Screening and Design of Inhibitor Scaffolds for the Antibiotic Resistance Oxacillinase-48 (OXA-48) through Surface Plasmon Resonance Screening

Article abstract of DOI:10.1021/acs.jmedchem.6b00660

The spread of antibiotic resistant bacteria is a global threat that shakes the foundations of modern healthcare. β-Lactamases are enzymes that confer resistance to β-lactam antibiotics in bacteria, and there is a critical need for new inhibitors of these

Full text of DOI:10.1021/acs.jmedchem.6b00660

Subscriber access provided by - Access paid by the | UCSB Libraries
Article
Screening and design of inhibitor scaffolds for the antibiotic resistance
oxacillinase-48 (OXA-48) through surface plasmon resonance screening
Bjarte Aarmo Lund, Tony Christopeit, Yngve Guttormsen, Annette Bayer, and Hanna-Kirsti S. Leiros
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00660 • Publication Date (Web): 11 May 2016
Downloaded from http://pubs.acs.org on May 16, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Screening and design of inhibitor scaffolds for the
antibiotic resistance oxacillinaseꢀ48 (OXAꢀ48)
through surface plasmon resonance screening
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Bjarte Aarmo Lund1, Tony Christopeit1, Yngve Guttormsen2, Annette Bayer2 and HannaꢀKirsti S.
Leiros1*
1The Norwegian Structural Biology Centre (NorStruct), Department of Chemistry, UiT The
Arctic University of Norway, Tromsø, Norway
2Department of Chemistry, UiT The Arctic University of Norway, Tromsø, Norway
ABSTRACT: The spread of antibiotic resistant bacteria is a global threat that shakes the
foundations of modern healthcare. βꢀLactamases are enzymes that confer resistance to βꢀ
lactam antibiotics in bacteria and there is a critical need for new inhibitors of these enzymes
for combination therapy together with an antibiotic. With this in mind, we have screened a
library of 490 fragments to identify starting points for the development of new inhibitors of
the class D βꢀlactamase oxacillinaseꢀ48 (OXAꢀ48), through Surface Plasmon Resonance
(SPR), doseꢀrate inhibition assays and Xꢀray crystallography. Furthermore, we have
uncovered structureꢀactivity relationships and used alternate conformations from a
crystallographic structure to grow a fragment into a more potent compound with a KD of 50
ꢀM and an IC50 of 18 ꢀM.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 51
1
2
3
4
Introduction
5
Each year infections with antibiotic resistant bacteria cause an estimated 25 000 deaths in
Europe.1 Thus there is a pressing need for new strategies to treat such infections. Today, the
most commonly used antibiotics are βꢀlactam antibiotics such as penicillins, cephalosporins
and carbapenems. However, there have not been any major developments since the
introduction of carbapenems in 1976,2 and many bacteria have become resistant to virtually
all βꢀlactam antibiotics. A major resistance mechanism to βꢀlactam antibiotics is the
expression of βꢀlactamases, enzymes able to hydrolyze the βꢀlactam ring, which renders the
drug inactive.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A prevalent βꢀlactamase which has caused outbreaks of antibiotic resistant bacteria all over
the world is OXAꢀ48, an oxacillinase belonging to the Ambler class D of βꢀlactamases.3,4 βꢀ
Lactamases belonging to this group have a serine in the active site similarly to the class A
and C βꢀlactamases,5 but show distinct properties such as dimerization6 and a carboxylated
lysine in the active site.7 The carboxylated lysine is believed to play an important role in the
βꢀlactam hydrolyze by acting as the general base responsible for activating the nucleophilic
hydroxylꢀgroup of the catalytic serine as well as activating the water responsible for
deacetylating the complex.8 The hydrophobic environment in the interior of the protein
stabilizes the carboxylated lysine, but the pH has to be close to neutral or higher for the
lysine to be nonꢀprotonated and amenable to carboxylation.7 Without the carboxylation of the
lysine there is a significant decrease in activity, and the decarboxylated enzyme appears to be
inhibited by chloride ions.7 OXAꢀ48 is able to hydrolyze a wide range of antibiotics like
penicillins and carbapenems, but is less efficient against expandedꢀspectrum cephalosporins.3
The coꢀadministration of βꢀlactamase inhibitors together with βꢀlactam antibiotics has been a
successful strategy to overcome resistance caused by class A and C βꢀlactamases.9 The recent
ACS Paragon Plus Environment

Page 3 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
approval of avibactam, a non βꢀlactam βꢀlactamase inhibitor binding covalently to the active
site serine, showed that this strategy is also successful for class D βꢀlactamases. 10–13
However, reports of avibactamꢀresistance for class A and C βꢀlactamases have been
published14–16 and even though no avibactamꢀresistance for OXAꢀ48 has been reported yet,
there is a clear need for novel βꢀlactamase inhibitors as additional treatment method when
resistance arises.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Fragment based drug discovery (FBDD) has been a successful approach in generating drug
leads for several now marketed drugs.17 In FBDD, a set of soꢀcalled fragments, small
molecules with favorable physicochemical properties and a molecular weight below 300 Da,
is screened. Generally, the affinity of fragment hits is low compared to hits from traditional
highꢀthroughput screenings. Hence, fragment hits have to be further optimized into potent
inhibitors by linking, evolving or growing.18 An additional advantage of FBDD is that
derivatives and analogous compounds are often commercially available, enabling rapid
determination of structureꢀactivity relationships.19
Surface plasmon resonance (SPR)ꢀbased biosensor assays allow direct measurements of
proteinꢀligand interactions. Furthermore, they are a useful tool to characterize such
interactions by determining the kinetic parameters and the affinity.19 SPRꢀbased methods are
also easily automated and the consumption of both protein and ligand is low.20 Hence, SPR
based assay are widely used in drug discovery to screen fragment libraries. A common
challenge in FBDD are PanꢀAssay Interference compounds (PAINS) that causes interference
across different targets while being unlikely to progress into a potent and selective
compound.21 SPRꢀbased assays are suitable to identify promiscuous binders and hence can
avoid problems caused by PAINS. By using an orthogonal screening strategy, such as
combining a biochemical assay together with an SPR assay, it is possible to identify likely
PAINS.20,22,23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 51
1
2
3
4
5
6
7
8
9
In this study, we report the results of screening a library of 490 fragments using SPR as the
primary assay. Fragment hits from this primary screen were confirmed in a secondary
biochemical enzymatic assay using the substrate nitrocefin. Furthermore, for several of the
identified fragments, the crystal structure in complex with OXAꢀ48 was determined. One of
the fragments was further optimized into an inhibitor with increased potency.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Results and discussion
Cloning, protein expression and purification of HisꢀTEV OXAꢀ48
Two different OXAꢀ48 gene constructs were used throughout this study. The first construct,
referred to as nOXAꢀ48, was expressed with the native signal sequence and purified from the
periplasm as previously described.24
In order to increase the yield of OXAꢀ48, a second gene construct was designed and
expressed, and is henceforth referred to as tOXAꢀ48. In this construct, the signal peptide
(residues 1ꢀ22) as predicted by SignalP25 was removed and a Hisꢀtag and a Tobacco Etch
Virus (TEV) cleavage site were added. The tOXA48 construct was cloned and expressed in
Escherichia coli. The protein was purified using the Hisꢀtag. The incorporation of a TEV
cleavage site allowed for removal of the Hisꢀtag by TEV protease, leaving only one
additional alanine at the Nꢀterminus of tOXAꢀ48. A second purification step separated
cleaved protein from the uncleaved protein, followed by polishing by cation exchange
chromatography. As judged by SDS PAGE analysis the protein was purified to homogeneity.
Expression of tOXAꢀ48 yielded approximately 10 mg of protein per liter of culture.
Furthermore, the use of autoinductionꢀmedia reduced the handsꢀon time during expression
since there was no need to monitor the optical density prior to induction.
ACS Paragon Plus Environment

