In Silico Design and Enantioselective Synthesis of Functionalized Monocyclic 3-Amino-1-carboxymethyl-β-lactams as Inhibitors of Penicillin-Binding Proteins of Resistant Bacteria

Article abstract of DOI:10.1002/chem.201801868

As a complement to the renowned bicyclic β-lactam antibiotics, monocyclic analogues provide a breath of fresh air in the battle against resistant bacteria. In that framework, the present study discloses the in silico design and unprecedented ten-step synthesis of eleven nocardicin-like enantiomerically pure 2-{3-[2-(2-aminothiazol-4-yl)-2-(methoxyimino)acetamido]-2-oxoazetidin-1-yl}acetic acids starting from serine as a readily accessible precursor. The capability of this novel class of monocyclic 3-amino-β-lactams to inhibit penicillin-binding proteins (PBPs) of various (resistant) bacteria was assessed, revealing the potential of α-benzylidenecarboxylates as interesting leads in the pursuit of novel PBP inhibitors. No deactivation by representative enzymes belonging to the four β-lactamase classes was observed, while weak inhibition of class C β-lactamase P99 was demonstrated.

Full text of DOI:10.1002/chem.201801868

A Journal of
Accepted Article
Title: In silico design and enantioselective synthesis of functionalized
monocyclic 3-amino-1-carboxymethyl-β-lactams as inhibitors of
penicillin-binding proteins of resistant bacteria
Authors: Lena Decuyper, Sari Deketelaere, Lore Vanparys, Marko
Jukic, Izidor Sosic, Eric Sauvage, Ana Maria Amoroso, Olivier
Verlaine, Bernard Joris, Stanislav Gobec, and Matthias
D'hooghe
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201801868
Link to VoR: http://dx.doi.org/10.1002/chem.201801868
Supported by

10.1002/chem.201801868
Chemistry - A European Journal
In silico design and enantioselective synthesis of functionalized monocyclic
3-amino-1-carboxymethyl-β-lactams as inhibitors of penicillin-binding proteins
of resistant bacteria
Lena Decuyper,[a] Sari Deketelaere,[a] Lore Vanparys,[a] Marko Jukič,[b] Izidor Sosič,[b] Eric
Sauvage,[c] Ana Maria Amoroso,[c] Olivier Verlaine,[c] Bernard Joris,[c] Stanislav Gobec,[b]
Matthias D'hooghe*[a]
Abstract
As a complement to the renowned bicyclic β-lactam antibiotics, monocyclic analogues provide a breath
of fresh air in the battle against resistant bacteria. In that framework, the present study discloses the
in silico design and unprecedented ten-step synthesis of eleven nocardicin-like enantiomerically pure
2-{3-[2-(2-aminothiazol-4-yl)-2-(methoxyimino)acetamido]-2-oxoazetidin-1-yl}acetic acids starting
from serine as a readily accessible precursor. The capability of this novel class of monocyclic 3-amino-
β-lactams to inhibit penicillin-binding proteins (PBPs) of various (resistant) bacteria was assessed,
revealing the potential of -benzylidenecarboxylates as interesting leads in the pursuit of novel PBP
inhibitors. No deactivation by representative enzymes belonging to the four -lactamase classes was
observed, while weak inhibition of class C -lactamase P99 was demonstrated.
[a] L. Decuyper, S. Deketelaere, L. Vanparys, Prof. Dr. M. D’hooghe
SynBioC Research Group
Department of Green Chemistry and Technology
Faculty of Bioscience Engineering
Ghent University, Coupure Links 653, 9000 Ghent (Belgium)
E-mail: Matthias.Dhooghe@UGent.be
[b] Dr. M. Jukič, Dr. I. Sosič, Prof. Dr. S. Gobec
Department of Pharmaceutical Chemistry
Faculty of Pharmacy
University of Ljubljana, Aškerčeva 7, 1000 Ljubljana (Slovenia)
[c] Dr. E. Sauvage, Dr. A. M. Amoroso, O. Verlaine, Prof. Dr. B. Joris
Center for Protein Engineering
Faculty of Sciences
University of Liège, Quartier Agora, Allée du 6 Août 13, Bât B6a, 4000 Liège-Sart Tilman (Belgium)
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/chem.2018xxxxx.
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Introduction
β-Lactam antibiotics undoubtedly represent a cornerstone in global health care. Through their
effectiveness in inhibiting penicillin-binding proteins (PBPs), the D,D-transpeptidase enzymes that
catalyze the crosslinking of peptidoglycan (the major component of the bacterial cell wall), β-lactam
antibiotics have gained worldwide attention.[1] More specifically, by functional mimicry of the natural
substrate of these serine enzymes, i.e. the C-terminal D-alanyl-D-alanine moiety of the stem peptide,
a stable acyl-enzyme complex is formed with β-lactam antibiotics. These act as suicide inhibitors and
cause missteps in the cell wall biosynthesis, thus provoking growth inhibition or lysis.[2],[3] This powerful
mechanism has made penicillins 1 and their analogues the most widely used antibiotics for any
bacterial infectious disease over the past 70 years.[4] Disappointingly, however, β-lactam and other
antibacterial agents have been challenged by the propagation of drug-resistant bacterial strains,
developing alarmingly fast, while the pace of antibiotic discovery has dropped significantly.[5] In that
framework, the development of innovative chemical entities is of paramount importance in order to
effectively address and counteract microbial resistance. While some pathogens evade β-lactam action
by the production of β-lactamases, others impede the uptake of antibiotics via decreased permeability
of the outer membrane, or remove the antibiotics from the bacterium via efflux pumps. Finally, certain
Gram-positive bacteria produce drug-insensitive PBPs.[6] Via so-called active-site distortion, the active-
site size and hence the β-lactam acylation efficiency is dramatically reduced (e.g. PBP2x of
Streptococcus pneumoniae, PBP2a of Staphylococcus aureus (MRSA), PBP5fm of Enterococcus
faecium).[4],[7] In order to overcome this intrinsic low acylation susceptibility, we envisioned the design
of β-lactam compounds that are smaller than classic penicillins 1 and cephalosporins 2, but still retain
their acylation potential: monocyclic β-lactams. Indeed, as we demonstrated recently via a three-
decade literature overview concerning the structure-activity relationships of diverse classes of
monocyclic β-lactams (e.g. monobactams 3), β-lactams do not require a conformationally constrained
bicyclic scaffold to exert their antibacterial properties, suggesting that a suitably functionalized
azetidin-2-one ring constitutes an adequate pharmacophore.[8],[9] Such a milestone achievement in the
quest for monocyclic β-lactam inhibitors involved the development of aztreonam 4, the first, and so
far only, synthetic monobactam approved by the FDA in 1986.[10] Isolated from Nocardia uniformis in
the late 70s, the first monocyclic β-lactams to be discovered, though, concerned the nocardicins 5.[11]
Originally, they possessed only limited antibacterial activity and required chemical modifications to
increase their bioactive potential.[12] Still, however, this class of monocyclic β-lactams did not receive
that much of attention.[8] In that respect, the promising potential of the nocardicins, combined with
the clinical success of aztreonam 4, prompted us to pursue the development of novel monocyclic 3-
amino-1-carboxymethyl-β-lactam analogues with possibly enhanced biological activities. Within this
2
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
larger framework, our first challenge comprised the successful synthesis of a library of new nocardicin-
like analogues 5 (Figure 1) in which R3 represents alkylic, olefinic, acylic or enolic side chains, in order
to expand the library of aromatic features inherent to the nocardicin class.
Figure 1. Penicillins 1, cephalosporins 2, monobactams 3, aztreonam 4 and nocardicins 5.
Results and discussion
1. In silico design
The vast structural modifications previously included in numerous penicillins 1 and cephalosporins 2
signify the possibility of incorporating these side chains in monocyclic azetidin-2-ones as well.
