Structure-based efforts to optimize a non-β-lactam inhibitor of AmpC β-lactamase

Article abstract of DOI:10.1016/j.bmc.2014.04.051

β-Lactams are the most widely prescribed class of antibiotics, yet their efficacy is threatened by expression of β-lactamase enzymes, which hydrolyze the defining lactam ring of these antibiotics. To overcome resistance due to β-lactamases, inhibitors that do not resemble β-lactams are needed. A novel, non-β-lactam inhibitor for the class C β-lactamase AmpC (3-[(4-chloroanilino)sulfonyl]thiophene-2-carboxylic acid; K_i 26 μM) was previously identified. Based on this lead, a series of compounds with the potential to interact with residues at the edge of the active site were synthesized and tested for inhibition of AmpC. The length of the carbon chain spacer was extended by 1, 2, 3, and 4 carbons between the integral thiophene ring and the benzene ring (compounds 4, 5, 6, and 7). Compounds 4 and 6 showed minimal improvement over the lead compound (K_i 18 and 19 μM, respectively), and compound 5 inhibited to the same extent as the lead. The X-ray crystal structures of AmpC in complexes with compounds 4, 5, and 6 were determined. The complexes provide insight into the structural reasons for the observed inhibition, and inform future optimization efforts in this series.

Full text of DOI:10.1016/j.bmc.2014.04.051

Bioorganic & Medicinal Chemistry xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Structure-based efforts to optimize a non-b-lactam inhibitor of AmpC
b-lactamase
Jenna M. Hendershot a, , Uma J. Mishra b, Robert P. Smart b, William Schroeder b, Rachel A. Powers a,b,
⇑
a Cell and Molecular Biology Program, Grand Valley State University, Allendale, MI 49401, United States
b Department of Chemistry, Grand Valley State University, Allendale, MI 49401, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
b-Lactams are the most widely prescribed class of antibiotics, yet their efficacy is threatened by expres-
sion of b-lactamase enzymes, which hydrolyze the defining lactam ring of these antibiotics. To overcome
resistance due to b-lactamases, inhibitors that do not resemble b-lactams are needed. A novel, non-b-lac-
tam inhibitor for the class C b-lactamase AmpC (3-[(4-chloroanilino)sulfonyl]thiophene-2-carboxylic
Received 14 February 2014
Revised 16 April 2014
Accepted 25 April 2014
Available online xxxx
acid; Ki 26 lM) was previously identified. Based on this lead, a series of compounds with the potential
to interact with residues at the edge of the active site were synthesized and tested for inhibition of AmpC.
The length of the carbon chain spacer was extended by 1, 2, 3, and 4 carbons between the integral thio-
phene ring and the benzene ring (compounds 4, 5, 6, and 7). Compounds 4 and 6 showed minimal
Keywords:
Antibiotic resistance
Structure-based drug design
X-ray crystallography
improvement over the lead compound (Ki 18 and 19 lM, respectively), and compound 5 inhibited to
the same extent as the lead. The X-ray crystal structures of AmpC in complexes with compounds 4, 5,
and 6 were determined. The complexes provide insight into the structural reasons for the observed inhi-
bition, and inform future optimization efforts in this series.
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
structure allows bacteria to recognize the lactam ring and easily
evolve resistance to these similar molecules. Inhibitors that do
Antibiotic resistance has emerged as one of the leading public
health crises of the 21st century.1–3 A crucial part of this problem
is the increasingly widespread resistance to b-lactam antibiotics,
such as the penicillins and cephalosporins (Fig. 1), mainly due to
their extensive use and misuse.4 Several mechanisms can contrib-
ute to the resistant phenotype. However, the most prevalent is the
expression of b-lactamase enzymes. These enzymes hydrolyze the
four-membered lactam ring of the antibiotic (Fig. 1A), rendering it
ineffective and unable to inhibit its intended target, the cell wall
transpeptidase enzyme.
To overcome resistance due to b-lactamases, compounds that
inhibit these enzymes were introduced and are generally co-
administered with a partner b-lactam antibiotic.5 Most are derived
from natural sources and contain the same defining b-lactam ring
as their antibiotic counterparts (Fig. 1B). Reliance on this same core
not resemble b-lactams would require bacteria to develop novel
resistance mechanisms.
Previous research identified
inhibitor for the class C b-lactamase AmpC (3-[(4-chloroanili-
no)sulfonyl]thiophene-2-carboxylic acid; Ki 26 M) that was found
a non-covalent, non-b-lactam
l
to bind competitively and reversibly (Fig. 1C).6 The X-ray crystal
structure of AmpC in complex with this lead compound suggested
numerous features of the inhibitor were essential for the binding
affinity. Several analogs of this lead were tested in inhibition
assays to better understand the SAR of this compound. Based on
these previous studies, some features of this inhibitor were deter-
mined to contribute to recognition by key active site residues on
the enzyme. The carboxylate group was shown to bind in the puta-
tive oxyanion hole,7 making hydrogen bonds with main chain
atoms of the catalytic serine (Ser64) and Ala318. The sulfonamide
group and its orientation were important for binding and recogni-
tion by R1 side chain recognition residues (Asn152 and Ala318).
Finally, the chlorophenyl ring was shown to stack with the con-
served residue Tyr221.