Page 5 of 51
Journal of Medicinal Chemistry
1
2
3
4
Enzyme characterization and DMSO tolerance of nOXAꢀ48
5
6
7
8
Analysis of the kinetic parameters of nOXAꢀ48 with five βꢀlactam substrates and two
reporter substrates (Table 1) show that nOXAꢀ48 is able to hydrolyze carbapenems and
penicillins. There has been several reports with steadyꢀstate kinetics for OXAꢀ48,26,27 and our
results mirror the overall trends. The penicillins are quickly hydrolyzed, while the
cephalosporins cefepime and CENTA are poorly hydrolyzed. nOXAꢀ48 is less active against
the carbapenems than against the penicillins. An exception is that nOXAꢀ48 is very active
against the cephalosporin substrate nitrocefin. However, nitrocefin is known to be easily
hydrolysable due to its conjugated ringsystems.28 For nOXAꢀ48, the kinetic parameters were
determined for the hydrolyzation of nitrocefin. The determined kcat and Km were similar to
the values determined for tOXAꢀ48, indicating that there were no significant differences in
enzyme activity for the two different gene constructs nOXAꢀ48 and tOXAꢀ48.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nitrocefin and CENTA are common reporter substrates to monitor the activity of β ꢀ
lactamases. As nitrocefin is expensive and not stable in aqueous solutions, CENTA has been
recommended for its favorable chemical properties and lower price.29 However, our
comparison (Table 1) shows that nOXAꢀ48 is most efficient against nitrocefin, and that
higher concentration of the CENTA substrate were necessary for detection, limiting the
sensitivity of the assay. Hence, it was decided to use inhibition assay based on nitrocefin for
the following fragment screening.
Since all fragments were solved in DMSO, it was necessary to investigate the influence of
DMSO on the activity of nOXAꢀ48. In our hands, nOXAꢀ48 was tolerated up to 2.5% DMSO
without significant loss in activity towards the hydrolysis of nitrocefin (data not shown).
Dilutionꢀseries for the βꢀlactamase inhibitor avibactam showed 50% inhibition at 1.5 µM
after 5 minutes of incubation. Avibactam is known to bind covalently,10 and covalent
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 51
1
2
3
4
5
6
inhibitors are timeꢀdependent in their doseꢀresponse, however the results do validate that the
assay is sensitive to inhibitor measurements and can be used for the fragment screening.
7
8
9
Establishing SPR assay and the influence of bicarbonate
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To our knowledge, we present the first SPRꢀexperiments for OXAꢀ48. Since SPR allows for
both measurements of affinity and determination of kinetic parameters, it is a very useful
technique for drug design. The tOXAꢀ48 protein was immobilized to the senor surface by
amine coupling resulting in a stable surface. Although immobilization levels up to 10000 RU
could be reached, only approximately 5000 RU were immobilized to reduce bulk effects and
avoid mass transport limitations.30
To verify the enzyme integrity on the chip, the interaction with the substrate meropenem was
investigated (Figure 1). Meropenem clearly bound to the enzyme showing that the enzyme
retains activity after immobilization. However, the substrate release was very slow. Based on
the report of oxacillinases dependence on bicarbonate7 we hypothesized that the absence of
bicarbonate in the running buffer leads to a decarboxylation of Lys73 causing the slow
dissociation of the substrate. By including 50 mM bicarbonate (NaHCO3) in the running
buffer the rate of disassociation increased dramatically as shown in Figure 1. This indicates
that adding bicarbonate to the buffer is essential to maintain the enzyme active through a
carboxylated Lys73.
With bicarbonate included in the running buffer, we were able to determine the affinity and
kinetic parameters for meropenem using a 1:1 binding model Figure 1 yielding a ka of 4 ×
104 Mꢀ1sꢀ1, a kd of 0.3463 sꢀ1 and KD of 0.85 µM. This KD value compares well with the Km of
1.0 µM for meropenem found in the characterization of nOXAꢀ48 (Table 1). Thus, the
similar KD and Km determined by the two different methods supporting that the protein
coupled to the CM5 SPR chip in the SPR experiment is active.
ACS Paragon Plus Environment

Page 7 of 51
Journal of Medicinal Chemistry
1
2
3
4
Orthogonal fragment screening with a primary biophysical screening and a
5
6
secondary biochemical screening
7
8
9
The screening strategy was designed to minimize the risk of false positives by using two
orthogonal assays. An SPR based assay was used as primary screen to identify binders from
our 490 compound subset from the Maybridge Ro3 collection and the secondary screen
measured the inhibition potency against the substrate nitrocefin. All the fragments adhere to
the “rule of three”.31
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The primary screening at a single concentration (300 ꢀM) identified 84 fragments, with
binding to the protein. These 84 fragments were analyzed in a dilutionꢀseries (118 – 900
µM). Among these, 77 compounds showed concentrationꢀdependent and reversible binding
to the tOXAꢀ48 surface. The other fragments exhibited very slow disassociation, no clear
concentration dependency or superstoichometric binding which are common signs of
unspecific binding19,22 and were therefore rejected.
The SPR based assay only identifies binders and in order to determine whether the binders
were inhibitors, we performed a secondary screening for all the 77 binders using a
biochemical inhibition assay against nitrocefin using nOXAꢀ48. We were able to determine
IC50 values (Figure 3) for 10 of the fragments. The remaining compounds did not have the
inhibitory potency necessary to give reliable IC50ꢀvalues in our assay, and were not
considered good starting points for further inhibitor optimization.
The KD values were calculated based on the SPR results for 8 of the fragments with highest
potency (Figure 2). For compound 4 and 5, the steady state values showed a linear
concentration dependency and hence, it was not possible to determine the KD. The linear
concentration dependency is most likely caused by a combination of specific and nonꢀspecifc
interaction with the protein. However, the signal intensities were in the range expected for a
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 51
1
2
3
4
5
6
7
8
9
1:1 interaction based on Equation 1 and therefore the compounds were not rejected. Some
fragments may also bind to other sites on the enzyme without influencing the activity, which
could explain the higher number of compounds binding in the SPR experiments without the
corresponding inhibition effect in the competitive nitrocefin experiment.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀꢁꢂꢃ
=
ꢄꢅꢂꢆꢁꢇꢈꢉ ꢊꢋ × ꢎꢏꢏꢐꢑꢎꢒꢎꢓꢔꢕꢎꢐꢖ ꢒꢗꢘꢗꢒ
Equation 1
ꢇꢈꢌꢍꢁꢇ ꢊꢋ
Overall there is a good agreement between the affinities (KD) measured by SPR and the IC50
values determined from the biochemical assay (Table 2), except for compounds 8 and 9
which had higher IC50 values than expected from the KD values. There were differences in the
buffers that may explain the differences. We kept the buffer for the biochemical assay
identical to the kinetics buffer, except for the addition of 2.5% DMSO. Whereas for the SPR
assay, we prioritized having the conditions similar to the crystallization conditions to
increase the probability of obtaining Xꢀray crystallographic complexꢀstructures of hits.
Ligand efficiency (LE) is a simple parameter for hit assessment, which scales the affinity
with the number of nonꢀhydrogen atoms to overcome the bias towards ligands of high
molecular weight.32 Overall, the identified hits had good affinities (Table 2), and the
calculated LEꢀvalues are within the range of what is expected of fragmentꢀhits.33
In summary, the ten fragments are all interesting starting points for further inhibitor
development after the screening.
Xꢀray structures of tOXAꢀ48 with compound 1, 2 and 3 and structure activity
relationships
Atomic resolution details from the inhibition mode of tOXAꢀ48 were found by soaking apo
tOXAꢀ48 crystals with the 10 fragment hits. By this procedure, we obtained crystal structures
ACS Paragon Plus Environment

