Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced antibacterial activity against multidrug resistant bacterial strains

Article abstract of DOI:10.1016/j.ejmech.2017.10.045

Antimicrobial resistance (AMR) is a major health problem worldwide, because of ability of bacteria, fungi and viruses to evade known therapeutic agents used in treatment of infections. Aryldiketo acids (ADK) have shown antimicrobial activity against several resistant strains including Gram-positive Staphylococcus aureus bacteria. Our previous studies revealed that ADK analogues having bulky alkyl group in ortho position on a phenyl ring have up to ten times better activity than norfloxacin against the same strains. Rational modifications of analogues by introduction of hydrophobic substituents on the aromatic ring has led to more than tenfold increase in antibacterial activity against multidrug resistant Gram positive strains. To elucidate a potential mechanism of action for this potentially novel class of antimicrobials, several bacterial enzymes were identified as putative targets according to literature data and pharmacophoric similarity searches for potent ADK analogues. Among the seven bacterial targets chosen, the strongest favorable binding interactions were observed between most active analogue and S. aureus dehydrosqualene synthase and DNA gyrase. Furthermore, the docking results in combination with literature data suggest that these novel molecules could also target several other bacterial enzymes, including prenyl-transferases and methionine aminopeptidase. These results and our statistically significant 3D QSAR model could be used to guide the further design of more potent derivatives as well as in virtual screening for novel antibacterial agents.

Full text of DOI:10.1016/j.ejmech.2017.10.045

European Journal of Medicinal Chemistry xxx (2017) 1e15
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Design, synthesis and biological evaluation of novel aryldiketo acids
with enhanced antibacterial activity against multidrug resistant
bacterial strains
,
**
c
c
ꢀ a
ꢀ b
ꢁ
Ilija N. Cvijetic
, Tatjana Z. Verbic , Pedro Ernesto de Resende , Paul Stapleton ,
c
,
, *
Simon Gibbons , Ivan O. Juranic , Branko J. Drakulic 1, Mire Zloh e
ꢀ d
ꢀ d
a Innovation Center of the Faculty of Chemistry, University of Belgrade, Studentski trg 16, Belgrade, Serbia
b Faculty of Chemistry, University of Belgrade, Studentski trg 16, Belgrade, Serbia
c UCL School of Pharmacy, University College London, London, UK
d
ꢁ
Institute of Chemistry, Technology and Metallurgy, Department of Chemistry, University of Belgrade, Njegoseva 12, Belgrade, Serbia
e University of Hertfordshire, College Lane, Hatfield, AL10 9AB, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Antimicrobial resistance (AMR) is a major health problem worldwide, because of ability of bacteria, fungi
and viruses to evade known therapeutic agents used in treatment of infections. Aryldiketo acids (ADK)
have shown antimicrobial activity against several resistant strains including Gram-positive Staphylo-
coccus aureus bacteria. Our previous studies revealed that ADK analogues having bulky alkyl group in
ortho position on a phenyl ring have up to ten times better activity than norfloxacin against the same
strains. Rational modifications of analogues by introduction of hydrophobic substituents on the aromatic
ring has led to more than tenfold increase in antibacterial activity against multidrug resistant Gram
positive strains.
To elucidate a potential mechanism of action for this potentially novel class of antimicrobials, several
bacterial enzymes were identified as putative targets according to literature data and pharmacophoric
similarity searches for potent ADK analogues. Among the seven bacterial targets chosen, the strongest
favorable binding interactions were observed between most active analogue and S. aureus dehy-
drosqualene synthase and DNA gyrase. Furthermore, the docking results in combination with literature
data suggest that these novel molecules could also target several other bacterial enzymes, including
prenyl-transferases and methionine aminopeptidase. These results and our statistically significant 3D
QSAR model could be used to guide the further design of more potent derivatives as well as in virtual
screening for novel antibacterial agents.
Received 3 July 2017
Received in revised form
4 October 2017
Accepted 15 October 2017
Available online xxx
Keywords:
Aryl diketo acid
Antimicrobial activity
Multidrug resistance
Gram-positive
3D QSAR
Molecular docking
Dehydrosqualene synthase
© 2017 Elsevier Masson SAS. All rights reserved.
1. Introduction
arises from the fact that the majority of recently discovered anti-
biotics belong to already known classes of antibiotics. With the
Growing bacterial resistance toward commonly used antibiotics
represents a major health problem for the community. After the
discovery of penicillin, the boost in antibiotics discovery was be-
tween the 1950s and 1960s, but after 1985 the number of newly
discovered antibiotics rapidly decreased [1]. Another problem
occurrence of Staphylococcus aureus strains resistant even to the
drugs of last resort such as vancomycin [2] it is clear that we need
antibiotics with an alternative mode of action.
Isoprenoid biosynthesis, catalyzed by enzymes prenyl trans-
ferases, is an interesting new target for antibacterial drug discovery.
