Rational design of new bifunctional inhibitors of type II dehydroquinase

Article abstract of DOI:10.1039/b507156a

Selective inhibitors of type II dehydroquinase were rationally designed to explore a second binding-pocket in the active-site. The molecular modelling, synthesis, inhibition studies and crystal structure determination are described. The Royal Society of C

Full text of DOI:10.1039/b507156a

View Online / Journal Homepage / Table of Contents for this issue
C O M M U N I C A T I O N
Rational design of new bifunctional inhibitors of type II
dehydroquinase†
Miguel D. Toscano,a Kirsty A. Stewart,b John R. Coggins,c Adrian J. Lapthornb and
Chris Abell*a
a University Chemical Laboratory, Lensfield Road, Cambridge, UK CB2 1EW.
E-mail: ca26@cam.ac.uk; Fax: 44 1223 336362; Tel: 44 1223 336405
b Dept. of Chemistry, Joseph Black Building, University of Glasgow, Glasgow, UK G12 8QQ
c Institute of Biomedical and Life Sciences, Division of Biochemistry and Molecular Biology,
University of Glasgow, Glasgow, UK G12 8QQ
Received 23rd May 2005, Accepted 19th July 2005
First published as an Advance Article on the web 1st August 2005
Selective inhibitors of type II dehydroquinase were ratio-
nally designed to explore a second binding-pocket in the
active-site. The molecular modelling, synthesis, inhibition
studies and crystal structure determination are described.
second binding site, a range of substituted benzyl groups were
attached to the C-1 and C-4 hydroxyl groups of 3.12 Molecular
docking suggested that the substituted benzyl groups would bind
at the “glycerol” pocket, and the most potent compound (6,
KI = 8 lM) included a 2-nitrobenzyl substituent. We have now
combined these strategies and designed compounds 7–9 to have
pendant aromatic groups attached to side chains attached to C-3
of the quinate core.
The shikimate pathway is involved in the biosynthesis of
aromatic amino acids and other key metabolites, and has been
found in plants, bacteria, fungi and some parasites, but not
in mammals.1 It is a recognised target in the development of
novel herbicides and antimicrobial agents. The third step in
this pathway is the dehydration of 3-dehydroquinate (1) to 3-
dehydroshikimate (2) (Scheme 1), catalysed by dehydroquinase
(EC 4.2.1.10, 3-dehydroquinate dehydratase).2 This step is also
part of the quinate pathway in fungi.1 There are two types of
dehydroquinase, which are structurally distinct and catalyse
the same transformation by different mechanisms.3,4 Type I
dehydroquinases are dimeric proteins with a 26–28 kDa subunit
and catalyse the syn dehydration of 1 through the formation
of a Schiff base.2,5 Type II dehydroquinases are dodecamers of
subunit 12–18 kDa and catalyse the anti elimination of water via
an enolate intermediate.6 Since only type II dehydroquinases
are present in important pathogens such as Mycobacterium
tuberculosis7 and Helicobacter pylori,8 specific inhibitors of these
enzymes have therapeutic potential.
Molecular docking with the program GOLD (version 2.1)13
was used to predict the binding modes of 7–9. These ligands were
docked to the type II dehydroquinase structure (1GU1), which
was prepared using Sybyl6.5.14 The result from 7 is shown in
Fig. 1a, compared with the position of 3 in the crystal structure.
Both 8 and 9 docked in the same orientation (see ESI†). The
quinate ring docks in a similar position to 3, and the side-
chain reaches to the second binding-pocket, which appears to be
large enough to accommodate the phenyl ring. The modelling
suggests ring-stacking between the ligands’ phenyl group and
the side-chain of Tyr-28, and a possible interaction of any ortho-
substituent with Arg-23.
The target compounds (7–9) were synthesised from quinic
acid (10) according to the method outlined in Scheme 2.
The intermediate 11 was prepared in three steps using known
methods16 and the free hydroxyl was acetylated with acetic
anhydride in pyridine in 96% yield to give 12. The anti-
Markovnikov bromination of the allyl side-chain to form 13
was achieved in high yield (97%) by bubbling HBr through
a solution of 12 in carbon tetrachloride in the presence of
catalytic AIBN. This terminal bromide was the key branch-
point in the synthetic strategy. It was reacted with phenol, 2-
nitrophenol and methyl 2-hydroxybenzoate, in the presence of
sodium hydride and potassium iodide (catalytic), and heated
Scheme 1 The reaction catalysed by dehydroquinase.
The first generation of mechanism-based inhibitors was based
on mimicking the flattened enolate intermediate.9 The anhydro
compound 3 had a KI of 30 lM against Streptomyces coelicolor
type II dehydroquinase and was significantly more potent than
the reduced analogue 4 (KI = 600 lM). A crystal structure of
S. coelicolor type II dehydroquinase with 3 bound in the active
site revealed a second binding-pocket, which was occupied by a
glycerol molecule (PDB code: 1GU1) (Fig. 1a).10 Subsequently,
bifunctional compounds with a quinate core and side-chains
from C-3 to reach into this second binding site were designed.
The primary alcohol 5 was the most potent of these compounds
but had a modest KI of 180 lM against S. coelicolor type II
dehydroquinase.11 In order to derive more benefit from this
† Electronic supplementary information (ESI) available: experimental
details. See http://dx.doi.org/10.1039/b507156a
3 1 0 2
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 0 2 – 3 1 0 4
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
©

