A focused fragment library targeting the antibiotic resistance enzyme - Oxacillinase-48: Synthesis, structural evaluation and inhibitor design

Article abstract of DOI:10.1016/j.ejmech.2017.12.085

β-Lactam antibiotics are of utmost importance when treating bacterial infections in the medical community. However, currently their utility is threatened by the emergence and spread of β-lactam resistance. The most prevalent resistance mechanism to β-lactam antibiotics is expression of β-lactamase enzymes. One way to overcome resistance caused by β-lactamases, is the development of β-lactamase inhibitors and today several β-lactamase inhibitors e.g. avibactam, are approved in the clinic. Our focus is the oxacillinase-48 (OXA-48), an enzyme reported to spread rapidly across the world and commonly identified in Escherichia coli and Klebsiella pneumoniae. To guide inhibitor design, we used diversely substituted 3-aryl and 3-heteroaryl benzoic acids to probe the active site of OXA-48 for useful enzyme-inhibitor interactions. In the presented study, a focused fragment library containing 49 3-substituted benzoic acid derivatives were synthesised and biochemically characterized. Based on crystallographic data from 33 fragment-enzyme complexes, the fragments could be classified into R¹ or R² binders by their overall binding conformation in relation to the binding of the R¹ and R² side groups of imipenem. Moreover, binding interactions attractive for future inhibitor design were found and their usefulness explored by the rational design and evaluation of merged inhibitors from orthogonally binding fragments. The best inhibitors among the resulting 3,5-disubstituted benzoic acids showed inhibitory potential in the low micromolar range (IC₅₀ = 2.9 μM). For these inhibitors, the complex X-ray structures revealed non-covalent binding to Arg250, Arg214 and Tyr211 in the active site and the interactions observed with the mono-substituted fragments were also identified in the merged structures.

Full text of DOI:10.1016/j.ejmech.2017.12.085

Accepted Manuscript
A focused fragment library targeting the antibiotic resistance enzyme -
Oxacillinase-48: Synthesis, structural evaluation and inhibitor design
Sundus Akhter, Bjarte Aarmo Lund, Aya Ismael, Manuel Langer, Johan Isaksson,
Tony Christopeit, Hanna-Kirsti S. Leiros, Annette Bayer
PII:
S0223-5234(17)31116-9
10.1016/j.ejmech.2017.12.085
EJMECH 10063
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 6 July 2017
Revised Date: 24 December 2017
Accepted Date: 26 December 2017
Please cite this article as: S. Akhter, B.A. Lund, A. Ismael, M. Langer, J. Isaksson, T. Christopeit, H.-K.S.
Leiros, A. Bayer, A focused fragment library targeting the antibiotic resistance enzyme - Oxacillinase-48:
Synthesis, structural evaluation and inhibitor design, European Journal of Medicinal Chemistry (2018),
doi: 10.1016/j.ejmech.2017.12.085.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
1
2
3
A focused fragment library targeting the antibiotic
resistance enzyme - oxacillinase-48: synthesis, structural
evaluation and inhibitor design
4
5
Sundus Akhter1,#,Bjarte Aarmo Lund2,#, Aya Ismael1, Manuel Langer1, Johan Isaksson1, Tony
Christopeit2, Hanna-Kirsti S. Leiros2,*, Annette Bayer1,*
6
7
1 Department of Chemistry, Faculty of Science and Technology, UiT The Arctic University of Norway,
2
8
N-9037 Tromsø, Norway. The Norwegian Structural Biology Centre (NorStruct), Department of
9
Chemistry, Faculty of Science and Technology, UiT The Arctic University of Norway, N-9037 Tromsø,
Norway.
10
11
12
* Corresponding authors: Annette Bayer, E-mail: annette.bayer@uit.no , Phone +47 77 64 40 69;
Hanna-Kirsti S. Leiros, E-mail: hanna-kirsti.leiros@uit.no , Phone +47 77 64 57 06;
13
# These authors have contributed equally to this work.
14
15
Highlights:
•
•
Synthesis and biochemical analysis of a focused fragment library consisting of 3-substituted
16
17
18
19
20
21
benzoic acids against the class D β-lactamase oxacillinase-48.
33 fragment-enzyme complex structures were structurally analyzed and fragment-enzyme
interactions useful for future drug design were identified.
•
•
•
Merged inhibitors were rationally designed by overly of fragments-enzyme structures.
A synthetic method for unsymmetrically 3,5-disubstituted benzoic acids was developed.
Merged inhibitors with IC50 of 2.9 µM were found and structurally analyzed.
22
23
Keywords: Crystal structure, inhibition properties, benzoic acid derivatives, serine-β-lactamase
inhibitors, fragments, structure-guided drug design.
24
25
26
Abbreviations: DMSO, dimethyl sulfoxide; OXA, oxacillinase; IC50,half maximal inhibitory
concentration; LE, ligand efficiency; MBL, metallo-β-lactamase; NMR, nuclear magnetic resonance;
SBL, serine-β-lactamase; SPR, surface plasmon resonance.
27
28
29
30
31
Author contributions: Designed the experiments: AB, BAL, HKSL, SA, TC. Performed the organic
synthesis: SA, AI, ML. Determined IC50 values and Kd-values: BAL. Prepared and solved crystal
structures: BAL. Analyzed 3D structures: AB, BAL, SA. NMR studies: BAL, JI. Analyzed data and wrote
the paper: AB, BAL, HKSL, JI, SA, TC. All authors have given approval to the final version of the
manuscript.
-1-

ACCEPTED MANUSCRIPT
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
1
Abstract
β-Lactam antibiotics are of utmost importance when treating bacterial infections in the
medical community. However, currently their utility is threatened by the emergence and
spread of β-lactam resistance. The most prevalent resistance mechanism to β-lactam
antibiotics is expression of β-lactamase enzymes. One way to overcome resistance caused by
β-lactamases, is the development of β-lactamase inhibitors and today several β-lactamase
inhibitors e.g. avibactam are approved in the clinic. Our focus is the oxacillinase-48 (OXA-48),
an enzyme reported to spread rapidly across the world and commonly identified in
Escherichia coli and Klebsiella pneumoniae. To guide inhibitor design, we used diversely
substituted 3-aryl and 3-heteroaryl benzoic acids to probe the active site of OXA-48 for
useful enzyme-inhibitor interactions. In the presented study, a focused fragment library
containing 49 3-substituted benzoic acid derivatives were synthesised and biochemically
characterized. Based on crystallographic data from 33 fragment-enzyme complexes, the
fragments could be classified into R1 or R2 binders by their overall binding conformation in
relation to the binding of the R1 and R2 side groups of imipenem. Moreover, binding
interactions attractive for future inhibitor design were found and their usefulness explored
by the rational design and evaluation of merged inhibitors from orthogonally binding
fragments. The best inhibitors among the resulting 3,5-disubstituted benzoic acids showed
inhibitory potential in the low micromolar range (IC50 = 2.9 µM). For these inhibitors, the
complex X-ray structures revealed non-covalent binding to Arg250, Arg214 and Tyr211 in the
active site and the interactions observed with the mono-substituted fragments were also
identified in the merged structures.
54
55
56
57
58
59
60
61
2
Introduction
Years of overuse of antibiotics have selected for antibiotic resistant strains (1), and today
medical personnel are frequently forced to administer last-resort antibiotics. However, the
number of cases where last-resort antibiotics fail in treatment are increasing (2) and deaths
due to antibiotic resistant infections are expected to surpass cancer deaths by 2050 (3).
Bacterial resistance towards clinically important β-lactam antibiotics (4) like penicillins,
cephalosporins and carbapenems originates most often from the occurrence of β-lactam-
hydrolysing enzymes – the β-lactamases.
62
63
64
65
66
67
68
The β-lactamase enzymes are of ancient origin (5) and today over 2600 enzymes spanning
four classes of β-lactamases are known (6-8). β-Lactamases are grouped into two super
families based on the enzyme mechanism for β-lactam hydrolysis: the serine dependent β-
lactamases (SBLs; Amber class A, C, and D) and metallo-β-lactamases (MBLs; Amber class B)
(7,9). SBLs are characterized by a serine residue in the active site, while MBLs require a metal
co-factor, usually one or two zinc ions, for enzyme activity. This work focuses on the class D
SBLs – also called oxacillinases (OXAs) – and in particular on the oxacillinase-48 (OXA-48).
69
70
71
72
The class D SBLs are characterized by a hydrophobic environment in the active site, that
facilitates the carboxylation of a lysine residue. The N-carboxylated lysine plays a critical role
in the substrate hydrolysis (10). Originally, the OXAs were believed to have a limited
substrate profile only hydrolysing penicillins, but with the emergence of carbapenem-
-2-

ACCEPTED MANUSCRIPT
73
74
75
76
hydrolysing OXA variants, e.g. OXA-23, OXA-24 and OXA-48, their clinical relevance has
increased (11). OXA-48 was reported for the first time in 2001 and has since then spread
rapidly across the world. (11) It is commonly identified in Escherichia coli and Klebsiella
pneumoniae.
77
78
79
80
81
82
83
84
85
86
One strategy to circumvent resistance in β-lactamase producing pathogens is the use of β-
lactamases inhibitors (4,12) in combination with the β-lactam antibiotic. Inhibitors of class A
SBLs like clavulanic acid, sulbactam and tazobactam became clinically available from the
1980s (13), but only a few class D β-lactamases are inhibited by these β-lactamase inhibitors
e.g. OXA-2 and OXA-18 (14). In 2015, a new SBL inhibitor, avibactam, targeting class A, C and
some class D SBLs, including OXA-48, was approved by the FDA for treatment of complicated
urinary tract and intra-abdominal infections (15). However, the inhibition level of different
class D β-lactamases by avibactam varies (16,17). With the first reports of resistance to
avibactam published (18), one can speculate that it will only be a matter of time before class
D β-lactamases show resistance to avibactam as well.
87
88
89
90
91
92
93
94
The development of new OXA inhibitors, either with a different enzyme-inhibition profile
compared to existing inhibitors, or as alternative when resistance to existing inhibitors
arises, is of importance. We have previously reported a fragment-based screening approach
to identify weak inhibitors of OXA-48 (19). The most interesting hit was 3-(pyridin-4-
yl)benzoic acid 1 with an IC50 of 250 µM and a ligand efficiency (LE) of 0.32. Crystallographic
data from enzyme-fragment complexes indicated two overlapping binding conformations of
the fragment. Merging of the two conformations of 1 into one molecule 2 (Fig. 1) gave a 10-
fold increase in binding affinity improving the IC50 from 250 µM to 18 µM (19).
95
96
97
98
99
Figure 1: The two alternate conformations of fragment 1 (light grey) in complex with OXA-48
(dark grey surface) (A and B), the merged compound 2 (pink) in complex with OXA-48 (dark
grey surface) (C), and a schematic view of the merging approach described in previous work
(D) (19).
100
101
102
103
104
105
106
In this study, we describe the use of small mono-substituted fragments - analogues of
fragment 1 - as probes to explore the OXA-48 binding site. The aim was to identify fragment-
enzyme interactions in the two alternate binding pockets of the active site of OXA-48, which
could be of general interest for the design of OXA-48 inhibitors. We wanted to exploit the
ability of small fragments to efficiently explore the binding pocket as they are less restricted
by size and more flexible compared to more elaborated inhibitors. Moreover, the smaller
fragments generally have the advantage of being more easily prepared making the discovery
-3-

