As observed with compounds 21 and 30, increasing lipophilicity led to a dramatic increase
in cellular potency; the ethoxy analogue 37 was 10-fold more potent than the methoxy compound 36,
suggesting that, within this series of carboxylic acids, the lipophilicity of 37 may approach the value
required for optimum cellular penetration.
Table 5: Profiles of key benzoxazolones
R
KMO Enzyme15
mean pIC50 (n)
KMO Cell3
mean pIC50
(n)
MW /
HT Sol
HSA
binding19
(%)
(µM)18
ChromlogD
pH 7.417
32
33
34
35
36
37
38
39
CN
CF3
7.0 (5)
7.4 (5)
8.0 (7)
7.0 (6)
7.9 (8)
8.3 (5)
7.1 (2)
8.5 (5)
4.1 (2)
4.8 (4)
6.5 (4)
5.7 (2)
5.3 (2)
6.4 (7)
5.8 (5)
5.9 (8)
267 / 0.5
310 / 2.1
270 / 1.9
284 / 2.4
272 / 0.8
286 / 1.5
300 / 1.9
298 / 1.6
>= 441
296
78.2
96.9
97.4
98.6
89.0
88.7
92.2
94.3
CH3CH2
(CH3)2CH
CH3O
295
299
422
CH3CH2O
(CH3)2CHO
>= 332
>= 299
>= 375
Data are given for the 2-amino-(2-hydroxymethyl)-propane-1,3-diol salt except for 32 and 36 which are given for the parent acids
The cyclopropyl analogue 39 regained approximately 30-fold increase in enzyme potency compared
to the isopropyl analogue 38, mirroring SAR found within the pyrimidine series 5 reported by CHDI /

Evotec.14 The difference in potency can possibly be rationalised from the P. fluorescens KMO co-
crystal structure of 39, which showed that the cyclopropyl group lies immediately adjacent to the
FAD isoalloxazine ring C6, whereas the isopropyl hydrogens in 38 would clash with the C6
hydrogen. In line with our cellular data, Evotec also found that the cellular potencies of the isopropyl
and cyclopropyl pyrimdine analogues were surprisingly similar given the large difference in the
enzyme inhibitory activity. Such discrepancies in the biochemical and cell potencies may reflect
subtle differences between the FAD region of the protein in the recombinant KMO construct and that
expressed in the HEK cells.
30 offered the preferred overall profile from this series, displaying good cell inhibition and free
fraction in human plasma (2.4%). 30 formed a crystalline salt with 2-amino-(2-hydroxymethyl)-
propane-1,3-diol, which possessed excellent solubility commensurate with intravenous administration
(>25mg/mL in 0.9%w/v sodium chloride solution). Further in vitro profiling showed negligible
inhibition of any of the related enzymes in the tryptophan pathway (<20% inhibition at 100µM
against indoleamine-2,3-dioxygenase (IDO), tryptophan-2,3-dioxygenase (TDO), kynurenine
aminotransferase-1 and -2 (KAT1 and KAT2), and kynureninase (KYNU)) and against a panel of
more than 50 unrelated proteins (IC50 > 30µM). 30 was not mutagenic in the Ames assay or in the
mouse lymphoma L5178Y TK+/- screen when tested in the presence and absence of S9-mix.
Structural and mechanistic insights on the mode of action of 30 will be presented elsewhere.23
30 demonstrated good metabolic stability across species in microsomes (< 0.01 mL/min/mg) and
hepatocytes (<0.01 mL/min/million of cells). Following a single bolus intravenous dose of the 2-
amino-(2-hydroxymethyl)-propane-1,3-diol salt (0.9mg/kg) in the rat, 30 showed limited volume of
distribution (0.52 L/kg), low clearance (4.7ml/min/kg) and a half-life of 3.7h. 30 had slightly lower
cellular potency against the rat enzyme than the human enzyme (pIC50 5.9).
30 is largely excluded from erythrocytes, giving a mean blood:plasma ratio in rat of 0.56, and is
weakly bound by rat plasma proteins, with a free fraction of 9.5%. These parameters suggest that a
total blood concentration of 6µM should deliver half maximal inhibition of KMO activity in vivo.