Page 9 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of tOXAꢀ48 in complex with 3 of the fragments. All structures have four molecules in the
asymmetric unit, but crystalized in two different space groups (P212121 and P21). For the new
tOXAꢀ48 structures the RMSD was below 0.4 Å compared to the published apo structure of
nOXAꢀ48 (PDB ID: 3HBR) as calculated by the protein structure comparison service
PDBeFold at EBI (http://www.ebi.ac.uk/msdꢀsrv/ssm).34 The low RMSD between our tOXAꢀ
48 and the apo nOXAꢀ4827 indicate very similar structures, and that our structures equally
well describe the OXAꢀ48 drug target. Refinement statistics are summarized in Table 4. Our
new tOXAꢀ48 structures all have carboxylated Lys73 residues (Kcx73) similar to e.g. nOXAꢀ
48 (PDB ID: 3HBR) as one would expect with the neutral pH.27
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
All the compounds are anchored near the active site Ser70, which explains the competitive
inhibition mode of these fragments.
From the crystal structure complexes of tOXAꢀ48 and compounds 1, 2 and 3, it is clear that
the carboxylate groups have ionic interactions with the sidechain of Arg250 and a charge
induced hydrogen bond to sidechain oxygen of Thr209. This is consistent with the role
Arg250 has in substrate binding,35,36 and similar to the interaction between the negatively
charged sulfate group of avibactam to Arg250 in nOXAꢀ48 (PDB ID: 4WMC) and the
equivalent Arg261 in OXAꢀ24 (PDB ID:4WM9).12,37 All the ten hits (compounds 1ꢀ10) are
carboxylates, and since the SPR and IC50 experiments are performed at pH 7.0, the
compounds are all likely to be negatively charged. The removal of the carboxylate group
abolished binding, as seen for compound 11, which had no activity. Compound 14, which has
a ketone functionality instead of the carboxylate group, also showed weak inhibition (Table
3). This shows that the carboxyl group forms interactions crucial for the fragment binding.
Compound 1, 2 and 3 had the highest affinity. These fragments are structurally related to
each other with 3ꢀsubstituted benzoic acids as a common feature. Compounds 4 and 10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 51
1
2
3
4
resemble compounds 1ꢀ3, but contain an additional linker, which appears to be detrimental to
5
6
the inhibition.
7
8
9
Compound 1 bound to tOXAꢀ48 (tOXAꢀ48:1; Figure 4A) was observed in multiple
conformations. The conformation shown in Figure 4B is facing out of the activeꢀsite and has
hydrophobic interactions with Trp105, Thr209, Gly210, Tyr211 and Leu247. The
carboxylateꢀgroup makes ionic bonds to Arg250 (d(Oligand…NR250) = 2.5ꢀ3.1Å). The
pyridineꢀnitrogen also makes a hydrogenꢀbond to a waterꢀmolecule (d(Oligand…Owater) = 2.5
Å). The conformation shown in Figure 4C is more buried, and has hydrophobic interactions
with Val120, Leu158 and Tyr211. The carboxylate group in this conformation makes only a
single ionic bond with Arg250 (d(Oligand…NR250) = 3.4Å), but also makes a waterꢀmediated
hydrogenꢀbond to Thr209.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
However, compound 2 (tOXAꢀ48:2; Figure 4D), which only differs in the position of the N
atom of the pyridine ring, crystallized in a single conformation in the 4 protein monomers in
the asymmetric unit corresponding to the conformation in Figure 4B. The compound makes
several hydrophobic interactions with Ser70, Ile102, Ser118, Gly210 and Tyr211. The
carboxylate group forms ionic bonds with Arg250 (d(Oligand…NR250) = 3.1ꢀ3.2Å), and a
hydrogenꢀbonds with Thr209 (d(Oligand…OT209) = 3.1Å), as well as waterꢀmediated
hydrogenꢀbonds with both Arg250 and Thr209.
Compound 3 (tOXAꢀ48:3; Figure 4F) also shares the same overall conformation and has ΠꢀΠ
interactions with Tyr211 (5.3 Å) and hydrophobic interactions with Ser70, Ile102, Trp105,
Ser118, Thr209, Gly210 and Leu247. The carboxylate forms ionic bonds with Arg250
((d(Oligand…NR250) = 2.5ꢀ3.0 Å).
There is no significant difference in activity between compound 1, 2 and 3 (Table 2).
Furthermore, compound 12 (Table 3) also shows that the nitrogen on the pyridineꢀring is not
ACS Paragon Plus Environment