ACCEPTED MANUSCRIPT
107
108
109
process more work-efficient. Furthermore, we wanted to translate the knowledge gained
into the rational design of di-substituted inhibitors related to compound 2 circumventing the
laborious preparation of a large library of elaborated inhibitors.
110
111
112
113
114
115
116
117
118
119
Towards this goal, we prepared a focused fragment library containing 3-aryl benzoic acids
decorated with a wide range of polar groups and a number of 3-heteroaryl benzoic acid
derivatives. In total 49 fragments were tested for inhibitory activity against OXA-48 and the
binding conformations of 33 fragment-enzyme complexes were analysed by X-ray
crystallography. Based on the structural information, fragments could be classified according
to their preferred binding pocket and useful fragment-enzyme interactions e.g. hydrogen
bonds were identified. Moreover, several new orthogonally binding fragments were found
leading to the design of symmetrically and unsymmetrically di-substituted inhibitors with
improved IC50 in the low micromolar range. The structural data from enzyme-inhibitor
complexes was compared with enzyme-fragment complexes.
120
121
3
Results and discussion
Synthesis
3.1
122
123
124
125
126
127
128
129
130
131
3.1.1
Synthesis of 3-substituted benzoic acids
A fragment library containing 49 3-substituted benzoic acid analogues 3a–35 was prepared
(Table 1). The fragments generally fulfilled the demands of libraries for fragment-based
ligand design (MW < 300, clogP < 3, hydrogen bond acceptor/donors < 3) (20). For the
synthesis, a strategy based on the Suzuki-Miyaura (SM) cross-coupling reaction to join two
sp2–hybridized carbons was employed (21). Two alternate coupling strategies were
successful starting with either 3-bromobenzoic acid (Table 1, strategy A) or 3-
carboxyphenylboronic acid pinacol ester (Table 1, strategy B) as starting materials allowing
for the utilisation of a wide range of aryl boronic acids or aryl bromides to introduce diversity
in the library.
132
133
134
135
136
137
138
139
Many of the required aryl boronic acids and bromides were commercial available, while the
aryl bromides used as starting materials for fragments 17-20, 24, 29 and 30 were prepared
according to standard acylation and sulphonylation protocols. The NH-tetrazol-5-yl-
substituted arylbromides (starting material for fragments 26a and 26b) were prepared by a
[3+2] intermolecular cycloaddition of 3- or 4-bromobenzonitrile with trimethyl silyl azide in
the presence of dibutyltin oxide in anhydrous 1,4-dioxane. The reaction mixture was
subjected to microwave irradiation in a tightly sealed vessel for 50 min at 150 °C to afford 3-
or 4-bromobenzotetrazole in 86% and 82% yield, respectively.
-4-

ACCEPTED MANUSCRIPT
140
141
Table 1: Preparation strategy and inhibitor activities of a library of 3-substituted benzoic
acids analogues against OXA-48 (IC50, Kd and LE).
142
Comp. Ar =
ID
Strateg.
Yield
IC50
(μM)
KD
(μM)
LEd
Comp.
ID
Ar =
Strateg.
Yield
IC50
(μM)
KD
(μM)
LEd
B
78%
A
97%
3a*
3b*
90
170 0.35
300 0.33
11b*
12a*
180
120
350 0.29
150 0.29
B
67%
A
82%
170
A
94%
A
90%
4a*
50
175 0.38
110 0.35
12b
13*
380
330
361 0.25
330 0.29
A
98%
B
35%
4b*
110
A
39%
A
95%
4c*
470
170 0.29
14*
390
600
220 0.27
800 0.27
A
84%
B
36%
5*
900
250
230 0.25
123 0.30
15a
15b
A
98%
B
86%
6a*
1400 550 0.23
110 300 0.31
1000 970 0.23
A
98%
B
15%
6b*
6c*
360
150
226 0.28
250 0.31
16a
16b
A
86%
B
67%
A
91%
Ba, c
41%
7
400 1000 0.28
17*
18
370
60
100 0.24
210 0.24
A
68%
Ba, c
65%
8a*
130
130
360
210
260
380
260
170 0.34
240 0.34
312 0.30
200 0.27
144 0.26
280 0.27
220 0.28
A
98%
Ba, c
26%
8b*
8c*
9a
19a
19b
20
110
450
370
35
110 0.26
240 0.22
200 0.22
100 0.33
290 0.25
130 0.27
A
78%
Ba, c
10%
Aa, c
57%
Ba, c
11%
A
54%
A
98%
9b*
10
21a*
21b*
22
A
98%
A
98%
450
130
A
98%
Ba, b
87%
11a
a
143
144
145
* X-ray structure of fragment-enzyme complex available. Reaction in anhydrous THF instead of
b
c
dioxane:water as solvent; XPhos-Pd G2 as catalyst instead of PdCl2(PPh3)2; PdCl2(dppf) as catalyst
instead of PdCl2(PPh3)2. d LE = − logꢀꢁ ICꢂꢁ/HeavyAtomCount
.
146
-5-

ACCEPTED MANUSCRIPT
147
Table 1 continues:
Comp. Ar =
ID
Strateg.
Yield
IC50
(μM)
KD
(μM)
LEd
Comp.
ID
Ar =
Strateg.
Yield
IC50
(μM)
KD
(μM)
LEd
Ba, c
46%
B
36%
23a
23b
230
520
170
190
0.24
0.22
29
30
170
800
130 0.33
900 0.29
Ba, c
34%
B
45%
Aa, b
34%
B
67%
24*
25
250
140
0.25
31
32
33
350
500
800
113 0.28
590 0.31
900 0.31
B
15%
A
6%
1300 ˃1000 0.20
B
98%
B
24%
26a*
60
70
0.30
B
98%
B
20%
26b
27*
28*
36
70
0.30
0.30
0.27
34
310
35
400 0.27
159 0.42
B
67%
A
98%
110
240
400
160
35*
B
87%
a
148
149
150
* X-ray structure of fragment-enzyme complex available. Reaction in anhydrous THF instead of
b
c
dioxane:water as solvent; XPhos-Pd G2 as catalyst instead of PdCl2(PPh3)2; PdCl2(dppf) as catalyst
instead of PdCl2(PPh3)2. d LE = − logꢀꢁ ICꢂꢁ/HeavyAtomCount.
151
152
153
154
155
156
157
158
159
160
In general, couplings under standard aqueous conditions using PdCl2(PPh3)2 as catalyst (5–10
mol%), K3PO4 as base (5 equiv.) in dioxane/water gave good yields. The couplings leading to
fragments 9, 17–20 and 22–24 were not successful under these standard conditions. More
efficient catalysts (XPhos-Pd G2 or PdCl2(dppf)) and water-free conditions (anhydrous THF
instead of dioxane/water) were successfully employed to solve reactivity and solubility
problems and to prevent hydrolysis for base sensitive products (9 and 24). However, for
some products (19a+b and 20) the yields were still low (< 20%). Generally, the reactions
were easily purified by automated C18 flash chromatography to provide compounds of high
purity (> 95% as determined by UHPLC). For some compounds (15, 16, 19, 23, 24, 32 and 34),
additional silica flash chromatography was necessary to provide sufficiently pure products.
161
162
163
164
165
166
167
168
169
170
3.1.2
Synthesis of 3,5-disubstituted benzoic acid derivatives.
To study inhibitor properties like activity and enzyme interactions of merged fragments, a
small series of symmetrical and unsymmetrical 3,5-disubstituted benzoic acids was designed
(vide infra) and prepared. The synthesis of symmetrical 3,5-disubstituted compounds 36 and
38 was achieved under the conditions established for the coupling of mono-substituted
fragments using Pd2(dba)3/XPhos or XPhos-Pd G2 as catalysts (Scheme 1) (19). The di-
substituted coupling products 36 and 38 were obtained from 3,5-dibromobenzoic acid as
starting material and an increased amount of the boronic acid derivative (2 equiv.) in 54%
and 65% yield, respectively. Compound 37 was isolated in 11% yield as by-product in an
attempt to selectively mono-substituted 3,5-dibromobenzoic acid (vide infra).
-6-

ACCEPTED MANUSCRIPT
171
172
173
174
175
176
Scheme 1. Preparation of symmetrical 3,5-disubstituted benzoic acids. Reagents and
conditions: 36: 3-acetamidophenylboronic acid (1.5 equiv.), Pd2(dba)3•CHCl3 (5 mol%), XPhos
(5 mol%), dioxane:water (1:1), 60 °C, 54%; 37: 4-acetamidophenylboronic acid (0.75 equiv.),
PdCl2(PPh3)2 (10 mol%), dioxane:water (1:1), 95 °C, 11%; 38: quinolin-6-ylboronic acid pinacol
ester (2.0 equiv.), XPhos-Pd G2 (5 mol%), tert-butanol, 60 °C, 65%.
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
For the synthesis of unsymmetrical 3,5-disubstituted benzoic acids 39, the sequential
addition of two different aryl boronic acids under the previously established conditions gave
only 15% isolated yield (Scheme 2). In addition, the procedure involved tedious HPLC
purifications as the reaction mixture was difficult to purify due to occurrence of symmetrical
by-products with similar properties. To improve the selectivity of the reaction, we changed
the starting material from 3,5-dibromobenzoic acid to 3-iodo-5-bromobenzoic acid in order
to take advantage of the faster coupling reaction of aryl iodides when compared with aryl
bromides and thereby to prevent formation of symmetrical disubstituted by-products
(Scheme 2). Investigation of the chemoselective coupling of 3-iodo-5-bromobenzoic acid
with quinolin-6ylboronic acid pinacol ester to form mono-substituted int-40 showed that a
second, unwanted coupling was not easily prevented and a careful fine tuning of catalyst
(RuPhos-Pd G3, XantPhos-Pd G3, Sphos/Pd2(dba)3, Xphos/Pd2(dba)3, SPhos-Pd G3, XPhos-Pd
G2, Pd2(dppf)Cl2), solvent (toluene/water, anhydrous THF, dioxane/water, tert-butanol),
reaction temperature (40–80 °C) and time (10–48 h) was initiated (Table SI1, see supporting
information). The composition of the crude reaction mixtures with respect to mono- and
disubstituted products as well as unreacted starting material was determined by mass
spectrometry (MS). The most chemoselective catalysts were XantPhos-Pd G3, Pd2(dppf)Cl2
and SPhos/Pd2(dba)3 showing good selectivity for the aryl iodide when the reaction was
performed with K3PO4 as base in dioxane/water at 60 °C for 24 hours (Scheme 2). At this
conditions with SPhos/Pd2(dba)3 as catalyst, the monosubstituted intermediate int-40 was
obtained as main product together with small amounts of the disubstituted by-product (8–
10%). Careful purification to remove any traces of the disubstituted compound provided int-
40 in moderate yield (45%). The mono-substituted int-40 was further subjected to a second
coupling with XPhos-Pd G2 (5 mol%) as catalyst to provide 40 in good yields (90%).
-7-