Abbreviations: DMSO, dimethylsulfoxide; DMF, dimethylformamide; NMP,
N-methylpyrrolidone; SAR, structure activity relationships.
⇑
Corresponding author. Tel.: +1 616 331 2853; fax: +1 616 331 3230.
Despite its modest inhibition, the 26 lM inhibitor was able to
overcome resistance in bacterial cell culture, reducing the mini-
mum inhibitory concentration (MIC) of the b-lactam ampicillin
E-mail address: powersra@gvsu.edu (R.A. Powers).
Current address: Biological Chemistry Department, University of Michigan, Ann
Arbor, MI 48109, United States.
http://dx.doi.org/10.1016/j.bmc.2014.04.051
0968-0896/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Hendershot, J. M.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.051

2
J. M. Hendershot et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
Table 1
Kinetic characterization of thiophene-carboxy derivatives against AmpC
Compound Structure
Ki
Compound Structure
Ki
(lM)
(lM)
Lead
26a
4
18
31
1
ND
5
(26c)
Figure 1. Structures of several ligands that bind to b-lactamase enzymes.
(A) b-Lactam antibiotic (amoxicillin) with the four-membered ring highlighted.
(B) Clavulanic acid, a b-lactamase inhibitor. (C) Lead compound (3-[(4-chloroani-
1.7
1A
6
19
(1.1b)
lino)sulfonyl]thiophene-2-carboxylic acid; Ki 26 lM).
four-fold in a resistant strain of Escherichia coli that overexpresses
AmpC.6 The novel chemical structure of this lead, along with its
weak biological activity, made it a good candidate for optimization
efforts to improve affinity and ultimately resulted in the identifica-
tion of a 1.1 lM inhibitor (compound 1A; Table 1) with improved
activity against resistant bacteria.8 This series of analogs explored
the addition of different functional groups to the aryl ring. The best
inhibitors (which include compound 1A) share a carboxylate group
addition to the aryl ring. The X-ray crystal structure of 1A in com-
plex with AmpC (PDB 1XGJ)8 reveals that this carboxylate group
points toward the solvent exposed entrance to the active site
(Fig. 2).
2
58
7
ND
Comparison of the structures of AmpC in complexes with both
the 26 lM lead and compound 1A reveals that a large portion of
the active site remains open for possible interactions with an
inhibitor, which could potentially improve the binding affinity.
Asp123, Arg204, and Ser212 are located at the lip of the active site
and are almost completely conserved in class C b-lactamases with
observed substitutions maintaining the same charge or similar
polarity. These polar residues could potentially allow for favorable
interaction with inhibitors (Fig. 2). The carboxylate group addition
to compound 1A extended toward this area of the active site but
3
112
Powers (2002).6
a
b
c
Reported in Tondi (2005).8
made only
a
water-mediated hydrogen bond with Arg204
Calculated by Lineweaver–Burk analysis.
(Fig. 2). Increasing the size and flexibility of the compounds might
allow for a direct ionic interaction between the carboxylate and
Arg204 to be formed. It could also provide a more extensive sam-
pling of this distal region of the active site, and possibly identify
other residues at the edge of the active site for interaction with a
more extended inhibitor.
Based on the previous set of analogs, the addition of a one-car-
bon linker between the sulfonamide and the phenyl ring provided
an increase in flexibility and also preserved the same binding affin-
ity as the lead.8 In many of the previous analogs, the addition of a
carboxylate to the aryl ring greatly improved the binding affinity,
up to 30-fold in some cases.8 We hypothesized that addition of a
carboxylate to a more flexible compound might result in similar
improvement in binding affinity. Additionally, a more flexible
analog might be able to interact with distal residues in the open
active site. We wondered what impact this would have on the
binding affinity. A series of compounds based on the lead were
synthesized where the length of the carbon chain between the sul-
fonamide group and the phenyl ring was extended by 1, 2, 3, and 4
carbons. The increased flexibility of the analogs would explore
more of the active site and could probe for favorable interactions
with active site lip residues to improve affinity.
The structure–activity relationship of this series was explored.
Despite this design effort, the inhibitory activity of the compounds
against AmpC did not drastically improve over the lead. Compound
5, the compound that increased the chain length by two carbons
Please cite this article in press as: Hendershot, J. M.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.051

J. M. Hendershot et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
3
Figure 2. Surface representation of the AmpC/1A complex. Key active site residues important for recognition of the thiophene/sulfonamide core of the inhibitor (white
labels). Lip residues (black labels) that were targeted for interaction with the extended, flexible series of analogs (Table 1). Carbon atoms for protein residues are colored
green, carbons for the inhibitor are pink, nitrogens are blue, oxygens red and sulfurs yellow. All figures were created with PyMOL.26
had a Ki value of 26
pounds 4 and 6, the one and three carbon compounds, were
slightly better than the lead, Ki values of 18 M and 19 M, respec-
tively. To investigate the structural basis for inhibition, we co-crys-
tallized each of the compounds 4, 5, and 6 with AmpC. The
structures of each of these complexes provide information for
future structure-based drug design against AmpC and underscore
the difficulties in optimization of novel leads.
l
M, the same as the lead compound. Com-
chloride with the free amine in anhydrous pyridine (Scheme 1).