View Online
Scheme 2 Reagents and conditions: (i) ref. 16; (ii) Ac2O, pyridine, RT;
(iii) HBr, CCl4, RT; (iv) 2-RPhOH, NaH, NaI, MeCN, D; (v) 1. NaOH,
H2O–MeCN (1 : 1), 2. HPLC purification (reverse phase, H2O–MeCN
gradient).
The analogues 7–9 were tested for inhibition against type I
(Salmonella typhi) and type II (S. coelicolor) dehydroquinases
by monitoring the formation of product at 232 nm.17 The results
are shown in Table 1. None of the compounds showed any
inhibition against the type I enzyme within the sensitivity of
the assay, which is consistent with all the results obtained on C-
3 substituted quinate analogues. However all three compounds
showed competitive reversible inhibition against type II dehy-
droquinase. Compound 7 is the best inhibitor of the S. coelicolor
type II enzyme, with a KI of 33 lM. This inhibitor has a
similar KI to 3 and is 20 times more potent than the reduced
analogue 4, and 6 times more potent than 5, which emphasises
the importance of the benzyl ring. The analogues with a 2-
substituted phenyl ring, 8 and 9, had KI values of 84 and 220 lM
respectively, which is in contrast with the positive effect of the
2-nitro group on 6. These results show a trend where increase
of hydrophilicity at the 2-position in the ring is unfavourable,
suggesting that there is no positive interaction with the Arg-23
side-chain.
Fig. 1 15 (a) Active-site view of the S. coelicolor type II dehydroquinase
crystal structure with 3 and glycerol (purple) bound (PDB code:
1GU110). Comparison between the position of 3 and the docking result
of 7 (blue); (b) active-site view of the S. coelicolor type II dehydroquinase
crystal structure with 7 (blue) bound (PDB code: 2BT4).
to reflux in acetonitrile to give 14 (69% yield), 15 (11%) and
16 (23%), respectively. The yields reflect the nucleophilicity of
the phenols (PhOH > 2-MeO2CPhOH > 2-O2NPhOH). This
strategy potentially allows the attachment of a wide range of
nucleophiles, such as other alcohols, amines or thiols, to the
propyl side-chain of these quinate analogues. Full deprotection
of all compounds (14–16) was achieved by treatment with
sodium hydroxide in a MeCN–H2O (1 : 1) solution at room-
temperature and the products 7–9 were obtained as a 1 : 1
mixture with benzoate. The pure ammonium salts of the target
compounds were obtained after HPLC purification through a C-
18 reverse-phase column eluting with a gradient of H2O–MeCN
and treatment with ammonium bicarbonate.
The best inhibitor of this set, 7, was crystallised with S.
coelicolor type II dehydroquinase and the crystal structure
Table 1 Inhibition constants (lM) for type I (S. typhi) and type II (S. coelicolor) dehydroquinases
Compounds
3
4
5
6
7
8
9
Type I (S. typhi)a
3000 1000b
30 10b
4500 500b
600 200b
>20 000c
—
8
>20 000
33 5.4
>20 000
84 13
>20 000
220 50
Type II (S. coelicolor)a
180 20c
2d
a KM values of 16 lM (S. typhi type I dehydroquinase) and 250 lM (S. coelicolor type II dehydroquinase) were obtained, using the standard assay.17
b Ref. 9. c Ref. 11. d Ref. 12.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 0 2 – 3 1 0 4
3 1 0 3