ACCEPTED MANUSCRIPT
201
202
203
204
205
206
207
Scheme 2: Preparation of unsymmetrical 3,5-disubstituted benzoic acids. Reagents and
conditions: 39: i. X = Br, 3-acetamidophenylboronic acid (0.75 equiv.), PdCl2(PPh3)2 (10 mol%),
dioxane:water (1:1), 60 °C; ii. pyridin-4-ylboronic acid (1.2 equiv.), PdCl2(PPh3)2 (10 mol%),
dioxane:water (1:1), 60 °C; int-40: X = I, quinolin-6-ylboronic acid pinacol ester (2.0 equiv.),
Pd2(dba)3*CHCl3 (5 mol%), SPhos (5 mol%), dioxane:water (1:1), 60 °C; 40: 3-
acetamidophenylboronic acid (1.5 equiv.), XPhos-Pd G2 (5 mol%), tert-BuOH, 60 °C.
208
3.2
Evaluation of 3-substituted benzoic acids
209
210
211
212
213
214
215
216
217
218
219
3.2.1
Inhibitor activity of 3-substituted benzoic acids
The mono-substituted fragments 3–35 were initially investigated for their inhibitory activity
against OXA-48 in an enzymatic assay and by SPR. Inhibition and binding data are given in
Table 1 along with the associated ligand efficiencies (LE). The original hit fragment 1 had an
IC50 of 250 µM and an LE of 0.32. Most of the fragments in this study showed inhibition at a
similar level with IC50 > 200 µM and LE ≤ 0.30. Fragments 4a (IC50 (µM)/LE: 50/0.38), 18 (IC50
(µM)/LE: 60/0.24), 21a (IC50 (µM)/LE: 35/0.33), 26b (IC50 (µM)/LE: 36/0.30) and 35 (IC50
(µM)/LE: 35/0.42) showed an order of magnitude stronger inhibition and were the most
potent fragments. Even though there are some discrepancies between the inhibition and
binding data, the same trends are maintained when comparing similar compounds,
indicating that the compounds indeed bind specifically to one site of the enzyme.
220
221
222
223
224
225
226
3.2.2
Structural analysis of 3-substituted benzoic acids
To evaluate the binding poses of our fragments, enzyme-fragment complexes for x-ray
crystallographic analysis were prepared. Rewardingly, 33 out of 49 fragments were
successfully soaked with OXA-48 and yielded crystal structures with resolution high enough
to warrant placement of the inhibitor in the electron density (Table 1). In addition, a crystal
structure of OXA-48 in complex with the substrate imipenem was obtained to better
understand substrate binding and to compare substrate and fragment binding interactions.
227
228
229
230
231
232
233
234
235
The crystal structure of the acyl-enzyme complex of OXA-48 with imipenem (Fig. 2A)
revealed a conformation close to previously observed conformations with OXA-13 (PDB-ID:
1h5x). In the complex the ring-opened imipenem was bound to OXA-48 covalently with
continuous electron density from the hydroxyl group of Ser70. There was an ionic bond from
the carboxylate group of imipenem to the guanidine group of Arg250. The carbonyl-group of
the now ring-opened β-lactam ring was positioned in the oxyanion-hole forming hydrogen
bonds to the main chain amides of Tyr211 and Ser70. The 6α-hydroxyethyl group (R1) of
imipenem was positioned towards the hydrophobic residues Trp105, Val120 and Leu158 and
in the following discussion this region will be called the R1 site. The amidine group (R2) was
-8-

ACCEPTED MANUSCRIPT
236
237
238
239
situated in the cleft defined by Ile102, Tyr211, Leu247 and Thr213 and this region will be
called the R2 site. The R1 and R2 side chains of imipenem (Fig. 2A) had the same overall
directions as the pyridinyl substituents in the two overlapping binding conformations
observed with our initial hit 3-pyridin-4-ylbenzoic acid 1 (19).
240
241
242
243
244
245
In all our structures of OXA-48 in complex with fragments, an ionic bond between the
carboxylate group of the fragments and the guanidine group of Arg250 was observed, which
resembled the interaction of the carboxylate group of imipenem or the sulfamate group of
avibactam with Arg250.(17,22) In some cases, the carboxylate group was oriented in such a
way that also Thr209 (fragments 9b, 28, 35), Lys208 (fragment 34) or both (fragment 26a)
participated in binding.
246
247
248
249
250
251
Another common feature found in almost all crystal structures, except for fragments 21a
and 26b, was a π-π stacking interaction of the 3-aryl substituents attached to the benzoic
acid scaffold with Tyr211. This is consistent with the binding of imipenem, where the R2 side
chain was oriented towards Tyr211 (Fig. 2C). The importance of Tyr211 as a non-polar patch
that contributes in binding substrate side-chains has been recognised before (23). We also
observed this interaction with our unsubstituted pyridyl benzoic acids previously. (19)
252
253
254
255
256
257
258
259
Figure 2: The crystal structure of imipenem in complex with OXA-48 (A) shows that the two
side chains of imipenem extends in separate directions. The carbapenem substrates of OXA-
48 have small R1 side chains. We were however able to fit larger groups in the R1 site like the
N-acetamide substituted phenyl ring in compound 21a (B). Yet, most of the tested 3-
substituted benzoic acids bind towards the larger R2 site, like the quinolin-7-yl substituted
compound 28 (C).
260
261
262
263
264
The weaker binding fragments (3a+b, 4a–c, 5, 6a–c, 8a–c, 9b, 11b, 12a, 13, 14, 17, 24) all
bound in nearly the same conformation with the ionic bond of the benzoic acid and Arg250
and the π-π stacking interaction with Tyr211 as major interactions. In these structures, the 3-
aryl substituent on the benzoic acid was directed towards the R2 pocket (Fig. 2C). Only minor
conformational differences were observed as described in the following. To help the reader
-9-

ACCEPTED MANUSCRIPT
265
266
267
in the following discussion, we will describe the fragments by the identity of the Ar groups
(Table 1), as the structural differences of the fragments relate to this group i.e. 3-(2-
methyl)phenylbenzoic acid 3a will be described as 2-methylphenyl substituted fragment.
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
The methylphenyl substituted fragments 3a (IC50 (µM)/LE: 90/0.35) and 3b (IC50 (µM)/LE:
170/0.33) had similar conformations, however, the 2-methyl group in 3a was facing towards
the hydrophobic Cβ of Ser244 explaining the more favourable binding. Fragments 4a–c (IC50
(µM)/LE: 50/0.38, 110/0.35 and 470/0.29, respectively) also had very similar conformations,
but again we saw that more favourable van der Waals interactions gave higher affinity for
the 2-hydroxyphenyl substituted 4a. The 4-hydroxy isomer 4c had an unfavourable solvent
exposure of the hydroxyl group. Adding a methylene bridge yielding 3-hydroxymethylphenyl
5 (IC50 (µM)/LE: 900/0.25) did not lead to any favourable interactions. The methoxyphenyl
fragments 6a–c (IC50 (µM)/LE: 250/0.30, 360/0.28 and 150/0.31) shared the canonical R2
binding pose. The methoxy group of the 2-substituted 6a appeared more shielded from
solvent exposure than in 6b and 6c, yet the methoxy group did not seem to make any strong
contacts. The weak inhibition seen with methyl thioether 7 (IC50 (µM)/LE: 400/0.28)
corresponded to the results observed with the methoxy ethers 6. The fluorophenyl
substituted 8a–c (IC50 (µM)/LE: 130/0.34, 130/0.34 and 360/0.30) had nearly identical
binding poses. The 4-substituted 8c gave the highest IC50 value, most likely due to the
solvent exposed fluorine. The 2-substituted 8a seemed more favourable based on the
decreased solvent exposure of the fluorine atom, however, the difference to 8b was
negligible only observed by SPR.
286
287
288
289
290
291
292
293
294
295
The methoxyacetylphenyl esters 9a+b (IC50 (µM)/LE: 210/0.27 and 260/0.26) showed no
clear additional interactions in the complex structures with OXA-48, and the methyl group
appeared to be unfavourably exposed to the solvent. The corresponding 4-acetylphenyl
substituted 10 (IC50 (µM)/LE: 380/0.27) and carbamoylphenyl substituted 11a+b (IC50
(µM)/LE: 260/0.28 and 180/0.29) gave generally weak inhibition indicating that a carbonyl
group attached to the aromatic ring was not contributing to binding. No complex structures
are available for 10 and 11a, but the complex structure of 4-carbamoylphenyl 11b was
similar in conformation to the esters 9a+b. Slightly tighter binding was observed with the
meta-substituted sulfone 12a (IC50 (µM)/LE: 120/0.29), which also shares the same overall
conformation.
296
297
298
299
300
301
302
303
304
The 4-aminophenyl substituent of 13 (IC50 (µM)/LE: 330/0.30) did not appear to make any
interaction with the enzyme, and the inhibition was weak. The complex structure of the
corresponding N,N-dimethyl-4-aminophenyl substituted 14 (IC50 (µM)/LE: 390/0.27) showed
that the two methyl groups are solvent exposed, and this is reflected in the poor inhibition
by this compound. Similar to the complex structure of 14, the methyl 4-sulphonamidophenyl
group of 17 (IC50 (µM)/LE: 370/0.24) was seemingly pushed out of the active site and appears
completely exposed to the solvent. The larger phenyl 4-sulphonamidophenyl substituted
fragment 18 (IC50 (µM)/LE: 60/0.24) showed lower IC50 values probably driven by the
increase in hydrophobicity, and no complex structure was obtained.
-10-

