Optimisation of conoidin A, a peroxiredoxin inhibitor

Article abstract of DOI:10.1002/cmdc.200900391

Lead optimisation: Interest in the inhibition of peroxiredoxin has been revitalised by their recently identified role in signalling cascades. Here, the synthesis and analysis of novel analogues of the peroxiredoxin inhibitor conoidin A is described. Computational methods are used to rationalise the generated SAR data. These studies lead to a proposed binding mode for this class of compounds that will aid the design of second generation inhibitors. (Figure Presented)

Full text of DOI:10.1002/cmdc.200900391

DOI: 10.1002/cmdc.200900391
Optimisation of Conoidin A, a Peroxiredoxin Inhibitor
Gu Liu, Catherine H. Botting, Kathryn M. Evans, Jeffrey A. G. Walton, Guogang Xu, Alexandra M. Z. Slawin, and
Nicholas J. Westwood*[a]
We recently reported that conoidin A (1) (Scheme 1)
inhibits the function of the peroxiredoxin TgPrxII,[1]
an understudied protein from the apicomplexan par-
asite Toxoplasma gondii.[2] 1 targets TgPrxII both in
vitro and in live parasites and also inhibits the hy-
peroxidation of two mammalian peroxiredoxin ho-
mologues (PrxI and PrxII) in cells.[2] As the biology
and chemistry associated with mammalian peroxire-
doxins has entered a new phase, with the discovery
that these proteins are involved in complex intracel-
lular signalling cascades and are overexpressed in
several cancers, inhibitors of peroxiredoxins are of
increasing interest.[3,4] To date, there are only a few
examples of Prx inhibitors in general[5] and the Toxo-
plasma enzyme TgPrxII in particular. To our knowl-
edge, 1 is the first reported inhibitor of TgPrxII.
As no crystal structure of TgPrxII exists, we decid-
ed to explore how changes to the structure of con-
oidin A (1) affected its biological activity. Our goal
was to develop a model for the binding mode of 1
to TgPrxII, thus aiding future inhibitor design and
optimisation. Whilst peroxiredoxins remain contro-
versial drug targets, mostly due to their high abun-
dance in cells,[6] optimised Prx inhibitors that can be
used to study Prx function are certainly of interest.[7]
Scheme 1. Synthetic routes used to prepare analogues of conoidin A (1). The numbering
Our previous studies with conoidin A (1) showed
that only one of the two available alkylating groups
reacts with TgPrxII.[2] This led us to prepare conoi-
din B (2) (Scheme 1), a mono-bromo-analogue of 1.
Additional analogues were also prepared to gener-
ate further structure–activity relationship (SAR) data.
Here we report our results culminating in proposed
binding modes for the conoidins with TgPrxII.
system used is indicated. Reagents and conditions: a) CuBr2, 18-crown-6, CHCl3, reflux,
60 h, 37%; b) 1,2-phenylene diamine, THF, 08C!RT, 17 h, 37%; c) mCPBA (2 equiv),
CH2Cl2, RT, 20 h, 50%; d) mCPBA (1 equiv), CH2Cl2, RT, 20 h, 75% (6:7 formed in ratio 6:1);
e) 1,2-phenylene diamine, AcOH, reflux, 1.5 h, 90%; f) mCPBA (1 equiv), CH2Cl2, RT, 20 h,
65%; g) Br2 (0.9 equiv), CH2Cl2, reflux, 2 h, 75%.
1
groups to be explored. H NMR analysis of the crude reaction
mixture indicated that oxidation occurred preferentially at N4
The synthesis of compound 2 and its analogues 6 and 7 was
achieved as shown in Scheme 1. Compound 2 was prepared
by conversion of butane-2,3-dione (3) to the brominated dike-
tone 4.[8,9] Subsequent condensation of 4 with 1,2-phenylene-
diamine gave high yields of the mono-bromo quinoxaline 5
that was oxidised (mCPBA, 2 equiv) to give 2. The use of limit-
ed amounts of mCPBA led to a mixture of the two mono-N-
oxides 6 and 7,[10] as it was thought that pure samples of these
two analogues would enable the role of the N-oxide functional
in compound 5 to give 6 as the major product (6/7; 6:1). This
presumably results from the lowered electron density at N1
due to the nearby bromine. However, pure 6 was obtained by
recrystallisation, and an alternative route to 7 was devised that
relied on the selective bromination of mono-N-oxide quinoxa-
line 8, readily accessed from 3 via 9. The structural assignment
of 7 was based on X-ray crystallographic analysis.[11] The bromi-
nation presumably occurs adjacent to the N-oxide through re-
action of bromine with the likely more abundant tautomer 8a,
in which the N-hydroxyenamine has increased reactivity to-
wards electrophiles compared to the enamine present in the
alternative tautomer 8b. The increased reactivity of the enam-
ine in 8a can be rationalised in an analogous manner to the
alpha effect.[12]
[a] G. Liu, C. H. Botting, K. M. Evans, J. A. G. Walton, G. Xu, A. M. Z. Slawin,
N. J. Westwood
School of Chemistry and Biomedical Sciences Research Complex,
University of St Andrews, North Haugh, St Andrews, Fife, KY16 9ST (UK).