One of prenyl transferases, undecaprenyl diphosphate synthase
(UPPS), is involved in early stages of bacterial cell wall biosynthesis
and represents an excellent target for drug design since UPPS is not
produced in humans [3]. Another interesting target is dehy-
drosqualene synthase (CrtM), an enzyme involved in the formation
of S. aureus virulence factor staphyloxanthin. Inhibition of CrtM
* Corresponding author.
** Corresponding author.
ꢀ
E-mail addresses: ilija@chem.bg.ac.rs (I.N. Cvijetic), zloh@live.co.uk (M. Zloh).
1
ꢀ
Authors regret to inform that Branko Drakulic has passed away since comple-
tion of the research for this paper.
https://doi.org/10.1016/j.ejmech.2017.10.045
0223-5234/© 2017 Elsevier Masson SAS. All rights reserved.
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
2
makes bacteria more sensible to host-derived reactive oxygen
species [4,5] and it is one of the potential targets for development of
antimicrobial agents.
high-pressure,
bomb-type
reactor,
5
mmol
of
4-
fluoroacetophenone, 20 mmol of corresponding cyclic secondary
amine and 3 ml of acetone were added. After heating for 3 h at
180 ꢀC, the reaction mixture was poured into water, and extracted
with ethyl acetate or chloroform. Combined organic layers were
dried over sodium sulphate, concentrated under vacuum, and pu-
rified by crystallization or column chromatography.
Diketo acids are known inhibitors of HIV-1 integrase (IN).
Structural similarity between fluoroquinolone antibiotics and
aryldiketo acids (ADK, Fig. 1) inspired initial research of ADKs as
antibacterial agents, initially reported by our group [6] that
revealed an antibacterial activity of ADK previously not reported for
this class of compounds. It was discovered that bulky hydrophobic
substituents on the aromatic ring of the ADK were responsible for
the increase of their antibacterial activity.
Compounds 1a-3a were synthesized through the Friedel-Crafts
acylation of corresponding benzene with succinic anhydride. The
identity and purity of compounds were confirmed using 1H and 13
C
NMR spectroscopy (Figs. S1eS25), IR spectroscopy, melting point
determination, and LC-HR/ESI-MS. Synthesis and characterization
of compounds 1e3, 5e7 was described previously [10].
IN active site contains Mg2þ/Asp motif and diketo acids IN in-
hibitors are able to sequestrate Mg2þ ion from the active site [7].
Since all prenyl transferases also possess Mg2þ/Asp motif (Fig. 2), it
was hypothesized and confirmed that diketo acids having a large
hydrophobic moiety inhibit UPPS and CrtM, and also have notable
activity against many Gram-positive bacteria [8].
In this study, we report rational design, synthesis and antibac-
terial activity of novel ADK derivatives, designed using previously
reported findings [6] against several MDR bacterial strains. Addi-
tional derivatives similar to ADK, without parts of diketo-acid
moiety, were also synthesized and tested in order to confirm the
importance of this part of the molecule for antibacterial activity.
Along with this, structure-activity relationship (SAR) study high-
lighted the structural features important for the antibacterial ac-
tivity of ADKs against MDR S. aureus strain. In this study, several
physicochemical properties important for pharmacokinetics
assessment (human serum albumin binding, acidity constants, and
lipophilicity) are determined for the most potent ADK derivative.
(Z)-2-hydroxy-4-oxo-4-(o-tolyl)but-2-enoic
acid
(2):
C
11H10O4, Mr
¼
206.19, white powder, (1.25 g, 76%),
mp ¼ 106e108 ꢀC, dcc (crystallized from EtOH); ESI-MS: Calculated
for C11H9O4 [MeH]- 205.0501, found 205.0509; IR ( , cmꢁ1): 2991 br
n
(COOeH), 1710 (C¼O), 1635 (C¼O), 1279 (CeO); 1H NMR (200 MHz,
DMSO-d6) d (ppm): 2.46 (s, 3 H), 4.48 (s, br, CH2 from diketo form),
6.71 (s, 1 H, CH of enol form), 7.35 (t, 1 H, overlapped with signal at
7.31), 7.31 (d, 1 H), 7.48 (t, J ¼ 7.30 Hz, 1 H), 7.68 (d, J ¼ 8,42 Hz, 1 H);
13C NMR (50 MHz, DMSO-d6)
d (ppm): 20.62, 51.78, 101.69, 126.32,
129.04, 131.83, 132.13, 135.86, 137.50, 163.35, 168.48, 194.81.