ACCEPTED MANUSCRIPT
305
306
307
308
309
310
311
312
313
314
315
The corresponding 4-acetamidophenyl 21b (IC50 (µM)/LE: 450/0.25) showed weak inhibition,
likely due to the solvent exposure of the hydrophobic methyl group. The 3-acetamidophenyl
containing fragment 21a (Figure 3), however, showed a 10-fold increased inhibition (IC50
(µM)/LE: 35/0.33). The complex structure of OXA-48 with fragment 21a revealed that the
carbonyl of the acetate formed a hydrogen bond to the guanidine group of Arg214, which
directs the 3-acetamidophenyl substituent to the R1 site (Fig. 2B) and lead to a T-shaped π-π-
stacking interaction of the 3-acetamidophenyl substituent with Trp105. The π-π stacking of
the 3-acetamidophenyl substituent to Tyr211 normally observed with these fragments was
not observed; instead Tyr211 interacted with the benzoic acid by T-shaped π-π-stacking. The
interaction of an acetamide with Arg214 has been described previously for the avibactam
analogue FPI-1523 in complex with OXA-48 (PDB-ID: 5fas) (22).
316
317
Figure 3: Compound 21a was one of the most potent 3-substituted benzoic acid derivatives
we found. The IC50-value (A) was determined to be 35 µM, while the Kd was found to be 100
µM (B). The crystal structure of the complex OXA-48:21a with an omit-type polder-map
(2.5σ) (C) and its 2D-representation (D) shows that the carbonyl of the acetamido-group
forms a hydrogen bond with the guanidine of Arg214. The interaction with Arg214 causes the
B-ring to move away from Tyr211, introducing a new interaction with Trp105.
318
319
320
321
322
323
324
325
326
Encouraged by the results for fragment 21a, we designed a series of fragments incorporating
a hydrocarbon linker between the phenyl ring and the amino, sulphonamido or acetamido
groups of 13, 18 and 21. The amines 15 and 16, the sulfonamides 19 and 20, the amides 22,
23a+b and the acetate 24 are more flexible, thus, increasing the potential of hydrogen
-11-

ACCEPTED MANUSCRIPT
327
328
329
330
bonding. However, none of these fragments showed substantially improved binding (IC50:
110–1000; LE: 0.19–0.30). Moreover, the crystal structures of the amides 22, 23a+b and the
acetate 24 (IC50 (µM)/LE: 230/0.24, 520/0.22 and 250/0.25) did not show any specific
interactions for the functional groups.
331
332
333
334
335
336
337
338
339
340
In fragments 26a and 26b NH-tetrazole substituted phenyl rings were investigated as Ar
substitutents. Introducing the weakly acidic tetrazol-5-ylphenyl substituent in either 3-
position 26a (IC50 (µM)/LE: 60/0.30) or 4-position 26b (IC50 (µM)/LE: 36/0.30) yielded good
binding for both fragments. However, the binding poses for the two compounds were very
different. The 3-tetrazol-5-ylphenyl substituted 26a bound in two alternate positions. The π-
π-stacking with Tyr211 was maintained for both conformations, but the tetrazoles appeared
completely solvent exposed with no interactions with the enzyme. The 4-tetrazol-5-ylphenyl
substituted 26b formed a hydrogen bond with the guanidine group of Arg214 (Fig. 4),
interrupting the π-π-stacking with Tyr211. Fragment 26b occupied the R1 site rather than the
more common R2 site.
341
342
343
344
345
Figure 4: The IC50-value of compound 26b (A) was determined to be 36 µM, while the KD was
found to be 70 µM (B). The crystal structure of the complex OXA-48:26b with an omit-type
polder-map (2.5σ) (C) and a 2D-representation of the protein:compound complex
interactions. (D).
-12-

ACCEPTED MANUSCRIPT
346
347
348
349
350
Figure 5: The IC50-value of compound 28 (A) was determined to be 240 µM, while the KD was
found to be 160 µM (B). The crystal structure of the complex OXA-48:28 with an omit-type
polder-map (2.5σ) (C) and a 2D-representation of the protein:compound complex
interactions. (D).
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
A number of heterocyclic aryl substituents were also evaluated (fragments 25, 28–35). With
some exceptions of the pyridinyls 29 and 35 (IC50 (µM)/LE: 170/0.33 and 35/0.42) most of
these fragments showed only weak inhibition. The quinolin-7-yl substituted fragment 28
(IC50 (µM)/LE: 240/0.30) did maintain the overall conformation of the previous R2 binding
fragments (Figure 5), and so did the corresponding naphtalen-2-yl substituted fragment 27
(IC50 (µM)/LE: 110/0.29). In the same manner the indol-5-yl substituted fragment 34 (IC50
(µM)/LE: 310/0.27) did show acceptable binding, yet no specific interaction except for the π-
stacking with Tyr211. In our previous paper, we investigated pyridin-4-yl and pyridin-3-yl
substituted fragments (19) , and both inhibited OXA-48 with the same potency (IC50 (µM)/LE:
250/0.32). The pyridin-2-yl substituted fragments 35 (IC50 (µM)/LE: 35/0.41) showed a 10-
fold improvement in binding (Fig. 6A and B). In the crystal structure, two alternative
conformations were observed (Fig. 6C). One conformation was the canonical with π-stacking
of the pyridinyl ring with Tyr211 occupying the R2 site (Fig. 6E), but in the other
conformation the pyridinyl ring was orientated to the R1 site. The second conformation
showed a hydrogen bond from the protonated N atom in the pyridine ring to the backbone
carbonyl of Tyr117, which represents a unique interaction for the fragments in the library
(Fig. 6D). Only the protonated pyridinyl-nitrogen would be able to form hydrogen bonds to
-13-

ACCEPTED MANUSCRIPT
368
369
Tyr117, which may explain the slower on/off-rates observed for fragment 35 in the SPR-
experiments (Fig. 6B).
370
371
372
373
374
375
376
377
In the discussion above most fragments were identified as R2 binders with fragment 4a (IC50
(µM)/LE: 50/0.38) being the strongest binder among them. For R2 binders, the edge-to-face
π-π-stacking with Tyr211 appears to be an important interaction in accordance with previous
analyses (23). Fragment 35 showed the best ligand efficiency (IC50 (µM)/LE: 35/0.42), but
could not be classified as a R1 or R2 binder as both binding pockets showed useful
interactions (Fig. 6C–E). Only two R1 binders – fragments 21a and 26b - were identified, both
showing hydrogen bonds with Arg214 as cause for the fragments orientation towards the R1
site.
378
379
380
381
382
383
Figure 6: Compound 35 bound in the two alternate conformations. The IC50-value (A) was
determined to be 35 µM, while the KD was found to be 159 µM (B). The crystal structure of
the complex OXA-48:35 with an omit-type polder-map (2.5σ) (C) and a 2D-representation of
the protein:compound complex interactions. (D for green colored conformation, E for
magenta colored conformation).
384
385
386
387
388
389
3.2.3
NMR studies
In order to evaluate the fragment-enzyme binding in solution, a 13C NMR experiment for
OXA-48 was developed based on previous studies (24,25). OXA enzymes can be selectively
carbamylated with bicarbonate at an active site lysine to provide the corresponding
carbamic acid (24,26,27). For OXA-48 the carbamylated residue is Lys73, which is situated in
the R1 site (Fig. 2B). By using 13C-labeled sodium bicarbonate (NaH13CO3), a 13C atom was
-14-

ACCEPTED MANUSCRIPT
390
391
introduced in the R1 site of OXA-48, which can be used as a reporter probe for fragment
binding in 13C NMR studies.
392
393
394
395
Fragments binding in the R1 site were expected to change the local environment of the 13C
labelled Lys73, which results in a change of the 13C chemical shift of Lys–NH–13CO2H, while
ligands binding in the R2 site are further than ~9 Å away from the Lys73 carbamic acid, and
are therefore not expected to directly affect the 13C chemical shift.
396
397
398
399
400
401
402
403
404
NMR experiments were performed by equilibrating OXA-48 with 13C-labeled sodium
bicarbonate followed by the addition of inhibitor 2 and selected fragments 21a, 28 and 35
with known binding modes from X-ray analysis. The results are shown in Fig. 7. The 13C NMR
spectrum of OXA-48 after equilibration with NaH13CO3 showed the carbamate resonance at
163.95 ppm as a broad signal (Fig. 7E), which is in good agreement with the reported
chemical shift for carbamylated OXA-48 (28). In addition, two unassigned signals were
observed at 164.04 ppm similar to the results reported for carbamylation of OXA-58 (27).
Here the authors speculated that the unassigned signal may be related to a second
carbamylation site (27).
405
406
407
408
409
410
411
412
413
414
415
416
417
On addition of R1 binding fragment 21a and inhibitor 2, the 13C chemical shifts of the
carbamate signal were consistently deshielded in both experiments (δ = 164.25, ∆δ = 0.30
ppm, Fig. 7C and 7D). These findings support that the compounds bind competitively in the
active site. Moreover, the observed chemical shift perturbation indicates that the
compounds occupy the R1 site as found in the crystal structures. The R2 binding fragment 28
showed a similar deshielding of the carbamate signal though at a smaller amplitude (δ =
164.?, ∆δ = 0.? ppm, Fig. 7X) supporting that the fragment binds in the active site, while
fragment 35, which was identified as R1 or R2 binder, only slightly affected the chemical shift
(δ = 164.00, ∆δ = 0.04 ppm, Fig. 7B). The observed chemical shift perturbations for
fragments 28 and 35 may indicate that fragment 28 has an effect on carbamylated Lys73,
while fragment 35 do not interact with the R1 site, which is not consistent with the X-ray
structures. However, a more detailed study of the NMR conformations would be needed to
be conclusive about the binding poses in solution.
418
419
420
421
422
423
424
425
426
427
428
429
430
The small amplitudes of the observed chemical shift perturbations indicated that the effect
is not caused by direct hydrogen bonding of the carbamic carbonyl, for which a ∆δ of several
ppm would be expected, even for a µM binder (29). This was supported by the crystal
structures of OXA-48 indicating that the Lys73 carbamic acid was preoccupied in hydrogen
bonding to Trp157 and was not affected by ligand binding. The observed consistent, but
rather subtle, deshielding of the Lys73 carbamic acid (δ = 164.25, ∆δ = 0.30 ppm, Fig. 7C and
7D) for our R1 binding fragments can possibly be explained by an anisotropic magnetic
deshielding by the edge of the aromatic rings of these fragments, which were positioned
roughly 5 Å away from the reporter carbon for R1 binding fragments. Moreover, amplitude
of the chemical shift perturbation observed with R1 binding fragments 21a and inhibitor 2
(Fig. 7C and 7D) were in line with the reported changes observed for OXA enzymes on
coordination with inhibitors like β-hydroxyisopropylpenicillanates (24), cyclic boronates (25)
and avibactam (28).
-15-