Fax: (+44)1334-462595
Subsequently the inhibition of TgPrxII by 2 and 5–7 was as-
sessed. The activity of recombinant TgPrxII (rTgPrxII) is as-
sessed using a glutamine synthetase protection assay.[1a] Using
E-mail: njw3@st-andrews.ac.uk
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.200900391
ChemMedChem 2010, 5, 41 – 45
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
41

MED
this assay with a 5 min pre-incubation time of compound 2
with the enzyme,[13] 2 was shown to inhibit rTgPrxII in vitro
with an IC50 value of 42.3ꢀ2.5 mm (figure S1A in the Support-
ing Information, c.f. IC50 (1)=25.1ꢀ0.8 mm[2]). Analysis of the
incubation of rTgPrxII with 2 using electrospray ionisation
mass spectrometry led to an observed increase in molecular
weight of rTgPrxII of 187.9 Da corresponding to the addition of
one molecule of 2 accompanied by loss of a bromine atom
(figure S1B in the Supporting Information; theoretical mass
shift for formation of rTgPrxII:2 complex with loss of bro-
mine=188.0 Da). In addition, tryptic digest and MALDI MS/MS
experiments supported the view that 2 inhibits rTgPrxII by co-
valent modification of the active site Cys47 (figure S1C in the
Supporting Information), as does 1.[2,14]
computational modelling techniques to provide some insights
into the nature of these interactions.
Whilst no crystal structure of TgPrxII exists, two computa-
tional models of TgPrxII in the “closed” and “open” state have
been reported in the literature.[1a] Regeneration of the “closed”
model for monomeric TgPrxII based on the crystal structure of
human PrxVI (PDB code: 1PRX) was carried out by us using the
SWISS-MODEL server. Overall the model of TgPrxII was very
similar to human PrxVI with a superimposition of the two
structures showing a root mean square deviation (rmsd) of
0.54 ꢁ over 221 Ca atoms. Figure 1a shows a view of the
model emphasising the active site containing the key Cys47
residue.
The program GOLD was used to dock the conoidin ana-
logues in the active site of TgPrxII. In brief, a 10 ꢁ radius
sphere around the Cys47 residue was defined as the ligand-
binding pocket and the sulfur atom of Cys47 was specified as
the covalent linking atom. For each analogue, a bromine atom
in the inhibitor was replaced by the sulfur atom of Cys47. Ten
independent optimisation runs were then carried out and
yielded ten GOLD docking solutions, which were ranked ac-
cording to their GOLD score. For all the analogues assessed,
the highest scoring poses were found to be very similar to
that shown in Figure 1b obtained using conoidin B (2). The in-
hibitor is oriented in such a way that the N1-oxide group is
placed in a shallow pocket close to the active site Cys47 resi-
due. The model suggests that the N1 oxygen may accept two
hydrogen bonds from the hydroxy group of Thr44 and the
backbone NH of Val46 (the highest scoring pose obtained for
2 was very similar in this region to that shown in Figure 1d for
12). In the case of 6, if the N4-oxide oxygen atom is positioned
in this pocket, the reactive alkylating group is forced to be
remote from the active site Cys47 sulfhydryl group. The C3
methyl group of 2 may reside inside the active site, with the
C6 and C7 carbon atoms pointing out into solution. It is inter-
esting that the model also predicts that the N4-oxide is posi-
tioned in a hydrophobic region of the protein consisting of the
residues Ile125 and Val127 (Figure 1b). An inability to stabilise
the highly electron-rich N4-oxide oxygen atom or the inability
of the protein to accommodate, for example, water molecules
that may be hydrogen bonded to this functional group in bulk
solvent, may explain why the presence of the N4-oxide was
found to be detrimental to activity (c.f., activity of conoidin B
(2) vs conoidin C (7), Table 1).