Each intermediate ester was characterized either through thin
layer chromatography (TLC) or 1H NMR. Each methyl ester was
then hydrolyzed to the final carboxylic acid derivative using sapon-
ification conditions of sodium hydroxide and precipitation with
HCl (Scheme 1; Table 1). Reactions were monitored using TLC,
and final products were characterized by 1H NMR, and 13C NMR.
The creation of compounds 5, 6, and 7 involved a multi-step syn-
thesis of the free amine followed by coupling to sulfonyl chloride.
The stepwise process to synthesize the free amine included Gabriel
synthesis9 followed by a Friedel Crafts acylation10 and finally oxi-
dation to the carboxylic acid by the haloform reaction (Scheme 2).
Once the amine salt was synthesized, Schotten–Baumann11 reac-
tion conditions were used in order to couple the free amine to sul-
fonyl thiophene chloride.
l
l
2. Results
2.1. Synthesis
The synthesis of methyl ester intermediates 1–4 was accom-
plished by reacting commercially available sulfonyl thiophene
Scheme 1. Synthesis of 3-((4-chloroanilin)sulfonyl)thiophene-2-carboxylic acid derivatives compounds 1–4. X = COOH (for 1, 1A), CH3 (for 2) or Cl (for 3). n = 1 (for 4).
Please cite this article in press as: Hendershot, J. M.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.051

4
J. M. Hendershot et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
Scheme 2. Synthesis of 3-[(4-chloroanilin)sulfonyl]thiophene-2-carboxylic acid derivatives compounds 5–7. n = 2, 3, or 4 for compounds 5, 6, and 7, respectively.
2.2. Enzymology
inhibitors were present in the active sites of the structures. The
inhibitors were built into the structures using Coot14 and refined
with Refmac515 in the CCP4 Suite.16 Overlays of each of these
structures with the lead compound indicate a similar binding
mode for the thiophene ring/sulfonamide core structure that is
bound near the active site serine residue (Ser64). In every inhibitor
conformation, the electron density for the core is very well defined.
However, in the distal portion beyond the sulfonamide group, the
electron density is much weaker, indicating the increased flexibil-
ity of this part of the inhibitor as it approaches solvent (Fig. 3).
With increasing carbon spacer length, the distal portion of the
inhibitor does indeed extend further out of the active site, away
from the catalytic serine and toward residues Arg204, Asp123,
and Ser212 that mark the edge of the active site.
In the AmpC/4 complex, the position of the inhibitor in the
active site was unambiguously identified in the B monomer only.
The addition of a carboxylate to the benzene ring of compound 4
added four hydrogen bond interactions; three to water molecules
(W78, W432, W433) and one to the main chain nitrogen of
Ser212 (Fig. 4A). The interaction with Ser212 is a result of the
increased length and flexibility afforded by the one-carbon exten-
Using an established enzyme assay,6 each of the compounds
was tested as an inhibitor of the class C b-lactamase AmpC
(Table 1). The ester precursors were observed to be insoluble in
assay buffer (50 mM Tris, pH 7.0) and were unable to be character-
ized kinetically. The seven free acids (1A and 2–7) were evaluated,
and IC50 and Ki values were determined (Table 1). Inhibition Ki val-
ues were obtained from IC50 plots assuming competitive inhibi-
tion.12 The Ki value for compound 1A was 1.7
similar to the Ki determined by Tondi et al. (1.1
lM, which is
l
M).8
Varying the substituent on the phenyl ring showed improved Ki
values in the initial lead as compared to compound 1A. Here, the
replacement of the chlorine with a carboxylate group improved
binding ꢀ13-fold. With compounds 2, 3, and 4, we further explored
the impact of substitutions at this position. The methyl (2), chlo-
rine (3) and carboxylate (4) substituents had Ki values of 58, 112
and 18 lM, respectively. Compounds 2 and 3 with the more non-
polar substituents bound with ꢀ3-fold and 6-fold worse binding
affinity as compared with compound 4 with the polar carboxylate
substituent at the phenyl ring. Therefore, the remaining analogs
contained a carboxylate substituent at this position and were fur-
ther pursued to explore the effect of linker length.
sion to the linker. The slightly improved binding affinity (18
lM
compared to 26 M for the lead) may be attributed to the two
l
With compound 4, the linker length was increased by one car-
bon as compared with 1A, and 4 was determined to have a Ki value
additional hydrogen bonds made to the enzyme (Ser212 and
Gln120). The favorable quadrupole stacking interaction with
Tyr221 is also maintained in this complex.
of 18
fold reduction in Ki when compared to compound 1A (Ki 1.1
Compound 5 contained a two-carbon linker and was determined
to have a Ki value of 31 M, an approximate 2-fold reduction when
l
M. Increasing linker length by one carbon resulted in a ꢀ10-
lM).
