Structural basis for the broad-spectrum inhibition of metallo-β- lactamases by thiols

Article abstract of DOI:10.1039/b802311e

The development of broad-spectrum metallo-β-lactamase (MBL) inhibitors is challenging due to structural diversity and differences in metal utilisation by these enzymes. Analysis of structural data, followed by non-denturing mass spectrometric analyses, identified thiols proposed to inhibit representative MBLs from all three sub-classes: B1, B2 and B3. Solution analyses led to the identification of broad spectrum inhibitors, including potent inhibitors of the CphA MBL (Aeromonas hydrophila). Structural studies revealed that, as observed for other B1 and B3 MBLs, inhibition of the L1 MBL thiols involves metal chelation. Evidence is reported that this is not the case for inhibition of the CphA enzyme by some thiols; the crystal structure of the CphA-Zn-inhibitor complex reveals a binding mode in which the thiol does not interact with the zinc. The structural data enabled the design and the production of further more potent inhibitors. Overall the results suggest that the development of reasonably broad-spectrum MBL inhibitors should be possible.

Full text of DOI:10.1039/b802311e

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Structural basis for the broad-spectrum inhibition of metallo-b-lactamases
by thiols†
d
d
Benoˆıt M. R. Lie´nard,a Gianpiero Garau,‡b Louise Horsfall,c Andreas I. Karsisiotis,§ Christian Damblon,§
Patricia Lassaux,c Cyril Papamicael,¶ Gordon C. K. Roberts,d Moreno Galleni,c Otto Dideberg,b
a
Jean-Marie Fre`rec and Christopher J. Schofield*a
Received 25th February 2008, Accepted 25th March 2008
First published as an Advance Article on the web 7th May 2008
DOI: 10.1039/b802311e
The development of broad-spectrum metallo-b-lactamase (MBL) inhibitors is challenging due to
structural diversity and differences in metal utilisation by these enzymes. Analysis of structural data,
followed by non-denturing mass spectrometric analyses, identified thiols proposed to inhibit
representative MBLs from all three sub-classes: B1, B2 and B3. Solution analyses led to the
identification of broad spectrum inhibitors, including potent inhibitors of the CphA MBL (Aeromonas
hydrophila). Structural studies revealed that, as observed for other B1 and B3 MBLs, inhibition of the
L1 MBL thiols involves metal chelation. Evidence is reported that this is not the case for inhibition of
the CphA enzyme by some thiols; the crystal structure of the CphA–Zn–inhibitor complex reveals a
binding mode in which the thiol does not interact with the zinc. The structural data enabled the design
and the production of further more potent inhibitors. Overall the results suggest that the development
of reasonably broad-spectrum MBL inhibitors should be possible.
a mono-zinc protein but is non-competitively inhibited by the
presence of a second Zn(II).4 Crystal and solution structures
of representatives from the three subclasses reveal significant
structural differences in the active site pocket5–7 and in a loop
involved in substrate binding.8,9
Introduction
b-Lactamases (BLs) are classified into those employing a nucle-
ophilic serinyl residue (classes A, C and D) and those binding one
or two zinc ions (metallo-b-lactamases, MBLs, class B) at their
active site. Medicinal attention has primarily focused on the serine
BLs, and inhibitors of these enzymes are widely used. However,
MBLs hydrolyze most b-lactam antibiotics and are an increasing
clinical problem (for a review see ref. 1). MBLs can be divided
into three subclasses on the basis of their substrate selectivity.
Sub-classes B1 and B3 display a broad substrate selectivity
whereas sub-class B2 most efficiently hydrolyze carbapenems.2,3 B1
MBLs, including BcII (Bacillus cereus) and IMP-1 (Pseudomonas
aeruginosa), can employ either one or two active site Zn(II) ions,
whereas B3 MBLs, e.g. FEZ-1 (Legionella gormanii) and L1
(Stenotrophomonas maltophilia), are only active with two Zn(II)
ions. The B2 MBL CphA (Aeromonas hydrophilia) is active as
The active site metal-binding chemistry and structural vari-
ations make the development of broad-spectrum MBL in-
hibitors challenging. At present, there are no reports of indi-
vidual MBL inhibitors with sub-micromolar Ki values against
all three MBL sub-classes; however, such a property will be
highly desirable, if not a prerequisite, for a clinically useful
compound. Thiols are the most studied MBL inhibitors, with
various derivatives having been described (e.g. thiomandelic acid,10
D-captopril,11 6-(mercaptomethyl)penicillinates,12 homocysteinyl-
containing peptides13 and derivatives of mercaptobenzoic acid).14
( )-Thiomandelic acid (1) and D-captopril (2) are reported to be
potent inhibitors of sub-classes B1 and B3. However, 1 and 2 fail to
potently inhibit members of the B2 sub-class (Ki = 144 lM for 1,
and 72 lM for 2), and thus are probably not useful broad-spectrum
inhibitors.
aChemistry Research Laboratory and OCISB, University of Oxford, 12
Mansfield Road, Oxford, OX1 3TA, UK. E-mail: Christopher.schofield@
chem.ox.ac.uk; Fax: +44 (0) 1865 275 674; Tel: +44 (0) 1865 275 625
bInstitut de Biologie Structurale Jean-Pierre Ebel (CNRS/CEA/UJF), 41
Rue J. Horowitz, Grenoble, 38100, France
Here we report how the application of MS-based screening
coupled to structural analyses led to the identification of broad-
spectrum inhibitors of clinically relevant metallo-b-lactamases.
cCentre d’Inge´nie´rie des Prote´ines, Universite´ de Lie`ge, Alle´e de 6 Aout B6,
Sart-Tilman, Lie`ge, Belgium
dHenry Wellcome Laboratories of Structural Biology, Biochemistry De-
partment, University of Leicester, Rm 1/02, Henry Wellcome Building,
Lancaster Road, Leicester, LE1 9HN, UK
Initial screening and assays
† Electronic supplementary information (ESI) available: X-Ray crystallo-
graphic data; additional NMR data. See DOI: 10.1039/b802311e
‡ Current address: Biocrystallography Unit, San Raffaele Scientific Insti-
tute, DIBIT, via Olgettina 58, 20132 Milano, Italy.
With the aim of identifying molecules capable of inhibiting
MBLs from all three sub-classes, we analysed X-ray structures
for representative MBLs from the three sub-classes: BcII (B1),
IMP-1 (B1), CphA (B2), L1 (B3) and FEZ-1 (B3) (Fig. 1).
Based on the structural analyses, selected thiols were then
synthesised (Scheme 1) and screened for binding to representative
MBLs (BcII, CphA and FEZ-1) using electrospray ionisation-MS
§ Current address: Centre de Biophysique Mole´culaire, UPR 4301 CNRS,
rue Carles Sadron, 45071 Orle´ans Cedex 2, France.
¶ Current address: Laboratoire de Chimie Fine et He´te´rocyclique UMR
6014 IRCOF, CNRS, Universite´ et INSA de Rouen, BP 08, 76131, Mont-
Saint-Aignan Cedex, France.
2282 | Org. Biomol. Chem., 2008, 6, 2282–2294
This journal is
The Royal Society of Chemistry 2008
©

View Online
Fig. 1 Superimposition of views derived from the crystal structures of BcII (sub-class B1, orange, PDB code 1BVT),15 IMP-1 (sub-class B1, yellow,
PDB code 1DDK),8 CphA (sub-class B2, blue, PDB code 18XI),5 L1 (sub-class B3, pink, PDB code 1SML)16 and FEZ-1 (sub-class B3, grey, PDB
code 1K07).17 (a) Overall folds and (b) close-up of selected residues at the active sites (nitrogen in blue, oxygen in red and sulfur in orange). The figure
highlights the overall structural similarities between these MBLs, including between the dizinc-binding enzymes (BcII, IMP-1, L1 and FEZ-1) and the
monozinc-binding MBLs (e.g. CphA, in blue). The numbering scheme is that for the BcII enzyme. Figure made using LSQMAN and PyMOL.
under non-denaturing ionisation conditions (Fig. 2). Although
there are limitations to the use of MS for screening for enzyme
inhibitors,18–21 ESI-MS has been productively used to analyse the
binding of inhibitors to MBLs including in dynamic thiol exchange
methodology.14
Compound 3 was prepared in three steps as reported (16%
overall yield)22 (Scheme 1a). a-Mercaptoketone 4 was prepared
from commercially available a-bromoketone (65% overall yield;
all yields unoptimised) (Scheme 1b). a-Bromoketone 16 was
alkylated using potassium thioacetate to give 17, then hydrolysed
to give 4. 5a,b and 8a–c were prepared, using an alternative to
the reported methods,22,23 in three steps from the corresponding
amino acids (overall yields: 12 to 48%) (Scheme 1c). Thus 3-
bromopropionyl chloride was treated with the methyl ester of
suitable amino acid hydrochlorides to give bromides 18–22.
Substitution using potassium thioacetate gave acetylthiolates 23–
27; subsequent hydrolysis of the protecting groups led to the
formation of mercaptocarboxylates 5a,b and 8a–c in moderate
Fig. 2 Deconvoluted ESI-MS spectra under non-denaturing conditions showing a) BcII–Zn2, b) BcII–Zn2 + 3 (1 eq.), c) BcII–Zn2 + 4 (1 eq.),
d) BcII–Zn2 + 5a (1 eq.), e) CphA–Zn, f) CphA–Zn + 3 (1 eq.), g) CphA–Zn + 4 (1 eq.), h) CphA–Zn + 5a (1 eq.), i) FEZ-1–Zn2, j) FEZ-1–Zn2 + 3
(1 eq.), k) FEZ-1–Zn2 + 4 (1 eq.) and l) FEZ-1–Zn2 + 5a (1 eq.). * FEZ-1–Zn2 + methionine (N-terminal methionine residue was confirmed by tryptic
digestion followed by MALDI-MS analyses).
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2282–2294 | 2283
©