ACCEPTED MANUSCRIPT
431
13
13
432
433
434
435
436
437
438
Figure 7: C NMR of the buffer alone including C labeled bicarbonate (A); OXA-48 without
13C labeled bicarbonate (B), OXA-48 with C labeled bicarbonate and fragment 35 (C); OXA-
13
13
13
48 with C labeled bicarbonate and fragment 28 (D); OXA-48 with C labeled bicarbonate
13
and fragment 21a (E); OXA-48 with C labeled bicarbonate and 3,5-di(4-pyridinyl)benzoic
acid 2 (F) and OXA-48 with C labeled bicarbonate and no fragment (G). Two unassigned
signals were observed at 164.1 ppm, and are believed to originate in a second carboxylated
site of OXA-48.
13
439
440
441
442
443
3.3
Inhibitor activity and structural analysis of 3,5-disubstituted benzoic acids.
In an attempted to design more potent inhibitors from our fragments, the mono-substituted
benzoic acids were evaluated for a merging approach (Fig. 8). By overlaying X-ray structures,
promising combinations showing orthogonal binding poses were identified and some of the
combined structures were prepared and evaluated with good results.
444
445
446
Figure 8: Strategy for substitution of the Ar1 and Ar2 groups in the focused fragment library of
3-substituted benzoic acids analogues.
447
448
449
450
451
452
An overlay of fragment 21a as well as 26b with several R2 binders identified the
combinations of fragments 21a/28, 21a/1 and 26b/35 as interesting partners (Fig. 9). The
combination 21a/1 and 21a/28 were synthetically feasible and gave compounds 39 and 40
(Scheme 2), respectively. In addition, the symmetrical 3,5-disubstituted benzoic acids 36–38
representing the symmetrical combinations of fragments 21a, 21b and 28 were included in
this study (Scheme 1).
-16-

ACCEPTED MANUSCRIPT
453
454
455
456
Figure 9: Superimpositions of the binding poses observed for 21a/28 (A), 21a/1 (B, 1: PDB-
ID:5dva) and 26b/35 (C) showing some of the possible combinations for 3,5-disubstituted
benzoic acids.
457
458
459
460
461
462
463
464
465
466
467
468
469
470
The 3,5-disubstituted compounds 36–40 were evaluated for their inhibitory activity against
OXA-48 as measured by their IC50, Kd and LE and complex structures with OXA-48 and
compounds 36, 38 and 40 were obtained (Table 2). The merged compounds 37, 38 and 39
(IC50 (µM)/LE: 110/0.19, 48/0.21, 100/0.22) failed to adequately maintain the binding
interactions as the IC50 values were at a similar level as the corresponding mono-substituted
fragments 28, 1 and 21a (IC50 (µM)/LE: 240/0.33, 250/0.32 and 35/0.33). When comparing
the IC50 values of compounds 36, 37 and 40 (IC50 (µM)/LE: 2.9/0.27, 48/0.21 and 2.9/0.27)
with the corresponding fragments 21a, 21b and 28 (IC50 (µM)/LE: 35/0.33, 450/0.26,
240/0.3), a 10-fold decrease of the IC50 value was observed. Nevertheless, the improved
binding was associated with a decrease in LE showing that the fragment-enzyme interactions
are less efficient with the merged compounds. The reduction in LE probably relates to the
ridged structure of the merged compounds allowing for little conformational freedom.
Overall, the strongest inhibitors in this study are compounds 36 and 40 with IC50 values of
2.9 µM and LE of 0.27.
471
472
Table 2. Inhibitor activities of 3,5-disubstituted benzoic acids analogues against OXA-48 (IC50,
KD and LE).
IC50
(μM) (μM)
KD
Ar1
Ar2
ID
LEa
36*
2.9
48
20
70
0.27
37
0.21
38*
39
110
100
70
70
0.19
0.22
40*
2.9
49
0.27
-17-

ACCEPTED MANUSCRIPT
473
* X-ray structure of fragment-enzyme complex available. a LE = − logꢀꢁ ICꢂꢁ/HeavyAtomCount
474
475
476
477
478
479
480
Figure 10: Compound 36 maintained the interaction with Arg214 as we observed for the 3-
substituted benzoic acid derivate. The IC50-value (A) was determined to be 2.9 µM, while the
KD was found to be 30 µM (B). For the higher concentrations of compound 36 some unspecific
binding was observed. The crystal structure of the complex OXA-48:36 with an omit-type
polder-map (2.5σ) (C) and its 2D-representation (D) shows one of the acetamide-groups
interacted with the guanidine group of Arg214, while the other group was solvent exposed.
481
482
483
484
485
486
487
The structural analysis of the OXA-48 complexes with 36, 38 and 40 showed that the
interaction of the carboxylic acid with Arg214 is maintained. For compound 36, a near
perfect overlay was obtained with the complex structure of fragment 21a showing that all
interactions seen with the fragments were preserved in the larger compound (Fig. 10). The
second 3-N-acetamidophenyl group forms a not previously observed hydrogen bond with
Ser244. In the SPR sensorgrams some concentration dependent aggregation was observed.
(30)
488
489
490
491
492
Interestingly, the conformation of compound 38 in complex with OXA-48 was changed
compared with the mono-substituted fragment 28. In the OXA-48:38 complex, one
quinolinyl group bound in the R1 site similar to fragment 21a. The other quinolinyl group
positions itself in a conformation similar to the alternative conformation observed with
fragment 35 (Fig. 6). No specific interactions were observed, but this conformation shielded
-18-

ACCEPTED MANUSCRIPT
493
494
the hydrophobic quinoline ring from solvent exposure by burying the compound deep in the
hydrophobic cleft.
495
496
497
498
The complex structure of the unsymmetrical compound 40 (Fig. 11) that was composed of
the quinoline ring of fragment 28 and the 3-N-acetamidophenyl substituent of fragment 13a
shared the key interactions of both mono-substituted fragments validating our approach,
with an IC50 of 3 µM.
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
4
Conclusion
A targeted fragment library consisting of 49 diversely 3-substituted benzoic acid derivatives
was prepared and biochemically analysed for their inhibitory activity against OXA-48.
Enzyme-fragment complexes for crystallographic studies were obtained for 33 fragments. By
systematically changing the substituent-groups of the benzoic acid derivatives we were able
to identify inhibitory fragments with IC50 < 40 µM (21a, 26b, 35). Based on the structural
information, fragments could be classified according to their preferred binding pocket. Most
fragments were orientated towards the R2 site induced by a π-π-stacking with Tyr221.
Unfortunately, no further interactions in the R2 site could be identified from our library. The
strongest binding fragments 21a and 26b were binding in the R1 site due to a hydrogen bond
to Arg214 and for fragment 35 a hydrogen bond to the amide backbone of Tyr117 was
observed. By overlaying the complex crystal structures of fragments 1, 21a, 26b, 28 and 35,
the design of five new 3,5-disubstituted inhibitors evolved. The strongest 3,5-disubstituted
inhibitors 36 and 40 showed IC50 values as low as 2.9 µM, thus have improved inhibitory
potential. The complex crystal structures of 36 and 40 revealed that the interactions of the
individual fragments were mainly retained in the merged structures. In addition, for inhibitor
36 a previously not observed hydrogen bond from the 3-N-acetamidophenyl group in the R2
site to Ser244 was found, which is interesting as we otherwise found few interaction in this
region. Future work will focus on the evaluation of fragments with increased flexibility e.g.
by introducing a CH2 or heteroatom linker bridging the aromatic ring systems to further
explore the active site.
-19-

ACCEPTED MANUSCRIPT
520
521
522
523
524
525
526
Figure 11: Compound 40 maintained the interaction with Arg214 as we observed for the 3-
substituted benzoic acid derivate. The IC50-value (A) was determined to be 2.9 µM, while the
KD was found to be 49 µM (B). The crystal structure of the complex OXA-48:40 with an omit-
type polder-map (2.5σ) (C) and its 2D-representation (D) shows that the acetamide-group
interacted with the guanidine group of Arg214, while the quinoline-ring was partially solvent
exposed.
527
528
5
Experimental
5.1
5.1.1
Synthesis
529
530
Synthesis of 3-substituted benzoic acids (complete data for all procedures and compounds is
found in the Supporting Information)
531
532
533
534
535
536
537
538
539
540
5.1.1.1
General procedure A – Aqueous conditions:
The halo aryl (1.0 equiv) was dissolved in a mixture of water:dioxane (1:1). The boronic acid
or ester (1.5 equiv) and potassium phosphate (5.0 equiv) were added. The solution was
degassed by vacuum/Argon cycles (10 times) before addition of PdCl2(PPh3)2 (10 mol%) and
further degassed (5 times). The resulting mixture was stirred at 95 °C under argon
atmosphere for 16-20 hours. The reaction mixture was filtered through Celite and diluted
with water (approx. 30 mL) before washing with chloroform (3 x 30 mL). If not stated
otherwise, the aqueous phase was concentrated under reduced pressure and applied to a
C18 precolumn before purification on a 10 g or 60 g C18 column with a gradient of
acetonitrile in water (10-100%) to yield the desired product.
-20-