Therefore, as a starting point for this study, in silico techniques were applied to construct a virtual
combinatorial library of novel monocyclic 3-amino-1-carboxymethyl-β-lactams with potential activity
against PBP5fm (Enterococcus faecium) (Figure 2, for more details concerning these computational
techniques, see Supporting Information). The designed monocyclic β-lactam compounds all comprised
scaffold 6, as shown in Figure 2a. The pharmacophore consisted of i) a 4-unsubstituted 3-amino-1-
carboxymethyl-β-lactam nucleus, securing the limited size of the core skeleton and mimicking the
natural substrate of the PBPs, D-alanyl-D-alanine; the carboxylic acid moiety enables crucial molecular
interactions with the enzyme’s active site and the possibility to form a destabilizing and thus activating
hydrogen bond with the β-lactam carbonyl group, which can, in a second stage, stabilize the ring-
opened azetidin-2-one after acylation of the active site serine residue, as was postulated by
Dobrowolski et al.[13]; ii) a renowned 2-(5-amino-1,2,4-thiadiazol-3-yl)-2-(hydroxyimino)acetamido
substituent linked to the β-lactam C3-position (this element will eventually be replaced by a similar 2-
(2-aminothiazol-4-yl)-2-(methoxyimino)acetamido group for synthetic reasons)[8]; and finally, iii) a side
chain ‘R’, a variable item in our combinatorial library which was carefully examined during the
subsequent virtual screening process. Following further fine filtering and conformer generation, the
3
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
resulting dataset was available for ligand-based drug design. A query pharmacophore was constructed
from a substructure 8 of the fifth-generation cephalosporin ceftobiprole 7 and its monocyclic β-lactam
analogue 9 (Figure 2b). The virtual structures were superimposed and a shared pharmacophore was
calculated, after which each ligand was scored for its electronic and steric similarity with the query
(‘virtual screening’, Figure 2c). This virtual screening campaign afforded 64 hits, which were visually
inspected for their synthetic feasibility.
Additionally, the highest scoring compounds were selected for in silico covalent docking studies using
PBP5fm, co-crystallized with a covalently bound ceftobiprole 7 molecule, in order to evaluate their
binding mode (structure-based drug design, Figure 2d).[14] The resulting β-lactam hits, of which a small
selection is presented in Figure 2e, show that a high variety of side chains is tolerated in the PBP’s
active site. In general, most side chains have in common that they reveal the preference for a side
chain (‘R’) embedding a lipophilic linker and a terminal hydrogen bond-donating or -accepting
functionality (‘HBD/HBA’), enabling additional non-covalent interactions with the enzyme. In a next
stage, keeping the latter in mind, the synthesis of a broad, model library of target compounds was
accomplished at first instance, paving the way for potential fine-tuning of the selected side chains after
preliminary biological evaluation.
4
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Figure 2. Ligand- and structure-based drug design. a: Virtual combinatorial library; b: Ceftobiprole 7 query, indicating the
3-amino side chain in green and the D-alanyl-D-alanine mimicry in red; c: Virtual screening, superimposing the virtual library
compounds on the pharmacophore model, constructed using ceftobiprole substructures 8 and 9; d: 3D-representation of
covalent docking of a modeled inhibitor (grey) and ceftobiprole 7 (green) in PBP5fm; e: Design rationale and/or synthetic
target of novel functionalized monocyclic 4-unsubstituted 3-amino-1-carboxymethyl-β-lactams and a selection of hits
resulting from the virtual screening campaign.
2. Chemical synthesis
Inspired by the molecular structures discovered via virtual screening techniques, a new synthetic
pathway was devised for the organic synthesis of these highly complex compounds, as presented in
Scheme 1. The absence of a C4-β-lactam substituent complicates the synthesis of the contemplated 3-
aminoazetidin-2-ones, as only few methodologies are described for their formation.[15] The proposed
synthetic strategy to afford -substituted 2-{3-[2-(2-aminothiazol-4-yl)-2-(methoxyimino)acetamido]-
5
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
2-oxoazetidin-1-yl}acetic acids 6 adopts Miller’s hydroxamate synthesis of 4-unsubstituted 3-amino-β-
lactams 12, starting from the amino acid serine 14 through Mitsunobu (N1-C4)-cyclization after the
required modification and protection steps (Scheme 1).[15e] N1-Deprotection and -alkylation then
enables derivatization of key building block 10 through -functionalization, amino group deprotection
and -acylation and final release of the essential carboxylic acid functionality.
Scheme 1. Retrosynthetic strategy for the synthesis of -substituted 2-{3-[2-(2-aminothiazol-4-yl)-2-
(methoxyimino)acetamido]-2-oxoazetidin-1-yl}acetic acids 6. PG = protecting group; LG = leaving group.
As such, upon thorough optimization and modification of literature procedures, key intermediates 10
were obtained after six reaction steps in up to 65% overall yield (Scheme 2). The biologically active
form of final compounds 6 is supposed to have the 3’S-configuration. Nevertheless, both enantiomers
were synthesized here for comparison of their bioactivity and to assess their optical purity. In that
respect, after quantitative tert-butyloxycarbonyl (Boc) protection of easily available precursors L- and
D-serine 14a-b by means of Boc anhydride in a (1/1)-mixture of dioxane and aqueous sodium hydroxide
(1 M),[16] hydroxamate synthesis was performed. To that end, after testing several alternatives (based
on the use of DCC, EDC/HOBt and EDC/Oxyma Pure), a TBTU (O-(benzotriazol-1-yl)-N,N,N',N'-
tetramethyluronium tetrafluoroborate)-mediated coupling reaction was performed using the
hydrochloric acid salt of O-benzylhydroxylamine and N-methylmorpholine (NMM) as a base, affording
(S)- and (R)-N-benzyloxy-2-(tert-butyloxycarbonylamino)-3-hydroxypropanamides 16a-b (Scheme
2).[17] Subsequently, cyclization-prone substrates 16a-b were subjected to Mitsunobu conditions
(DEAD, PPh3), resulting in β-lactam ring formation.[18] As a second amino protecting group proved to
be necessary during later reaction steps, a second Boc-group was introduced under DMAP catalysis (4-
(dimethylamino)pyridine) to produce -lactams 18.[19]
A
final Raney Nickel-mediated
hydrogenolysis[17a] and alkylation with methyl bromoacetate using caesium carbonate as a base,
furnished 3-(di-Boc-amino)-1-(methoxycarbonylmethyl)azetidin-2-ones 20a-b as central building
blocks for further functionalization en route to the target β-lactams 6.
6
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Scheme 2. Six-step synthesis of di-Boc-protected 3-amino-1-methoxycarbonylmethyl-β-lactams 20 in 44-65% overall yield.
A crucial reaction step toward the desired 3-amino-1-carboxymethyl-β-lactams 6 concerned the
introduction of a side chain in -position with respect to the ester carbonyl group via enolate alkylation
or acylation (Scheme 3). An overview of the final target structures with the various side chains
introduced, obtained in this way, is presented in Figure 3. As the scientific literature only provides a
few reports on the coupling of similar substrates with low-molecular-weight and highly reactive
electrophiles such as short-chain aliphatic acid chlorides, alkyl halides and aldehydes,[20] significant
optimization seemed obvious and necessary. Indeed, only after elaborate screening of i) numerous
commercially available or newly synthesized electrophiles (alkyl halides, acid chlorides, aldehydes); ii)
different bases (bis(trimethylsilyl)amide bases such as LiHMDS, KHMDS, NaHMDS); iii) temperatures (-
96 °C up to reflux); iv) concentrations (0.01-0.1 M); v) additives (e.g. NaI to accomplish a Finkelstein
reaction,[21] or lithium solvating agents such as HMPA (hexamethylphosphoramide), crown ethers or
DMPU (1,3-dimethyl-3,4,5,6-tetrahydro-2(1H)-pyrimidinone)) and vi) the order of addition and the
number of equivalents of the reagents, optimal reaction conditions were established. Firstly, some
model derivatives with less functionalized but chemically interesting side chains in -position of the
carboxylate were envisioned and considered as anchor points to validate the new synthetic
methodology. Despite the absence of H-bond-donors or -acceptors, but with the potential to provoke
π-stacking interactions in the enzyme’s active site cavity, a benzyl derivative seemed convenient for
that purpose. Unfortunately, it appeared not to be possible to selectively form the desired
monoalkylated products 21A-a-b (as a diastereomeric mixture), as a smaller amount of the
corresponding dialkylated products 22A-a-b was present in all cases (Table 1). In additional studies, it
soon became clear that the choice of optimal reaction conditions for further application would be a
trade-off between sufficient conversion of the starting material on the one hand, and a satisfying ratio
of mono- vs. dialkylated product on the other hand. Nonetheless, the major monoalkylated substrates
7
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
21A-a-b could be isolated in an analytically pure form in moderate yields via (reversed phase) column
chromatography. In Table 1, the successful synthesis of the latter and eight other, gradually more
complex derivatives containing the envisioned functionalities (i.e. lipophilic linker with HBD/HBA), is
presented. Indicated by the isomeric ratios of the obtained monoalkylated adducts 21, no effect of the
stereochemistry at the C3-position of the β-lactam nucleus on the preferential formation of one of
both isomers could be observed. The non-crystalline physical state of these products did not allow X-
ray analysis to offer a decisive answer regarding the stereochemical identity with respect to the -
position of each of both isomers, nor could Nuclear Overhauser Effect (NOE) analysis, probably
attributable to the free rotatability of the N1-substituent in these compounds. However, for -
unsaturated Boc-feruloyl (enol tautomer) and benzylidene analogues 21H-I (Table 1), NOESY analysis
provided clarity with respect to the configuration of the -positioned double bond, as will be discussed
in more detail in a next section.