Page 11 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
critical for activity; while the nitrogen is useful to maintain solubility. This gives an
indication that the pyridine/thiazole ring may be modified.
7
8
9
To investigate if the pyridine/thiazole ring was necessary for binding, as it was not clear
which interactions the second ring made, compound 15 and 16 were both tested and gave
very weak inhibition (Table 3). This shows that while we could modify the pyridine/thiazole
ring without losing activity, it is not beneficial to remove them. However, the addition of an
amine group in compound 13 did not significantly alter the activity, which shows that the
second ring may have good exit vectors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compounds 5ꢀ9 contain indolꢀ or quinoline moieties as a common feature. Among these,
compound 5 and 6 have the highest potency, but the LE is higher for compounds 7, 8 and 9.
Unfortunately, no clear density could be observed for these compounds in soakingꢀ
experiments, indicating that the binding mode is too flexible to be observed by Xꢀray
crystallography.
Hitꢀoptimization by organic synthesis
In the crystal structure of compound 1 bound to tOXAꢀ48 we found two different fragment
conformations as illustrated in Figure 4A and Figure 4B. By structurally merging the two
different conformations of the pyridine ring, compound 17 was designed. Docking using
GLIDE indicated that compound was able to bind to the enzyme with higher affinity as
compared to compound 1, and decided to pursue the synthesis of compound 17 based on
these results. A Suzuki coupling of 3,5ꢀdibromobenzoic acid with 4ꢀpyridylboronic acid
provided compound 17 (Scheme 1) in excellent yield (96%) and good purity (>99%) as
determined by HPLC analysis.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 51
1
2
3
4
5
6
7
8
The SPR based assay was used to determine the KD of 50 µM (Figure 5), with a LE of 0.29
kcal per mole per heavy atom, and the enzyme inhibition assay was used to determine the
IC50 of 18 µM for compound 17.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The crystal structure of compound 17 bound to tOXAꢀ48 (tOXAꢀ48:17; Figure 5D) shows
that it overlaps with two of the conformations observed for compound 1, which validated our
approach of merging alternate conformations. Compound 17 forms hydrophobic interactions
with Ser70, Ile102, Val120, Tyr211, Ser244 and Leu247, and there is an additional waterꢀ
mediated (W3) hydrogen bond to Arg214. This waterꢀnetwork was not observed for the other
compounds, but this could be due to the lower resolution of these structures. Compound 17
forms ionic bonds to Arg250 ((d(Oligand…NR250) = 2.8ꢀ2.9 Å), and waterꢀmediated hydrogen
bonds to both Ser118 (W1) and Thr209 (W2) and with an intricate waterꢀnetwork revealed
by the high resolution of this structure.
There is still a need to investigate the possible exit vectors of compound 17, however
compound 17 shows that for this class of compounds it is possible to optimize inhibitor
propensity in terms of higher enzyme affinity and demonstrates that alternate conformations
may be used as a tool in structure guided drug design.
Based on a sequence alignment of the class D βꢀlactamase family38 it is reasonable to believe
that compound 17 would have potential to inhibit the majority of the family. Ala69, Trp105,
Val120, Leu158, Thr209 and Arg250 are all conserved (>80% of the familyꢀmembers), and
in many of the cases the amino acid substitutions might be compatible with the same
interactions.
It is interesting to compare the results of this study with the inhibitor scaffolds identified for
the Ambler class B metalloꢀβꢀlactamase VIMꢀ2 3,4 from similar experiments.39 Several of the
hits overlap and more than half of the hits identified are carboxylates. While there is no
ACS Paragon Plus Environment

Page 13 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
significant sequence identity between the different classes of βꢀlactamases, both enzymes
OXAꢀ48 and VIMꢀ2 are capable of hydrolyzing the same substrates, and could potentially be
inhibited by the same compounds. Incidentally, compound 17 also shows some activity
against VIMꢀ2 (data not shown), opening the possibility of a pan βꢀlactamase inhibitor.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Conclusions
In this study, we have done a fragment screening for new inhibitor scaffolds of the class D βꢀ
lactamase OXAꢀ48. From a total of 490 screened fragments, 10 compounds could be
confirmed by two orthogonal screens, SPR and enzyme inhibition assay, to bind and inhibit
OXAꢀ48. Three Xꢀray crystallographic structures established the binding modes of these
fragments. One compound was further optimized by merging two observed conformations in
two different protein chains in the Xꢀray structure of fragment 1 (tOXAꢀ48:1). This resulted
in a more potent inhibitor with a KD of 50 ꢀM and IC50 value 18 ꢀM compared to 280 ꢀM
(KD) and 250 ꢀM (IC50) respectively for compound 1, thus validating that our reported
inhibitor scaffolds based on heterocycles with a single carboxylate group have potential for
further development into more potent inhibitors for the antibiotic resistance enzyme OXAꢀ
48.
Experimental Section
General procedures: As long as not otherwise stated chemicals were purchased from Sigmaꢀ
Aldrich, and used as received. Compound 13 were purchased from Fluorochem (UK) and
compound 14 from Enamine (USA). All chemicals were of analytical quality. Residues are
numbered according to the apo OXAꢀ48 structure (PDB ID:3HBR).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 51
1
2
3
4
5
6
7
8
Cloning, protein expression and purification of HisꢀTEV OXAꢀ48(tOXAꢀ48)
Native fullꢀlength OXAꢀ48 with the native leader sequence was expressed and purified from
the periplasm as previously described,24 and is referred to as nOXAꢀ48.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In addition, a new HisꢀTEV OXAꢀ48 gene construct was cloned by introducing a hexaꢀ
histidine tag together with a TEV protease cleavage site while removing the predicted native
leader sequence (residues 1ꢀ22) from SignalP.25 This construct is henceforth called tOXAꢀ48.
The cloning was done similarly to the native fullꢀlength OXAꢀ48 using exponential
megapriming PCR cloning,24,40 however, the insertion of the TEV protease site required
multiꢀcomponent assembly.41 All primers (SigmaꢀAldrich, Table 5) were desalted and
dissolved to 100 µM in nucleaseꢀfree water.
The new tOXAꢀ48 construct was transformed into in house modified E. coli strain BL21 Star
(DE3) pRARE and expressed in the autoinduction media ZYPꢀ505242 with ampicillin (100
mg/L) and chloramphenicol (34 mg/L). 3ꢀ4 hours after inoculation at 37 °C, the temperature
was reduced to 25 °C to allow low temperature induction overnight. Pelleted cells were
frozen, later thawed and sonicated. The lysate was cleared by centrifugation (50 000×g, 4 °C,
30 min), then loaded on a 1 mL HisTrap FF crude column equilibrated with 25 mM HEPES
pH 6.5 and 50 mM K2SO4. The tOXAꢀ48 was eluted with a linear gradient over 30 column
volumes to 500 mM imidazole, 25 mM HEPES pH 6.5 and 50 mM K2SO4. Fractions with
tOXAꢀ48 were cleaved overnight with 1:10 ratio of inꢀhouse produced Hisꢀtagged TEV
protease to tOXAꢀ48, during dialysis against 25 mM HEPES pH 6.5, 25 mM K2SO4, 1 mM
EDTA and 2 mM 2ꢀmercaptoethanol. The dialyzed sample was loaded on a second HisTrap
column, and the flowꢀthrough examined for cleaved tOXAꢀ48, with the Hisꢀtagged TEVꢀ
protease, uncleaved protein and contaminants bound to the column. A 5 mL cation exchange
ACS Paragon Plus Environment