View Online
Scheme 1 Reagents and conditions: (a) a) NBS, dibenzoyl peroxide, CCl4, reflux, 1.5 h, 28%; b) AcSK, EtOAc, rt, 14 h, 69%; c) NaOH (1 N), 70 ◦C, 3 h,
82%. (b) a) AcSK, CH2Cl2, rt, 14 h, 94%; b) NaOH (1 N), rt, 6 h, 69%. (c) a) 3-bromopropionyl chloride, NaHCO3 (5% aq.), CH2Cl2, rt, 1–3 h, 22–99%;
b) AcSK, EtOAc, rt, 12 h, 54–99%; c) NaOH (1 N), 70 ◦C, 1.5 h, 88–97%. (d) a) TBDPSCl, imidazole, DMF, rt, 12 h, 65%; b) LiOH, THF–MeOH, 0 ◦C
→ rt, 20 h, 89%; c) (i) (COCl)2, benzene, 0 ◦C, 3 h; (ii) D-proline, Et3N, rt, 18 h, 76%; d) HF·pyridine, THF, rt, 12 h, 50%. (e) a) Et3N, dioxane–H2O,
−10 ◦C → rt, 18 h, 73%. (f) a) NaOH (4 N), AcCl, 0 ◦C, 1 h, 41%. (g) a) b-propiolactone, AcONa, EtOH, 0 ◦C→rt, 15 h, 21%; b) NaOH (1 N), rt, 15 h,
53%.
yields. D-Phenylglycine derivative 8a was obtained as a racemic
mixture in the last deprotection step. Compound 9 was prepared
in two steps from D-phenylalanine methyl ester hydrochloride (22%
overall yield) (Scheme 1d). Protection of commercially available
alcohol 29 using tert-butyldiphenylsilyl chloride (TBDPSCl) and
imidazole to give 30, followed by methyl ester hydrolysis (LiOH)
gave 31 in good yield. Carboxylic acid 31 was converted into
the corresponding acyl chloride (oxalyl chloride), which was then
reacted with D-proline to give carboxylic acid 32 in good overall
yield. Deprotection of the silyl group using fluoridric acid in
pyridine afforded the desired hydroxycarboxylic acid in reasonable
yield. 10 was prepared in one step from D-proline (Scheme 1e).
11 was prepared in two steps from D-phenylalanine methyl ester
hydrochloride (11% overall yield) (Scheme 1g). 12 was prepared
in one step from D-phenylalanine in 41% yield under Schotten–
Baumann conditions (Scheme 1f).24
With the exceptions of 5a and 5b, which were not observed
to bind significantly to CphA and FEZ-1 by ESI-MS under the
applied experimental conditions, two of the tested compounds
(3 and 4, Fig. 3) showed a significant ability to form enzyme–
metal–ligand complexes with all the tested MBLs; CphA–Zn–5a
or 5b and FEZ-1–Zn2–5a or 5b complexes were observed when
an excess of 5a,b was used (data not shown). Thiols 5a,b are less
functionalised derivatives of the reported MBL inhibitors 6 and
7 (see below).8,25 Thiols 3 and 4 formed significant CphA–Zn–
ligand complexes (Fig. 2d,e), but did not induce the binding of
a second zinc ion in the active site of CphA as reported for 1.14
These results implied that the distance between the thiol and the
2284 | Org. Biomol. Chem., 2008, 6, 2282–2294
This journal is
The Royal Society of Chemistry 2008
©

View Online
found to significantly inhibit sub-classes B1 and B3 MBLs but
not the sub-class B2 CphA MBL, in partial agreement with the
ESI-MS studies. For comparison, 6 displays a Ki value of 0.3 lM
vs. BcII25 and 7 Ki values of 0.1–0.5 lM vs. IMP-1, BcII and L1
(Fig. 3).8
Modelling and NMR studies
To investigate the broad-spectrum nature of the inhibition of 4,
docking studies were carried out with representatives of each
sub-class (IMP-1, CphA and L1). Initially, to test the utility
of the modelling technique, 7 was docked into IMP-1–Zn–78
(with 7 extracted) using different zinc coordination geometries
(data not shown). A close correlation between the structure of
the IMP-1–Zn–7 complex in the reported X-ray structure8 and
the highest scoring docked conformation was obtained using a
tetrahedral geometry for both active site zinc cations. A trigonal
bipyramidal geometry was used for the monozinc CphA based on
the previously reported crystallographic data for the complexed
CphA structure.5 In all cases, the choice of the metal coordination
type of the enzyme as well as the parametrisation of the ligand
(i.e. ionisation state of the thiol/carboxylate groups and the
tautomeric form of the amide bond) were found to be important
for good correlation between the binding mode observed in the
crystal structures and the modelling data. Generally, for dizinc
MBLs, the lowest rms deviations between the MBL–Zn2–ligand
crystal structures and the highest scoring modelled structures
were obtained when using deprotonated SH/CO2H groups and
the amide tautomer form of the inhibitor. Conversely, the use
of deprotonated SH/CO2H and the imidic acid tautomer form
achieved the most consistent results (and the highest docking
scores) for the monozinc CphA. These observations highlight the
influence of the preparation of molecules for docking studies with
metalloenzymes.
Compound 4 was predicted to interact with the dizinc MBLs
(IMP-1 and L1) via a metal bridging chelation involving its thiolate
and a hydrogen-bond between its carbonyl group and the con-
served MBL hydrogen-bond donor region (Fig. 1, Fig. 4a,c). The
proposed bridging metal chelation of 4 is precedented by various
crystallographic and solution studies on dizinc MBLs.7,8,11,17 In
contrast, the modelling studies led to the prediction that both
the thiolate and carbonyl groups of 4 chelate the monozinc ion
CphA (Fig. 4b), indicating a potential for the same inhibitor
to adopt different (predominant) binding modes to different
MBLs. Docking of 3 resulted in a more complex analysis due
to several possible predicted structures for the MBL–3 complexes
(data not shown). However, the results suggested that the key
Fig. 3 Structures of the thiols and their derivatives under investigation in
this study.
carboxylate groups of inhibitors not only influences the potency of
MBL inhibition by mercaptocarboxylate, as proposed,10 but can
also influence the mechanism of inhibition via modification of the
MBL–metal–ligand stoichiometry, i.e. 1 binds as a CphA–Zn2–
ligand complex but 3 and 4 bind as CphA–Zn–ligand complexes
(at least by MS analyses).
Compounds 3–5 were then screened for inhibition against
MBLs from all three sub-classes (Table 1). Strikingly, 3 and 4 were
found to inhibit all the tested MBLs. 3 is the first reported broad
spectrum MBL inhibitor with Ki values <1 lM for all MBLs. 3
(Ki = 90 nM) and 4 (Ki = 50 nM) were the most potent reported
inhibitors for the monozinc CphA MBL (Table 1).
The results also imply that mercaptocarboxylate compounds
that bind as CphA–Zn complexes rather than CphA–Zn2 com-
plexes are more potent CphA inhibitors; thus, thiol 4, which
appears to bind as a monozinc complex to CphA (by ESI-MS), is
ca. 3000 times more potent than 1, which binds to the CphA MBL
preferentially as a dizinc complex.14 Compounds 5a and 5b were
Table 1 Competitive inhibition constants Ki (lM)
B1
B2
B3
L1
Compound
IMP-1
BcII
CphA
FEZ-1
3
0.36 0.01
0.67 0.09
68 30
0.97 0.2
2.7 0.2
5.6 0.3
0.09 0.004
0.05 0.02
0.21 0.01
0.24 0.01
0.3a
1a
240 50
4
b
5a
5b
—
6.5
2
b
7.5
5
37
1
—
37 0.7
42
1
a IC50. b Not determined.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2282–2294 | 2285
©

View Online
Fig. 4 Results of docking studies (using GOLD) of 4 (carbon atoms in grey) complexed with a) IMP-1 (PDB ID: 1DDK),8 b) CphA (PDB ID: 1X8G)5
and c) L1 (PDB ID: 1SML)16 MBLs, highlighting the potential for different binding modes. Results of docking studies (using GOLD) for 5a (carbon
atom in grey) complexed with d) BcII (PDB ID: 1BVT),15 e) CphA (PDB ID: 1X8G)5 and f) L1 (PDB ID: 1SML)16 MBLs. Results of NMR analyses of
the BcII–Zn2 MBL with 5a showing g) the most significantly affected BcII residues (dark blue) by the presence of 5a mapped onto the predicted structure
of BcII–Zn2–5a complex (Fig. 4d) (left panel). This was determined using minimum chemical shift index calculations26 on BcII–Zn2–5a as well as on
BcII–Zn2–5b for comparison (right panels). Superimposition of the results of the docking studies of (h) 5a, 8a, 8b and 8c (carbon atoms are in grey) in
complex with CphA (PDB ID: 1X8G)5 and (i) 5a and 8a,c in complex with L1 (PDB ID: 1SML).16 Only the D-isomer of 8c was tested. The zinc ions are
represented as green spheres. Figure made using PyMOL.
2286 | Org. Biomol. Chem., 2008, 6, 2282–2294
This journal is
The Royal Society of Chemistry 2008
©

View Online
enzyme–inhibitor interactions involve the carboxylate group of 3
with the conserved hydrogen-bond region (Fig. 1) and its thiolate
group with the two zinc ions. 5a was predicted to bind similarly to
the dizinc MBLs with similar interactions predicted for 3 and 4,
i.e. the C-7 carboxylate of 5a interacting with the MBL hydrogen-
bond donor region and the thiolate with the zinc ions (Fig. 4d
and Fig. 5f). In contrast, when 5a was docked into the monozinc
MBL CphA the results suggested that 5a may chelate to the zinc
cation via its C-7 carboxylate and imidic acid nitrogen (Fig. 4e).
In this mode the C-4 carbonyl of 5a can interact with the side
chains of Asp120 and His118 and the thiolate group is positioned
to make hydrogen bonds with the hydroxyl side chains of Thr119
and Thr157, i.e. not to the zinc ions. Although not precedented by
any structural work on thiol MBL inhibitors, this binding mode is
Fig. 5 a) Active site view from a crystal structure of L1–Zn2 in complex with 5a (carbon atoms in yellow, PDB ID: 2QDT). Only water molecules
(red spheres) forming H-bonds with 5a are shown; the figure highlights the good correlation between the binding mode of 5a to L1–Zn2 from the
crystallographic data (carbon atoms coloured in yellow) and the binding mode for 5a predicted from modelling studies (carbon atoms coloured in grey).
View from the active sites of b) BlaB (PDB ID: 1M2X),27 in complex with 2 (carbon atoms in grey) c) CphA (PDB ID: 2QDS) in complex with 2 (carbon
atoms in grey) and d) Angiotensin Converting Enzyme (ACE) (PDB ID: 1UZF)28 in complex with 2 (carbon atoms in grey). The zinc ions are in green.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2282–2294 | 2287
©