Prior to decorating the 3-amino functional groups with the convenient side chain, deprotection of the
monoalkylated 21A-I-a-b and some dialkylated intermediates 22D-E-a-b in trifluoroacetic acid (TFA)-
environment for 16 hours at room temperature afforded 3-amino-β-lactams 23a-b, the diastereomeric
ratios and yields of which are displayed in Table 2. Subsequently, the free amino groups were acylated
with commercially available 2-amino-α-(methoxyimino)-4-thiazoleacetic acid (predominantly syn)
using TBTU as a coupling reagent and NMM as a base. Stirring in DMF at room temperature for 1 to 16
hours
furnished
3-[2-(2-aminothiazol-4-yl)-2-(methoxyimino)acetamido]-1-
(methoxycarbonylmethyl)azetidin-2-ones 24a-b, which were either further purified via conventional
purification techniques resulting in isomeric mixtures, or additionally separated via preparative HPLC
affording isolated isomers (Table 2).[22]
Scheme 3. Functionalization of 3-amino-1-methoxycarbonylmethyl-β-lactams 20 toward target products 25.
8
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Following the docking experiments and crystallization studies of PBP5fm in the presence of
ceftobiprole 7, it was verified that the presence of a carboxylic acid group is beneficial, if not crucial,
with respect to potential biological activity. Furthermore, an adequate distance between the latter
functional group and the β-lactam core is indispensable in order to mimic the D-alanyl-D-alanine
terminus of the natural substrate of the PBPs. In that respect, diastereomerically pure methyl ester
24A-a-2 was hydrolyzed as a final reaction step by means of lithium hydroxide monohydrate in a
solution of methanol and water (2/1) at room temperature for five minutes to two days (follow-up via
LC-MS analysis, Method A). The hydrolyzing reagent, however, caused epimerization at the -position
of the carboxylate, resulting again in a mixture of diastereomers 25A-a-1 and 25A-a-2 (dr = 43/57, as
determined via LC-MS analysis). Multiple attempts toward separation of the isomers using preparative
HPLC failed. Similar results were obtained upon treatment of methyl esters 24A-a-2 with one
equivalent of sodium hydroxide (0.1 M) in THF/water (1/1) or 1.05 equivalents of tetrabutylammonium
hydroxide in methanol/water (2/1), both at room tempature for 30 minutes. To circumvent the need
for cumbersome purification steps at the level of these carboxylates, an alternative synthetic protocol
was explored, relying on lithium iodide as a non-basic SN2-type dealkylating agent (Method B).
However, hydrolysis by means of lithium iodide seemed to demand harsh reflux conditions in
acetonitrile to drive the reaction to completion, combined with a large excess of the reagent (up to 5
equivalents) and long reaction times. Nonetheless, thorough optimization efforts of both hydrolytic
procedures gave rise to the final lithium carboxylates 25A-E,I, either as single isomers or as mixtures.
For further biological assessment, β-lactams 25A-E,I were subjected to one equivalent of HCl (0.1 M)
to afford final carboxylic acids 6 (Figure 3).
For benzoyl-substituted compounds 24F-G, multiple hydrolysis attempts resulted in chemoselectivity
issues (Table 2). Therefore, analogues 24F-a-b were converted to their reduced analogues via
treatment with sodium borohydride resulting in compounds 24J-a-b, introducing an additional
hydrogen bond-donating group (Scheme 4). Similar reducing conditions resulted in a difficult
purification in the case of compound 24G-a. Therefore, reduction of precursor 21G-a and subsequent
Boc-deprotection were performed, which furnished compound 23K-a that could subsequently be
acylated (Table 2). Despite removal of the problematic β-ketone moiety, hydrolysis of β-hydroxyl-
containing derivatives 24J-K-a-b was still not straightforward, as opening of the β-lactam ring occurred
under all hydrolyzing conditions tested, which might be attributed to the presence of an activating
hydrogen bond between the β-lactam carbonyl and the β-hydroxyl group. Similarly, deprotection of
feruloyl esters 24H-a-b failed after multiple attempts. Therefore, we decided to submit methyl esters
24H,J,K to preliminary biological evaluation instead, as the presence of the hydroxyl group might
replace the carboxylic acid as activating moiety.[13]
9
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Scheme 4. Reduction of benzoyl-substituted analogues 24F and 21G to prepare hydroxy(phenyl)methyl analogues 24J and
23K.
Figure 3. Overview of the newly synthesized 3-acetamidoazetidin-2-ones 6 and 24.
10
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Table 1. Optimized synthesis of -substituted 3-amino-1-(methoxycarbonylmethyl)azetidin-2-ones 21.
Entry[a] R’’’
Reaction conditions[b]
Crude analysis
20/21/22[c]
21 (%)[d]
dr before purif.
dr after purif.