ACCEPTED MANUSCRIPT
541
542
543
544
545
546
547
548
549
550
5.1.1.2
General procedure B – Anhydrous conditions:
The halo aryl (1.0 equiv) was dissolved in anhydrous THF. The aryl boronic acid or aryl
boronic ester (1.5 equiv) and inorganic base (5.0 equiv) were added. The solution was
degassed by vacuum/Argon cycles (10 times), before addition of a palladium catalyst (10
mol%) and further degassed (5 times). The resulting mixture was stirred at 75–90 °C under
an inert atmosphere for 16-20 hours. The reaction mixture was filtered through Celite and
diluted with water (approx. 30 mL) before washing with ethyl acetate (3 x 30 mL). If not
stated otherwise, the aqueous phase was concentrated under reduced pressure and applied
to a C18 precolumn before purification on a 10 g or 60 g C18 column with a gradient of
acetonitrile in water (10–80%) to yield the desired molecule.
551
5.1.2
Screening of catalysts (for results see Table SI1)
552
553
554
555
556
557
558
559
560
561
562
563
5.1.2.1
General procedure:
3-Bromo-5-iodobenzoic acid (0.03–0.06 mmol, 1.0 equiv.) was dissolved in the indicated
solvent (0.5–1 mL/0.01 mmol substrate). The boronic acid or ester (1.5 equiv.) and base (5.0
equiv.) were added. The solution was degassed by vacuum/Ar cycles (10 times) before
addition of the palladium catalyst and further degassed (5 times). The resulting mixture was
stirred at the indicated temperature under an inert atmosphere for the indicated reaction
time. The crude reaction mixture was analysed by HRMS to determine the ratio of int-39 :
disubstituted 38 : starting material. The reaction mixture was filtered through Celite bed and
diluted with water (approx. 30 mL) before washing with chloroform (3 x 30 mL). The aqueous
phase was concentrated under reduced pressure and applied to a C18 precolumn before
purification on a 60 g C18 column with a gradient of acetonitrile in water (0–5% over 15 min)
to yield the product.
564
5.1.3
Synthesis of symmetrical 3,5-disubstituted benzoic acid derivatives
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
5.1.3.1
3,5-Di(3-acetamidophenyl)benzoic acid 36:
3-Bromo-5-iodobenzoic acid (0.30 mmol, 100 mg, 1.0 equiv), 3-acetamidophenylboronic acid
(0.45 mmol, 816 mg, 1.5 equiv), potassium phosphate (1.5 mmol, 324 mg, 5.0 equiv) were
dissolved in a mixture of water/dioxane (1:1). The solution was degassed by vacuum/Ar
cycles (10 times) before addition of Pd2(dba)3•CHCl3 (15 mg, 5 mol%), and XPhos (7.2 mg, 5
mol%) and further degassed (5 times). The resulting mixture was stirred at 60 °C for 20–24
hours. The reaction mixture was filtered through Celite bed and diluted with water (approx.
30 mL) before washing with chloroform (3 x 30 mL). The aqueous phase was concentrated
under reduced pressure and applied to a C18 precolumn before purification on a 60 g C18
column with a gradient of acetonitrile in water (0–5% over 15 min) to provide 36 (60 mg,
54%) as white powder. 1H NMR (400 MHz, methanol-d4) δ 8.21 (s, 2H), 7.90 (t, J = 1.7 Hz, 1H),
7.81 (t, J = 1.7 Hz , 2H), 7.68 (d, J = 8 Hz, 2H), 7.43 (s, 1H), 7.49-7.46 (m, 2H), 7.43-7.39 (m,
2H), 2.16 (s, 6H). 13C NMR (101 MHz, methanol-d4) δ 175.0, 171.8, 142.9, 142.3, 140.5, 132.2,
130.4, 128.2, 128.1, 123.9, 120.3, 119.7, 24.0. HRMS (ESI): Calcd. for C23H19N2O4 [M-H]-
387.1350; found 387.1342. UPLC: purity = 97.5 %
-21-

ACCEPTED MANUSCRIPT
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
5.1.3.2
3,5-di(4-acetamidophenyl)benzoic acid 37:
3,5-Dibromobenzoic acid (1.01 mmol, 300 mg, 1.0 equiv), 3-acetamidophenylboronic acid
(0.81 mmol, 178 mg, 0.75 equiv), potassium phosphate (3.76 mmol, 0.80 g, 3.5 equiv) and
PdCl2(PPh3)2 (0.11 mmol, 77 mg, 10 mol%) were stirred in a mixture of water/dioxane (1:1)
for 24 hours at 95 °C under argon atmosphere. The crude reaction mixture was filtered
through Celite and diluted with water (approx. 30 mL) before washing with chloroform (3 x
30 mL). The aqueous phase was concentrated under reduced pressure and applied to a C18
precolumn before purification on a 60 g C18 column with a gradient of acetonitrile in water
(0–100 % over 12 minutes). The fractions were analysed by MS and fractions containing 37
were combined. The product was purified by reverse-phase automated flash
chromatography before being subjected to purification by HPLC, to yield 37 (0.09 mmol, 34
mg, 11%) as a white solid. 1H NMR (400 MHz, methanol-d4) δ 8.24 (s, 2H), 7.98 (d, J = 7.8 Hz,
2H), 7.85 (d, J = 7.9 Hz, 2H), 7.68-7.66 (m, 2H), 7.63-7.60 (m, 2H), 7.57-7.53 (m, 1H), 2.16 (s,
6H). 13C NMR (101 MHz, methanol-d4) δ 175.2, 171.7, 142.0, 140.2, 139.4, 137.9, 131.7,
128.4, 128.2, 127.6, 127.4, 123.3, 121.4, 116.2, 23.9. HRMS (ESI): Calcd. for C23H19N2O4 [M-
H]- 387.1350; found 387.1340. UPLC: purity >99.5 %
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
5.1.3.3
3,5-diquinolin-6-ylbenzoic acid 38:
3,5-Dibromobenzoic acid (0.11 mmol, 33 mg, 1.0 equiv), 6-quinolinylboronic acid pinacol
ester (0.23 mmol, 60 mg, 2.0 equiv), potassium phosphate (0.58 mmol, 125 mg, 5.0 equiv)
were dissolved in tert-butanol. The solution was degassed by vacuum/Ar cycles (10 times)
before addition of XPhos-Pd G2 (5 mol%, 5 mg) and further degassed (5 times). The resulting
mixture was stirred at 60 °C for 20–24 hours. The reaction mixture was filtered through
Celite bed and diluted with water (approx. 30 mL) before washing with chloroform (3 x 30
mL). The aqueous phase was concentrated under reduced pressure and applied to a C18
precolumn before purification by C18 RP flash chromatography with a gradient of
acetonitrile in water (0–5% over 15 min) to yield 38 (0.08 mmol, 29 mg, 65%) as white
powder. 1H NMR (400 MHz, methanol-d4) δ 8.87-8.86 (m, 2H), 8.52 (s, 1H), 8.50 (s, 1H), 8.46
(m, 2H), 8.38 (m, 2H), 8.29-8.26 (m, 3H), 8.18 (s, 1H), 8.16 (s, 1H), 7.61-7.58 (dd, J = 8.3, 4.2
Hz, 2H). 13C NMR (101 MHz, methanol-d4) δ 174.4, 151.1, 148.0, 141.5, 140.5, 138.6, 130.6,
130.1, 129.5, 128.7, 126.9, 122.8. HRMS (ESI): Calcd. for C25H15N2O2 [M-H]– 375.1139; found
375.1133. UPLC: purity = 99.1 %
611
5.1.4
Synthesis of unsymmetrical 3,5-disubstituted benzoic acid derivatives
612
613
614
5.1.4.1
3-(3'-Acetamidophenyl)-5-pyridin-4-ylbenzoic acid 39: attempted synthesis from 3,5-
dibromobenzoic acid
3,5-Dibromobenzoic acid (1.01 mmol, 300 mg, 1.0 equiv), 3-acetamidophenylboronic acid
(0.81 mmol, 178 mg, 0.75 equiv), potassium phosphate (3.76 mmol, 0.80 g, 3.5 equiv) and
PdCl2(PPh3)2 (0.11 mmol, 77 mg, 10 mol%) were stirred in a mixture of water/dioxane (1:1)
for 24 hours at 95 °C under argon atmosphere. The crude reaction mixture was filtered
through Celite and diluted with water (approx. 30 mL) before washing with chloroform (3 x
30 mL). The aqueous phase was concentrated under reduced pressure and applied to a C18
precolumn before purification by C18 RP flash chromatography with a gradient of
acetonitrile in water (10–100 % over 12 minutes). The fractions were analysed by MS and
615
616
617
618
619
620
621
-22-