View Online
consistent with the reported mode of zinc chelation in the complex
between hydrolysed biapenem and CphA.5
interacts significantly with the b3–b4 loop and the MBL conserved
hydrophobic region, probably via its Ca methyl group.
These modelling studies suggested that the inhibition of CphA
may be enhanced by hydrophobic interactions with residues such
as Val67, possibly explaining the lack of significant inhibition of
5a,b (at 100 lM) vs. CphA. Significantly they also suggest that
mercaptocarboxylates do not always inhibit monozinc MBLs via
a thiol–zinc interaction.
Derivatives of 5a
We then sought to identify more potent inhibitors. Superimposi-
tion of the predicted L1–Zn2–5a complex with the IMP-1, BcII,
CphA and FEZ-1 crystal structures revealed that (i) the methyl
group of 5a could be positioned to make an interaction with the
hydrophobic region conserved in all MBL active sites (Fig. 1) and
(ii) derivatisation of 5a at the C-2 and/or C-3 positions (such as
in compounds 6 and 7) would likely lead to structural clashes
with the sub-class B2 CphA active site. These analyses prompted
preparation of derivatives of 5a with bulkier hydrophobic groups
at C-6; it was proposed that increasing the hydrophobicity at
the C-6 position as well as restricting substitution at the C-
2/C-3 positions would increase the potency vs. the B1 and
B3 sub-classes and activity vs. the CphA MBL. Starting from
the appropriate amino acids, 8a–c (Fig. 3) were synthesised in
three steps from DL-phenylglycine, D-phenylalanine and D-valine,
respectively (Scheme 1c).
Compounds 8a–c inhibited all the MBLs tested with Ki values
as low as 19 nM (Table 2). CphA was also inhibited by 8a–c with
Ki values in the low micromolar range, i.e. these compounds are
up to 20 times more potent at inhibiting CphA than 2. ESI-MS
analyses (see ESI†) conducted with 8a–c and CphA also indicated
the preferential formation of CphA–Zn–ligand complexes (and
not CphA–Zn2–ligand such as observed with 1), supporting the
importance of the monozinc-binding inhibition mode for potent
inhibition of the CphA MBL by this thiol series (8a–c were up to
40 times more potent than 1 vs. CphA).
To investigate how 8a–c bind to MBLs, with both one and two
active site zinc ions, further modelling analyses were then carried
out. On this basis, all derivatives 8a–c were predicted (i) to interact
with the zinc ions through chelation of the thiolate function (except
for CphA), (ii) to have a near-superimposable backbone with the
lead molecule 5a, (iii) to interact with the conserved hydrophobic
region of the MBL active sites (Fig. 1) and (iv) to have a similar
binding mode vs the dizinc enzymes BcII, IMP-1, L1 and FEZ-1
(Fig. 4h). The docking of 8a–c with monozinc CphA resulted in
high docking scores and suggested a near-identical binding mode
to that of 5a to CphA, i.e. without a thiol–metal interaction, with
an apparent increase in the hydrophobic interactions with the side
chain of Val67 (Fig. 4i).
To obtain structural information on the binding of 5a,b to the
sub-class B1 MBL and to test the modelling studies, NMR studies
of BcII–Zn2 with 5a and 5b were then carried out. The imidazole
resonances of the metal binding histidine residues (His86, His88,
His149 and His210)6 are useful probes of the effects of inhibitors.
In the absence of inhibitors, relatively sharp resonances were
observed for His86 and His88 and a broad resonance for His210,
but the signal of His149 was too broad to be detected. Titration of
BcII–Zn2 with 5a resulted in a progressive decrease in the intensity
of the imidazole NH signals of the free enzyme and a progressive
increase in a new set of signals assigned to the enzyme–inhibitor
complex (see ESI†). The intensity of the new peaks increased with
the concentration of the inhibitor but their chemical shift was not
affected. This behaviour is typical of a slow exchange between
complexed and free enzyme, as observed for other compounds of
the same class.10
Through the use of 1H–15N HSQC experiments optimized
to detect long range 1H–(C)–15N couplings in the imidazole
ring the resonance assignments for the histidine Hd2 and He1
protons (C–H), along with the Hd1 and He2 (N–H) protons
and the corresponding nitrogens, can be obtained.6 Analysis of
the imidazole resonances for His88 and His210, which belong to
different zinc binding sites, suggests that 5a binds predominantly
in a single conformation, whereas 5b binds in two, or perhaps
three, conformations (see ESI†).
The inhibitor titrations for 5a and 5b reveal that 5a binds more
tightly to BcII than does 5b, an observation consistent with their
respective Ki values. The chemical shift changes, in the form of
minimum chemical shift indices,26 indicate that the binding of both
compounds affects residues in the same loops near the active site
(Fig. 4g, right panels). Both 5a and 5b affect backbone amide
resonances corresponding to the loops b6–a2 (84–88), b9–b10
(144–154), b11–a4 (168–188); in general, the minimum chemical
shift difference index is somewhat greater for 5b, particularly for
residues 83–91, 142–149 and 186–192. In terms of the chemical
shift index, Asp90 and Arg91 are some of the BcII residues
most affected by binding of 5a,b, as previously observed for
mercaptocarboxylate inhibitors.10
Compound 5a appears to make significantly more interactions
with the “mobile” b3–b4 BCII loop, compared to 5b (Fig. 4g,
left panel). Potent MBL inhibition has already been linked to the
ability of an inhibitor to induce the b3–b4 loop closure.8,9 5a but
not 5b was observed to make a significant interaction with Val39
from the b3–b4 loop in solution, consistent with the modelling
studies (Fig. 4d) and with our hypothesis for the design of 5a
derivatives, i.e. the increase of the interactions with this conserved
hydrophobic region.
Crystallographic analyses
To investigate the inhibition mechanism of 5a,b and derivatives
8a–c as well as to further test the modelling studies, we then co-
crystallised 5a in complex with the sub-class B3 L1 MBL (Fig. 5a).
Table 2 Competitive inhibition constants Ki (lM)a
Compound IMP-1
BcII
7.7 0.7
CphA
L1
8a
8b
8c
0.019 0.002
0.088 0.010 0.85 0.08 15.0 5.0
0.063 0.009 0.32 0.01
5.7 2.0
1.8 0.4
0.96 0.08
3.6 0.3 0.082 0.002
Compound 5a apparently makes significant interactions with
metal-coordinating histidines at both zinc coordination sites,
and therefore its binding likely involves the previously observed
bridging interaction between its thiolate and the two zincs. 5a also
a For experimental conditions see Kinetic analyses section below.
2288 | Org. Biomol. Chem., 2008, 6, 2282–2294
This journal is
The Royal Society of Chemistry 2008
©