22 (%)
21-1/21-2[e]
21-1/21-2[e]
1a
1b
2a
2b
3a
3b
4a
4b
5a
5b
6a
6b
7a
7b
8a
8b
9a
9b
Benzyl
1 eq. 20, 1.5 eq. LiHMDS, 1.5 eq. HMPA, 1.5 eq. benzyl bromide,
THF, -78 °C, 3 h
1 eq. 20, 1.5 eq. LiHMDS, 1.5 eq. HMPA, 1.5 eq. allyl bromide,
THF, -78 °C, 8 h → rt, 16 h
1 eq. 20, 1.5 eq. LiHMDS, 1.5 eq. HMPA, 1.5 eq. 4-methoxybenzyl bromide,
THF, -78 °C, 4 h → rt, 16 h
5/82/13
3/78/19
3/56/41
19/72/9
0/57/43
4/87/9
21A-a (30)
21A-b (41)
21B-a (24)
21B-b (13)
21C-a (11)
21C-b (25)
21D-a (5)
21D-b (13)
21E-a (10)
21E-b (7)
43/57
48/52
49/51
41/59
44/56
47/53
45/55
49/51
46/54
47/53
39/61
41/59
48/52
47/53
43/57
48/52
46/54
47/53
44/56
46/54
22A-a (7)
22A-b (5)
22B-a (7)
22B-b (3)
22C-a (19)
22C-b (1)
22D-a (15)
22D-b (11)
22E-a (2)[i]
22E-b (2)[i]
Allyl
4-MeO-benzyl
4-SO2NHMe-benzyl
Pyridin-4-ylmethyl
4-MeO-benzoyl
1 eq. 20, 3.15 eq. LiHMDS, 2 eq. HMPA, 1.5 eq. 4-bromomethyl-N-methylbenzenesul- 39/39/22
fonamide 26[f], THF, -78 °C → rt, 7 h
28/55/17
1 eq. 20, 1.5 eq. LiHMDS, 1.5 eq. HMPA, 1.5 eq. 4-(bromomethyl)pyridine hydrobromide[g]
THF, -78 °C, 4 h → rt, 16 h
1 eq. 20, 3 eq. LiHMDS, 3 eq. HMPA, 1.5 eq. p-anisoyl chloride,
THF, -78 °C → rt, 16 h
,
79/15/6
81/8/11
4/79/16
-
[h]
49/51
43/57
41/59
[j]
21F-a (40)
-
-
[j]
7/84/9
21F-b (72)
21G-a (34)
21G-b (26)
21H-a (17)
21H-b (18)
21I-a (13)
21I-b (14)
[k]
Benzoyl
1 eq. 20, 3 eq. LiHMDS, 3 eq. HMPA, 3 eq. benzoyl chloride,
THF, -78 °C, 3 h
1/69/30
8/68/24
0/100/0
0/10l/0
0/100/0
0/100/0
-
-
41/59 (keto)[k]
41/59 (keto)[k]
0/100[m]
22G-a (22)
22G-b (17)
-
-
-
-
[k]
Boc-Feruloyl
Benzylidene[n]
1 eq. 20, 3 eq. LiHMDS, 3 eq. HMPA, 3 eq. Boc-feruloyl chloride 28[l],
THF, -84 °C, 3 h
0/100[m]
0/100[m]
0/100[o]
0/100[o]
0/100[m]
1 eq. 20, 3 eq. LiHMDS, 3 eq. HMPA, 2 eq. benzaldehyde
THF, -78 °C → rt, 24 h
0/100[o]
0/100[o]
[a] Entries a and b refer to the 3’S- and 3’R-configuration, respectively. [b] LiHMDS: 1 M solution in THF; addition of electrophile after 30 min. [c] Determined via LC-MS analysis of the crude reaction mixture (MS signal
[d]
[e]
1
[f]
ratio). After automatic column chromatography (C18). Determined via H NMR analysis (CDCl3). Prepared via radical bromination of N,4-dimethylbenzenesulfonamide using 4.1 eq. NBS and 0.046 eq. AIBN
(CH2Cl2, , 4 d; 41% yield).[23] [g] Deprotected prior to addition using 1.6 eq. DIPEA (CH2Cl2, 0 °C, 30 min). [h] Mixture of 27% di-Boc- and 73% mono-Boc-protected product 21E-a, both as mixtures of diastereomers. The
dr’s are 37/63 and 40/60, respectively. [i] 22E-a: Mixture of 10% di-Boc- and 90% mono-Boc-protected product; 22E-b: Mixture of 66% di-Boc- and 34% mono-Boc-protected product. [j] Product not isolated. [k] dr before
purification could not be determined due to complexity regarding keto-enol tautomerism: keto/enol 21G-a: 86/14; 21G-b: 87/13. [l] Prepared via Boc protection of trans-ferulic acid (27, 37% yield), followed by activation
of 27 with oxalyl chloride (28, 99% yield). [m] Isomeric ratio (100% enol tautomer with (2E,4E)-stereochemistry); determination of stereochemistry via NOESY analysis, see below. [n] Formed via hydroxy(phenyl)methyl
intermediate, followed by dehydration. [o] Isomeric ratio (100% (2Z)-stereochemistry); determination of stereochemistry via NOESY analysis, see below.
11
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Table 2. Yields and diastereomeric ratios of intermediate and target products 25.
Cpd[a]
R’’’
X
23 (%)
dr
24 (%)
dr before purif.
dr after purif.
Separation of Reaction conditions of hydrolysis
25 (%)
dr
23-1/23-2
24-1/24-2[b]
24-1/24-2[b]
diastereo-
Method A or B
25-1/25-2[b]
(/23-3/23-4)[b]
mers (%)[e]
21A-a
21A-b
21B-a
21B-b
Benzyl
H
23A-a (82)
23A-b (75)
23B-a (54)
23B-b (64)
46/54
47/53
49/51
47/53
24A-a (86)[c]
24A-b (77)[c]
24B-a (82)[c]
24B-b (89)[c]
43/57
48/52
49/51
46/54
-
-
-
-
24A-a-1 (26)
24A-a-2 (32)
24A-b-1 (25)
24A-b-2 (38)
24B-a-1 (4)
24B-a-2 (4)
24B-b-1 (4)
24B-b-2 (5)
-
25A-a-1 (95)
25A-a-2 (92)
25A-b-1 (94)
25A-b-2 (93)
25B (99)[f]
-
-
-
-
B: 3 eq. LiI, CH3CN, , 16 h
Allyl
H
A[f]
28/72
21C-a
21C-b
21D-a
21D-b
4-MeO-
benzyl
4-SO2-
NHMe-
benzyl
4-SO2-
H
H
23C-a (98)
23C-b (88)
23D-a (99)
23D-b (99)
48/52
42/58
49/51
46/54
24C-a (22)[d]
24C-b (20)[d]
24D-a (16)[d]
24D-b (25)[d]
43/57
43/57
43/57
41/59
46/54
46/54
45/55
43/57
A
A
25C-a (97)
25C-b (99)
25D-a (88)
25D-b (98)
39/61
36/64
39/61
35/65
-
-
-
22D-a
22D-b
4-SO2-
NHMe-
Benzyl
H
23D’-a (99)
23D’-b (99)
-
-
24D’-a (47)[d]
24D’-b (11)[d]
-
-
-
-
-
-
25D’-a (99)
25D’-b (99)
-
-
A or B: 4.6 eq. LiI, CH3CN, , 4 d
A
NHMe-
benzyl
21E-a
21E-b
22E-a
Pyridin-4-
ylmethyl
Pyridin-4- Pyridin-4- 23E’-a (99)
23E-a (99)
23E-b (99)
38/62
32/68
-
24E-a (20)[d]
24E-b (21)[d]
-
40/60
26/74
-
37/63
21/79
-
-
-
-
A
-
25E-a (99)
25E-b (99)
-
34/66
30/70
-
ylmethyl
4-MeO-
benzoyl
ylmethyl
H
[g]
21F-a
21F-b
23F-a (99)
23F-b (99)
47/53
50/50
24F-a (32)[d]
24F-b (56)[d]
49/51
42/58
46/54
43/57
-
-
A: 3 eq. LiOH.H2O, CH3OH/H2O (2/1), rt,
24 h or B: 5 eq. LiI, CH3CN, , 20 h
A: 1-10 eq. LiOH.H2O, CH3OH/H2O (2/1),
rt, 24 h or B: 1-3 eq. LiI, CH3CN, , 20 h
Multiple reagents and conditions:
e.g. LiOH, LiI, NaOH, HCl, CH3SO3H,
(CH3)3SnOH, (nBu3Sn)2O
-
-
-
-
-
-
-
-
-
-
-
-
[h]
[g]
[h]
[k]
21H-a
21H-b
Feruloyl
-
23H-a (99)
23H-b (99)
0/100[i] (2E,4E)
0/100[i] (2E,4E)
24H-a (23)[d]
24H-b (14)[d]
0/100[i] (2E,4Z)[j] 0/100[i] (2E,4Z)
0/100[i] (2E,4Z)[j] 0/100[i] (2E,4Z)
-
-
[k]
21I-a
21I-b
-
Benzyli-
dene
Hydroxy-
(4-MeO-
phenyl)-
methyl
Hydroxy-
(phenyl)-
methyl
-
23I-a (99)
23I-b (54)
-
0/100[l] (2Z)
0/100[l] (2Z)
-
24I-a (25)[m]
24I-b (18)[m]
24J-a (26)[d]
0/100[l] (2Z)
0/100[l] (2Z)
0/100[l] (2Z)
0/100[l] (2Z)
43/57
39/61
50/50
47/53
34/66
49/51
-
A
25I-a (99)
25I-b (99)
0/100[l] (2Z)
0/100[l] (2Z)
-
[n]
[o]
[o]
H
-
24J-a-1-2 (19)
24J-a-3-4 (7)
24J-b-1-2 (15)
24J-b-3-4 (12)
24K-a-1-2 (6)
24K-a-3-4 (8)
A
-
[o]
[o]
-
-
24J-b (27)[d]
24K-a (32)[d]
-
-
-
[n]
[o]
[o]
21K-a
H
23K-a (99)
23/23/26/28
-
Multiple reagents and conditions:
1 eq. LiOH.H2O or NaOH, CH3OH/H2O
(2/1), rt, 10 min or 3-12 eq. LiI, CH3CN, ,
24 h or 1.06 eq. K2CO3, THF/H2O (6/7), rt,
10 min
-
-
12
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
[a] Compounds a and b have the 3’S- and 3’R-configuration, respectively. [b] Determined via 1H NMR analysis (CDCl3, CD3OD or D2O). [c] Crude yield. [d] After automatic column chromatography (C18). [e] Separation of the
two diastereomers via preparative HPLC (Zorbax Eclipse XDB-C18 and Supelco Ascentis C18 column coupled in series) (24A-B) or reversed phase column chromatography (24J-K). [f] Hydrolysis performed on a mixture
of the four isomers 24B-a-1, 24B-a-2, 24B-b-1 and 24B-b-2 considering the small amount of the corresponding methyl esters available. [g] Hydrolysis of the ketone. [h] Decarboxylation of the β-ketocarboxylic acid. [i]
Isomeric ratio (100% enol tautomer). [j] Isomerization of double bond upon base-catalyzed acylation reaction: (4E)- to (4Z)-stereochemistry (based on decrease of vicinal coupling constants in 1H NMR spectra (MeOD-
d4): J = 15.6-15.8 → 8.1 Hz). [k] Complex reaction mixtures, product not isolated. [l] Isomeric ratio. [m] After preparative TLC (EtOAc/MeOH/H2O 30/5/4). [n] dr before purification could not be determined due to overlap
of the spectral signals of the four diastereoisomers in 1H NMR and LC-MS spectra. [o] Hydrolysis of the β-lactam ring.