ACCEPTED MANUSCRIPT
622
623
624
625
626
627
628
629
fractions containing int-39 were combined and reacted with pyridin-4-ylboronic acid (0.97
mmol, 119 mg, 1.2 equiv), potassium phosphate (4.05 mmol, 0.86 g, 5.0 equiv) and
PdCl2(PPh3)2 (0.08 mmol, 56 mg, 10 mol%). The product was purified by reverse-phase
automated flash chromatography before being subjected to purification by HPLC, to yield 39
1
(0.12 mmol, 39 mg, 15%) as a white solid. H NMR (400 MHz, methanol-d4) δ 8.22 (s, 1H),
7.92 (d, J = 7.6 Hz, 1H), 7.76 (s, 2H), 7.68-7.60 (m, 3H), 7.46-7.33 (m, 4H), 2.14 (s, 3H). 13C
NMR (101 MHz, methanol-d4) δ 175.3, 171.7, 143.0, 141.5, 140.4, 139.8, 130.3, 129.7, 129.3,
129.3, 128.9, 123.7, 120.1, 119.6, 23.9. UPLC: purity = 97.9%
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
5.1.4.2
3-Bromo-5-(quinolin-6-yl) benzoic acid int-40:
3-Bromo-5-iodobenzoic acid (0.15 mmol, 50 mg, 1.0 equiv), 6-quinolinylboronic acid pinacol
ester (0.22 mmol, 58 mg, 1.5 equiv) and potassium phosphate (0.76 mmol, 162 mg, 5.0
equiv) were dissolved in a mixture of water/dioxane (1:1). The solution was degassed by
vacuum/Ar cycles (10 times) before addition of Pd2(dba)3.CHCl3 (5 mol%, 7.5 mg), and SPhos
(5 mol%, 3.1 mg) and further degassed (5 times). The resulting mixture was stirred at 60 °C
for 20–24 hours. The reaction mixture was filtered through a Celite bed and diluted with
water (approx. 30 mL) before washing with chloroform (3 x 30 mL). The aqueous phase was
concentrated under reduced pressure and applied to a C18 precolumn before purification on
a 60 g C18 column with a gradient of acetonitrile in water (0–5% over 20 min). Product int-
40 (0.07 mmol, 23 mg, 45%) was obtained as a white powder. 1H NMR (400 MHz, methanol-
d4) δ 8.92-8.91 (m,1H), 8.49-8.46 (m, 1H), 8.35 (s, 1H), 8.28 (s, 2H), 8.10 (s, 2H), 8.02-8.01 (m,
1H), 7.97-7.96 (m,1H), 7.59-7.56 (dd, J = 8.3, 4.2 Hz, 1H). 13C NMR (101 MHz, DMSO-d6) δ
166.6, 150.8, 147.2, 143.6, 140.6, 136.8, 136.5, 131.7, 131.1, 129.6, 128.5, 128.2, 127.4,
126.5, 125.8, 121.9, 121.7; HRMS (ESI): Calcd. for C16H979BrNO2 [M-H]– 325.9822; found
325.9822.
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
5.1.4.3
3-(3'-Acetamidophenyl)-5-quinolin-6-ylbenzoic acid 40:
3-Bromo-5-(quinolin-6-yl) benzoic acid int-40 (0.039 mmol, 13 mg, 1.0 equiv), 3-
acetamidophenylboronic acid (0.55 mmol, 10 mg, 1.5 equiv) and potassium phosphate (0.20
mmol, 0.42 g, 5.0 equiv) were dissolved in tert-butanol. The solution was degassed by
vacuum/Ar cycles (10 times) before addition of Xphos-Pd G2 (5 mol%, 1.5 mg) and further
degassed (5 times). The resulting mixture was stirred at 60 °C for 20–24 hours. The reaction
mixture was filtered through Celite bed and diluted with water (approx. 30 mL) before
washing with chloroform (3 x 30 mL). The aqueous phase was concentrated under reduced
pressure and applied to a C18 precolumn before purification on a 60 g C18 column with a
gradient of acetonitrile in water (0–5% over 20 min). Product 40 (0.023 mmol, 9 mg, 90%)
1
was obtained as white powder. H NMR (400 MHz, methanol-d4) δ 8.87-8.83 (m, 1H), 8.56-
8.45 (m, 1H), 8.41-8.39 (m, 1H), 8.35-8.20 (m, 3H), 8.18-8.11 (m, 1H), 8.08 (t, J = 1.8 Hz, 1H),
7.87-7.86 (m, 1H), 7.72-7.68 (m, 1H), 7.62-7.56 (m, 1H), 7.56-7.49 (m, 1H), 7.46-7.42 (m, 1H),
2.17 (s, 3H). 13C NMR (101 MHz, DMSO-d6) δ 174.7, 171.8, 151.2, 148.2, 142.8, 142.5, 141.4,
140.8, 140.7, 140.5, 138.8, 130.8, 130.4, 130.3, 129.7, 128.6, 128.5, 128.5, 127.0, 123.9,
123.0, 120.3, 119.7, 23.9. HRMS (ESI): Calcd. for C24H18N2O3 [M-H]- 381.1245; found
381.1243.UPLC: purity = 96.4 %
-23-

ACCEPTED MANUSCRIPT
663
664
665
666
5.2
Protein production
For the biochemical assay OXA-48 was expressed with the native signal-peptide and purified
from the periplasm as described earlier.(31) For surface plasmon resonance assays, nuclear
magnetic resonance and crystallization a His-tagged construct was used.(19)
667
668
669
670
671
672
673
674
675
5.3
Biochemical assay
All experiments were performed using a Spectramax M2e at 25 °C in 100 mM sodium
phosphate (pH 7.0) supplemented with 50 mM NaHCO3 and 0.2 mg/ml bovine serum
albumin (BSA). Velocities from the linear range were determined in the SoftMax Pro
software (Molecular Devices). All experiments were done with a sample volume of 100 μL.
IC50 values were determined for all compounds in competition with 25 µM of the
chromogenic substrate nitrocefin. The log10 of the inhibitor concentrations to the response
with bottom and top constant based on controls were fitted nonlinearly in GraphPad Prism 6
(GraphPad Software) to determine the IC50 value.
676
677
678
679
680
681
682
5.4
Surface plasmon resonance
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.
The final running buffer
included 50 mM HEPES pH 7.0, 50 mM K2SO4, 0.5% Tween-20, 50 mM NaHCO3, and 2.5%
DMSO. The enzyme, tOXA-48, 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 CM5 chip using standard amine coupling.
683
684
685
686
687
688
689
Compounds were tested with 10 dilutions from 400 µM to 10.5 µM, with 30 s injection and
60 s dissociation time. Compounds exhibiting kinetic behavior had the dissociation time
extended to 300 s. Seven startup cycles with buffer were performed. Solvent correction was
performed every 48th cycle and a positive control was included every 24th cycle with 3.5-Di(4-
pyridinyl)benzoic acid as the control (19). Affinities were calculated from the steady-state
affinity model with a constant Rmax adjusted by the control and the molecular weight of the
compound.
690
691
692
693
694
695
696
697
698
699
5.5
13C nuclear magnetic resonance
A solution of NaH13CO3 in D2O (50 mM) was prepared. The NaH13CO3/D2O-mixture was
added to 1 mM OXA-48 in 50 mM sodium phosphate and 50 mM sodium bicarbonate pH 6.5
in a 1 : 9 ratio of bicarbonate to enzyme. Compounds were diluted from a 150 mM stock
solution in 100% DMSO to a final concentration of 3.75 mM (2.5% DMSO). Sample volumes
of 500 µL were used. We performed the experiment at 37 °C with a Bruker Avance III HD
with an inverse detected TCI probe with cryogenic enhancement for H, 13C and H,
operating at 599.90 MHz for protons and 150.86 MHz for carbon. 10 000 scans at 30° pulse
angle with 2 s relaxation delay were collected using 1D 13C NMR with power-gated
decoupling of protons (zgpg30 using waltz16).
1
2
700
701
702
703
5.6
Crystallization and data processing
Crystals of OXA-48 was grown from hanging drops containing 0.1 M HEPES pH 7.5, 8-11%
PEG 8000 and 4-8% 1-butanol as previously described.(17) Compounds were diluted to
3.75mM in the cryo solution with 0.1M HEPES pH 7.5, 10% PEG 8000, 5% 1-butanol, and 25%
-24-

ACCEPTED MANUSCRIPT
704
705
ethanediol, usually overnight. The exception was the crystal soaked in imipenem. Imipenem
was added to saturation in the cryosolution, and the crystal was just given a quick soak.
706
707
708
709
710
711
712
713
Crystals were flash cooled in liquid nitrogen. X-ray diffraction data were collected at BL 14.1
and BL14.2 at BESSY (Berlin, Germany) (32) and at ID23 and ID30 at ESRF (Grenoble, France).
In most cases the structures were solved by refining against the protein-atoms of previous
structures (P212121 PDB ID: 5DVA and P21 PDB ID: 5DTK), but in cases where the unit cells
were to different PHASER was used with chain A from PDB ID: 5dtk as the search model for
molecular replacement. In most cases images were collected autoprocessed using the tools
at the beamlines,(33-37) but in some cases we found it useful to reprocess using DIALS or
XDS together with AIMLESS.(38-40)
714
715
716
717
718
719
The compounds were built into difference density maps after initial refinement in
phenix.refine,(41) with waters deleted from the active site. Restraints for the compounds
were prepared using the GRADE Web Server.(42) Omit maps were calculated using the
phenix.polder-tool which excludes bulk-solvent from the volume surrounding the ligand.(43)
Figures were made using PyMOL.(44) Ligand-interaction diagrams were prepared using the
Maestro-suite from Schrödinger Release 2016-3 (Schrödinger, LLC, New York).
720
Acknowledgement:
721
722
723
724
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, and the MX beamlines at the
European Radiation Facility (ESRF), Grenoble, France are highly valued.
725
PDB accession codes:
726
727
728
729
730
731
732
Coordinates and structure factors for all OXA-48 complexes are deposited in the Protein
Data Bank. Accession numbers are listed with reference to the complexed compound. PDB
IDs: imipenem: 5QB4; 3a: 5QA4; 3b: 5QA5; 4a: 5QA6; 4b: 5QA7; 4c: 5QA8; 5: 5QA9; 6a:
5QAA; 6b: 5QAB; 6c: 5QAC; 8a: 5QAD; 8b: 5QAE; 8c: 5QAF;9a: 5QAG; 9b: 5QAH; 12a: 5QAI:
13: 5QAJ; 14: 5QAK; 11b: 5QAL; 17: 5QAM; 19a: 5QAN; 19b: 5QAO; 21a: 5QAP; 21b: 5QAQ;
23a: 5QAR; 23b: 5QAS; 24: 5QAT; 26a: 5QAU; 26b: 5QAV; 27: 5QAW; 28: 5QAX; 32: 5QAY;
34: 5QAZ; 35: 5QB0; 36: 5QB1; 38: 5QB2; 40: 5QB3.
733
Supplementary material:
734
735
736
Supplementary material containing synthetic procedures and analytical data for all
compounds and biophysical, biochemical and structural analysis of OXA-48:compound
complexes.
737
References:
738
739
740
1.
Guillard, T., Pons, S., Roux, D., Pier, G. B., and Skurnik, D. (2016) Antibiotic resistance and
virulence: Understanding the link and its consequences for prophylaxis and therapy.
Bioessays 38, 682-693
-25-