In the AmpC/5 complex, electron density for the inhibitor is
observed in both monomers. The thiophene/sulfonamide core
binds the same as the lead and 4, except the sulfonamide group
of 5 does not make a hydrogen bond with Gln120. In the A mono-
mer, a single conformation for 5 is present, with the trajectory of
the carboxyphenyl ring extending toward the edge of the active
site. The carboxylate group makes hydrogen bonds with Tyr221OH
and the main chain oxygen of Ser212 (Fig. 4B). A similar conforma-
tion is also observed in the B monomer, but in addition to this, an
alternate binding mode is adopted. The alternate conformation
also extends out to solvent, and explores a different area of the
active site. The carboxylate group now favorably interacts with
Arg204 (hydrogen bond to Arg204N (2.8 Å) and ionic interactions,
3.5 Å). Due to the two-carbon linker, this inhibitor extends too far
beyond Tyr221 to make the favorable quadrupole interactions
observed in the lead and compound 4, and the binding affinity
l
compared to 4. The Ki value for 5 was also obtained by varying
both substrate and inhibitor concentrations, and this full kinetic
analysis confirmed competitive inhibition (data not shown).
Interestingly, the Ki value for compound 6, which contained the
three-carbon linker, was approximately the same as the one-
carbon linker (4), with a Ki value of 19 lM. Overall the inhibitory
activity of the compounds against AmpC was not drastically
improved over the lead, but these results show that even with a
one-carbon extension, addition of a carboxylate on the phenyl ring
improves binding affinity (compound 4 vs compound 2, 3).
2.3. X-ray crystallographic structure determination
The X-ray crystal structures of AmpC in complexes with com-
pounds 4–6 were determined to 1.9, 1.8, and 2.3 Å, respectively.
The quality of the models was evaluated with MolProbity.13 For
the AmpC/4 complex, 100% of the residues were in the allowed
region of the Ramachandran plot, with 99.6% in the allowed region
for AmpC/5, and 99.9% for AmpC/6. In all cases, AmpC crystallized
in the space group C2 with a dimer in the asymmetric unit. Initial
(31 lM) was similar to the lead (26 lM).
With the longer linker of 6, the carboxyphenyl ring is able to
extend even further away from the center of the active site and
toward the edge of the active site. However, the carboxylate func-
tional group does not make any specific interactions with polar
residues at the lip of the active site (e.g., Arg204, Asp123, or
Ser212) in the B monomer, and in the A monomer, there is only
a single hydrogen bond made between this group and a main chain
Fo-Fc difference density, contoured at 3.0 r, indicated that the
Please cite this article in press as: Hendershot, J. M.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.051

J. M. Hendershot et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
5
Figure 3. Stereo view 2Fo–Fc electron density maps contoured at 1
r of the active site of AmpC in complexes with non-covalent inhibitors. (A) Compound 4. (B) Compound 5.
(C) Compound 6. Carbon atoms of the protein are colored green and of 4 are cyan, 5 are magenta and 6 are orange. Nitrogen atoms are blue, oxygen atoms red and sulfur
atoms yellow. All figures were created with PyMOL.26
oxygen of His210 (Fig. 4C). The binding affinity was not signifi-
cantly improved (Ki 19 M), and the even longer three-carbon lin-
3. Discussion
l
ker does not allow for quadrupole interactions with Tyr221. In the
complex of AmpC with 6, some evidence exists for a second confor-
mation in the A monomer, but the density is so poorly defined that
modeling it was not feasible. Compound 6 does bind differently in
monomer A when compared to monomer B.
b-Lactamases are the most widespread resistance mechanisms
to b-lactam antibiotics, and there is a pressing need for novel,
non-b-lactam drugs. A previously identified thiophene-based lead
(3-[(4-chloroanilino)sulfonyl] thiophene-2-carboxylic acid; Ki
26 l
M) was found to bind competitively and reversibly to AmpC,6
Please cite this article in press as: Hendershot, J. M.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.051

6
J. M. Hendershot et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
and its affinity was improved to 1.1
lM (compound 1A) by replac-
two carbon linker allows for hydrogen bonds to be formed with
ing the chlorine of the lead with a carboxylate on the aryl ring.8 The
improvement was tentatively attributed to a water-mediated
interaction between this carboxylate and Arg204 (1XGJ) at the
entrance of the active site, although other possibilities could not
be ruled out. We hypothesized that extension of the linker
between the core structure of the lead and the aryl ring, coupled
with the addition of the distal carboxylate, might result in a direct
interaction with Arg204 and thereby improve binding affinity.
Additionally, we wondered if increased flexibility in this series of
analogs might identify other potential binding partners for the dis-
tal carboxylate at the edge of the active site.
side chain atoms of both Arg204 and Ser212, located at the edge
of the active site, but only in one of the two conformations
observed. Despite the additional hydrogen bonds, this compound
has a worse Ki value than 4, which could be attributed to the loss
of the quadrupole interactions with Tyr221. Compound 6, with
the three-carbon linker fails to make quadrupole interactions with
Tyr221 or with any of the polar residues at the edge of the active
site, and only a single hydrogen bond is made with a main chain
carbonyl oxygen. Strangely, this lack of protein interactions with
6 translates into a Ki value similar to that of compound 4 that does
make additional interactions with the enzyme.