13
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
3. Determination of stereochemistry
In order to evaluate the optical purity of β-lactams 25 with respect to their crucial β-lactam-C3-
position, the four stereoisomers of 3-(acylamino)azetidin-2-ones 24 were synthesized and isolated,
whenever possible or necessary. The enantiomeric excess (ee) values of a selection of compounds 24,
belonging to the various types synthesized, clearly demonstrate that the C3-protons remain untouched
during the entire nine-step synthesis (Table 3).
Table 3. Enantiomeric excesses (ee) of 3-(acylamino)azetidin-2-ones 24A-E,H-I.
Cpd
a-1
(2R,3’S) or
(2S,3’S)
a-2
(2S,3’S) or
(2R,3’S)
b-1
(2S,3’R) or
(2R,3’R)
b-2
(2R,3’R) or
(2S,3’R)
Chromatographic conditions
ChiralPak® IA column
24A
24B
>92
>94
>99
>92
>97
>99
>99
>98
>99
>98
>97
>99
1 mL/min; 35 °C; hexane/EtOH (70/30);
210.2 nm (24A) or 220.8 nm (24B-D)
24C[a]
24D[b]
24E[b]
[c]
-
>99
>98
>95
0.5 mL/min; 30 °C; hexane/EtOH (75/25); 210.8 nm
Cpd
a
(3’S)
b
(3’R)
Chromatographic conditions
ChiralPak® IA column
24D’
24H
24I
>97
>99
>99
>97
>99
>96
1 mL/min; 35 °C; hexane/EtOH (70/30); 220.8 nm
1 mL/min; 35 °C; hexane/EtOH (70/30); 368 nm
0.3 mL/min; 35 °C; hexane/EtOH (85/15); 220.8 nm
[a] The isolated diastereomers of 4-methoxybenzyl derivatives 24C-a-b have only been obtained on an analytical scale after prep. HPLC using
a Phenomenex® Luna C18(2) column. [b] Diastereomers have not been isolated. [c] Not to be determined due to overlap of diastereomers and
enantiomers using different chromatographic conditions.
The optical purity of the target β-lactams 25, however, could not be determined via chiral HPLC, even
after testing multiple chiral columns and chromatographic conditions. Nevertheless, based on the
reaction mechanism behind the conversion of methyl esters 24 to lithium carboxylates 25 by means of
lithium iodide as a dealkylating reagent, assuming the SN2-reaction type and considering the non-basic
reaction conditions, the enantiopurity follows as a direct result and no change in ee values could be
expected. Per contra, the conservation of the C3 stereochemical integrity could not be guaranteed
during lithium hydroxide-mediated hydrolysis.[24] Therefore, another strategy was pursued, using a
chiral resolving agent. (-)-Menthyloxyacetic acid, for example, has a fixed stereochemistry at three
aliphatic positions, and is hence insensitive to the alkaline conditions prevailing during the hydrolysis
reaction.[25] Bearing this in mind, C3-menthyloxyacetamido-linked analogues 30 were synthesized
lacking a substituent in -position of the methyl ester. Upon LiOH-mediated hydrolysis, a change in
the stereochemistry at the C3-position would now easily be observed through the formation of two
diastereomers 31, as this is the only acidic proton left in the scaffold. A comprehensive report on the
synthesis of the latter derivatives 30 and demonstration of their optical purity before and after LiOH-
mediated hydrolysis is provided in the Supporting Information. This study therefore provides an
indirect proof for the stereochemical purity of the final carboxylates 25 as well.
14
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Scheme 5. Synthesis of (-)-menthyloxyacetamido coupled analogues 31. For synthetic details, see supporting information).
As mentioned earlier, methyl esters 24A-E were synthesized as pairs of diastereomers, which could be
separated by means of preparative HPLC. For example, the four different isomers of benzylic
derivatives 24A were obtained as such in 25-38% isolated yield. In order to have a clue of the spatial
configuration and, more specifically, of the stereochemistry of the unknown chiral center in -position
of the ester moiety, numerous attempts were made to create crystalline material for X-ray diffraction
analysis. None of these, regrettably, proved to be successful and the same holds for all other
derivatives 24-25. As an alternative strategy, 2D NOESY analysis was performed on diastereomers 24A-
1
b-1 and 24A-b-2 as model compounds in combination with inspection of the H NMR chemical shifts
and coupling constants, and energy-minimized molecular models, in analogy with literature reports.[26]
Based on the coupling constants of the multiplet signals in the 1H NMR spectra (CDCl3) of compounds
24A-b-1 and 24A-b-2, corresponding to literature data,[18b] the relative stereochemistry of the β-lactam
protons Ha-c could be derived and a clear distinction could be made regarding their chemical shift
(Figure 4, Figure 5). Moreover, protons Hd-f form an AMX system of which He and Hf are diastereotopic.
There is a large difference in their vicinal coupling constant (Jde = ±10 Hz and Jdf = ±5 Hz), indicating that,
presumably, conformational rotation around the (CHd)-(CHeHfPh)-bond is restricted by high energy
barriers (Figure 4).[26d],[27]
Figure 4. NMR coupling constants in benzylic derivatives 24A-b-1 and 24A-b-2.
Upon inspection of the NOESY spectroscopic data of diastereomer 24A-b-2 (for NOESY spectrum, see
supporting information), a clear interaction could be observed between proton Hb and the phenyl ring
as well as between Hb and He. After OPLS3 force field minimization and ab initio QM optimization of a
conformer,[28] arbitrarily assigned the (2R,3’R)-configuration and with its aromatic ring pointing to the
15
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
H(c-b)-side of the β-lactam ring system, Hb was indeed located within NOE-distance of He (2.97 Å) and
the aromatic ring (2.77 Å). This was not the case for a 3D model of the other diastereomer.
Furthermore, a moderate shielding effect on protons Hb and Hc, most plausibly caused by their
1
proximity to the aromatic ring system (anisotropic effect), could be noted in the H NMR spectra of
this diastereomer, in comparison with the corresponding proton signals in diastereomer 24A-b-1
(Figure 5).