Page 15 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
(HiTrap SP HP) column was used for a final polishing step with a gradient from 25 mM
HEPES pH 6.5 to 25 mM HEPES pH 8.5 with 150 mM K2SO4.
7
8
9
Fragment library
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
490 fragments were purchased from the Maybridge Ro3 collection. All compounds were
dissolved in 100% DMSO to a concentration of 150 mM. These stock liquids and powders
were kept at 4 °C, while aliquot samples were kept at ꢀ20 °C. All fragments in the library had
a molecular weight between 80 and 300 Da, a cLogP ≤ 3, the number of Hꢀbond acceptors
and donors was ≤ 3 and the number of rotatable bonds was ≤ 3.
Biochemical studies: enzyme kinetics, DMSO tolerance and inhibition studies
All experiments were performed using a Spectramax M2e at 25 °C. Initial velocities were
determined in the SoftMax Pro software (Molecular Devices). All experiments were done
with a volume of 100 µL.
MichaelisꢀMenten kinetics were first established for the chromogenic βꢀlactam substrate
nitrocefin in the kinetics buffer with 100 mM Na2HPO4/NaH2PO4 (pH 7.0), 50 mM NaHCO3
and 0.2 mg/ml BSA as previously described.26 Then additional substrates were examined:
ampicillin (ꢁε235 nm = ꢀ820 Mꢀ1 cmꢀ1, 16ꢀ2000 µM, 1nM nOXAꢀ48), benzylpenicillin (ꢁε235 nm
= ꢀ775 Mꢀ1 cmꢀ1, Panpharma,176ꢀ3000 µM, 100pM nOXAꢀ48), cefepime (ꢁε260 nm = ꢀ10000
Mꢀ1 cmꢀ1, 26ꢀ450 µM, 1nM nOXAꢀ48), CENTA (ꢁε405 nm = 6400 Mꢀ1 cmꢀ1, Calbiochem, 2ꢀ
300 µM, 20 nM nOXAꢀ48), meropenem (ꢁε300 nm = ꢀ6500 Mꢀ1 cmꢀ1, 1ꢀ11 µM, 140 nM
nOXAꢀ48), nitrocefin (ꢁε482 nm 17400 Mꢀ1 cmꢀ1, Calbiochem, 10ꢀ300 µM, 1 nM nOXAꢀ48
and tOXAꢀ48) and imipenem (ꢁε300 nm = ꢀ9000 Mꢀ1 cmꢀ1, 1ꢀ100 µM, 1 nM nOXAꢀ48). The
reactions were started by the addition of enzyme. The linear phase of the reaction time course
was used to determine the steadyꢀstate velocities. The observed rate constants (ꢙꢚꢛꢜ
=
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 51
1
2
3
4
ꢘ
) as a function of the substrate concentrations were used to determine kcat and Km with
ꢞ
[ꢝ]
5
6
nonꢀlinear regression in GraphPad Prism 6 (GraphPad Software).
7
8
9
DMSO tolerance was established with dilutionꢀseries of DMSO concentrations from 0.15ꢀ
10% using 25 µM nitrocefin as the reporter substrate.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Avibactam (AstraZeneca) in dilutions from 50 nM – 20 µM were tested in a competition
experiment with 25 µM nitrocefin and 100 pM nOXAꢀ48 to validate the sensitivity of the
assay for inhibitors. Fragments identified in the primary screen by stronger signal than the
reference surface, were tested at a single concentration of 750 µM with 25 µM nitrocefin and
100 pM nOXAꢀ48. Duplicates were done, and for fragments with more than 10% deviation
triplicates were run. Avibactam and the fragments were incubated with the enzyme for
approximately 5 minutes before the experiments were started by the addition of nitrocefin.
For the inhibition testing DMSO was included in the kinetics buffer to a final concentration
of 2.5%.
Doseꢀresponse IC50 values were determined for fragments with ≥50% inhibition at 750 µM,
with fragmentꢀconcentrations from 7.3 µMꢀ1.9 mM in 9 dilutions in a competition
experiment with 25 µM nitrocefin. The log10 of the inhibitor concentrations to the response
with bottom and top constant based on controls were fitted nonꢀlinearly in GraphPad Prism 6
(GraphPad Software) to determine the IC50ꢀvalue. The Hillꢀconstant was kept constant at 1.
ACS Paragon Plus Environment

Page 17 of 51
Journal of Medicinal Chemistry
1
2
3
4
Surface plasmon resonance (SPR) interaction studies: primary screening and KD
5
6
measurements
7
8
9
All SPR experiments were performed on a Biacore T200 at 25°C. The data were analyzed
using Biacore T200 Evaluation Software 2.0 (GE Healthcare). The sensorgrams were double
reference subtracted using a reference surface and blank injections.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
tOXAꢀ48 was diluted to 25 µg/mL in 10 mM MES pH 5.5. The enzyme was immobilized to
a level of around 5000 RU on a CM5ꢀchip using standard amine coupling.
For the interaction studies with Meropenem, the final running buffer included 25 mM
HEPES pH 7.0, 50 mM K2SO4, 0.5% Tweenꢀ20, 50 mM NaHCO3 and 2.5% DMSO. The KD
was calculated from the sensorgrams recorded with 62.5 nMꢀ32 µM meropenem using the
1:1 binding kinetics model. The flow rate was 30 µL/min and with a contact time of 60 s and
240 s disassociation time.
In the primary screening, each fragment was tested twice in two independent experiments,
with reverse sample order in the second run. The fragment concentration was 300 µM with
30 s contact time and 15 s disassociation time. The fragments were diluted with running
buffer without DMSO to maintain a constant DMSO concentration of 2.5%. Flow rate was
30 µL/min and the flow system was washed with 50% DMSO between each sample.
Positive controls (meropenem) were tested every 12th cycle to ensure active enzyme together
with blanks (running buffer), and solventꢀcorrection with a DMSOꢀgradient (1.5ꢀ4%) was
performed every 48th cycle. Seven startup cycles with running buffer were performed. No
surface regeneration was done, and protein was immobilized to a new flowꢀchannel when the
ratio of the response to control surface was significantly lowered.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 51
1
2
3
4
5
6
7
8
9
For testing fragments in concentration series, they were tested in a dilutionꢀseries with six
concentrations from 118ꢀ900 µM. Seven startup cycles with buffer were performed. A
positive control (meropenem) was used for every 30th cycle and a negative control with
buffer was used for every 30th cycle. Solvent correction with varying DMSO concentration
was performed every 45th cycle. Affinities were calculated from the steadyꢀstate affinity
model with a constant Rmax adjusted by the control and molecular weight of the fragment.
The same protocol was used for the compounds 11ꢀ17, with the exception that compound 17
had 10 dilutions with concentrations from 1.8ꢀ900 µM.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Crystallization, soaking and Xꢀray data collection
Crystals of tOXAꢀ48 were obtained by vapor diffusion technique in a manual hanging drop
setup. tOXAꢀ48 at 8 mg/mL was mixed with reservoir solution containing 0.1 M HEPES pH
7.5, 10% PEG 8000 and 4ꢀ8% 1ꢀbutanol as previously published12 with a dropꢀratio of
1.5:2.5 µL for protein:reservoir.
Fragments were diluted to 3.75 mM in the cryo solution with 0.1 M HEPES pH 7.5, 10%
PEG 8000, 5% 1ꢀbutanol and 25% ethanediol and crystals were soaked for two hours (OXAꢀ
48:1) or 16ꢀ18 hours (OXAꢀ48:2, OXAꢀ48:3 OXAꢀ48:17). Then the crystals were flash
cooled in liquid nitrogen. In total 10 fragments were soaked (all from Table 2) but tOXAꢀ48
complex structures were only obtained for compound 1, 2, 3 and later with compound 17.
Xꢀray diffraction data were collected at the BL 14.1 and 14.2 beamlines at BESSY (Berlin,
Germany). The data was integrated in XDS.43 The datasets were scaled and merged in
AIMLESS.44
Molecular replacement was performed with PHASER45 with the nOXAꢀ48 apo structure
(PDB ID: 3HBR)27 as search model to obtain initial phases.
ACS Paragon Plus Environment