The kinetic characterization coupled with representative crystal
structures suggests that increasing the linker length in this series
of analogs is not a rational approach to improving the affinity of
this series. Other groups17,18 have explored the inclusion of linkers,
both rigid and flexible, as a way to improve binding affinity with
varying degrees of success. In each instance, optimization of the
length of the linker was essential. In our case, the addition of a
one carbon linker to the lead resulted in compound 3, which bound
ꢀ4.5-fold worse than the lead. Replacing the chlorine in 3 with a
carboxylate resulted in a 6-fold reduction in the Ki value from
The addition of a 1–4 carbon linker introduced length and flex-
ibility within the series. We initially hypothesized that this flexibil-
ity would allow compounds to explore other possible binding
partners present at the active site entrance, for example Asp123
and Ser212 (Figs. 2 and 5). All three of these residues are almost
completely conserved in class C b-lactamases with observed sub-
stitutions maintaining the same charge or similar polarity (e.g.,
Arg204 ? Lys; Asp123 ? Glu; Ser212 ? Thr/Asn). The increased
flexibility of these molecules is evident in the crystal structures
(Fig. 4). Multiple conformations are observed for the substituted
aryl ring of compounds 5 and 6, but not compound 4. The longer
linkers did allow these molecules to extend to the far edge of the
active site and interact with Arg204 and Ser212 (Fig. 5). However,
the affinity of these compounds derives from a balance between
the favorable interactions formed with these polar residues and
the entropic penalty for binding more flexible compounds. Overall,
if these lip residues are to be exploited as binding sites in inhibitor
design, a more rigid linker scaffold may be desired.
A longer linker in the presence of the distal carboxylate posed a
trade-off between flexibility and solubility in our optimization
attempts. Our data indicate that the three-carbon linker was the
maximum length that was soluble in buffer. Within this series of
new analogs, the one carbon linker was the optimal length for inhi-
bition (compound 4). This is a compromise that allowed additional
hydrogen bonds with Gln120, but not so flexible that many binding
modes are allowed. Overall, it appears that molecules with
increased length do not behave as predicted.
In summary, our optimization efforts highlight the design chal-
lenges faced in attempting to improve a novel lead compound in a
large, relatively open binding site. Increased flexibility due to
entropic destabilization coupled with the loss of quadrupole inter-
actions with Tyr221 appear to be detrimental to binding affinity in
this series of analogs. Additionally, interactions with lip active site
residues (such as Arg204, Ser212, and Asp123) are not significant,
and do not translate to improved binding affinity as initially pro-
posed. Future optimization efforts in this series of novel, non-cova-
lent b-lactamase inhibitors should maintain edge-to-face aromatic
interactions with Tyr221. Exploring the impact of adopting a more
hindered, rigid linker system might also improve affinity by plac-
ing the trajectory of the inhibitor in a single conformation to favor-
ably interact with residues in the AmpC active site.
112 lM (3) to 18 lM (4). This trend was also observed when the
chlorine of the lead was substituted for a carboxylate to produce
compound 1A.8 Increasing linker length by another one or two car-
bons (compounds 5 and 6, respectively) gives mixed results: the
affinity of compound 5 is about two-fold worse than compound
4, and the affinity of compound 6 returns to that of compound 4.
So while the affinity of this series of inhibitors can be improved
by addition of a distal carboxylate, none inhibit as well as 1A. To
more fully explore why the design of this series was unsuccessful,
the structures of three of these analogs were determined in com-
plexes with AmpC.
Unfortunately, the structural basis for inhibition in this series is
not easily explained by these complexes, although there are a few
indications that might inform future design efforts. The structure
of the lead molecule in complex with AmpC highlighted the impor-
tance of the thiophene/sulfonamide core and its hydrogen bonding
interactions with Ser64 and Asn152, and quadrupole stacking with
Tyr221.6 In each of the complexes determined here, the interac-
tions with Ser64 and Asn152 were maintained, and the electron
density for the thiophene/sulfonamide core was much better
defined than the portion beyond the sulfonamide group.
The quadrupole interactions between the aryl ring of the inhib-
itors and Tyr221 are not maintained in all of the complexes deter-
mined. These interactions are only observed in the complex with 4,
the one carbon extension. Increasing the carbon linker length any
further, as with compounds 5 and 6, results in the loss of the quad-
rupole interaction with Tyr221. Attractive nonbonded interactions
between aromatic rings are prevalent in protein-ligand complexes;
the energetic contributions for this type of interaction in protein/
ligand complexes is suggested to be on the order of 1–2 kcal/
mol.19 This type of interaction has been determined to contribute
to the binding affinity of other b-lactamase inhibitors, such as
the arylboronic acids,20–23 in addition to the lead compound in this
series, and it appears to be advantageous to maintain this interac-
tion in any future analogs.
As for unique interactions between the analogs and the enzyme,
compounds 4, 5, and 6 make relatively few new interactions with
AmpC, compared to the lead. In the AmpC/4 complex, additional
hydrogen bonds were formed between the core sulfonamide group
and Gln120, and also the distal carboxylate and Ser212, thus pro-
viding one possible explanation for improved affinity. However,
the residue that was targeted for direct interaction with the distal
carboxylate, Arg204, does not interact with the carboxylate due to
altered trajectory of 4 as compared to the lead. Compound 5 with a
4. Materials and methods
4.1. General information
All reagents were purchased from Sigma–Aldrich and used as
received. Reactions were performed with dry glassware under an
atmosphere of nitrogen or argon unless otherwise noted. 1H and
13C NMR spectra were measured on a JEOL JNM-ECP 300 FT-NMR
(300 MHz) spectrometer using DMSO-d6 as a solvent. The values
of the chemical shifts (d) are given in ppm and coupling constants
(J) in hertz. Peak multiplicities were given as follows: s, singlet; d,
Please cite this article in press as: Hendershot, J. M.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.051

J. M. Hendershot et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
7
Figure 4. Hydrogen bonding interactions between AmpC and non-covalent inhibitors. (A) Compound 4. (B) Compound 5. (C) Compound 6.