Likewise, a correlation exists between both Ha and N(Hi)2 with respect to the aromatic carbocycle as
observed in the 2D NOESY NMR spectra of the other diastereomer 24A-b-1 (for NOESY spectrum, see
supporting information). Considering the additional NOESY interactions, two conformers ((2S,3’R) and
(2R,3’R), respectively) were designed and energy-minimized. Only in the 3D-model with the former
configuration, Ha and the aromatic ring were indeed positioned closely at only 2.79 Å apart from each
other. The same holds for protons Hh and Hc (3.94 Å) which, considering the energetically possible
rotation of the substituted thiazole moiety (superimposed in Figure 5), confirms the imposed
stereochemistry based on the NOESY signals. Furthermore, Ha and the 3-amino (Hg) 1H NMR signal are
slightly shifted upfield in comparison with the second diastereomer 24A-b-2, again due to the
proximity of the aromatic system, although we realize that the effect on the latter is not fully reliable.
An additional observation concerns the (CHd)-(CHeHfPh)-bond in the energy-minimized and ab initio
optimized models of both diastereomers. In each of them, Hd adopts an anti-relationship with respect
to He with dihedral angles of 56.4° and 71.5° between Hd and Hf (gauche), which is reflected by very
similar vicinal coupling constants in the 1H NMR spectra of both diastereomers (Jde = ±10 Hz and Jdf =
±5 Hz). To summarize, by means of evaluation of the energy-minimized molecular models of the two
possible diastereomers and their compatibility with experimental NMR coupling constants and NOESY
analysis, a credible suggestion could be made with respect to the stereochemical identity of both
isomers 24A-b-1 (2S,3’R) and 24A-b-2 (2R,3’R).
16
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Figure 5. Ab initio optimized 3D-models of diastereomers 24A-b-1-2, calculated using Jaguar software (version 9.5,
Schrödinger, Inc., New York, NY, 2017),[29] corresponding molecular structures with suggested stereochemistry and
corresponding 1H NMR spectra.
Additionally, a similar strategy was used to determine the configuration of the -positioned double
bond in feruloyl- and benzylidene-substituted derivatives 21H-24H and 21I-24I. In case of compound
21I, a significant NOESY interaction (CDCl3) was observed between β-lactam proton Ha ( = 3.88 ppm,
CDCl3) and the aromatic protons, as well as between double bond proton Hd ( = 7.49 ppm, CDCl3) and
the methoxy protons (for NOESY spectrum, see supporting information). Translation of these
interactions to OPLS3 energy-minimized and ab initio optimized 3D models of both the (3’S)- and (3’R)-
enantiomer 21I-a-b resulted in a preference for the (2Z)-stereochemistry based on smaller spatial
distances between the relevant protons in the models (Figure 6). To confirm this unexpected outcome,
the procedure was repeated for the more complex aminothiazoleoxime-substituted analogues 24I-a-
b, again showing that only the (2Z)-stereochemistry could reflect the experimental NMR interactions.
Similar observations were made for feruloyl derivatives 21H, suggesting a (2E,4E)-configuration. In a
later stage, however, isomerization of the non-enolic double bond occurred during the base-catalyzed
acylation reaction to form compounds 24H, as proven by a significant decrease in the value of the
vicinal coupling constants in the 1H NMR spectra (J = 15.6-15.8 Hz for compounds 21H-23H; J = 8.1 Hz
for compound 24H, MeOD-d4).
17
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Figure 6. Ab initio optimized 3D-models of (E)- and (Z)-isomer of compound 21I-a, calculated using Jaguar software (version
9.5, Schrödinger, Inc., New York, NY, 2017),[29] indicating preference for the (Z)-isomer.
4. Biological evaluation
In order to provide some preliminary data on their biological potential, the synthesized set of
compounds was subjected to a PBP binding and competition assay. The purified PBP3 of E. coli K12,
PBP5fm of Enterococcus faecium D63r and R39 DD-carboxypeptidase of Actinomadura spp.[30] were
therefore incubated with compounds 6 and some of the corresponding methyl esters 24, after which
their residual activity (RA) was determined by labelling the free enzyme with Bocillin FL, a fluorescent
reporter molecule and penicillin V analogue (for more information, see supporting information).[31]
Although no exceptional residual activity values were found, careful inspection of the results presented
in Table 4 revealed the inhibitory potential of derivatives 24A, 6E (albeit a chemical instability issue
was detected in this case) and 6I (Figure 3). Overall, benzylidene derivative 6I-a showed a promising
inhibitory potential on PBP3, a lethal target of E. coli (28% residual activity). Slight inhibition of the
other two PBPs (88% RA of PBP5fm and 62% RA of R39) and a moderate PBP3 IC50 value in the
micromolar range prompted us to promote this derivative to become the first hit of this study. Note
that the IC50 value of 130 µM (54 µg/mL) fits in a recently reported series of IC50 values belonging to
various commercial antibiotics ranging from 0.01 to >1000 µg/mL on the same PBP of E. coli DC2, a
hypersusceptible mutant.[32] On the other hand, commercial monobactam aztreonam 4 has a very high
affinity for PBP3 of E. coli strains DC2 and MC4100 (IC50 of 0.02 and 0.03 µg/mL, respectively).[33]
18
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Table 4. Residual enzymatic activities after incubation of PBP3, PBP5fm and R39 (2.5 µM) with selected compounds 24/6
(1 mM, pH 7, 30 °C, 3 h) and IC50 values on PBP3.
Compound
24A-a-1-2
6A-a-1
%RA PBP3 (IC50 (µM))
%RA PBP5fm
%RA R39
50 (~1000)
87
100
92
94
>100
6A-a-2
94
>100
>100
6A-b-1
96
>100
99
6A-b-2
93
>100
100
6B-a-b-1-2
24C-a-1
24C-a-2
6C-a-1-2
6C-b-1-2
24D-a-b-1-2
6D-a-1-2
6D-b-1-2
24D’-a
80
-
-
71
-
-
82
-
-
100
79
-
90
>100
92
>100
87
-
91
>100
74
79
-
95
>100
83
-
6D’-a
87
>100
73
77
-
6D’-b
96
>100
24E-a-1-2
24E-b-1-2
6E-a-1-2
6E-b-1-2
24F-a-1-2
24H-a
73
-
65
-
-
33-96[a] (983)
93
79
-
80
-
73
-
-
80
71
-
-
24H-b
-
-
24I-a
71
-
-
24I-b
81
-
88
100
-
-
6I-a
28 (130)
47 (~1000)
74
62
100
-
6I-b
24J-a-1-2
24K-a-1-2
24K-a-3-4
Blanc
78
-
-
82
-
-
100
100
100
[a] Variability due to stability issue upon conservation of the compound.
Albeit not the only requisite, our presumption regarding an increased electron flow through the
additional double bond and the aromatic ring upon acylation by the active site serine residue indeed
seems to elevate the inhibitory activity of compound 6I-a compared to all other and even very close
derivatives, such as for example benzyl-substituted compound 6A-a. Similar observations of elevated
biological activity are made upon comparison of 3-cephalosporins 2a and their unnatural 2-isomers
2b, which lack a conjugated, -positioned double bond and are almost completely devoid of activity
(Figure 7).[34] Besides, though still a controversial topic, if the cephalosporin C3’-substituent ‘X’
represents an adequate leaving group (e.g. pyridinium, acetate, heterocyclic thiols), it can act as an
19
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
electron sink. Its elimination, whether or not concerted with -lactam acylation and ring opening, can
then further enhance the antibacterial activity or -lactamase inhibitory potential of certain
cephalosporins, but is certainly not essential.[35]
Figure 7. Comparison of double bond position in cephalosporins 2a-b and monocyclic -lactam 6I-a.
Additionally, the inhibitory activity of benzylidene--lactam 6I-a on -lactamases was determined
using a colorimetric assay. The residual activity of one or two representative enzymes of each of the
four classes of -lactamases was measured, after incubation in the presence of 1 mM of compound
6I-a, by following the rate of hydrolysis of nitrocefin as a chromogenic cephalosporin reporter
substrate.[36] Compound 6I-a was found not to be a substrate for most -lactamases tested. In addition,
class C -lactamase P99 was inhibited under the conditions tested, with about 40% loss of catalytic
activity (Table 5). The specific inhibition mechanism, either suicidal or competitive, of P99, a
chromosomally encoded, clavulanate-resistant cephalosporinase from Enterobacter cloacae, by 6I-a is
subject of ongoing research.