View Online
Analysis of the interactions between 5a and the L1 enzyme active
of CphA and IMP-1. Similarly, non-thiol derivatives of 8b were
also synthesised (Fig. 3 and Scheme 1 compounds 11 and 12) and
screened under the same conditions.
site showed that the inhibitor thiolate bridges the two Zn(II) ions
−
−
˚
˚
(S –Zn1 = 2.3 A; S –Zn2 = 2.4 A). The 5a inhibitor carboxylate
group is positioned to form a hydrogen-bond with the Ser225 side
Compound 9 retained a significant level of inhibition of the
CphA MBL (Ki = 189 12 lM), suggesting that the thiol group
of 2 is not essential for the inhibition of CphA, in support of
the crystallographic and modelling studies, indicating the thiol
of 2 does not chelate to the zinc. In contrast, 9 was inactive (at
100 lM) against IMP-1, demonstrating that the replacement of
the thiol group of 2 is detrimental to the inhibition of IMP-1,
consistent with the apparently general mechanism of inhibition
of the sub-classes B1 and B3 involving a thiol–zinc chelation. No
inhibition was observed when 11 and 12 (at 100 lM) were tested
against CphA and IMP-1, possibly reflecting the necessity of the
thiolate group for potent inhibition of both monozinc and dizinc
MBLs.
˚
chain (2.6 A). Hydrophobic contacts apparently occur between
˚
the inhibitor methyl group in C-6 and His263 (4.1 A) and Trp39
˚
˚
(4.2 A), and between C4–C7 and the side chains of His118 (3.7 A),
˚
˚
Phe156 (4.1 A) and Pro227 (3.5 A). Furthermore, the crystal
structure of L1–Zn2–5a is in agreement with and supports the
docking studies (Fig. 5a).
There is no reported X-ray diffraction study on the mode
of binding of thiols to mono-zinc MBLs. To gain insight into
the proposed binding mode of thiols such as 5a,b and 8a–c
with the monozinc CphA MBL, where the modelling analyses
suggested that the thiol does not chelate to the zinc, we attempted
crystallisation of CphA–ligand complexes. We were able to obtain
˚
crystals of CphA–Zn in complex with 2, which diffracted to 1.7 A
resolution (Fig. 1c). As revealed from the crystal structure of BlaB–
Zn2–2 complex (Chryseobacterium meningosepticum),27 2 acts as
an inhibitor of sub-class B1 di-zinc MBLs with bridging of its
thiolate group between the two active site metal ions (Fig. 5b), as
previously established for other thiolate inhibitors.7,8 2 has also the
ability to interact with monozinc enzymes via its thiolate group as
reported with an Angiotensin Converting Enzyme (ACE)–Zn–2
crystal structure in which the thiolate function occupies the single
tetrahedral coordination site left by the zinc-binding triad residues
(Fig. 5d).28
In contrast to the crystal structures of the L1–Zn2–5a complex
(Fig. 5a) and the BlaB–Zn2–2 complex (Fig. 5b) and in agreement
with the modelling predictions (Fig. 4h), the binding of 2 in the
crystalline state was not by chelation to the zinc ion. Instead
the main metal interactions between 2 and the monozinc CphA
occur through the inhibitor carboxylate group, which coordinates
Conclusions
This work has demonstrated that it should be possible to develop
individual compounds that are potent inhibitors of all three MBL
sub-classes. An important structural point that has arisen is that
the same compound may bind in different modes to different MBL
sub-classes whilst still retaining inhibition activity. Specifically,
crystallographic analyses revealed that whilst mercaptocarboxy-
late was observed to bind to the zinc ion via its thiolate group in
the cases of the B1 and B3 sub-classes, for the B2 sub-class this
binding mode was not observed. Instead, for the B2 CphA enzyme
the thiolate group of 2 was observed to bind to the alcohol side
chains of Thr119 and Thr157. In some cases, whether or not the
thiol inhibitors preferentially bind with one or two zinc ions can be
important to the potency of inhibitor observed. Thus, compounds
that interact preferentially with monozinc CphA–Zn appear to be
more potent than those that induce binding of a second zinc to
give CphA–Zn2 complexes (e.g. thiomandelate). The useful role
of non-denaturing ESI-MS in rapidly providing data on metal
binding stoichiometry, data that is often not easily obtained for
metallo-proteins, is highlighted in this aspect of the work.
˚
a tetrahedrally cooordinated zinc ion (O2–Zn 2.1 A) (Fig. 5c).
The carboxylate O1 of 2 is positioned to hydrogen-bond with
˚
˚
the side chains of His196 (3.1 A) and Asn233 (2.7 A). These
residues approach the carboxylate group upon inhibitor binding
via “closure” of the mobile loop, observed previously for the CphA
enzyme.5 The sulfhydryl group of 2 likely forms hydrophobic
Overall, although the work reveals further complexities in the
way that (thiol) inhibitors bind to different MBL sub-classes, it
also suggests that utilisation of different binding modes for the
same inhibitor binding to different MBL sub-classes may be a
productive route to achieving potent broad-spectrum inhibition.
˚
˚
contacts with Phe156 (3.6 A) and the side chain of Arg233 (3.8 A)
˚
˚
˚
as well as with Trp87 (3.9 A), Leu161 (3.9 A) and Val67 (4.0 A). It
is notable that, in contrast to the data for the Zn(II) complex
reported here, evidence from EXAFS and PAC indicates that
in the Cd(II)-substituted form of CphA, 2 appears to bind the
zinc ion preferentially via its thiolate group.11 Inhibitors such as
Enalapril or Lisinopril (non-thiol analogues of 2) are reported to
interact with the active site zinc ion of ACE via their carboxylate
group, which occupies a similar position relative to the metal as
the sulfhydryl group of 2 in complex with ACE–Zn.29 It therefore
seems likely that mercaptocarboxylate inhibitors of zinc hydrolases
can interact with the zinc ion(s) of their target enzymes via either
their carboxylate or their thiolate groups. It cannot be excluded
that the mode of binding of 2 to CphA is influenced by the
experimental conditions, i.e. higher pH values may favor sulfhydryl
group ionisation, leading to better interaction with the metal ion.
To investigate the importance of the thiol group of 2 for inhibi-
tion of CphA and to test the results obtained by crystallographic
studies, non-thiol derivatives of 2 were then synthesised (Fig. 3,
and Scheme 1 compounds 9 and 10) and tested for inhibition
Experimental
Crystallography
Crystallizations of the wildtype CphA and L1 were performed
as described.5,16 The CphA–Zn–2 and L1–Zn–5a complexes were
obtained by incubation (24 hours) of protein crystals in a drop
of reservoir in which an excess of 2 or 5a was added. Before data
collection, crystals were transferred to a cryoprotectant solution
(reservoir solution containing 20% (v/v) glycerol), then mounted
rapidly in loops and flash-cooled. X-Ray data for the CphA–2
and L1–5a complexes were collected at the beamline BM30A of
the European Synchrotron Radiation Facility (Grenoble, France)
and in-house using a Nonius FR591 rotating anode X-ray
generator coupled to a Mar Research Imagine Plate detector,
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2282–2294 | 2289
©

View Online
respectively. Data were processed using the CCP4 programs (Table
S1†).30 Initial phases for the CphA–Zn–2 and L1–Zn–5a complex
structures were generated by molecular replacement using the
structure of the wildtype CphA and L1, respectively, as starting
models (PDB accession code 1X8G and 1SML). Refinement was
carried out using Coot31 and REFMAC (CCP4). The calculation
of the first (Fo − Fc) electron density map clearly showed the
presence of the inhibitor molecules in the active site. Inhibitors
were modelled in the map after most of the protein main chain
and side chain atoms and most of water molecules were built
and refined. Conformational torsion-angle restraints and charges
assignments for the 2 and 5a were obtained using CCP4i Libcheck.
Data collection and refinement statistics are shown in Table S1†.
Coordinates and structure factors have been deposited with the
Protein Data Bank32 with accession codes 2QDS and 2QDT.
verified that the low concentrations of DMSO present had no
inhibition effects; the rate remained the same upon addition
of 1% DMSO. The inhibitors 3, 4, 8a–c were prepared as
100 lM stock solutions directly into the buffer used. For these
inhibitors the buffer was 20 mM PIPES buffer pH 6.0 containing
20 lg mL−1 BSA (and 100 lM ZnCl2 where indicated). BcII, IMP-
1, CphA, L1 and FEZ-1 enzymes were used at fixed concentrations
between 0.03 and 0.7 nM. The enzyme and inhibitor were pre-
incubated, when found to be necessary (see optimisation below),
at room temperature before the substrate was added. Substrate
concentrations were varied between 20 and 200 lM at a minimum
of 2 inhibitor concentrations and in its absence. Hydrolysis of
imipenem and nitrocefin was monitored by following the variation
in absorbance at 300 nm or 482 nm respectively, using a Uvikon
860 spectrophotometer connected to a computer via a RS232 serial
interface. Cells of 2 mm or 10 mm path length were used depending
on substrate concentration. The experiments were performed at
30 ◦C and initial rate conditions were used to study the inhibition
with imipenem or nitrocefin using the Hanes linearisation of the
Henri–Michaelis equation and the KaleidaGraph 3.5 programme.
Since the FEZ-1 MBL did not show competitive inhibition
with these inhibitors, the IC50 was determined at a substrate
concentration of 200 lM nitrocefin.
NMR
Singly labelled (15N) BcII protein used in all the NMR experiments
was expressed and purified as described previously.6 Sequence-
specific resonance assignments for the BcII enzyme were obtained
by using triple resonance experiments and will be published
elsewhere. All NMR experiments were carried out in 20 mM
MES, 100 mM NaCl, 0.2 mM ZnCl2, pH 6.4 at 298 K. Samples
of the inhibitor–enzyme complexes were prepared by addition of
microlitre volumes of inhibitor solutions (100–200 mM in MES
buffer). 1D 1H and 2D 1H–15N heteronuclear single quantum
coherence (HSQC) spectra were used to establish when the
complex was fully formed. The NMR spectra were obtained
by using Bruker Avance DRX or DMX 600 MHz instruments.
To optimise preincubation times, an appropriate amount of
enzyme was incubated with 100 lM inhibitor in a total volume of
5 mL in 10 mM HEPES buffer pH 7.0 (containing 100 lM ZnCl2
◦
with all enzymes except for CphA) at 30 C. Samples of 490 lL
were removed at various time intervals, 10 lL of 5 mM imipenem
or nitrocefin was added and the activity measured at 30 ◦C.
1
1D H spectra were obtained by using a water flip-back pulse
combined with Watergate.33,34 Backbone NH resonances were
observed by 1H–15N HSQC with States TPPI and Watergate.
The imidazole 15N(C)H resonances were observed by 1H–15N
HSQC (heteronuclear single quantum coherence) as previously
described.6 In the absence of resonance assignments for the
inhibitor complexes, residues affected by inhibitor binding were
identified by using minimum chemical shift approach. Here, the
chemical shift difference between an amide cross-peak in the
1H–15N HSQC spectrum of the free protein (whose assignment
is known) and the position of the nearest cross-peak in the
corresponding spectrum of the enzyme–inhibitor complex is
calculated. The 1H and 15N chemical shift differences, DH and DN,
between each pair of cross-peaks provide the minimum chemical
shift index based on the following formula:35
Docking experiments
The program GOLD37 (version 3.0) was used for docking analyzes.
In GOLD, metal ions are considered to bind to hydrogen
bond acceptors in the ligand. Coordination points are added at
points around the metal, where coordination sites are missing.
Tetrahedral and trigonal bipyramidal geometries were used for
zinc metals in dizinc and monozinc MBLs respectively. These
coordination points can then bind to acceptor atoms in the ligand.
The ligands were prepared using MarvinSketch.38 The proteins
were prepared for docking using WHAT IF.39 This online program
generates proteins with the correct input for docking programs
such as GOLD. All water molecules and ligands were removed
from the proteins. For the dizinc MBLs the binding site of the
˚
ligands was defined as a 20 A sphere centered at the zinc ion
that is coordinated by the triad of zinc-binding histidine residues
and default settings were applied. For CphA, the binding site was
˚
defined as a 20 A sphere centered at the single zinc ion.
Whereas the true values of the chemical shift changes may be
underestimated, the pattern of interactions observed by this
approach has been found in a number of systems to be similar
to the pattern seen when assignments are available for both free
enzyme and the complex.26,36
ESI-MS under mild ionisation conditions
Samples were desalted using a Microcon YM-10 (cut-off = 10 000
Da) centrifugal filters (Millipore, Bedford MA, USA) in 15 mM
NH4Ac buffer (pH 7.5). Seven dilution/concentration steps were
◦
performed at 4 C and 14 000 g. The stock enzyme solution was
Kinetic analyses
diluted in NH4OAc buffer to a final concentration of 100 lM. Prior
to each experiment, individual thiols were freshly dissolved in
DMSO at a final concentration of 100 mM. Each thiol was then
diluted to 100 lM in 15 mM ammonium acetate buffer pH 7.5.
Solution of 5a and 5b were prepared as 10 mM DMSO solutions
before dilution with 20 mM HEPES buffer pH 7.0 containing
20 lg mL−1 BSA (and 100 lM ZnCl2 where indicated). Tests
2290 | Org. Biomol. Chem., 2008, 6, 2282–2294
This journal is
The Royal Society of Chemistry 2008
©