Analogous results were obtained when the docking was re-
peated using our original inhibitor, conoidin A (1). Interestingly,
the modelling studies suggest that the TgPrxII active site can
accommodate the additional large bromine atom in 1 (Fig-
ure 1c). We have shown previously that when 1 reacts with
TgPrxII, only one of the two bromine atoms is displaced. A
possible rationalisation of this result is that in the proposed
binding mode there are no nucleophilic residues positioned
correctly for a second SN2 reaction.[15] Our previous studies
with 1 also showed that the initially formed TgPrxII:1 complex
could react rapidly with an externally added thiol nucleo-
phile.[2] At present, it is difficult to rationalise this experimental
observation with the proposed binding mode unless the pro-
IC50 values were determined for analogues 5–7 and competi-
tion experiments with a model thiol, methyl mercaptoacetate
(Table 1), were carried out to assess differences in their inher-
Table 1. IC50 values for conoidin B (2) analogues against rTgPrxII, and the
relative reactivity of the analogues with a model thiol.
Compd
N1
N4
IC50[a] [mm]
Competition
assay[b]
2
5
6
7
N+Oꢁ
N
N
N+Oꢁ
N+Oꢁ
N
N+Oꢁ
N
42.3ꢀ2.5
32.9ꢀ1.5
105.6ꢀ5.2
16.6ꢀ0.9
=1
3
2
=1
[a] Pre-incubation time of inhibitor with enzyme=5 min in all cases. For
inhibition curves see Reference [14] [b] The thiol competition experiment
was carried out using the model thiol, methyl mercaptoacetate. Order of
reactivity: 1, fastest; 3, slowest.[14]
ent reactivity. Comparison of the IC50 values for 5 and the di-N-
oxide analogue 2 showed that by incorporating the two N-
oxide functional groups in 2, an overall reduction in biological
activity occurred even though 2 reacts significantly faster with
a model thiol than 5 does. In addition, removal of the N4-
oxide in 2 to give conoidin C (7) led to an increase in potency
against the enzyme despite the fact that 7 reacts at a very sim-
ilar rate to 2 with a model thiol.[14] The relative potency of 2
and 7 is consistent with a detrimental interaction of the N4-
oxide with the protein. This effect can also be seen when the
potency of 5 and 6 are compared. These compounds differ
only because 6 contains a N4-oxide moiety, but 5 is a signifi-
cantly more potent inhibitor of rTgPrxII than 6 despite the fact
that 5 is less reactive than 6 towards a model thiol.[14] A prefer-
ence for 6 to adopt an unproductive binding mode may also
explain the relative potencies of 5 and 6. These results clearly
indicated that the potency of the analogues cannot be ex-
plained exclusively based on their chemical reactivity and that,
as expected, the details of their interaction with the active site
of the protein are important. It was therefore decided to use
42
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 41 – 45

Figure 1. a) Molecular model of TgPrxII; b) Docking of conoidin B (2) in the active site of TgPrxII model; c) Docking of conoidin A (1) in the active site of
TgPrxII model; d) Docking of conoidin D (12) in the active site of TgPrxII model.
tein undergoes a significant conformational change after reac-
Table 2. IC50 values for conoidin A (1) analogues against rTgPrxII, and the
tion occurs at Cys47. A conformational change of this type has
been documented previously for the PrxI subfamily of peroxir-
edoxins[16a] and proposed for both TgPrxII[1a] and the PrxVI sub-
family.[16b]
relative reactivity of the analogues with a model thiol.