View Online
shows the compound binding at the active-site (Fig. 1b).
The quinate ring is in a similar position as 3 in the main binding-
pocket with the exception of C-2, which in the absence of a
References
1 C. Abell, in Enzymology and Molecular Biology of the Shikimate Path-
way; Comprehensive Natural Products Chemistry, ed. U. Sankawa,
Pergamon, Elsevier Science Ltd., Oxford, 1999, pp. 573–607.
2 C. Kleanthous, R. Deka, K. Davis, S. M. Kelly, A. Cooper, S. E.
Harding, N. C. Price, A. R. Hawkins and J. R. Coggins, Biochem. J.,
1992, 282, 687–695.
3 V. B. Patel, M. Schweizer, C. C. Dyskstra, S. R. Kushner and N. H.
Giles, Proc. Natl. Acad. Sci. USA, 1981, 78, 5783–5787.
4 D. G. Gourley, A. K. Shrive, I. Polikarpov, T. Krell, J. R. Coggins,
A. R. Hawkins, N. W. Isaacs and L. Sawyer, Nat. Struct. Biol., 1999,
6, 521–525.
5 (a) C. Kleanthous, M. Reilly, A. Cooper, S. Kelly, N. C. Price and J. R.
Coggins, J. Biol. Chem., 1991, 266, 10893–10898; (b) R. K. Deka, C.
Kleanthous and J. R. Coggins, J. Biol. Chem., 1992, 267, 22237–
22242; (c) A. P. Leech, R. James, J. R. Coggins and C. Kleanthous,
J. Biol. Chem., 1995, 270, 25827–25836.
6 (a) J. Harris, C. Kleanthous, J. R. Coggins, A. R. Hawkins and C.
Abell, J. Chem. Soc., Chem. Commun., 1993, 270, 1080–1081; (b) A.
Shneier, J. Harris, C. Kleanthous, J. R. Coggins, A. R. Hawkins and C.
Abell, Bioorg. Med. Chem. Lett., 1993, 3, 1399–1402; (c) J. M. Harris,
C. Gonzalez-Bello, C. Kleanthous, A. R. Hawkins, J. R. Coggins and
C. Abell, Biochem. J., 1996, 319, 333–336.
7 T. Garbe, S. Servos, A. Hawkins, G. Dimitriadis, D. Young, G.
Dougan and I. Charles, Mol. Gen. Genet., 1991, 228, 385–392.
8 J. R. Bottomley, C. L. Clayton, P. A. Chalk and C. Kleanthous,
Biochem. J., 1996, 319, 559–565.
9 M. Frederickson, E. J. Parker, A. R. Hawkins, J. R. Coggins and C.
Abell, J. Org. Chem., 1999, 64, 2612–2613.
10 A. W. Roszak, D. A. Robinson, T. Krell, I. S. Hunter, M. Fredrickson,
C. Abell, J. R. Coggins and A. Lapthorn, Structure, 2002, 10, 493–
503.
11 M. D. Toscano, M. Frederickson, D. P. Evans, J. R. Coggins, C. Abell
and C. Gonza´lez-Bello, Org. Biomol. Chem., 2003, 1, 2075–2083.
12 C. Gonza´lez-Bello, E. Lence, M. D. Toscano, L. Castedo, J. R.
Coggins and C. Abell, J. Med. Chem., 2003, 46, 5735–5744.
13 (a) G. Jones, P. Willet and R. C. Glen, J. Mol. Biol., 1995, 245, 43–53;
(b) G. Jones, P. Willet, R. C. Glen, A. R. Leach and R. Taylor, J. Mol.
Biol., 1997, 267, 727–748.
˚
double bond is moved by 0.85 A, the quinate ring adopting
a full chair conformation. The carboxyl interactions with the
backbond amides of Ile-107 and Ser-108 are essentially the
same, while the H-bond from the hydroxyl group at C-1 with
˚
the side chain of His-106 is longer (0.1 A). The quinate ring
˚
of 7 is moved away from the protein by ∼0.3 A and as
a result H-bonding distances between the hydroxyl groups
at C-4 and C-5 with His-85 and Arg-117 (not shown) are
˚
slightly changed (by +0.1 and −0.2 A, respectively). This
˚
movement is accompanied by a 0.5 A movement in the water
conserved in all S. coelicolor type II dehydroquinase structures
and implicated in the mechanism.10 This water movement is
caused by the side-chain at C-3 extending into the “glycerol”
binding-pocket, where there is ring-stacking between the benzyl
ring and Tyr-28. In addition, the benzene ring displaces two
water molecules and effects the position of a third buried
water.
Despite some disruption to the structure, the side-chain at
C-3, which extends into the “glycerol” binding-pocket must
account for the increased potency of 7. The position of 7 in
the active site agrees well with the binding position predicted
by molecular docking (Fig. 1a vs. 1b), validating the strategy
and methods used in this project. Further optimisation of
7, to reduce the unfavourable interactions identified from
the crystal structure should lead to significantly more potent
specific inhibitors of type II dehydroquinase than are currently
available.
In summary, we have rationally designed a set of compounds
to explore specific interactions at a second binding-pocket in the
active site of type II dehydroquinase, achieving an increase in po-
tency, with 7 in particular, and high selectivity. Structural infor-
mation showed 7 in a position consistent with molecular docking
experiments, and identified key interactions to optimise in future
compounds.
14 SYBYL. 6.5. 1699 South Hanley Road, St. Louis, Missouri, 63144,
USA: Tripos Inc.
15 WebLab ViewerPro4.0. 9685 Scranton Road, San Diego, CA 92121,
USA: Accelrys Inc.
16 T. Widlanski, S. L. Bender and J. R. Knowles, Biochemistry, 1989,
28, 7572–7582.
17 UV assay monitored the formation of 3-dehydroshikimate at 234 nm
using black-walled quartz cuvettes. All assays were done in 1 ml
reaction volumes, at 25 ◦C and pH 7.0 (50 mM potassium phosphate
buffer for type I dehydroquinase; 50 mM Tris-HCl buffer for type II
dehydroquinase).
Acknowledgements
We would like to thank Fundac¸a˜o para a Cieˆncia e a Tecnologia
for financial support.
3 1 0 4
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 3 1 0 2 – 3 1 0 4