View Online
Samples were prepared by mixing the thiols and enzyme to a
final concentration of 15 lM. An aliquot of this mixture was
placed in a 96-well plate and analysed. ESI-MS analyses used a Q-
TOF mass spectrometer (Q-TOFmicro Micromass, Altrincham,
UK) interfaced with a NanoMateTM chip-based nano-ESI source
(Advion Biosciences, Ithaca, NY, USA). Samples were infused
to the Q-TOF through the ESI chip (estimated flow rate ca. 100
nL min−1). Typically a spraying voltage of 1.70 0.1 kV depending
on the “sprayability” of the sample and a sample pressure of
0.25 psi were applied. The instrument was equipped with a
standard Z-spray source block. Clusters of Cs(n+1)In (1 mg mL−1
CsI in 100% methanol) were used for calibration. Calibration and
sample acquisitions were performed in the positive ion mode in
the range of m/z 500–5000. Operating conditions for the mass
spe◦ctrometer were: sample cone voltage 50 V, source temperature
40 C. Acquisition and scan time were 30 s and 1 s, respectively.
The pressure at the interface between the atmospheric source and
the high vacuum region was fixed at 6.6 mbar (measured with the
roughing pump Pirani gauge) by throttling the pumping line using
an Edwards Speedivalve to provide collisional cooling.
HPLC and a Phenomenex Luna 5 l C18 (250 × 4.60 mm) column
for preparative HPLC.
Methyl 3,5-bis(bromomethyl)benzoate (14). Pre◦pared as previ-
◦
1
ously described.22 M.p. 96–97 C (lit.22 100–101 C); H NMR
(400 MHz, CDCl3): d = 8.01 (d, 2H, J = 1.5 Hz, 2 × ArCH),
7.63 (s, 1H, ArCH), 4.51 (s, 4H, 2 × CH2), 3.94 ppm (s, 3H,
13
=
CH3); C NMR (100.6 MHz, CDCl3): d = 165.9 (C O), 138.9
(2 × ArC), 133.9 (ArCH), 131.4 (ArC), 130.0 (2 × ArCH), 52.4
(OCH3), 31.9 (2 × CH2) ppm; HRMS (CI+): [M + NH4]+ calcd.
for C10H14NO279Br81Br, 339.9371; found, 339.9377.
Methyl 3,5-bis[(acetylsulfanyl)methyl]benzoate (15). Prepared
as previously described.40 1H NMR (400 MHz, CDCl3): d = 7.84
(d, 2H, J = 1.5 Hz, 2 × ArCH), 7.41 (s, 1H, ArCH), 4.12 (s,
4H, 2 × CH2), 3.91 (s, 3H, CH3), 2.36 ppm (s, 6H, 2 × CH3);
13
=
C NMR (100.6 MHz, CDCl3): d = 194.7 (2 × CH3C O), 166.5
=
(ArC O), 138.6 (2 × ArC), 133.7 (ArCH), 130.9 (ArC), 128.9 (2 ×
ArCH), 52.3 (OCH3), 32.9 (2 × CH2), 30.4 ppm (2 × CH3CO);
HRMS (ES+): [M + NH4]+ calcd. for C14H20NO4S2, 330.0834;
found, 330.0833.
3,5-Bis(sulfanylmethyl)benzoic acid (3). Prepared as previ-
ously described.22 M.p. 121–123 C (lit.22 133 ◦C); 1H NMR
◦
Synthesis
(400 MHz, DMSO-d6): d = 7.90 (d, 2H, J = 1.5 Hz, 2 × ArCH),
7.59 (s, 1H, ArCH), 3.80 (s, 4H, 2 × CH2); 13C NMR (125.8 MHz,
General. All solvents used were either anhydrous solvents
purchased from Aldrich (Sigma-Aldrich Chemical Co., Dorset,
UK) or dried by passing over an alumina column under nitrogen
pressure. Reagents were used as obtained from commercial sources
unless otherwise stated. Measurement of pH was carried out
using Prolabo RotaTM pH 1–10 paper. Flash chromatography
was performed using silica gel (0.125–0.25 mm, 60–120 mesh)
as the stationary phase. Thin layer chromatography (TLC) was
performed on aluminium plates pre-coated with silica gel (Merck
silica gel 60 F254), which were visualized by the quenching of
UV fluorescence (using an irradiation wavelength k = 254 nm),
and/or by staining with iodine or 10% ammonium molybdate
in 2 M sulfuric acid, followed by heating. Melting points were
obtained using a Bu¨chi 510 Cambridge Instruments Gallen III
hot stage melting point apparatus. Infrared (IR) spectra were
recorded as thin films between NaCl plates or as KBr discs on a
Tensor 27 FT-IR Bru¨ker spectrometer. Only selected absorbances
are reported. Proton magnetic resonance spectra (1H NMR) were
recorded on a Bruker DPX 250 (250 MHz), Bruker DQX 400
(400 MHz), or Bruker AMX500 (500 MHz) spectrometers at am-
=
DMSO-d6): d = 167.1 (C O), 142.4 (2 × ArC), 132.5 (ArCH),
131.1 (ArC), 127.5 (2 × ArCH), 27.3 ppm (2 × CH2); MS (ES−):
m/z (%): 213 (100) [M − H]−.
(S)-(2-Oxo-2-phenylethyl) ethanethioate (17). Prepared as pre-
viously described.40 1H NMR (400 MHz, CDCl3): d = 8.01–7.98
(m, 2H, 2 × ArCH), 7.63–7.58 (m, 1H, ArCH), 7.51–7.47 (m,
2H, 2 × ArCH), 4.41 (s, 2H, CH2), 2.41 ppm (s, 3H, CH3); 13C
=
=
NMR (100.6 MHz, CDCl3): d = 194.2 (C O), 193.2 (C O), 135.5
(ArC), 133.8 (ArC), 128.8 (2 × ArCH), 128.5 (2 × ArCH), 36.7
(CH2), 30.3 ppm (CH3); MS (ES+): m/z (%): 195 (100) [M + H]+.
1-Phenyl-2-sulfanylethanone (4). An aqueous solution of 1 N
NaOH (8.25 mL, 8.25 mmol, 4 eq) was added to 17 (400 mg,
2.06 mmol, 1 eq) under a nitrogen atmosphere. The resulting
mixture was stirred at 22 ◦C for 6 h. The reaction mixture was
acidified with HCl (10 N) and extracted with EtOAc (2 × 20 mL).
The combined organic extracts were washed with brine (2 ×
20 mL), dried over MgSO4 and the solvent removed under reduced
pressure. Purification by flash chromatography using a mixture
of CH2Cl2–petroleum ether (40–60)–AcOH (20 : 80 : 1) as eluent
afforded 216 mg (69%) of 4 as a colourless oil. 1H NMR (400 MHz,
CDCl3): d = 7.98–7.96 (m, 2H, 2 × ArCH), 7.63–7.59 (m, 1H,
ArCH), 7.52–7.48 (m, 2H, 2 × ArCH), 3.98 (d, 2H, J = 7.5 Hz,
CH2), 2.15 ppm (t, 1H, J = 7.5 Hz, SH); 13C NMR (100.6 MHz,
1
bient temperature. H NMR spectral assignments are supported
by 1H–1H COSY experiments where necessary. Coupling constant
values (J) are reported to the nearest 0.5 Hz. Carbon magnetic
resonance spectra (13C NMR) were recorded on a Bruker DPX 250
(62.9 MHz), Bruker DQX 400 (100.6 MHz) or Bruker AMX500
(125.8 MHz) spectrometers at ambient temperature. 13C NMR
assignments were made using DEPT-135 along with HMQC, and
HMBC correlation experiments. High-resolution mass spectra
were recorded on a VG Autospec spectrometer by chemical
ionization or on a Micromass LCT electrospray ionization mass
spectrometer operating at a resolution of 5000 full width half
height. High performance liquid chromatography (HPLC) used
a Waters 996 photodiode array detector, a Waters 600E system
controller and a Waters 717 plus autosampler, with a Phenomenex
Synergy 4 l MAX RP80A (250 × 4.60 mm) column for analytical
=
CDCl3): d = 194.8 (C O), 135.0 (ArC), 133.7 (ArC), 128.9 (2 ×
ArCH), 128.5 (2 × ArCH), 31.2 ppm (CH2); HRMS (CI+): [M +
H]+ calcd. for C8H9OS, 153.0374; found, 153.0370.
(R)-Methyl 2-(3-bromopropanamido)-3-phenylpropanoate (18).
Prepared as previously described.41 IR mmax (film): 3055, 2953,
−1
1
=
=
1739 (C O), 1681 (C O), 1266, 739 cm ; H NMR (400 MHz,
CDCl3): d = 6.19 (brs, 1H, NH), 4.64 (p, 1H, J = 7 Hz, CHa),
3.77 (s, 3H, OCH3), 3.66–3.63 (m, 2H, CH2), 2.81–2.78 (m, 2H,
CH2), 1.44 ppm (d, 3H, J = 7.0 Hz, CH3); 13C NMR (100.6 MHz,
=
=
CDCl3): d = 173.8 (C O), 169.5 (C O), 53.0 (OCH3), 48.6 (CHa),
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2282–2294 | 2291
©