Bolus injection of four escalating doses (1, 3, 10 and 30mg/kg) of 30 gave increases in the Kyn
concentration that increased in magnitude and duration with dose (Figure 4). These profiles suggest
that drug levels of 5-9 µM are required in blood in vivo to inhibit the KMO activity by 50%, exactly in
line with the potency predicted from in vitro data described above.

18
16
14
12
10
8
6
4
2
0
0
2
4
6
8
10
12
time (hr)
Figure 4: Kynurenine concentrations following intravenous
administration of 30 (1, 3, 10 and 30mg/kg) to rats
Compounds described herein were prepared using standard procedures, following previously
described chemistry, as illustrated below for compound 30.24
Scheme 1. Synthesis of 30. Reagents, conditions and yields. (a) 70% nitric acid, acetic acid, -30 °C then rt, 1 h.
47%; (b) Iron trichloride, active C, methanol, NH2NH2.H2O, 80 °C for 15 h, 79%; (c) 1,1'-carbonyldiimidazole,
THF, 66 °C, 1 h, 86%; (d) 3-bromopropanoic acid, potassium carbonate, acetonitrile, 80 °C, 18 h, 65%.
In conclusion, a series of potent, competitive and highly selective KMO inhibitors have been
discovered via a substrate-based approach. 30 demonstrated good cellular potency and clear
pharmacodynamic activity in vivo. Importantly, this class of KMO inhibitors have excellent aqueous
solubility amenable for intravenous administration and treatment of AP in a critical care setting. Our
subsequent efforts to optimise cell activity and to deliver a clinical candidate will be reported in due
course.25

Acknowledgements
The authors thank Paul Homes and Clare Hobbs for preparation of the Pseudomonas KMO protein.
This work was funded by GlaxoSmithKline. The human biological samples were sourced ethically
and their research use was in accord with the terms of the informed consents. All animal studies were
ethically reviewed and carried out in accordance with Animals (Scientific Procedures) Act 1986 and
the GSK Policy on the Care, Welfare and Treatment of Animals. All animal studies were ethically
reviewed and carried out in accordance with European Directive 86/609/EEC and the GSK Policy on
the Care, Welfare and Treatment of Animals.
References and notes
1. Mole, D. J.; Gungabissoon, U.; Johnston, P.; Cochrane, L.; Hopkins, L.; Wyper, G. M. A.; Skouras, C.; Dibben,
C.; Sullivan,F.; Morris, A.; Ward, H. J. T.; Lawton, A. M.; Donnan, P. T. BMJ Open 2016;6:e011474.
doi:10.1136/bmjopen-2016-011474.
2. Christos, S.; Zheng, X.; Binnie, M.; Homer, N. Z. M.; Murray, T. B, J.; Robertson, D.; Briody, L.; Paterson, F.;
Spence, H.; Derr, L.; Hayes, A. J.; Tsoumanis, A.; Lyster, D.; Parks, R. W.; Garden, O. J.; Iredale, J. P.; Uings, I.
J.; Liddle, J.; Wright, W. L.; Dukes, G.; Webster, S. P.; Mole, D. J. Scientific Reports 6, Article number: 33951
(2016) doi:10.1038/srep33951
3. Mole, D. J. ; Webster, S. P.; Uings, I.; Zheng, X.; Binnie, M.; Wilson, K.; Hutchinson, J. P.; Mirguet, O.; Walker,
A.; Beaufils, B.; Ancellin, N.; Trottet, L.; Bénéton, V.; Mowat, C. G.; Wilkinson, M.; Rowland, P.; Haslam, C.;
McBride, A.; Homer, N. Z. M.; Baily, J. E.; Sharp, M. G. F.; Garden, O. J.; Hughes, J.; Howie, S. E. M.; Holmes,
D. H.; Liddle, J.; Iredale, J. P. Nature Medicine 2016, 22 (2), 202-209.
4. Crozier-Reabe, K. R.; Phillips,R.S.; Moran, G. R. ‡Biochemistry 2008, 47, 12420--12433
5. Smith, J. R.; Jamie, J. F.; Guillemin, G. J. Drug Discovery Today 2015, 21(2), 315-24
6. Amaral, M.; Levy, C.; Heyes, D. J.; Lafite, P.; Outeiro, T. F.; Giorgini, F.; Leys, D.; Scrutton, N. S. Nature 2013,
496, 382-385.