Page 19 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Electron density could be seen for three fragments and compound 17 after initial refinement
in PHENIX46 when automatically added waters were removed from the active site and the
difference Fourier electron density maps recalculated. Restraints for the fragments were
prepared using the GRADE webserver,47 the fragments were placed manually using COOT,48
and the final refinements were done in PHENIX. Omitꢀmaps were calculated using the
phenix.polderꢀtool.46,49 Figures were made using PyMOL50 and LigPlot+51.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Docking of compound 17
The structure of tOXAꢀ48:1 was used as a starting model. The protein preparation wizard in
Maestro was used to add hydrogens, assign bond orders, delete waters and optimize the
hydrogen bond network. A docking grid was prepared using the coordinates of compound 1
as the centroid. Compound 17 was prepared using the Ligprep tool to generate 3D
coordinates, optimize geometries and assign ionization states. Compound 17 was docked
using the induced fit protocol52 in the Schrödinger suite 2013ꢀ1 with no constrains defined
and with residues within 5 Å of the binding pose refined.
Organic synthesis of compound 17
Compound 17 was prepared in analogy to literature (Scheme 1).53 3,5ꢀDibromobenzoic acid
(228 mg, 0.81 mmol, 1 equiv), 4ꢀpyridylboronic acid (300 mg, 2.4 mmol, 3 equiv) and
potassium phosphate (1.04 g, 4.9 mmol, 6 equiv) were dissolved in dioxane (6 mL) and water
(6 mL). The solution was degassed by vacuum/Argon cycles (10 times) before addition of
PdCl2(PPh3)2 (57 mg, 81 µmol, 10 mol%) and further degassing (5 times). The resulting
mixture was stirred at 95 °C under an Argon atmosphere for 16 hours, after which HRMS in
negative mode showed only biꢀcoupled benzoic acid. The reaction mixture was filtered and
diluted with water (30 mL) before washing with chloroform (3×30 mL). The aqueous phase
was concentrated on a rotary evaporator and applied to a Biotage snaplet precolumn. On a
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 51
1
2
3
4
5
6
7
8
9
Biotage SP1 automated flash system, the compound was purified by a 60 g C18 Biotage
SNAP column with a gradient of 0ꢀ100% acetonitrile in water to yield the compound 17 as a
white solid (215 mg, 0.78 mmol, 96%). An aliquot of the solid was redissolved in ethyl
acetate and QuadraSil AP metal scavenger was added and stirred for 30 minutes. The
suspension was filtered, evaporated and subjected to HPLC purification to give the
compound in >99 % purity. 1H NMR (400 MHz, D2O) δ 8.08 (d, J = 6.2 Hz, 4H), 7.61 (d, J =
1.7 Hz, 2H), 7.02 – 6.98 (m, 5H). 13C NMR (101 MHz,) δ 172.9, 148.6, 146.7, 137.6, 136.71,
127.7, 126.4, 121.1. HRMS (ESI) m/z: [MꢀH+] calculated for C17H12O2N2 275.0826; found
275.0826.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Author Contributions
Conceived and designed the experiment: HKSL BAL TC AB. Performed the experiments:
BAL TC YG HKSL Analyzed the data: BAL TC HKSL YG AB. Wrote the paper: BAL
HKSL TC. All authors have given approval to the final version of the manuscript.
Acknowledgements
This study was supported by The National graduate school in structural biology, BioStruct,
and The Norwegian Research Council (FRIMEDBIO Project number 213808). Provision of
beam time at BL14.1 and BL14.2, Bessy II, Berlin, Germany, is highly valued. Avibactam
was a generous gift from AstraZeneca. Thanks to Ronny Helland for assisting in collection of
datasets at the beamline and Gro Elin Kjæreng Bjerga for designing the cloning experiments.
We would like to thank Ørjan Samuelsen for providing genomic DNA for the clinical isolate
with OXAꢀ48.
ACS Paragon Plus Environment

Page 21 of 51
Journal of Medicinal Chemistry
1
2
3
4
Abbreviations
5
AMP, ampicillin; CHL, chloramphenicol; TEV, Tobacco Etch Virus; NHC, Nꢀ
hydroxysuccinimide; EDC, 1ꢀethylꢀ3ꢀ(3ꢀdimethylaminopropyl)carbodiimide; SPR, surface
plasmon resonance; OXA, oxacillinase; RU, Response Units; PAINS, PanꢀAssay
Interference compounds.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
PDB Accession codes
Coordinates and structure factors of tOXAꢀ48:1, tOXAꢀ48:2, tOXAꢀ48:3 and tOXAꢀ48:17
have been deposited in the Protein Data Bank with accession numbers 5DVA, 5DTS, 5DTT
and 5DTK.
Corresponding author information
HannaꢀKirsti S. Leiros, Eꢀmail: hannaꢀkirsti.leiros@uit.no, Phone: (+47) 77 64 57 06
Supporting information Available:
HPLC analysis of compound 17, enzyme kinetic steadyꢀstate plots, SPR sensorgrams and
steadyꢀstate plots for compounds 12 and 13, and logarithmic binding affinities and inhibition
constants with standard errors.
References
(1)
ECDC/EMEA Joint Working Group. The bacterial challengeꢁ: time to react.
EMEA/576176/2009: Stockholm, Sweden 2009.
(2)
PappꢀWallace, K. M.; Endimiani, A.; Taracila, M. A.; Bonomo, R. A. Carbapenems:
past, present, and future. Antimicrob. Agents Chemother. 2011, 55, 4943–4960.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 51
1
2
3
4
5
6
7
8
9
(3)
(4)
Poirel, L.; Potron, A.; Nordmann, P. OXAꢀ48ꢀlike carbapenemases: the phantom
menace. J. Antimicrob. Chemother. 2012, 67, 1597–1606.
Mathers, A. J.; Hazen, K. C.; Carroll, J.; Yeh, A. J.; Cox, H. L.; Bonomo, R. A.; Sifri,
C. D. First clinical cases of OXAꢀ48ꢀproducing carbapenemꢀresistant Klebsiella
pneumoniae in the United States: the “menace” arrives in the new world. J. Clin.
Microbiol. 2013, 51, 680–683.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(5)
(6)
Drawz, S. M.; Bonomo, R. a. Three decades of βꢀlactamase inhibitors. Clin.
Microbiol. Rev. 2010, 23, 160–201.
Danel, F.; Paetzel, M.; Strynadka, N. C. J.; Page, M. G. P. Effect of Divalent Metal
Cations on the Dimerization of OXAꢀ10 and ꢀ14 Class D βꢀLactamases from
Pseudomonas aeruginosa. Biochemistry 2001, 40, 9412–9420.
(7)
Vercheval, L.; Bauvois, C.; di Paolo, A.; Borel, F.; Ferrer, J.ꢀL.; Sauvage, E.;
Matagne, A.; Frère, J.ꢀM.; Charlier, P.; Galleni, M.; Kerff, F. Three factors that
modulate the activity of class D βꢀlactamases and interfere with the postꢀtranslational
carboxylation of Lys70. Biochem. J. 2010, 432, 495–504.
(8)
(9)
Golemi, D.; Maveyraud, L.; Vakulenko, S.; Samama, J. P.; Mobashery, S. Critical
involvement of a carbamylated lysine in catalytic function of class D βꢀlactamases.
Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 14280–14285.
Toussaint, K. A.; Gallagher, J. C. β ꢀLactam/βꢀlactamase inhibitor combinations:
from then to now. Ann. Pharmacother. 2015, 49, 86–98.
(10) Ehmann, D. E.; Jahic, H.; Ross, P. L.; Gu, R.ꢀF.; Hu, J.; DurandꢀRéville, T. F.; Lahiri,
S.; Thresher, J.; Livchak, S.; Gao, N.; Palmer, T.; Walkup, G. K.; Fisher, S. L.
ACS Paragon Plus Environment