Inhibition of TgPrxII by C6/C7-substituted analogues of con-
oidin A (1) was subsequently considered. Whilst the proposed
binding modes must be viewed with some caution, additional
modelling studies led us to two interesting predictions that
could be addressed experimentally. First, previous studies
using a model thiol nucleophile had shown that in analogue
10 (Scheme 1) the initial substitution reaction occurs at the C2
bromine atom adjacent to the N1-oxide functional group
rather than at the methylene bromide substituent at C3 (data
not shown). It would therefore seem likely that 10 should fit
well with the predicted binding mode due to the presence of
the N1-oxide and absence of this functional group at N4. Inter-
estingly, this prediction turned out to be correct as analysis of
10 in the glutamine synthetase protection assay showed that
10 was twice as active as 1 (IC50 (10)=10.3ꢀ1.7 mm vs IC50
(1)=23.1ꢀ0.8 mm;[2] Table 2). In addition, it was observed that
the absence of both N-oxide functional groups in 11 was detri-
mental to activity probably reflecting both the reduced chemi-
Compd
N1
N4
R
IC50[a] [mm]
Competition
assay[b]
1
10
11
12
13
N+Oꢁ
N+Oꢁ
N
N+Oꢁ
N+Oꢁ
N+Oꢁ
N
H
H
H
23.1ꢀ0.8
10.3ꢀ1.7
41.4ꢀ0.6
8.1ꢀ0.1
3
4[14]
n.d.
2
N
N+Oꢁ
N+Oꢁ
NHBoc
Br
19.9ꢀ0.2
1
[a] Pre-incubation time of inhibitor with enzyme=5 min in all cases. For
inhibition curves see Reference [14]. [b] The thiol competition experiment
was carried out using the model thiol, methyl mercaptoacetate. Order of
reactivity: 1, fastest; 4, slowest.[14] n.d. not determined.
cal reactivity and the lack of an N-oxide functionality that
helps correctly orientate the inhibitor in the binding site.
The second prediction was that if the proposed binding
mode is correct then it may be possible to prepare C6/C7-sub-
ChemMedChem 2010, 5, 41 – 45
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
43

MED
cooled and washed with 10% [wt/vol] Na2S2O3 (3ꢂ20 mL). The or-
ganic phase was dried (MgSO4), filtered and concentrated in vacuo
to give a pale yellow solid. Purification by flash column chromatog-
raphy on silica gel (EtOAc:petroleum ether 40–60; 1:19!1:9) gave
the desired product 7 as a pale yellow solid (0.54 g, 2.15 mmol,
75%); mp: 117–118 8C; IR (NaCl, Nujol): nmax =1506 (w), 1341 (m),
stituted analogues of conoidin A (1) that have increased activi-
ty resulting from additional interactions with residues remote
from the active site. For example, the GOLD score associated
with an analogue of 1 in which an additional NHBoc substitu-
ent had been incorporated at the C6/C7 position of 1 (conoi-
din D, 12) was predicted to be better than that of 1 itself (data
not shown). We have previously reported the synthesis of 12
and 13,[17] the C6/C7-NHBoc and -bromine substituted ana-
logues of 1, respectively. Whilst there was little difference in
the biological activity of 13 compared to 1, the incorporation
of the NHBoc functionality did indeed lead to an increase in
potency (IC50 (12)=8.1ꢀ0.1 mm vs IC50 (1)=23.1ꢀ0.8 mm;[2]
Table 1), making 12 the most potent analogue prepared in this
series to date. This result is currently rationalised by the poten-
tial for 12 to form an additional hydrogen bond with the hy-
droxy group in Thr150 that cannot be achieved by 1 (Fig-
ure 1d). Interestingly, competition studies with model thiols
again showed that there was little correlation between inher-
ent reactivity with a model thiol and observed activity against
TgPrxII as 13 was found to be slightly more reactive towards
methyl mercaptoacetate than 12 (Table 2).
1
1065 (m), 774 (m), 629 (m) (C-Br) cmꢁ1; H NMR (400 MHz, CDCl3):
d=8.92 (d, J=8.0 Hz, 1H), 8.02 (d, J=8.3 Hz, 1H), 7.67–7.78 (m,
2H), 4.78 (s, 2H), 2.68 ppm (s, 3H); 13C NMR (100 MHz, CDCl3): d=
163.0, 142.8, 139.6, 135.5, 130.6, 129.2, 118.8, 45.8, 13.9 ppm; MS
(ES+): m/z (%): 252 (100%) [M+H]+; HRMS (ES+): m/z [M]+ calcd
C
for C10H9BrN2O: 251.9898, found: 251.9900.
Acknowledgements
We thank Professor Sylke Mꢀller and Dr. Janet Storm for supply-
ing rTgPrxII and for their help and advice. We thank Professors
Garry Taylor, Malcolm Walkinshaw, Rupert Russell and Gary Ward
for advice and encouragement. We thank Dr. Russell Pearson.