doublet; t, triplet; q, quartet; m, multiplet; and br, broad. Infrared
spectra were recorded on a JASCO FTIR 4100. Mass spectrometry
was performed on a single quadrupole MSQ mass spectrometer
(Thermo Scientific) connected to the HPLC system (C18 column:
HPLC water, acetonitrile and acetonitrile/formic acid). Elemental
Analyses were performed by Atlantic Microlab, Inc, in Norcross,
GA.
4.2.1. 3-(4-Carboxybenzylsulfamoyl)thiophene-2-carboxylic
acid (Compound 4)
Step 1-To a solution of 4-aminomethylbenzoic acid (0.7 g) and
sodium carbonate (1.0 g) in water (25 mL) was added 2-carbome-
thoxy-3-chlorosulfoxylthiophene (1.2 g) in 1,2-dimethoxyethane
(25 mL). The reaction mixture was stirred at room temperature
for 30 min followed by an additional 30 min of stirring at 55 °C.
Water (100 mL) was added and the resulting solution was filtered.
The filtrate was acidified with concentrated HCl to give a precipi-
tate of the methylester-acid. The crystals were collected by vac-
uum filtration, rinsed with water, and dried at 95 °C.
Step 2-A mixture of 3-(4-carboxybenzylsulfamoyl)thiophene-2-
carboxylic acid methyl ester (0.7 g), potassium hydroxide (1.0 g),
and water (10 mL) was heated at 85 °C for 40 min. After cooling
to room temperature an additional 5 mL of water was added. The
product was precipitated by the drop wise addition of concen-
trated hydrochloric acid. After drying at 95 °C, 3-(4-carbo-
xybenzylsulfamoyl)thiophene-2-carboxylic acid was obtained in
4.2. Synthesis
The mono methyl ester intermediates were synthesized by acyl-
ation of the amino acids with 2-carboxy-3-chlorosulfonyl thio-
phene. 4-Aminomethylbenzoic acid and 4-aminoethyl benzoic
acid were purchased from Sigma–Aldrich. 4-Aminopropylbenzoic
acid and 4-aminobutylbenzoic acid were prepared by published
procedures. Saponification of the methyl esters afforded the dia-
cids which were characterized by melting point, IR, 1H NMR, 13C
NMR, ESI-MS, and elemental analysis.
Please cite this article in press as: Hendershot, J. M.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.051

8
J. M. Hendershot et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
Figure 5. Overlay of the AmpC complexes with compounds 4, 5, and 6. Carbon atoms of the protein are colored green and of 4 are cyan, 5 are magenta and 6 are orange.
Nitrogen atoms are blue, oxygen atoms red and sulfur atoms yellow.
37% yield. Mp 184–187 °C. 1H NMR: 7.9(d, J = 5.2 Hz, 1H); 7.8(d,
J = 8.4 Hz, 2H); 7.4(d, J = 5.2 Hz, 1H); 7.3(d, J = 8.4 Hz, 2H), and
4.3(d, J = 6.6 Hz, 2H). 13C NMR: 169, 163, 147, 134, 134, 134, 130,
128, 127, and 46. ESI-MS: 342 (M+1, positive mode). Anal. Calcd
for C13H11NO6S2: C, 45.74; H, 3.25; N, 4.10. Found: C 45.47; H
3.24; N 4.18.
384 (M+1, positive mode). Anal. Calcd for C16H17NO6S2: C 50.12; H
4.47; N 3.65. Found: C 50.39; H 4.45; N 3.72.
4.3. Enzymology
AmpC was expressed and purified as previously described.7
Kinetic measurements were performed using nitrocefin as sub-
strate in 50 mM Tris, pH 7.0, and monitored on an Agilent HP
8453 UV–vis spectrophotometer at 482 nm. The Km of nitrocefin
4.2.2. 3-(4-Carboxyphenylethylsulfamoyl)thiophene-2-
carboxylic acid (Compound 5)
3-(4-Carboxyphenylethylsulfamoyl)thiophene-2-carboxylic
acid methyl ester underwent saponification with potassium
hydroxide and water. The product was precipitated by the drop
wise addition of concentrated hydrochloric acid. After drying, 3-
for AmpC in this buffer was determined to be 127
tration of enzyme in all reactions was 1.25 nM. IC50 values were
determined at a substrate concentration of 200 M. Ki values were
obtained by comparison of progress curves in the presence and
absence of inhibitor and assumed competitive inhibition.12,24 Suf-
ficient inhibitor was used to give at least 50% inhibition. Inhibitors
were dissolved in DMSO at a concentration of 50 mM; dilute stock
solutions were subsequently prepared as necessary, so that the
concentration of DMSO in assays was 5% or less. Dilution of com-
pound 7 from the stock solution into assay buffer resulted in pre-
lM. The concen-
l
(4-carboxyphenylethyl-sulfamoyl)thiophene-2-carboxylic
acid
was obtained in 55% yield. Mp 209–210 °C. 1H NMR: 7.9(d,
J = 5.2 Hz, 1H); 7.8(d, J = 8.4 Hz, 1H); 7.4(d, J = 5.2 Hz, 1H); 7.2(d,
J = 8.4 Hz, 1H), 3.2(m, 2H), and 2.8(m, 2H). 13C NMR: 169, 163,
145,134, 134, 134, 130, 128, 128, 127, 43, and 34. ESI-MS: 356
(M+1, positive mode). Anal. Calcd for C14H13NO6S2: C 47.31; H
3.69; N 3.94. Found: C 47.44; H 3.78; N 3.95.