Page 23 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
Kinetics of avibactam inhibition against Class A, C, and D βꢀlactamases. J. Biol.
Chem. 2013, 288, 27960–27971.
7
8
9
(11) Garber, K. A βꢀlactamase inhibitor revival provides new hope for old antibiotics. Nat.
Rev. Drug Discovery 2015, 14, 445–447.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(12) Lahiri, S. D.; Mangani, S.; Jahić, H.; Benvenuti, M.; DurandꢀReville, T. F.; De Luca,
F.; Ehmann, D. E.; Rossolini, G. M.; Alm, R. A.; Docquier, J.ꢀD.; Jahic, H.;
Benvenuti, M.; DurandꢀReville, T. F.; De Luca, F.; Ehmann, D. E.; Rossolini, G. M.;
Alm, R. A.; Docquier, J.ꢀD. Molecular basis of selective inhibition and slow
reversibility of Avibactam against class D carbapenemases: a structureꢀguided study
of OXAꢀ24 and OXAꢀ48. ACS Chem. Biol. 2014, 10, 591–600.
(13) Drawz, S. M.; PappꢀWallace, K. M.; Bonomo, R. a. New βꢀlactamase inhibitors: a
therapeutic renaissance in an MDR world. Antimicrob. Agents Chemother. 2014, 58,
1835–1846.
(14) Humphries, R. M.; Yang, S.; Hemarajata, P.; Ward, K. W.; Hindler, J. A.; Miller, S.
A.; Gregson, A. First report of ceftazidimeꢀavibactam resistance in a KPCꢀ3
expressing Klebsiella pneumoniae. Antimicrob. Agents Chemother. 2015, 59, 6605–
6607.
(15) PappꢀWallace, K. M.; Winkler, M. L.; Taracila, M. A.; Bonomo, R. A. Variants of βꢀ
lactamase KPCꢀ2 that are resistant to inhibition by Avibactam. Antimicrob. Agents
Chemother. 2015, 59, 3710–3717.
(16) Winkler, M. L.; PappꢀWallace, K. M.; Taracila, M. A.; Bonomo, R. A. Avibactam
and inhibitorꢀresistant SHV βꢀlactamases. Antimicrob. Agents Chemother. 2015, 59,
3700–3709.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 51
1
2
3
4
(17) Congreve, M.; Chessari, G.; Tisi, D.; Woodhead, A. J. Recent developments in
5
6
fragmentꢀbased drug discovery. J. Med. Chem. 2008, 51, 3661–3680.
7
8
9
(18) Roughley, S. D.; Hubbard, R. E. How well can fragments explore accessed chemical
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
space? A case study from heat shock protein 90. J. Med. Chem. 2011, 54, 3989–4005.
(19) Danielson, U. Fragment library screening and lead characterization using SPR
biosensors. Curr. Top. Med. Chem. 2009, 9, 1725–1735.
(20) Navratilova, I.; Hopkins, A. L. Emerging role of surface plasmon resonance in
fragmentꢀbased drug discovery. Future Med. Chem. 2011, 3, 1809–1820.
(21) Baell, J. B.; Holloway, G. A. New substructure filters for removal of pan assay
interference compounds (PAINS) from screening libraries and for their exclusion in
bioassays. J. Med. Chem. 2010, 53, 2719–2740.
(22) Giannetti, A. M.; Koch, B. D.; Browner, M. F. Surface plasmon resonance based
assay for the detection and characterization of promiscuous inhibitors. 2008, 4, 574–
580.
(23) Giannetti, A. M. FROM EXPERIMENTAL DESIGN TO VALIDATED HITS A
COMPREHENSIVE WALKꢀTHROUGH OF FRAGMENT LEAD
IDENTIFICATION USING SURFACE PLASMON RESONANCE. 1st ed.; Elsevier
Inc.: Cambridge, Massachusetts, 2011; Vol. 493.
(24) Lund, B. A.; Leiros, H.ꢀK. S.; Bjerga, G. E. A highꢀthroughput, restrictionꢀfree
cloning and screening strategy based on ccdBꢀgene replacement. Microb. Cell Fact.
2014, 13, 38.
ACS Paragon Plus Environment