This work was supported by a Royal Society University Research
Fellowship (NJW), a Public Health Service grant (AI054961) (GW/
NJW) and the Wellcome Trust (MS/MS instrumentation).
There has been a surge in interest in the biological activities
associated with the peroxiredoxins, both in mammalian and
parasite systems. While there is a wealth of genetic techniques
that enable the role of this protein family to be dissected, ad-
ditional information can also be gained through the use of
peroxiredoxin inhibitors. The relative lack of inhibitors of this
protein class may explain why the chemical genetic approach
is currently underused for the peroxiredoxins. In addition, the
lack of inhibitors is one factor in the limited number of drug
discovery reports linked to these proteins at present. Covalent
modifiers of proteins are often viewed as less useful than re-
versible inhibitors as chemical tools or drugs[18] as they cannot
be washed out of cell experiments or are too reactive to pro-
vide the required levels of selectivity. However, covalent inhibi-
tors of proteins can provide useful starting points for tool or
drug development.
Keywords: conoidin A · inhibitors · molecular modeling ·
peroxiredoxin · Toxoplasma gondii
[1] a) M. Deponte, K. Becker, Mol. Biochem. Parasitol. 2005, 140, 87–96;
b) S. E. Akerman, S. Mꢃller, Sylke, J. Biol. Chem. 2005, 280, 564–570;
c) K. Kim, I. H. Kim, K. Y. Lee, S. G. Rhee, E. R. Stadtman, J. Biol. Chem.
1988, 263, 4704–4711; d) P. Pino, B. J. Foth, L.-Y. Kwok, L. Sheiner, R.
Schepers, T. Soldati, D. Soldati-Favre, PLoS Pathog. 2007, 3, e115; e) L. Y.
Kwok, D. Schlꢃter, C. Clayton, D. Soldati, Mol. Microbiol. 2004, 51, 47–
61.
[2] J. D. Haraldsen, G. Liu, C. H. Botting, J. G. A. Walton, J. Storm, T. J.
Phalen, L. Y. Kwok, D. Soldati-Favre, N. H. Heintz, S. Mꢃller, N. J. West-
wood, G. E. Ward, Org. Biomol. Chem. 2009, 7, 3040–3048.
[3] a) Z. A. Wood, L. B. Poole, P. A. Karplus, Science 2003, 300, 650–653;
b) H. Z. Chae, K. Robison, L. B. Poole, G. Church, G. Storz, S. G. Rhee, G.
Sue, Proc. Natl. Acad. Sci. USA 1994, 91, 7017–7021; c) S. G. Rhee, H. Z.
Chae and K. Kim, Free Radical Biol. Med. 2005, 38, 1543–1552.
[4] a) S. C. Sautel, P. Ortet, N. Saksouk, S. Kieffer, J. Garin, O. Bastien, M.-A.
Hakimi, Mol. Microbiol. 2009, 71, 212–226; b) C. A. Neumann, Q. Fang,
Curr. Opin. Pharmacol. 2007, 7, 375–380; S. G. Rhee, S. W. Kang, W.
Jeong, T.-S. Chang, K.-S. Yang, H. A. Woo, Curr. Opin. Cell Biol. 2005, 17,
183–189.
[5] a) O. Bar-Am, O. Weinreb, T. Amit, M. B. Youdim, J. Mol. Neurosci. 2009,
37, 135–145; b) A. Koshkin, X. Zhou, C. N. Kraus, J. M. Brenner, P. Ban-
dyopadhyay, I. D. Kuntz, C. E. Barry, P. R. De Montellano, Antimicrob.
Agents Chemother. 2004, 48, 2424–2430.
The studies presented here describe our attempts to under-
stand how conoidin A (1) binds and covalently modifies
TgPrxII. A model of the binding mode of conoidin A (1) to
TgPrxII is presented. Evidence in support of the proposed bind-
ing mode comes from our SAR studies on a range of novel
conoidin A (1) analogues that have been prepared. Several
problems associated with the fit of 1 into the active site of
TgPrxII have been identified and, when corrected, have led to
the preparation of more active analogues (conoidins B–D (2, 7
and 12)). The ultimate goal of this programme is to use this in-
formation to convert the conoidins into a second generation
of inhibitors that are highly potent noncovalent peroxiredoxin
inhibitors. We believe these studies provide the first step along
this path.