Table 5. Residual enzymatic activities after incubation of different -lactamases with compound 6I-a (1 mM, 37 °C, 1.5 min).
Class
Enzyme
CTX-M15
Imp1
%RA
103
109
97
A
B
B
C
C
D
Imp4
P99
61
AmpC-HD
Oxa48
109
98
Based on the encouraging results regarding PBP3 inhibition, stability against class A, B and D -
lactamases and weak inhibition of class C -lactamase P99, additional structure-activity relationship
studies around benzylidene-like compounds, aiming at an increased electron flow and additional non-
covalent interactions with the PBP’s active site cavity, will be performed in due time, based on the
synthetic methodology that has been developed in the current work. In the best case, these efforts
will culminate in lead compounds with low micromolar to nanomolar inhibitory activities against the
target PBPs of resistant bacteria.
20
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Conclusion
The never faded, but – more than ever – upcoming attention for the need to tackle bacterial resistance,
has stimulated both governments and the scientific community to intervene and blow the whistle. To
address a critical resistance mechanism in Gram-positive bacteria, e.g. active-site distortion in mutated
PBPs, a possible strategy comprises the synthesis of monocyclic β-lactam analogues, which are more
restricted in volume, and thus more prone to enter a PBP active site, in contrast to their well-known
bicyclic counterparts. In that framework, the current study focused in the first place on the synthetic
development of a library of novel nocardicin-like analogues in the quest for new PBP inhibitors and
extension of the knowledge on their structure-activity relationships. In summary, employing
phamacophore modeling and covalent docking approaches as an onset, 4-unsubstituted 3-amino-1-
carboxymethyl-β-lactams were designed, functionalized in -position with a side chain comprised of a
lipophilic linker moiety and, in most cases, a terminal hydrogen bond-donating or -accepting functional
group. An unprecedented ten-step synthetic procedure was successfully elaborated toward this novel
class of highly functionalized, enantiomerically pure monocyclic β-lactams, starting from L- and D-
serine as easily available substrates. This protocol thus offers a valuable way to accomplish the
enantioselective synthesis of unprecedented 4-unsubstituted 3-amino-1-carboxymethyl-β-lactams,
which remain highly underexplored when compared to established monocyclic and of course bicyclic
β-lactam antibacterials. Both evident and more complex side chains can be introduced. The former,
however, offer multiple opportunities for chemical derivatization by means of e.g. oxidation,
electrophilic addition or aromatic cross-coupling reactions. Preliminary screening of the PBP inhibitory
potential of the synthesized β-lactams did not result in noteworthy residual activity values for all
compounds, though benzylidene-substituted azetidin-2-one 6I-a showed good inhibition of PBP3 of E.
coli, thereby emerging as a lead structure for further inhibitor design. A colorimetric assay showed
stability against representatives of the four classes of -lactamases, while a weak inhibition of class C
-lactamase P99 was observed. In summary, an elaborate synthetic protocol was established for the
construction of an introducing library of unprecedented 2-{3-[2-(2-aminothiazol-4-yl)-2-
(methoxyimino)acetamido]-2-oxoazetidin-1-yl}acetic acids, inspired by pharmacophore modelling and
covalent docking, and biological assessment revealed α-benzylidenecarboxylates to be eligible lead
structures en route to a novel class of antibacterials with PBP inhibitory activity. Encouraged by the
fact that an appropriate lipophilic linker has now been identified, a follow-up library will be synthesized
in due time using the established synthetic protocol and relying on the criteria set by the in silico design
study (additional introduction of HBD/HBA).
21
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
Acknowledgements
The authors are indebted to the Research Foundation - Flanders (FWO), the Slovenian Research Agency
(project L1-6745) and Belspo IAP-Project P44-iPros for financial support. Special thanks are devoted to
Asst. Prof. Marija Genčić of the University of Niš, Serbia, for the help with the energy-mimization
calculations and NOESY experiments and analysis thereof.
Conflict of interest
The authors declare no conflict of interest.
Keywords
Antibiotics • Drug design • Virtual screening • Lactams • Chiral pool • Configuration determination •
Biological activity
22
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
FULL PAPER
Lena Decuyper, Sari Deketelaere,
Lore Vanparys, Marko Jukič, Izidor
Sosič, Eric Sauvage, Ana Maria
Amoroso, Olivier Verlaine, Bernard
Joris, Stanislav Gobec, Matthias
D'hooghe
[Page No. – Page No.]
No action today, no cure tomorrow. Inspired by a virtual screening drug design
In
enantioselective synthesis of
functionalized monocyclic
3-amino-1-carboxymethyl-β-
lactams as inhibitors
silico
design
and
approach toward new inhibitors of penicillin-binding proteins of resistant bacteria,
a ten-step chiral pool strategy was elaborated for the synthesis of a library of
enantiomerically pure, highly functionalized monocyclic 3-amino-1-carboxymethyl-
β-lactams.
of
penicillin-binding proteins of
resistant bacteria
23
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
References
24
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
[1] W. Vollmer, D. Blanot, M. A. De Pedro, FEMS Microbiol. Rev. 2008, 32, 149-167.
[2] D. J. Tipper, J. L. Strominger, Proc. Natl. Acad. Sci. U.S.A. 1965, 54, 1133-1141.
[3] N. Georgopapadakou, S. Hammarstrom, J. L. Strominger, Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 1009-
1012.
[4] P. Macheboeuf, C. Contreras-Martel, V. Job, O. Dideberg, A. Dessen, FEMS Microbiol. Rev. 2006, 30,
673-691.
[5] M. Bassetti, M. Merelli, C. Temperoni, A. Astilean, Ann. Clin. Microbiol. Antimicrob. 2013, 12:22.
[6] S. M. Drawz, R. A. Bonomo, Clin. Microbiol. Rev. 2010, 23, 160-201.
[7] A. Zapun, C. Contreras-Martel, T. Vernet, FEMS Microbiol. Rev. 2008, 32, 361-385.
[8] L. Decuyper, M. Jukič, I. Sosič, A. Žula, M. D'hooghe, S. Gobec, Med. Res. Rev. 2018, 38, 426-503.
[9] G. S. Singh, Mini Rev. Med. Chem. 2004, 4, 69-92.
[10] R. B. Sykes, D. P. Bonner, Am. J. Med. 1985, 78, 2-10.
[11] H. Aoki, H. Sakai, M. Kohsaka, T. Konomi, J. Hosoda, J. Antibiot. 1976, 29, 492-500.
[12] D. P. Bonner, R. B. Sykes, J. Antimicrob. Chemother. 1984, 14, 313-327.
[13] J. Cz. Dobrowolsky, J. Sadlej, A. P. Mazulrek, J. Mol. Struct. 2003, 638, 229-234.
[14] E. Sauvage, unpublished work, 2018.
[15] a) C. A. Townsend, G. M. Salituro, L. T. Nguyen, M. J. DiNovi, Tetrahedron Lett. 1986, 27, 3819-3822;
b) H.-Y. Kang, A. N. Pae, H. Y. Koh, M. H. Chang, Bull. Korean. Chem. Soc. 1991, 12, 75-79; c) O.
Nakaguchi, T. Oku, H. Takeno, M. Hashimoto, T. Kamiya, Chem. Pharm. Bull. 1987, 35, 3985-3994; d)
G. H. Hakimelahi, S.-C. Tsay, J. R. Hwu, Helv. Chim. Acta 1995, 78, 411-420; e) M. J. Miller, P. G.
Mattingly, M. A. Morrison, J. F. Kerwin, J. Am. Chem. Soc. 1980, 102, 7026-7032; f) L. E. Overman, T.
Osawa, J. Am. Chem. Soc. 1985, 107, 1698-1701; g) S. Deketelaere, T. V. Nguyen, C. V. Stevens, M.
D’hooghe, ChemistryOpen 2017, 6, 301-319.