View Online
39.9 (CH2), 27.4 (CH2), 19.0 ppm (CH3); HRMS (ES+): [M + H]+
calcd. for C7H13NO3Br, 238.0079; found, 238.0081.
7.11 (m, 2H, 2 × ArCH), 6.24 (d, 1H, J = 7 Hz, NH), 4.92 (m,
1H, CHa), 3.73 (s, 3H, OCH3), 3.59 (m, 2H, CH2), 3.16 (AA’B,
dd, 1H, J = 14 Hz, J = 5.5 Hz, CH2b), 3.09 (AAꢀB, dd, 1H,
J = 14 Hz, J = 5.5 Hz, CH2b), 2.76 ppm (m, 2H, CH2); 13C
(R)-Methyl 2-(3-(acetylthio)propanamido)propanoate (23).
Prepared as previously described.40 IR mmax (film): 3055, 2987,
=
=
NMR (100.6 MHz, CDCl3): d = 171.9 (C O), 169.2 (C O), 135.7
(ArC), 129.3 (2 × ArCH), 128.6 (2 × ArCH), 127.2 (ArCH), 53.2
(CHa), 52.4 (OCH3), 39.4 (CH2), 37.8 (CH2b), 27.0 ppm (CH2);
HRMS (ES+): [M + H]+ calcd. for C13H17NO3Br, 314.0392; found,
314.0390.
−1
1
=
=
1742 (C O), 1686 (C O), 1266, 739 cm ; H NMR (400 MHz,
CDCl3): d = 6.20 (brs, 1H, NH), 4.59 (p, 1H, J = 7.5 Hz, CHa),
3.74 (s, 3H, OCH3), 3.13 (t, 2H, J = 7 Hz, CH2), 2.52 (t, 2H, J =
7 Hz, CH2), 2.32 (s, 3H, CH3COS), 1.40 ppm (d, 3H, J = 7.5 Hz,
CH3b); 13C NMR (100.6 MHz, CDCl3): d = 196.0 (C O), 173.4
=
=
=
(C O), 170.1 (C O), 52.5 (OCH3), 48.1 (CHa), 36.0 (CH2), 30.6
(CH3COS), 24.7 (CH2), 18.4 ppm (CH3b); HRMS (ES+): [M +
H]+ calcd. for C9H16NO4S, 234.0800; found, 234.0802.
(R)-Methyl 2-(3-(acetylthio)propanamido)-3-phenylpropanoate
(26). Prepared as previously described.40 IR mmax (film): 3055,
−1
2987, 1743 (C O), 1686 (C O), 1266, 737 cm ; 1H NMR
(400 MHz, CDCl3): d = 7.27 (m, 3H, 3 × ArCH), 7.10 (m, 2H, 2 ×
ArCH), 6.02 (d, 1H, J = 7.5 Hz, NH), 4.89 (m, 1H, CHa), 3.73 (s,
3H, OCH3), 3.12 (m, 4H, 2 × CH2), 2.49 (m, 2H, CH2), 2.32 ppm (s,
=
=
(R)-2-(3-Mercaptopropanamido)propanoic acid (5a). Prepared
as previously described.23 M.p. 78–79 ◦C (lit.23 79–81 ◦C); 1H NMR
(250 MHz, CD3OD): d = 4.41 (q, 1H, J = 7.5 Hz, CHa), 2.82–
2.78 (m, 2H, CH2), 2.57–2.54 (m, 2H, CH2), 1.42 ppm (d, 3H, J =
7.5 Hz); MS (ES−): m/z (%): 176 (100) [M − H]−; [a]2D5 = +39
(MeOH, c = 0.5).
3H, CH3COS); 13C NMR (100.6 MHz, CDCl3): d = 196.0 (C O),
=
=
=
171.9 (C O), 170.2 (C O), 135.7 (ArC), 129.3 (2 × ArCH), 128.6
(2 × ArCH), 127.2 (ArCH), 53.1 (CHa), 52.4 (OCH3), 37.8 (CH2),
36.0 (CH2), 30.6 (CH3COS), 24.7 ppm (CH2); HRMS (ES+): [M +
H]+ calcd. for C15H20NO4S, 310.1113; found, 310.1104.
(S)-2-(3-Mercaptopropanamido)propanoic acid (5b). 5b and its
reaction intermediates (19 and 24) were obtained following the
same procedure that leading to 5a. All analytical data for 5b were
identical to that of 5a with the exception of the optical rotation
(−39.5 (MeOH, c = 0.5)).
(R)-2-(3-Mercaptopropanamido)-3-phenylpropanoic acid (8b).
◦
Prepared as previously described.23 M.p. 102–104 C (lit.23 106–
◦
1
107 C); H NMR (250 MHz, CD3OD): d = 7.30–7.25 (m, 5H,
5 × ArCH), 4.75–4.69 (m, 1H, CHa), 3.29–3.20 (m, 1H, CH2b),
2.97 (dd, 1H, J = 9.5 Hz, J = 14 Hz, CH2b), 2.64 (m, 2H, CH2),
(R)-Methyl 2-(3-bromopropanamido)-2-phenylacetate (20).
2.46 (m, 2H, CH2); MS (ES−): m/z (%): 253(100) [M − H]−; [a]2D5
−10.0 (MeOH, c = 0.5).
=
Prepared as previously described.41 M.p. 61–62 ◦C; IR mmax (film):
−1
3019, 1741 (C O), 1679 (C O), 1216, 756 cm ; 1H NMR
(400 MHz, CDCl3): d = 7.38–7.33 (m, 5H, 5 × ArCH), 6.73 (d,
1H, J = 7 Hz, NH), 5.60 (d, 1H, J = 7 Hz, CHa), 3.74 (s, 3H,
=
=
(R)-Methyl 2-(3-bromopropanamido)-4-methylpentanoate (22).
Prepared as previously described.41 IR mmax (film): 3055, 2960,
OCH3), 3.61 (t, 2H, J = 7 Hz, CH2), 2.90–2.76 ppm (m, 2H,
−1
1
=
=
1742 (C O), 1677 (C O), 1266, 740 cm ; H NMR (400 MHz,
CDCl3): d = 6.10 (d, 1H, J = 7.5 Hz, NH), 4.70–4.66 (m, 1H,
CHa), 3.74 (s, 3H, OCH3), 3.64–3.62 (m, 2H, CH2), 2.83–2.81 (m,
2H, CH2), 1.68–1.66 (m, 2H, CH2), 1.57–1.56 (m, 1H, CH), 0.95–
0.94 ppm (m, 6H, 2 × CH3); 13C NMR (100.6 MHz, CDCl3): d =
13
=
CH2); C NMR (100.6 MHz, CDCl3): d = 171.2 (C O), 169.0
=
(C O), 136.2 (ArC), 129.0 (2 × ArCH), 128.6 (ArCH), 127.3
(2 × ArCH), 56.5 (CHa), 52.9 (OCH3), 39.3 (CH2), 26.8 ppm
(CH2); HRMS (ES+): [M + H]+ calcd. for C12H15NO3Br, 300.0235;
found, 300.0226; [a]2D5 = −145 (CHCl3, c = 0.45).
=
=
173.4 (C O), 169.4 (C O), 52.4 (OCH3), 50.8 (CHa), 41.7 (CH2),
39.5 (CH2), 27.1 (CH2), 24.8 (CH), 22.8 (CH3), 21.9 ppm (CH3);
HRMS (ES+): [M + H]+ calcd. for C10H19NO3Br, 280.0548; found,
280.0540.
(R)-Methyl 2-(3-(acetylthio)propanamido)-2-phenylacetate (25).
Prepared as previously described.40 M.p. 48–50 ◦C; IR mmax (film):
−1
1
=
=
1741 (C O), 1684 (C O), 909, 734 cm ; H NMR (400 MHz,
CDCl3): d = 7.34 (m, 5H, 5 × ArCH), 6.69 (d, 1H, J = 7 Hz,
NH), 5.57 (d, 1H, J = 7 Hz), 3.72 (s, 3H, OCH3), 3.11 (t, 2H, J =
7 Hz, CH2), 2.56 (m, 2H, CH2), 2.30 ppm (s, 3H, CH3COS); 13C
(R)-Methyl 2-(3-(acetylthio)propanamido)-4-methylpentanoate
(27). Prepared as previously described.40 IR mmax (film): 3055,
−1
2960, 1742 (C O), 1685 (C O), 1266, 739 cm ; 1H NMR
(400 MHz, CDCl3): d = 6.02 (d, 1H, J = 8 Hz, NH), 4.66–4.61 (m,
1H, CHa), 3.73 (s, 3H, OCH3), 3.13 (t, 2H, J = 7 Hz, CH2), 2.53
(t, 2H, J = 7 Hz, CH2), 2.32 (s, 3H, CH3COS), 1.65–1.60 (m, 2H,
=
=
=
=
NMR (100.6 MHz, CDCl3): d = 196.0 (C O), 171.2 (C O), 170.1
=
(C O), 136.3 (ArC), 129.0 (2 × ArCH), 128.6 (ArCH), 127.3 (2 ×
ArCH), 56.4 (CHa), 52.8 (OCH3), 35.9 (CH2), 30.5 (CH3COS),
24.6 ppm (CH2); HRMS (ES+): [M + H]+ calcd. for C14H18NO4S,
296.0957; found, 296.0959; [a]2D5 = −124 (CHCl3, c = 0.10).
CH2), 1.54–1.50 (m, 1H, CH), 0.94–0.93 ppm (m, 6H, 2 × CH3);
13
=
=
C NMR (100.6 MHz, CDCl3): d = 196.1 (C O), 173.5 (C O),
=
170.4 (C O), 52.3 (OCH3), 50.7 (CHa), 41.7 (CH2), 36.1 (CH2),
(RS)-2-(3-Mercaptopropanamido)-2-phenylacetic◦ acid (8a).
Prepared as previously described.23 M.p. 105–108 C (lit.42 102–
30.6 (CH3COS), 24.9 (CH), 24.8 (CH2), 22.8 (CH3), 22.0 ppm
(CH3); HRMS (ES+): [M + H]+ calcd. for C12H22NO4S, 276.1270;
found, 276.1261.
◦
1
106 C); H NMR (250 MHz, CD3OD): d = 7.49–7.34 (m, 5H,
5 × ArCH), 5.48 (s, 1H, CHa), 2.78–2.69 (m, 2H, CH2), 2.67–
2.56 ppm (m, 2H, CH2); MS (ES−): m/z (%): 238 (30) [M −
H]−.
(R)-2-(3-Mercaptopropanamido)-4-methylpentanoic acid (8c).
Prepared as previously described.23 1H NMR (400 MHz, CD3OD):
d = 4.50–4.44 (m, 1H, CHa), 2.82–2.74 (m, 2H, CH2), 2.59–2.55
(m, 2H, CH2), 1.79–1.62 (m, 3H, CH + CH2), 1.00–0.94 ppm (m,
(R)-Methyl 2-(3-bromopropanamido)-3-phenylpropanoate (21).
Prepared as previously described.41 M.p. 54–55 ◦C; IR mmax (film):
−1
1
3056, 2954, 1743 (C O), 1665 (C O), 1266, 738 cm ; H NMR
6H, 2 × CH3); MS (ES−): m/z (%): 218 (100) [M − H]−; [a]2D5
=
=
=
(400 MHz, CDCl3): d = 7.30–7.24 (m, 3H, 3 × ArCH), 7.13–
+32.0 (MeOH, c = 0.5).
2292 | Org. Biomol. Chem., 2008, 6, 2282–2294
This journal is
The Royal Society of Chemistry 2008
©