ACCEPTED MANUSCRIPT
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
2.
Chen, L., Todd, R., Kiehlbauch, J., Walters, M., and Kallen, A. (2017) Notes from the Field: Pan-
Resistant New Delhi Metallo-β-Lactamase-Producing Klebsiella pneumoniae - Washoe
County, Nevada, 2016. MMWR Morb. Mortal. Wkly. Rep. 66, 33-33
O’Neill, J. (2016) Tackling Drug-Resistant Infections Globally: final report and
recommendations. Review on Antimicrobial Resistance, London, UK
Bush, K., and Bradford, P. A. (2016) β-Lactams and β-Lactamase Inhibitors: An Overview. CSH
Perspect. Med. 6
Hall, B. G., and Barlow, M. (2004) Evolution of the serine β-lactamases: past, present and
future. Drug Resist. Update. 7, 111-123
Poirel, L., Naas, T., and Nordmann, P. (2010) Diversity, Epidemiology, and Genetics of Class D
β-Lactamases. Antimicrob. Agents Chemother. 54, 24-38
Bush, K., and Jacoby, G. A. (2010) Updated functional classification of β-lactamases.
Antimicrob. Agents Chemother. 54, 969-976
Naas, T., Oueslati, S., Bonnin, R. A., Dabos, M. L., Zavala, A., Dortet, L., Retailleau, P., and
Iorga, B. I. (2017) Beta-lactamase database (BLDB) – structure and function. Journal of
Enzyme Inhibition and Medicinal Chemistry 32, 917-919
Ambler, R. P. (1980) The structure of β-lactamases. Philos. Trans. R. Soc., B 289, 321-331
Golemi, D., Maveyraud, L., Vakulenko, S., Samama, J.-P., and Mobashery, S. (2001) Critical
involvement of a carbamylated lysine in catalytic function of class D β-lactamases. Proc. Natl.
Acad. Sci. U. S. A. 98, 14280-14285
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Poirel, L., Potron, A., and Nordmann, P. (2012) OXA-48-like carbapenemases: the phantom
menace. J. Antimicrob. Chemoth. 67, 1597-1606
Drawz, S. M., and Bonomo, R. a. (2010) Three decades of β-lactamase inhibitors. Clin.
Microbiol. Rev. 23, 160-201
Buynak, J. D. (2006) Understanding the longevity of the β-lactam antibiotics and of
antibiotic/β-lactamase inhibitor combinations. Biochem. Pharmacol. 71, 930-940
Antunes, N., and Fisher, J. (2014) Acquired Class D β-Lactamases. Antibiotics 3, 398
Liscio, J. L., Mahoney, M. V., and Hirsch, E. B. (2015) Ceftolozane/tazobactam and
ceftazidime/avibactam: two novel β-lactam/β-lactamase inhibitor combination agents for the
treatment of resistant Gram-negative bacterial infections. Int. J. Antimicrob. Agents 46, 266-
271
Ehmann, D. E., Jahić, H., Ross, P. L., Gu, R.-F., Hu, J., Durand-Réville, T. F., Lahiri, S., Thresher,
J., Livchak, S., Gao, N., Palmer, T., Walkup, G. K., and Fisher, S. L. (2013) Kinetics of Avibactam
Inhibition against Class A, C, and D β-Lactamases. J. Bio. Chem. 288, 27960-27971
Lahiri, S. D., Mangani, S., Jahic, H., Benvenuti, M., Durand-Reville, T. F., De Luca, F., Ehmann,
D. E., Rossolini, G. M., Alm, R. A., and Docquier, J. D. (2015) Molecular Basis of Selective
Inhibition and Slow Reversibility of Avibactam against Class D Carbapenemases: A Structure-
Guided Study of OXA-24 and OXA-48. ACS Chem. Biol. 10, 591-600
Shields, R. K., Chen, L., Cheng, S. J., Chavda, K. D., Press, E. G., Snyder, A., Pandey, R., Doi, Y.,
Kreiswirth, B. N., Nguyen, M. H., and Clancy, C. J. (2017) Emergence of Ceftazidime-
Avibactam Resistance Due to Plasmid-Borne bla(KPC-3) Mutations during Treatment of
Carbapenem-Resistant Klebsiella pneumoniae Infections. Antimicrob. Agents Chemother. 61
Lund, B. A., Christopeit, T., Guttormsen, Y., Bayer, A., and Leiros, H. K. S. (2016) Screening and
Design of Inhibitor Scaffolds for the Antibiotic Resistance Oxacillinase-48 (OXA-48) through
Surface Plasmon Resonance Screening. J. Med. Chem. 59, 5542-5554
14.
15.
16.
17.
18.
19.
20.
21.
Congreve, M., Carr, R., Murray, C., and Jhoti, H. (2003) A ‘Rule of Three' for fragment-based
lead discovery? Drug Discov. Today 8, 876--877
Lukyanov, S. M., Bliznets, I. V., Shorshnev, S. V., Aleksandrov, G. G., Stepanov, A. E., and
Vasil'ev, A. A. (2006) Microwave-assisted synthesis and transformations of sterically hindered
3-(5-tetrazolyl)pyridines. Tetrahedron 62, 1849-1863
22.
King, A. M., King, D. T., French, S., Brouillette, E., Asli, A., Alexander, J. A. N., Vuckovic, M.,
Maiti, S. N., Parr, T. R., Brown, E. D., Malouin, F., Strynadka, N. C. J., and Wright, G. D. (2016)
-26-

ACCEPTED MANUSCRIPT
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
Structural and Kinetic Characterization of Diazabicyclooctanes as Dual Inhibitors of Both
Serine-β-Lactamases and Penicillin-Binding Proteins. ACS Chem. Biol. 11, 864-868
Leonard, D. A., Bonomo, R. A., and Powers, R. A. (2013) Class D β-Lactamases: A Reappraisal
after Five Decades. Acc. Chem. Res. 46, 2407-2415
Maveyraud, L., Golemi-Kotra, D., Ishiwata, A., Meroueh, O., Mobashery, S., and Samama, J.-P.
(2002) High-Resolution X-ray Structure of an Acyl-Enzyme Species for the Class D OXA-10 β-
Lactamase. J. Am. Chem. Soc. 124, 2461-2465
Cahill, S. T., Cain, R., Wang, D. Y., Lohans, C. T., Wareham, D. W., Oswin, H. P., Mohammed, J.,
Spencer, J., Fishwick, C. W. G., McDonough, M. A., Schofield, C. J., and Brem, J. (2017) Cyclic
Boronates Inhibit All Classes of β-Lactamases. Antimicrob. Agents Chemother. 61
Li, J., Cross, J. B., Vreven, T., Meroueh, S. O., Mobashery, S., and Schlegel, H. B. (2005) Lysine
carboxylation in proteins: OXA-10 β-lactamase. Proteins 61, 246-257
Verma, V., Testero, S. A., Amini, K., Wei, W., Liu, J., Balachandran, N., Monoharan, T., Stynes,
S., Kotra, L. P., and Golemi-Kotra, D. (2011) Hydrolytic Mechanism of OXA-58 Enzyme, a
Carbapenem-hydrolyzing Class D β-Lactamase from Acinetobacter baumannii. J. Bio. Chem.
286, 37292-37303
Lohans, C. T., Wang, D. Y., Jorgensen, C., Cahill, S. T., Clifton, I. J., McDonough, M. A., Oswin,
H. P., Spencer, J., Domene, C., Claridge, T. D. W., Brem, J., and Schofield, C. J. (2017) 13C-
Carbamylation as a Mechanistic Probe for the Inhibition of Class D β-Lactamases by
Avibactam and Halide Ions. Org. Biomol. Chem., just accepted
Asakawa, N., Kuroki, S., Kurosu, H., Ando, I., Shoji, A., and Ozaki, T. (1992) Hydrogen-bonding
effect on carbon-13 NMR chemical shifts of L-alanine residue carbonyl carbons of peptides in
the solid state. J. Am. Chem. Soc. 114, 3261-3265
Giannetti, A. M., Koch, B. D., and Browner, M. F. (2008) Surface plasmon resonance based
assay for the detection and characterization of promiscuous inhibitors. J. Med. Chem. 51,
574-580
Lund, B. A., Leiros, H. K., and Bjerga, G. E. (2014) A high-throughput, restriction-free cloning
and screening strategy based on ccdB-gene replacement. Microb. Cell Fact. 13, 38
Mueller, U., Förster, R., Hellmig, M., Huschmann, F. U., Kastner, A., Malecki, P., Pühringer, S.,
Röwer, M., Sparta, K., Steffien, M., Ühlein, M., Wilk, P., and Weiss, M. S. (2015) The
macromolecular crystallography beamlines at BESSY II of the Helmholtz-Zentrum Berlin:
Current status and perspectives. Eur. Phys. J. Plus 130, 141
Gabadinho, J., Beteva, A., Guijarro, M., Rey-Bakaikoa, V., Spruce, D., Bowler, M. W.,
Brockhauser, S., Flot, D., Gordon, E. J., Hall, D. R., Lavault, B., McCarthy, A. A., McCarthy, J.,
Mitchell, E., Monaco, S., Mueller-Dieckmann, C., Nurizzo, D., Ravelli, R. B. G., Thibault, X.,
Walsh, M. A., Leonard, G. A., and McSweeney, S. M. (2010) MxCuBE: a synchrotron beamline
control environment customized for rnacromolecular crystallography experiments. J.
Synchrotron. Radiat. 17, 700-707
Incardona, M. F., Bourenkov, G. P., Levik, K., Pieritz, R. A., Popov, A. N., and Svensson, O.
(2009) EDNA: a framework for plugin-based applications applied to X-ray experiment online
data analysis. J. Synchrotron. Radiat. 16, 872-879
Delageniere, S., Brenchereau, P., Launer, L., Ashton, A. W., Leal, R., Veyrier, S., Gabadinho, J.,
Gordon, E. J., Jones, S. D., Levik, K. E., McSweeney, S. M., Monaco, S., Nanao, M., Spruce, D.,
Svensson, O., Walsh, M. A., and Leonard, G. A. (2011) ISPyB: an information management
system for synchrotron macromolecular crystallography. Bioinformatics 27, 3186-3192
Bourenkov, G. P., and Popov, A. N. (2010) Optimization of data collection taking radiation
damage into account. Acta Cryst. Section D 66, 409-419
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
Sparta, K. M., Krug, M., Heinemann, U., Mueller, U., and Weiss, M. S. (2016) XDSAPP2. 0. J.
Appl. Cryst. 49, 1085-1092
Waterman, D. G., Winter, G., Gildea, R. J., Parkhurst, J. M., Brewster, A. S., Sauter, N. K., and
Evans, G. (2016) Diffraction-geometry refinement in the DIALS framework. Acta Cryst. Section
D 72, 558-575
-27-

ACCEPTED MANUSCRIPT
845
846
847
848
849
850
851
852
853
854
855
856
39.
Evans, P. R., and Murshudov, G. N. (2013) How good are my data and what is the resolution?
Acta Cryst. Section D 69, 1204-1214
Kabsch, W. (2010) XDS. Acta Cryst. Section D 66, 125-132
Afonine, P. V., Grosse-Kunstleve, R. W., Echols, N., Headd, J. J., Moriarty, N. W.,
Mustyakimov, M., Terwilliger, T. C., Urzhumtsev, A., Zwart, P. H., and Adams, P. D. (2012)
Towards automated crystallographic structure refinement with phenix.refine. Acta Cryst.
Section D 68, 352-367
Smart, O. S., Womack, T. O., Sharff, A., Flensburg, C., Keller, P., Paciorek, W., Vonrhein, C.,
and Bricogne, G. (2014) grade, version 1.102. Global Phasing
40.
41.
42.
43.
44.
Liebschner, D. (2016) phenix.polder - A tool for calculating difference maps around atom
selections by excluding the bulk solvent mask. The Phenix Project, Berkeley, California
Schrodinger, LLC. (2015) The PyMOL Molecular Graphics System, Version 1.8.
857
-28-