[16] J. Bayardon, D. Sinou, Tetrahedron: Asymmetry 2005, 16, 2965-2972.
[17] a) R. I. Higuchi, T. C. Roberts, P. A. Smith, D. Campbell, P. Paraselli, PCT Int. Appl. WO 2013123456A1,
2013; b) A. Sliwa, G. Dive, J.-L. Habib Jiwan, J. Marchand-Brynaert, Tetrahedron 2010, 66, 9519-9527.
[18] O. Mitsunobu, M. Yamada, Bull. Chem. Soc. Jpn. 1967, 40, 2380-2382.
[19] G. Papageorgiou, D. Ogden, J. E. T. Corrie, J. Org. Chem. 2004, 69, 7228-7233.
[20] a) H. Wild, W. Hartwig, Synthesis 1992, 1099-1103; b) H. Mastalez, M. Menard, V. Vinet, J. Desiderio,
J. Fung-Tomc, R. Kessler, Y. Tsai, J. Med. Chem. 1988, 31, 1190-1196; c) O. Sakurai, H. Horikawa, T.
Iwasaki, J. Chem. Soc. Chem. Commun. 1995, 2527-2528; d) I. Ojima, T. Komata, X. Qiu, J. Am. Chem.
Soc. 1990, 112, 770-774.
[21] a) V. N. Yandovskii, B. V. Gidaspov, I. V. Tselinskii, Russ. Chem. Rev. 1980, 49, 237-248; b) R. A. Moss,
Acc. Chem. Res. 1974, 7, 421-427.
25
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
[22] M. F. Brown, M. J. Mitton-Fry, J. T. Arcari, R. Barham, J. Casavant, B. S. Gerstenberger, S. Han, J. R.
Hardink, T. M. Harris, T. Hoang, M. D. Huband, M. S. Lall, M. M. Lemmon, C. Li, J. Lin, S. P. McCurdy, E.
McElroy, E. McPherson, E. S. Marr, J. P. Mueller, L. Mullins, A. A. Nikitenko, M. C. Noe, J. Penzien, M. S.
Plummer, B. P. Schuff, V. Shanmugasundaram, J. T. Starr, J. Sun, A. Tomaras, J. A. Young, R. P. Zaniewski,
J. Med. Chem. 2013, 56, 5541-5552.
[23]
J. Liu, Q. Liu, X. Yang, S. Xu, H. Zhang, R. Bai, H. Yao, J. Jiang, M. Shen, X. Wu, J. Xu, Bioorg. Med.
Chem. 2013, 21, 7742-7751.
[24] a) J. W. Fisher, K. L. Trinkle, Tetrahedron Lett. 1994, 35, 2505-2508; b) P. D. G. Dean, J. Chem. Soc.
1965, 6655; c) E. Biron, H. Kessler, J. Org. Chem. 2005, 70, 5183-5189.
[25] D. Nasipuri, Stereochemistry of Organic Compounds: Principles and Applications, Ed. 2, New Age
International, New Delhi, 1994.
[26]
a) R. Lacret, I. Pérez-Victoria, D. Oves-Costales, M. de la Cruz, E. Domingo, J. Martin, C. Diaz, F.
Vicente, O. Genilloud, F. Reyes, Mar. Drugs 2016, 14, 188; b) S.-P. Yang, J.-G. Cheng, J. Huo, H.-L. Jiang,
K.-X. Chen, J.-M. Yue, Chin. J. Chem. 2005, 23, 1530-1536; c) W. M. Zin, S. Buttachon, T. Dethoup, C.
Fernandes, S. Cravo, M. M. M. Pinto, L. Gales, J. A. Pereira, A. M. S. Silva, N. Sekeroglu, A. Kijjoa, Mar.
Drugs 2016, 14, 136; d) S. M. Ali, S. Khan, G. Crowyn, Magn. Reson. Chem. 2012, 50, 299-304; e) G. A.
Molander, F. Cadoret, Tetrahedron Lett. 2011, 52, 2199-2202.
[27] a) P. M. Nair, J. D. Roberts, J. Am. Chem. Soc. 1957, 79, 4565-4566; b) E. R. Kennedy, R. S. Macomber,
J. Org. Chem. 1974, 39, 1952-1955.
[28] E. Harder, W. Damm, J. Maple, C. Wu, M. Reboul, J. Y. Xiang, L. Wang, D. Lupyan, M. K. Dahlgren,
J. L. Knight, J. W. Kaus, D. S. Cerutti, G. Krilov, W. L. Jorgensen, R. Abel, R. A. Friesner, J. Chem. Theory
Comput. 2016, 12, 281-296.
[29] A. D. Bochevarov, E. Harder, T. F. Hughes, J. R. Greenwood, D. A. Braden, D. M. Philipp, D. Rinaldo,
M. D. Halls, J. Zhang, R. A. Friesner, Int. J. Quantum Chem. 2013, 113, 2110-2142.
[30] a) E. Sauvage, A. Derouaux, C. Fraipont, M. Joris, R. Herman, M. Rocaboy, M. Schloesser, J. Dumas,
F. Kerff, M. Nguyen-Distèche, P. Charlier, PlosOne 2014, 9, e98042; b) E. Sauvage, F. Kerff, E. Fonzé, R.
Herman, B. Schoot, J.-P. Marquette, Y. Taburet, D. Prevost, J. Dumas, G. Leonard, P. Stefanic, J. Coyette,
P. Charlier, Cell. Mol. Life Sci. 2002, 59, 1223-1232; c) B. Granier, M. Jamin, M. Adam, M. Galleni, B.
Lakaye, W. Zorzi, J. Grandchamps, J. M. Wilkin, C. Fraipont, B. Joris, C. Duez, M. Nguyendisteche, J.
Coyette, M. Leyhbouille, J. Dusart, L. Christiaens, J. M. Frere, J. M. Ghuysen, Methods Enzymol. 1994,
244, 249-266.
[31] a) G. Zhao, T. I. Meier, S. D. Kahl, K. R. Gee, L. C. Blaszczak, Antimicrob. Agents Chemother. 1999, 43,
1124-1128; b) B. Lakaye, C. Damblon, M. Jamin, M. Galleni, S. Lepage, B. Joris, J. Marchand-Brynaert,
C. Frydrych, J. M. Frere, Biochem. J. 1994, 300, 141-145; c) S. Turk, O. Verlaine, T. Gerards, M. Zivec, J.
Humljan, I. Sosic, A. M. Amoroso, A. Zervosen, A. Luxen, B. Joris, S. Gobec, PLoS One 2011, 6, e19418.
26
This article is protected by copyright. All rights reserved.

10.1002/chem.201801868
Chemistry - A European Journal
[32] O. Kocaoglu, E. E. Carlson, Antimicrob. Agents Chemother. 2015, 59, 2785-2790.
[33]
a) A. Asli, E. Brouillette, K. M. Krause, W. W. Nichols, F. Malouin, Antimicrob. Agents Chemother.
2016, 60, 752-756; b) N A. C. Curtis, D. Orr, G. W. Ross, M. G. Boulton, Antimicrob. Agents Chemother.
1979, 16, 533-539; c) T. A. Davies, W. Shang, K. Bush, R. K. Flamm, Antimicrob. Agents Chemother.
2008, 52, 1510-1512.
[34] a) R. R. Chauvette, E. Flynn, J. Med. Chem. 1966, 9, 741-745; b) J. D. Cocker, S. Eardley, G. I. Gregory,
M. E. Hall, A. G. Long, J. Chem. Soc. 1966, 1142-1151.
[35] a) D. B. Boyd, W. H. W. Lunn, J. Med. Chem. 1979, 22, 778-784; b) D. B. Boyd, J. Org. Chem. 1985,
50, 886-888; c) C. H. O’Callaghan, S. M. Kirby, A. Morris, R. E. Waller, R. E. Duncombe, J. Bacteriol. 1972,
110, 988-991; d) M. I. Page, Acc. Chem. Res. 1984, 17, 144-151.
[36] M. Galleni, J. M. Frère, Biochem. J. 1988, 255, 119-122.
27
This article is protected by copyright. All rights reserved.