cipitation of 7. Preparation of stock solutions of
7
in
dimethylformamide (DMF), or N-methylpyrrolidone (NMP) was
also attempted, but 7 still precipitated when added to assay buffer.
Therefore, the inhibition of AmpC by this compound was not mea-
sured. For compound 5, the Ki value was also determined by
Lineweaver-Burk analysis of multiple substrate (50, 100, 200, and
4.2.3. 3-(4-Carboxyphenylpropylsulfamoyl)thiophene-2-
carboxylic acid (Compound 6)
3-(4-Carboxyphenylpropylsulfamoyl)thiophene-2-carboxylic
acid methyl ester underwent saponification as described above in
63% yield. Mp 190–193 °C. 1H NMR: 7.8(d, J = 5.2 Hz, 1H); 7.8(d,
J = 8.3 Hz, 2H); 7.4(d, J = 5.2 Hz, 1H); 7.2(d, J = 8.3 Hz, 2H), 2.9(t,
J = 7.7 Hz, 2H), 2.6(t, J = 7.7 Hz, 2H), and 1.7(m, 2H). 13C NMR:
169, 163, 147, 134, 134, 134, 130, 128, 128, 127, 42, 30, and 28.
ESI-MS: 370 (M+1, positive mode). Anal. Calcd for C15H15NO6S2:
C 48.77; H 4.09; N 3.79. Found: C 48.66; H 4.06; N 3.84.
285 lM) and inhibitor (0, 20, 30, 50 lM) concentrations. Linear
regression was performed in Microsoft Excel. The value obtained
in this manner is consistent with that obtained by using the IC50
value.
4.4. Crystal growth and structure determination
4.2.4. 3-(4-Carboxyphenylbutylsulfamoyl)thiophene-2-
carboxylic acid (Compound 7)
Co-crystals of AmpC with each of the inhibitors were grown by
vapor diffusion in hanging drops over 1.7 M potassium phosphate,
pH 8.7, at room temperature using microseeding. Crystals
appeared between 3–5 days after equilibration at room tempera-
ture. The concentration of AmpC was 3–4 mg/mL, and the concen-
tration of the inhibitors was 1 mM. Crystals were harvested with a
55% yield. Mp 155–157 °C. 1H NMR: 7.9(d, J = 5.2 Hz, 1H); 7.8(d,
J = 8.4 Hz, 1H); 7.4(d, J = 5.2 Hz, 1H); 7.3(d, J = 8.4 Hz, 1H), 2.9 (m,
2H), 2.5(m, 2H), 1.4(m, 2H), and 1.31(m, 2H). 13C NMR: 169, 163,
148, 134, 134, 134, 130, 128, 128, 127, 43, 35, 29, and 28. ESI-MS:
Please cite this article in press as: Hendershot, J. M.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.051

J. M. Hendershot et al. / Bioorg. Med. Chem. xxx (2014) xxx–xxx
9
Table 2
Crystallographic statistics for complexes of AmpC with compounds 4, 5, and 6
AmpC/4
AmpC/5
AmpC/6
Cell constants (Å, °)
Space group
Resolution (Å)
Unique reflections
Total observations
Rmerge (%)
a = 118.4, b = 76.7, c = 97.6; b = 115.6
a = 118.6, b = 77.2, c = 97.6; b = 115.4
C2
50–1.76 (1.82–1.76)
70,212
239,983
4.6 (21.3)
88.7 (86.4)
23.2
a = 118.9, b = 77.7, c = 98.4; b = 115.4
C2
50–2.3 (2.38–2.30)
34,262
127,338
6.7 (36.6)
94.7 (99.2)
16.6
C2
50–1.90 (1.97–1.90)a
62,237
230,181
9.2 (36.8)
99.9 (100)
10.2
Completeness (%)
<I/rI>
Rcryst, Rfree (%)
16.9, 21.2
20.5 (24.4)
18.1, 25.5
Average B factors
Protein atoms (Å2)
Inhibitor atoms (Å2)
Waters (Å2)
24.3
39.8b
34.3
31.6
53.7c
37.1
46.8
62.8d
47.1
a
b
c
Values in parentheses are for the highest resolution shell.
Calculated for all 22 atoms of inhibitor present in B monomer only.
Calculated for 69 inhibitor atoms, which is bound in a single conformation in the A monomer and in two conformations in the B monomer.