View Online
(R)-Methyl 2-(3-hydroxypropanamido)-3-phenylpropanoate (28).
To an ice-cooled suspension of D-phenylalanine methyl ester
hydrochloride (1.00 g, 4.64 mmol) in absolute EtOH (10 mL) was
added b-propiolactone (0.29 mL, 4.64 mmol) and sodium acetate
purification. 1H NMR (CDCl3, 200 MHz) d 1.03 (s, 9H), 1.16 (d,
J = 7.0 Hz, 3H), 2.72 (m, 1H), 3.69 (s, 3H), 3.76 (m, 2H), 7.3–7.45
(m, 6H), 7.64–7.69 (m, 4H).
(R)-4-(tert-Butyldiphenylsilyloxy)-2-methylbutanoic acid (31).
A solution of 3 N LiOH (4.7 mL) was added dropwise to a stirred
and cooled solution of 30 (1.0 g, 2.81 mmol) in 11 mL of THF
at 0 ◦C. To this was added 11 mL of MeOH and the mixture was
stirred for 20 h at room temperature. The solution was acidified
with 3 N HCl. It was extracted 5 times with diethyl ether. The
organic layer was washed with water and dried over MgSO4. The
solvent was removed under reduced pressure to yield the desired
product as colourless oil (89%), which was used without further
purification. 1H NMR (CDCl3, 200 MHz) d 1.07 (s, 9H), 1.18 (d,
J = 7.0 Hz, 3H), 2.75 (m, 1H), 3.82 (m, 2H), 7.30–7.47 (m, 6H),
7.64–7.73 (m, 4H).
◦
(380 mg, 4.64 mmol). The resulting mixture was stirred at 22 C
for 15 h. The reaction mixture was partitioned between EtOAc
(100 mL) and H2O (100 mL) and the aqueous phase was extracted
with EtOAc (2 × 100 mL). The combined organic extracts were
washed with brine (2 × 100 mL), dried over MgSO4 and the
solvent removed under reduced pressure. Purification by flash
chromatography using a gradient of EtOAc–MeOH (from 99 :
1 to 90 : 10) as eluent afforded 230 mg (21%) of 28 as a colourless
1
oil. H NMR (400 MHz, CDCl3): d = 7.32–7.23 (m, 3H, 3 ×
PhCH), 7.13–7.11 (m, 2H, 2 × PhCH), 6.34 (brs, 1H, J = 7.5 Hz,
NH), 4.93–4.88 (m, 1H, CHa), 3.83 (t, 2H, J = 5.5 Hz, CH2OH),
3.74 (s, 3H, OCH3), 3.18 (dd, 1H, J = 5.5 Hz, J = 14.0 Hz, CHb),
3.08 (dd, 1H, J = 6.5 Hz, J = 14.0 Hz, CHb), 2.44–2.41 ppm
(R)-1-((R)-4-(tert-Butyldiphenylsilyloxy)-2-methylbutanoyl)pyr-
rolidine-2-carboxylic acid (32). Oxalyl chloride (0.3 mL,
3.44 mmol) was added to a stirred and cooled solution of 31
(0.46 g, 1.34 mmol) in 3.9 mL of dry benzene at 0 ◦C under
an atmosphere of argon. The mixture was stirred for 1h at
(m, 2H, CH2C O); 13C NMR (100.6 MHz, CDCl3): d = 172.1
=
=
=
(C O), 170.7 (C O), 135.7 (PhC), 129.2 (2 × PhCH), 128.6 (2 ×
PhCH), 127.2 (PhCH), 58.8 (CH2O), 53.1 (Ca), 52.5 (OCH3), 38.1
−
−
=
(CH2C O), 37.8 ppm (Cb); HRMS (ES ): [M − H] calcd. for
C13H16NO4, 250.1074; found, 250.1079.
0
◦C. Then oxalyl chloride (0.18 mL, 2.06 mmol) was again
added and the stirring was continued at 0 ◦C for 2 h. The
mixture was concentrated in vacuo to give the acid chloride,
which was used without further purification. The acid chloride
in 10 mL of CH2Cl2 was added dropwise to a solution of the
corresponding proline (136 mg, 1.18 mmol) and triethylamine
(164 lL, 1.18 mmol) in 10 mL of CH2Cl2. After stirring at room
temperature for 18 h, the mixture was washed twice with water,
dried with MgSO4, and evaporated to give the desired product
(76%) as colourless oil. 1H NMR (CDCl3, 200 MHz) d 1.01–1.08
(m, 9H), 1.15–1.26 (m, 3H), 1.75–2.40 (m, 3H), 2.60–3.10 (m,
2H), 3.45–3.95 (m, 3H), 4.50–4.65 (m, 1H), 7.3–7.5 (m, 6H),
7.6–7.8 (m, 4H). Anal. Calcd for C25H33O4SiN (439.63): C, 68.30;
H, 7.57; N, 3.19. Found: C, 67.97; H, 7.49; N, 3.33.
(R)-2-(3-Hydroxypropanamido)-3-phenylpropanoic acid (11).
An aqueous solution of 1 N NaOH (2.40 mL, 2.40 mmol) was
added to 28 (200 mg, 0.80 mmol). The resulting mixture was
stirred at 22 ◦C for 15 h. The reaction mixture was acidified
with HCl (10 N) and extracted with EtOAc (5 × 10 mL). The
combined organic extracts were washed with brine (50 mL), dried
over MgSO4 and the solvent removed under reduced pressure.
Purification by flash chromatography using a mixture of EtOAc–
MeOH–AcOH (95 : 5 : 1) as eluent afforded 100 mg (53%) of 11 as
a colourless oil. 1H NMR (400 MHz, DMSO-d6): d = 8.14 (d, 2H,
J = 8.5 Hz, NH), 7.29–7.19 (m, 5H, 5 × PhCH), 4.45–4.41 (m,
1H, CHa), 3.56–3.50 (m, 2H, CH2OH), 3.04 (dd, 1H, J = 5.0 Hz,
J = 14.0 Hz, CHb), 2.86 (dd, 1H, J = 9.0 Hz, J = 14.0 Hz, CHb),
2.24 ppm (t, 2H, J = 7.0 Hz, CH2C O); 13C NMR (100.6 MHz,
(R)-1-((R)-4-Hydroxy-2-methylbutanoyl)pyrrolidine-2-carboxy-
lic acid (9). To a solution of 32 (256 mg, 0.58 mmol) in 10 mL of
THF was added 0.5 mL of HF·pyridine. The mixture was stirred
overnight at room temperature and then evaporated. Water and
AcOEt were added and the aqueous phase was freeze-dried to
give the crude compounds which was then purified by HPLC
(ODS column, k = 218 nm, NH4HCO3 25 mM, pH= 7.5) to
yield the desired product (50%) as a white solid. M.p. 250–280 ◦C
(dec.). 1H NMR (D2O, 400 MHz) d 0.82–0.83 and 0.89–0.98
(m, 3H), 1.69–1.95 (m, 3H), 2.05–2.19 (m, 1H), 2.27–2.37 and
2.40–2.61 and 2.70–2.87 (m, 1H), 3.30–3.66 (m, 4H), 3.95–4.11
and 4.28–4.34 (m, 1H). 13C NMR (CDCl3, 62.9 MHz) d 13.0,
13.1, 23.1, 24.7, 29.9, 31.7, 40.6, 41.1, 47.5, 48.2, 63.1, 64.3, 65.0,
172.9, 173.1, 177.1. IR (KBr) 3432, 1616, 1399 cm−1. LRMS
APCI− m/z 200.4 ([M − H]−, 100). [a]D25 = +26.5 (H2O pH = 7.3,
c = 0.52). Anal. Calcd for C9H15NO4 (201.22): C, 53.72; H, 7.51;
N, 6.96. Found: C, 53.64; H, 7.41; N, 7.11.
=
=
=
DMSO-d6): d = 173.5 (C O), 171.1 (C O), 138.1 (PhC), 129.6
(2 × PhCH), 128.6 (2 × PhCH), 126.9 (PhCH), 58.0 (CH2O), 53.8
−
−
=
(Ca), 39.3 (CH2C O), 37.3 ppm (CHb); HRMS (ES ): [M − H]
calcd. for C12H14NO4, 236.0917; found, 236.0918.
(R)-2-Acetamido-3-phenylpropanoic acid (12). Prepared as
previously described.24 1H NMR (400 MHz, CDCl3): d = 8.20
(d, 1H, J = 8.0 Hz, NH), 7.30–7.18 (m, 5H, 5 × PhCH), 4.43–4.37
(m, 1H, CHa), 3.04 (dd, 1H, J = 5.0 Hz, J = 14.0 Hz, CHb), 2.83
(dd, 1H, J = 9.5 Hz, J = 14.0 Hz, CHb), 1.78 ppm (s, 3H, CH3);
13
=
=
C NMR (100.6 MHz, CDCl3): d = 174.1 (C O), 170.1 (C O),
138.6 (ArC), 129.9 (2 × PhCH), 129.0 (2 × PhCH), 127.3 (PhCH),
54.4 (Ca), 37.6 (Cb), 23.2 ppm (CH3).
(R)-Methyl
4-(tert-butyldiphenylsilyloxy)-2-methylbutanoate
(30). To a solution of alcohol 29 (1.0 g, 8.46 mmol) in 2.5 mL of
DMF were added imidazole (1.15 g, 16.9 mmol) and TBDPSCl
(4.4 mL, 8.46 mmol). The mixture was stirred overnight and
poured into 50 mL of 1.2 M HCl. The aqueous phase was
extracted with diethyl ether (3 × 50 mL) and the organic layer
was washed with water and dried over MgSO4. The solvent was
removed under reduced pressure to yield (65%) to the desired
product as a colourless oil which was used without further
(R)-1-Isobutyrylpyrrolidine-2-carboxylic acid (10). To a solu-
tion of D-proline (0.2 g, 1.74 mmol) in CH2Cl2 (15 mL) and
triethylamine (183 lL, 1.74 mmol) was added dropwise at −10 ◦C
in an atmosphere of argon a solution of isobutyryl chloride
(183 lL, 1.74 mmol) in CH2Cl2 (15 mL). The mixture was stirred
at room temperature for 18 h and then washed twice with water
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2282–2294 | 2293
©