7. Courtney, S.; Scheel, A.; Top. Med. Chem. 2010, 6, 149-176.
8. Stone, T. W.; Darlington, L. G.; Nat. Rev. Drug Discov. 2002, 1, 609.
9. Pellicciari, R.; Natalini, B.; Costantino, G.; Mahmoud, M. R.; Mattoli, L.; Sadeghpour, B. M.; Moroni, F.;
Chiarugi, A.; Carpenedo, R. J. Med. Chem., 1994, 37 (5), 647–655.
10. Sapko, M. T;. Guidetti, P.; Yu, P.; Tagle, D. A.; Pellicciari, R.; Schwarcz, R. Exp Neurol. 2006, 197, 31-40.
11. Drysdale, M. J.; Hind, S. L.; Jansen, M; Reinhard, J. F. J. Med. Chem., 2000, 43, 123–127.
12. Giordani, A.; Pevarello, P.; Cini, M.; Bormetti, R.; Greco, F.; Toma, S.; Speciale, C.; Varasi, M. Bioorg. Med
Chem Lett. 1998, 8, 2907-2912.
13. Röver, S.; Cesura, A. M.; Huguenin, P.; Kettler, R.; Szente, A. J. Med Chem. 1997, 40, 4378-4385
14. Toledo-Sherman, L. M.; Prime, M. E.; Mrzljak, L.; Beconi, M. G.; Beresford, A.; Brookfield, F. A.; Brown, C. J.;
Cardaun, I.; Courtney, S. M.; Dijkman, U.; Hamelin-Flegg, E.; Johnson, P. D.; Kempf, V.; Lyons, K.; Matthews,
K.; Mitchell, W. L.; O’Connell, C.; Pena, P.; Powell, K.; Rassoulpour, A.; Reed, L.; Reindl, W.; Selvaratnam, S.;
Friley, W. W.; Weddell, D. A.; Went, N. E.; Wheelan, P.; Winkler, C.; Winkler, D.; Wityak, J.; Yarnold, C. J.;
Yates, D.; Munoz-Sanjuan, I.; Dominguez, C. J. Med. Chem. 2015, 58 (3), 1159-1183.
15. Lowe, D. M.; Gee, M.; Haslam, C.; Leavens, B.; Christodoulou, E.; Hissey, P.; Hardwicke, P.; Argyrou, A.;

Webster, S. P.; Mole, D. J.; Wilson, K.; Binnie, M.; Yard, B. A.; Dean, T.; Liddle, J.; Uings, I.; Hutchinson, J. P.
Journal of Biomolecular Screening, 2014, 19 (4), 508-515.
16. Abadzapatero, C.; Metz, J. Drug Discovery Today 2005, 10 (7): 464–469
17. Young, R. J.; Green, D. V. S.; Luscombe, C. N.; Hill, A. P. Drug Discovery Today, 2011, 16, 822-830.
18. 5ml of 10mM DMSO stock solution diluted to 100ul with pH7.4 phosphate buffered saline, equilibrated for 1 hour
at room temperature, filtered through Millipore MultiscreenHTS-PCF filter plates (MSSL BPC). The filtrate is
quantified by suitably calibrated flow injection Chemi-Luminescent Nitrogen Detection
19. Valko, K.; Nunhuck, S.; Bevan, C.; Abraham, M. H,; Reynolds, D. P.; J. Pharm. Sci. 2003, 92, 2236–2248
20. The protein production, crystallisation, crystal soaking and structure determination were carried out using the
methods described in Hutchinson et al. (manuscript in preparation). The X-ray data were collected at the European
Synchrotron Radiation Facility (ESRF). The coordinates and structure factors have been deposited in the Protein
Data Bank under the accession code 5N7T.
21. Wilcken, R.;.Zimmermann, M. O.; Lange, A.; Joerger, A. C; Boeckler, F. M. J. Med. Chem. 2013, 56, 1363-1388.
22. Bissantz, C.; Kuhn, B.; Stahl, M. J Med. Chem. 2010, 53, 5061–5084
23. Hutchinson et al. (manuscript in preparation).
24. Bouillot, A. M. J.; Mirguet, O.; Liddle, J.; Walker, A. L. 2015, WO 2015091647.
25. Walker et al. (manuscript in preparation).