[6] a) L. Flohꢄ, H. Budde and B. Hofmann, Biofactors 2003, 19, 3–10; b) T.
Jaeger, L. Flohꢄ, Biofactors 2006, 27, 109–120.
[7] For some recent reports and reviews on the development of novel
chemical tools, see: a) S. Lain, J. J. Hollick, J. Campbell, O. Staples, M.
Higgins, M. Aoubala, A. McCarthy, V. Appleyard, K. E. Murray, L. Baker, A.
Thompson, J. Mathers, S. J. Holland, M. J. R. Stark, G. Pass, J. Woods,
D. P. Lane, N. J. Westwood, Cancer Cell 2008, 13, 454–463; b) A. F.
Straight, A. Cheung, J. Limouze, I. Chen, N. J. Westwood, J. R. Sellers,
T. J. Mitchison, Science 2003, 299, 1743–1747; c) K. Carey, N. J. West-
wood, T. J. Mitchison, G. E. Ward, Proc. Natl. Acad. Sci. USA 2004, 101,
7433–7438; d) J. M. Thompson, N. J. Westwood, Philos. Trans. R. Soc., A
2004, 362, 2761–2774.
Experimental Section[14]
[8] R. J. Pearson, K. M. Evans, A. M. Z. Slawin, D. Philp, N. J. Westwood, J.
Org. Chem. 2005, 70, 5055–5061.
[9] a) L. C. King, G. K. Ostrum, G. Kenneth, J. Org. Chem. 1964, 29, 3459–
3461; b) J. F. Okonya, R. V. Hoffman, M. C. Johnson, J. Org. Chem. 2002,
2-(Bromomethyl)-3-methylquinoxaline 1-oxide (7): Br2 (0.41 g,
2.58 mmol) was added to a solution of 8 (0.50 g, 2.87 mmol) in
anhyd CH2Cl2 (20.0 mL). The reaction was heated at reflux (2 h),
44
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 41 – 45

67, 1102–1108; c) A. S. Elina, R. M. Titkoua, L. G. Tsyrulnikova, T. Y. Fili-
penko, Pharm. Chem. J., Engl. Transl. 1976, 10, 38–41.
[10] K. M. Evans, A. M. Z. Slawin, T. Lebl, D. Philp, and N. J. Westwood, J. Org.
Chem. 2007, 72, 3186–3193.
[11] CCDC-740449 contains the supplementary crystallographic data for for
7 (excluding structure factors). These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[12] For a discussion of the a effect: S. Wolfe, D. J. Mitchell, H. B. Schlegel, C.
Minot, O. Eisenstein, Tetrahedron Lett. 1982, 23, 615–618.
[13] A standard pre-incubation time was selected due to the relatively large
number of compounds studied here. As pointed out by a reviewer, we
accept that a study of the dependence of inhibition on pre-incubation
time would provide an improved comparison of the activity of the in-
hibitors.
volved (see B. Hofmann, H. J. Hecht, L. Flohꢄ, Biol. Chem. 2002, 383,
347–364) but is proposed to form a salt bridge with Glu50.
[16] a) Z. A. Wood, E. Schroder, J. R. Harris and L. B. Poole, Trends Biochem.
Sci. 2003, 28, 32–40; b) P. A. Karplus, A. Hall, Sub-cellular Biochemistry
2007, 44, 41–60.
[17] K. M. Evans, J. D. Haraldsen, R. J. Pearson, A. M. Z. Slawin, G. E. Ward,
N. J. Westwood, Org. Biomol. Chem. 2007, 5, 2063–2069.
[18] Drugs that covalently modify their biological targets, such as the b-lac-
tams, are typically only useful in cases where there is very high specific-
ity for the target, or in the case of antimicrobials, where the target is
either absent from the host or has a unique property with respect to
the host isoform. We thank a reviewer for the suggestion to include
this caveat.
Received: September 18, 2009
Revised: November 2, 2009
[14] For a more detailed discussion see the Supporting Information.
[15] Reference [1a] proposes that His39 deprotonates the active site Cys47
and hence would be expected to be protonated. Arg130 could be in-
Published online on November 27, 2009
ChemMedChem 2010, 5, 41 – 45
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
45