Calculated for 48 atoms of the inhibitor.
d
nylon loop and immersed in a cryoprotectant solution containing
1.7 M potassium phosphate, pH 8.7, 20% sucrose for approximately
30 s and then flash cooled in liquid nitrogen. Data were measured
from single crystals on the LS-CAT and DND-CAT beam lines (21-
ID-D and 5-ID-B) of the Advanced Photon Source at Argonne
National Lab at 100 K using a Mar CCD detector.
We thank Jim O’Keefe at the Annis Water Resources Institute for
performing the mass spectrometry analyses.
References and notes
1. Fisher, J. F.; Meroueh, S. O.; Mobashery, S. Chem. Rev. 2005, 105, 395.
2. Neu, H. C. Science 1992, 1064, 257.
Reflections were indexed, integrated, and scaled using
HKL200025 (Table 2). The space group was C2, with two AmpC mol-
ecules in the asymmetric unit. The structures of the complexes
were determined using a native apo AmpC structure (PDB entry
1KE4, with water molecules and ions removed), as the initial phas-
ing model. The structure was refined using REFMAC515 in the CCP4
program suite.16 Electron density maps were calculated with REF-
MAC and used to build the protein model in the program Coot.14
The inhibitors were each built into the initial observed difference
density in each active site of the asymmetric unit, and the struc-
tures of the complexes were further refined using REFMAC
(Table 2). In the final model, each molecule contained 358 residues,
with one exception in the A monomer of AmpC/4, which only con-
tained 352 residues. The structures have been deposited with the
PDB as 4JXS for AmpC/4, 4JXV for AmpC/5, and 4JXW for AmpC/6.
3. Davies, J.; Davies, D. Microbiol. Mol. Biol. Rev. 2010, 74, 417.
4. Bush, K. Ciba Found. Symp. 1997, 207, 152.
5. Sutherland, R. Trends Pharmacol. Sci. 1991, 12, 227.
6. Powers, R. A.; Morandi, F.; Shoichet, B. K. Structure 2002, 1013, 10.
7. Usher, K. C.; Blaszczak, L. C.; Weston, G. S.; Shoichet, B. K.; Remington, S. J.
Biochemistry 1998, 37, 16082.
8. Tondi, D.; Morandi, F.; Bonnet, R.; Costi, M. P.; Shoichet, B. K. J. Am. Chem. Soc.
2005, 127, 4632.
9. Gibson, M. S. ; Bradshaw, R. W. Angew. Chem., Int. Ed. Engl. 1968, 7, 919.
10. Heaney, H. Compr. Org. Synth. 1991, 2, 733.
11. Sonntag, N. O. V. Chem. Rev. 1953, 52, 237.
12. Waley, S. G. Biochem. J. 1982, 205, 631.
13. Chen, V. B.; Arendall, W. B., 3rd; Headd, J. J.; Keedy, D. A.; Immormino, R. M.;
Kapral, G. J.; Murray, L. W.; Richardson, J. S.; Richardson, D. C. Acta Crystallogr. D
Biol. Crystallogr. 2010, 66, 12.
14. Emsley, P.; Cowtan, K. Acta Crystallogr. D Biol. Crystallogr. 2004, 60, 2126.
15. Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta Crystallogr. D Biol. Crystallogr.
1997, 53, 240.
16. Collaborative Computational Project, Number
4 Acta Crystallogr. D Biol.
Crystallogr. 1994, 50, 760.
17. Bodill, T.; Conibear, A. C.; Mutorwa, M. K.; Goble, J. L.; Blatch, G. L.; Lobb, K. A.;
Klein, R.; Kaye, P. T. Bioorg. Med. Chem. 2013, 21, 4332.
18. LaFrate, A. L.; Carlson, K. E.; Katzenellenbogen, J. A. Bioorg. Med. Chem. 2009, 17,
3528.
19. Burley, S. K.; Petsko, G. A. Science 1985, 229, 23.
20. Powers, R. A.; Blazquez, J.; Weston, G. S.; Morosini, M. I.; Baquero, F.; Shoichet,
B. K. Protein Sci. 1999, 8, 2330.
21. Powers, R. A.; Shoichet, B. K. J. Med. Chem. 2002, 45, 3222.
22. Tondi, D.; Powers, R. A.; Caselli, E.; Negri, M. C.; Blazquez, J.; Costi, M. P.;
Shoichet, B. K. Chem. Biol. 2001, 8, 593.
23. Weston, G. S.; Blazquez, J.; Baquero, F.; Shoichet, B. K. J. Med. Chem. 1998, 41,
4577.
Acknowledgments
The authors would like to thank LS-CAT and DND-CAT at the
Advanced Photon Source (APS) for the use of synchrotron radiation.
Use of the APS was supported by the U.S. Department of Energy,
Office of Science, Office of Basic Energy Sciences, under Contract
No. DE-AC02-06CH11357. Use of the LS-CAT Sector 21 was sup-
ported by the Michigan Economic Development Corporation and
the Michigan Technology Tri-Corridor for the support of this
research program (Grant 085P1000817). We thank the NSF-
Advance program for a professional development grant (to RAP).
24. Chou, T. C. Mol. Pharmacol. 1974, 10, 235.
25. Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307.
26. DeLano, W. L.; Schrodinger LLC, 2010.
Please cite this article in press as: Hendershot, J. M.; et al. Bioorg. Med. Chem. (2014), http://dx.doi.org/10.1016/j.bmc.2014.04.051