Page 25 of 51
Journal of Medicinal Chemistry
1
2
3
4
(25) Petersen, T. N.; Brunak, S.; von Heijne, G.; Nielsen, H. SignalP 4.0: discriminating
5
6
signal peptides from transmembrane regions. Nat. Methods 2011, 8, 785–786.
7
8
9
(26) Antunes, N. T.; Lamoureaux, T. L.; Toth, M.; Stewart, N. K.; Frase, H.; Vakulenko,
S. B. Class D βꢀlactamases: are they all carbapenemases? Antimicrob. Agents
Chemother. 2014, 58, 2119–2125.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(27) Docquier, J.ꢀD.; Calderone, V.; De Luca, F.; Benvenuti, M.; Giuliani, F.; Bellucci, L.;
Tafi, A.; Nordmann, P.; Botta, M.; Rossolini, G. M.; Mangani, S. Crystal structure of
the OXAꢀ48 βꢀlactamase reveals mechanistic diversity among class D
carbapenemases. Chem. Biol. 2009, 16, 540–547.
(28) O’Callaghan, C. H.; Morris, A.; Kirby, S. M.; Shingler, A. H. Novel method for
detection of βꢀlactamases by using a chromogenic cephalosporin substrate.
Antimicrob. Agents Chemother. 1972, 1, 283–288.
(29) Bebrone, C.; Moali, C.; Mahy, F.; Rival, S.; Docquier, J. D.; Rossolini, G. M.;
Fastrez, J.; Pratt, R. F.; Frère, J. M.; Galleni, M. CENTA as a chromogenic substrate
for studying βꢀlactamases. Antimicrob. Agents Chemother. 2001, 45, 1868–1871.
(30) BiacoreTM Assay Handbook; GE Healthcare BioꢀSciences AB: Uppsala, Sweden,
2012; p 40.
(31) Congreve, M.; Carr, R.; Murray, C.; Jhoti, H. A “Rule of Three” for fragmentꢀbased
lead discovery? Drug Discovery Today 2003, 8, 876–877.
(32) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: a useful metric for lead
selection. Drug Discovery Today 2004, 9, 430–431.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 51
1
2
3
4
5
6
7
8
(33) Hopkins, A. L.; Keserü, G. M.; Leeson, P. D.; Rees, D. C.; Reynolds, C. H. The role
of ligand efficiency metrics in drug discovery. Nat. Rev. Drug Discovery 2014, 13,
105–121.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(34) Krissinel, E.; Henrick, K. Secondaryꢀstructure matching (SSM), a new tool for fast
protein structure alignment in three dimensions. Acta Crystallogr. Sect. D Biol.
Crystallogr. 2004, 60, 2256–2268.
(35) Beck, J.; Vercheval, L.; Bebrone, C.; HertegꢀFernea, A.; Lassaux, P.; Marchandꢀ
Brynaert, J. Discovery of novel lipophilic inhibitors of OXAꢀ10 enzyme (class D βꢀ
lactamase) by screening amino analogs and homologs of citrate and isocitrate. Bioorg.
Med. Chem. Lett. 2009, 19, 3593–3597.
(36) Docquier, J.ꢀD.; Benvenuti, M.; Calderone, V.; Giuliani, F.; Kapetis, D.; De Luca, F.;
Rossolini, G. M.; Mangani, S. Crystal structure of the narrowꢀspectrum OXAꢀ46 class
D βꢀlactamase: relationship between activeꢀsite lysine carbamylation and inhibition
by polycarboxylates. Antimicrob. Agents Chemother. 2010, 54, 2167–2174.
(37) King, D. T.; King, A. M.; Lal, S. M.; Wright, G. D.; Strynadka, N. C. J. Molecular
mechanism of Avibactamꢀmediated βꢀlactamase inhibition. ACS Infect. Dis. 2015, 1,
175–184.
(38) Szarecka, A.; Lesnock, K. R.; RamirezꢀMondragon, C. A.; Nicholas, H. B.; Wymore,
T. The class D βꢀlactamase family: residues governing the maintenance and diversity
of function. Protein Eng. Des. Sel. 2011, 24, 801–809.
(39) Christopeit, T.; Carlsen, T. J. O.; Helland, R.; Leiros, H.ꢀK. S. Discovery of novel
inhibitor scaffolds against the metalloꢀβꢀlactamase VIMꢀ2 by surface plasmon
resonance (SPR) based fragment screening. J. Med. Chem. 2015, 58, 8671–8682.
ACS Paragon Plus Environment

Page 27 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
(40) Ulrich, A.; Andersen, K. R.; Schwartz, T. U. Exponential megapriming PCR (EMP)
cloningꢀseamless DNA insertion into any target plasmid without sequence
constraints. PLoS One 2012, 7, e53360.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(41) Unger, T.; Jacobovitch, Y.; Dantes, A.; Bernheim, R.; Peleg, Y. Applications of the
Restriction Free (RF) cloning procedure for molecular manipulations and protein
expression. J. Struct. Biol. 2010, 172, 34–44.
(42) Studier, F. W. Protein production by autoꢀinduction in highꢀdensity shaking cultures.
Protein Expression Purif. 2005, 41, 207–234.
(43) Kabsch, W. XDS. Acta Crystallogr. D. Biol. Crystallogr. 2010, 66, 125–132.
(44) Evans, P. R.; Murshudov, G. N. How good are my data and what is the resolution?
Acta Crystallogr. D. Biol. Crystallogr. 2013, 69, 1204–1214.
(45) McCoy, A. J.; GrosseꢀKunstleve, R. W.; Adams, P. D.; Winn, M. D.; Storoni, L. C.;
Read, R. J. Phaser crystallographic software. J. Appl. Crystallogr. 2007, 40, 658–674.
(46) Afonine, P. V; GrosseꢀKunstleve, R. W.; Echols, N.; Headd, J. J.; Moriarty, N. W.;
Mustyakimov, M.; Terwilliger, T. C.; Urzhumtsev, A.; Zwart, P. H.; Adams, P. D.
Towards automated crystallographic structure refinement with phenix.refine. Acta
Crystallogr. D. Biol. Crystallogr. 2012, 68, 352–367.
(47) Smart, O. S.; Womack, T. O.; Sharff, A.; Flensburg, C.; Keller, P.; Paciorek, W.;
Vonrhein, C.; Bricogne, G. Grade, version 1.102. Global Phasing 2014.
(48) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features and development of
Coot. Acta Crystallogr. Sect. D Biol. Crystallogr. 2010, 66, 486–501.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 51
1
2
3
4
5
6
7
8
(49) Liebschner, D. Phenix.polder ꢀ A tool for calculating difference maps around atom
selections by excluding the bulk solvent mask. The Phenix Project: Berkeley,
California 2016.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(50) Schrödinger LLC. The PyMOL Molecular Graphics System. New York, NY August
2014.
(51) Laskowski, R. A.; Swindells, M. B. LigPlot+: multiple ligand–protein interaction
diagrams for drug discovery. J. Chem. Inf. Model. 2011, 51, 2778–2786.
(52) Sherman, W.; Day, T.; Jacobson, M. P.; Friesner, R. A.; Farid, R. Novel procedure
for modeling ligand/receptor induced fit effects. J. Med. Chem. 2006, 49, 534–553.
(53) Tominaga, M.; Suzuki, K.; Kawano, M.; Kusukawa, T.; Ozeki, T.; Sakamoto, S.;
Yamaguchi, K.; Fujita, M. Finite, spherical coordination networks that selfꢀorganize
from 36 small components. Angew. Chem. Int. Ed. Engl. 2004, 43, 5621–5625.
ACS Paragon Plus Environment

Page 29 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
Tables and Figures
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1 Steadyꢀ state enzyme kinetic parameters of nOXAꢀ48 with penicillins,
cephalosporin, carbapenems and reporter substrates, with standard errors. Kinetic parameters
for tOXAꢀ48 with the reporter substrate nitrocefin is also included as comparison.
Table 2 Molecular structure, binding affinity (KD) and IC50 values for the fragments
identified during the screening campaign. The LE is also calculated from the KD value using
the formula ꢟꢠ = (ꢡꢢꢣꢤꢥ) ÷ ꢦꢧꢨꢩꢪꢫ ꢩꢬꢭꢮ. n.f.: Not Found. Standard errors are given in
Table S1.
Table 3 Binding affinities (KD), IC50 values for compound 11ꢀ17. Standard errors are given
in Table S1.
Table 4 Data collection and refinement statistics for the complexꢀstructures of tOXAꢀ48 in
complex with compounds 1, 2, 3 and 17. Statistics for the highestꢀresolution shell are shown
in parentheses
Table 5 Primers used in the cloning of the new tOXAꢀ48 gene construct. The vectorꢀspecific
sequence is in bold, the TEV protease cleavage site sequence is in italic and OXAꢀ48 geneꢀ
specific sequence is underlined.
ACS Paragon Plus Environment