View Online
and dried over MgSO4. The solvent was filtered and evaporated
14 N. Selevsek, A. Tholey, E. Heinzle, B. M. Lienard, N. J. Oldham, C. J.
Schofield, U. Heinz, H. W. Adolph and J. M. Frere, J. Am. Soc. Mass
Spectrom., 2006, 17, 1000–1004.
15 A. Carfi, E. Duee, M. Galleni, J. M. Frere and O. Dideberg, Acta
Crystallogr., Sect. D: Biol. Crystallogr., 1998, 54, 313–323.
16 J. H. Ullah, T. R. Walsh, I. A. Taylor, D. C. Emery, C. S. Verma, S. J.
Gamblin and J. Spencer, J. Mol. Biol., 1998, 284, 125–136.
17 I. Garcia-Saez, P. S. Mercuri, C. Papamicael, R. Kahn, J. M. Frere,
M. Galleni, G. M. Rossolini and O. Dideberg, J. Mol. Biol., 2003, 325,
651–660.
under reduced pressure to give 10 as a white solid (73%). M.p.
◦
1
112 C (dec.). H NMR (D2O, 250 MHz) d 1.19 and 1.22 (2 ×
d, J = 7.0 Hz, 2 × 3H), 1.93–2.15 and 2.44–2.52 (2 × m, 4H),
2.74 (m, 1H), 3.50–3.73 (m, 2H), 4.63 (m, 1H). 13C NMR (D2O,
62.9 MHz) d 18.9, 19.2, 25.1, 28.5, 32.8, 47.7, 58.8, 174.4, 178.4.
IR (KBr) 3434, 1723, 1600 cm−1. LRMS APCI− m/z 184.4 ([M −
H]−, 100). Anal. Calcd for C9H15NO3 (185.22): C, 58.36; H, 8.13;
N, 7.56. Found C, 58.21; H, 8.25; N, 7.19.
18 J. Gao, X. Cheng, R. Chen, G. B. Sigal, J. E. Bruce, B. L. Schwartz,
S. A. Hofstadler, G. A. Anderson, R. D. Smith and G. M. Whitesides,
J. Med. Chem., 1996, 39, 1949–1955.
19 J. L. Benesch and C. V. Robinson, Curr. Opin. Struct. Biol., 2006, 16,
245–251.
Acknowledgements
20 A. J. Heck and R. H. Van Den Heuvel, Mass Spectrom. Rev., 2004, 23,
368–389.
We thank the Biotechnology and Biological Science Research
Council and the E.U. (contract no HPRN-CT-2002-00264) for
funding and Amura for a CASE award to BMRL. We also thank
Carine Bebrone for CphA.
21 J. A. Loo, Mass Spectrom. Rev., 1997, 16, 1–23.
22 H. A. Staab and R. G. H. Kirrstetter, Liebigs Ann. Chem., 1979, 6,
886–898.
23 M. A. Ondetti and D. W. Cushman, Ger. Pat. DE 2717548, 1977.
24 H. K. Chenault, J. Dahmer and G. M. Whitesides, J. Am. Chem. Soc.,
1989, 111, 6354–6364.
25 S. Bounaga, M. Galleni, A. P. Laws and M. I. Page, Bioorg. Med. Chem.,
2001, 9, 503–510.
26 R. A. Williamson, M. D. Carr, T. A. Frenkiel, J. Feeney and R. B.
Freedman, Biochemistry, 1997, 36, 13882–13889.
27 I. Garcia-Saez, J. Hopkins, C. Papamicael, N. Franceschini, G.
Amicosante, G. M. Rossolini, M. Galleni, J. M. Frere and O. Dideberg,
J. Biol. Chem., 2003, 278, 23868–23873.
References
1 T. R. Walsh, M. A. Toleman, L. Poirel and P. Nordmann, Clin.
Microbiol. Rev., 2005, 18, 306–325.
2 A. Felici, G. Amicosante, A. Oratore, R. Strom, P. Ledent, B. Joris, L.
Fanuel and J. M. Frere, Biochem. J., 1993, 291, 151–155.
3 B. Segatore, O. Massidda, G. Satta, D. Setacci and G. Amicosante,
Antimicrob. Agents Chemother., 1993, 37, 1324–1328.
28 R. Natesh, S. L. U. Schwager, H. R. Evans, E. D. Sturrock and K. R.
Acharya, Biochemistry, 2004, 43, 8718–8724.
4 M. Hernandez Valladares, A. Felici, G. Weber, H. W. Adolph, M.
Zeppezauer, G. M. Rossolini, G. Amicosante, J. M. Frere and M.
Galleni, Biochemistry, 1997, 36, 11534–11541.
5 G. Garau, C. Bebrone, C. Anne, M. Galleni, J. M. Frere and O.
Dideberg, J. Mol. Biol., 2005, 345, 785–795.
6 C. Damblon, C. Prosperi, L. Y. Lian, I. Barsukov, R. P. Soto, M.
Galleni, J. M. Frere and G. C. K. Roberts, J. Am. Chem. Soc., 1999,
121, 11575–11576.
7 C. Damblon, M. Jensen, A. Ababou, I. Barsukov, C. Papamicael, C. J.
Schofield, L. Olsen, R. Bauer and G. C. Roberts, J. Biol. Chem., 2003,
278, 29240–29251.
29 R. Natesh, S. L. U. Schwager, E. D. Sturrock and K. R. Acharya,
Nature, 2003, 421, 551–554.
30 S. Bailey, Acta Crystallogr., Sect. D: Biol. Crystallogr., 1994, 50, 760–
763.
31 P. Emsley and K. Cowtan, Acta Crystallogr., Sect. D: Biol. Crystallogr.,
2004, 60, 2126–2132.
32 H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H.
Weissig, I. N. Shindyalov and P. E. Bourne, Nucleic Acids Res., 2000,
28, 235–242.
33 H. Kuboniwa, S. Grzesiek, F. Delaglio and A. Bax, J. Biomol. NMR,
1994, 4, 871–878.
8 N. O. Concha, C. A. Janson, P. Rowling, S. Pearson, C. A. Cheever,
B. P. Clarke, C. Lewis, M. Galleni, J. M. Frere, D. J. Payne, J. H. Bateson
and S. S. Abdel-Meguid, Biochemistry, 2000, 39, 4288–4298.
9 J. H. Toney, G. G. Hammond, P. M. Fitzgerald, N. Sharma, J. M.
Balkovec, G. P. Rouen, S. H. Olson, M. L. Hammond, M. L. Greenlee
and Y. D. Gao, J. Biol. Chem., 2001, 276, 31913–31918.
10 C. Mollard, C. Moali, C. Papamicael, C. Damblon, S. Vessilier, G.
Amicosante, C. J. Schofield, M. Galleni, J. M. Frere and G. C. Roberts,
J. Biol. Chem., 2001, 276, 45015–45023.
11 U. Heinz, R. Bauer, S. Wommer, W. Meyer-Klaucke, C. Papamichaels,
J. Bateson and H. W. Adolph, J. Biol. Chem., 2003, 278, 20659–20666.
12 J. D. Buynak, H. Chen, L. Vogeti, V. R. Gadhachanda, C. A. Buchanan,
T. Palzkill, R. W. Shaw, J. Spencer and T. R. Walsh, Bioorg. Med. Chem.
Lett., 2004, 14, 1299–1304.
34 M. Piotto, V. Saudek and V. Sklenar, J. Biomol. NMR, 1992, 2, 661–665.
35 L. Y. Lian, I. Barsukov, A. P. Golovanov, D. I. Hawkins, R. Badii, K. H.
Sze, N. H. Keep, G. M. Bokoch and G. C. K. Roberts, Structure, 2000,
8, 47–55.
36 F. W. Muskett, T. A. Frenkiel, J. Feeney, R. B. Freedman, M. D. Carr
and R. A. Williamson, J. Biol. Chem., 1998, 273, 21736–21743.
37 G. Jones, P. Willett, R. C. Glen, A. R. Leach and R. Taylor, J. Mol.
Biol., 1997, 267, 727–748.
38 http://www.chemaxon.com/marvin/.
39 http://swift.cmbi.kun.nl/WIWWWI/.
40 G. J. Bodwell, J. N. Bridson, S. L. Chen and R. A. Poirier, J. Am. Chem.
Soc., 2001, 123, 4704–4708.
41 J. Geiwiz, E. Goetschi and P. Hebeisen, Synthesis, 2003, 1699–1704.
42 J. L. Stanton, N. Gruenfeld, J. E. Babiarz, M. H. Ackerman, R. C.
Friedmann, A. M. Yuan and W. Macchia, J. Med. Chem., 1983, 26,
1267–1277.
13 Q. Sun, A. Law, M. W. Crowder and H. M. Geysen, Bioorg. Med. Chem.
Lett., 2006, 16, 5169–5175.
2294 | Org. Biomol. Chem., 2008, 6, 2282–2294
This journal is
The Royal Society of Chemistry 2008
©