(Z)-2-hydroxy-4-oxo-4-(m-tolyl)but-2-enoic
acid
(3):
C
11H10O4, Mr ¼ 206.19, white powder (0.82 g, 73%), mp ¼ 98e99 ꢀC
(crystallized from EtOH), ESI-MS: Calculated for C11H9O4 [MeH]-
205.0501, found 205.0508; IR (n
, cmꢁ1): 3501 br (COOeH), 1683
(C¼O), 1628 (C¼O), 1257 (CeO); 1H NMR (200 MHz, DMSO-d6)
d
(ppm): 2.36 (s, 3 H), 4.53 (s, br, CH2 from diketo form), 7.06 (s, 1 H,
CH of enol form), 7.41 (t, J ¼ 7.99 Hz, 1 H), 7.45 (d, J ¼ 7,99 Hz, 1 H),
7.79 (d, J ¼ 7.99 Hz, 1 H), 7.83 (s, 1 H); 13C NMR (50 MHz, DMSO-d6)
2. Materials and methods
d
(ppm): 20.84, 53.13, 97.95, 125.15, 128.30, 129.08, 134.70, 134.78,
138.72, 163.30, 170.30, 190.70.
2.1. Synthesis and characterization
(Z)-4-(4-butylphenyl)-2-hydroxy-4-oxobut-2-enoic acid (4):
C
14H16O4, Mr
¼
248.27, white powder (4.82 g, 82%),
mp ¼ 145e147 ꢀC; ESI-MS: Calculated for C14H17O4 [MþH]þ
All chemicals for the synthesis of ADK were purchased from
Fluka, Aldrich, or Merck, having >98% purity, and were used as
received. For the thin-layer chromatography, silica gel pre-coated
plates with fluorescent indicator (254 nm) were used.
249.1121, found 249.1113; IR (n
, cmꢁ1): 3007 br, 1714, 1628, 1272; 1H
NMR (200 MHz, CDCl3)
d
(ppm): 0.93 (t, J ¼ 7.30 Hz, 3 H), 1.36 (sx,
J ¼ 7.30, 15.16 Hz, 2 H), 1,63 (qn, J ¼ 7.86 Hz, 2 H), 2.69 (t, J ¼ 7.86,
15.16 Hz, 2 H), 3.94 (s, br, CH2 from diketo form), 7.15 (s, 1 H, CH of
enol form), 7.31 (d, J ¼ 8.42 Hz, 2 H), 7.92 (d, J ¼ 7.86 Hz, 2 H); 13C
Melting points were determined in open capillary tubes on
Stuart SMP-10 apparatus and are uncorrected. ESI-MS analysis was
performed in methanol on an Agilent Technologies 6210-1210 TOF-
LC-ESI-MS instrument in positive or negative mode. IR spectra were
recorded on a Thermo Nicolet 6700 FT-IR spectrometer equipped
with ATR accessory. 1H and 13C NMR spectra were recorded in
DMSO-d6, CDCl3 or CD3OD at 298 K using a Varian Gemini 2000
200/50 MHz or Bruker Avance 500/125 MHz NMR instrument.
ADKs were synthesized through the Claisen condensation be-
tween substituted acetophenones and diethyl oxalate, as previ-
ously described (Fig. 3) [6,9].
Some substituted acetophenones, which were not commercially
available, were synthesized through the Friedel-Crafts acylation of
corresponding benzene.
For the synthesis of 17e20, para-amino substituted acetophe-
nones were synthesized according to the following procedure: in a
NMR (50 MHz, CDCl3)
d (ppm): 13.98, 22.42, 33.24, 35.93, 96.49,
128.29, 129.28, 131.36, 150.63, 163.65, 171.89, 188.82.
(Z)-4-(2,4-dimethylphenyl)-2-hydroxy-4-oxobut-2-enoic
acid (5): C12H12O4, Mr ¼ 220.22, pale yellow powder (5.21 g, 78%),
mp ¼ 123e124 ꢀC (crystallized from AcOEt/PhMe), ESI-MS: Calcu-
lated for C12H11O4 [MeH]- 219.0657, found 219.0663; IR ( , cmꢁ1):
n
2969 br (COOeH), 1710 (C¼O), 1633 (C¼O), 1271 (CeO); 1H NMR
(200 MHz, CDCl3) d (ppm): 2.38 (s, 3 H), 2.55 (s, 3 H), 6.96 (s, CH of
enol form, 1 H), 7.11e7.15 (m, 2 H, overlapped d and s from m-H
atoms), 7.60 (d, J ¼ 8.40 Hz, 1 H); 13C NMR (50 MHz, CDCl3)
d (ppm):
21.44, 100.77, 126.86, 129.66, 131.75, 133.08, 139.16, 143.63, 164.70,
168.50, 194.17.
(Z)-2-hydroxy-4-oxo-4-(2,4,5-trimethylphenyl)but-2-enoic
acid (6): C13H14O4, Mr ¼ 234.25, pale yellow powder (4.58 g, 68%),
mp ¼ 139e141 ꢀC (crystallized from AcOEt/PhMe), ESI-MS: Calcu-
lated for C13H13O4 [MeH]- 233.0819, found 233.0821; IR ( , cmꢁ1):
n
2925 br (COOeH), 1709 (C¼O), 1619 (C¼O), 1275 (CeO); 1H NMR
(200 MHz, DMSO-d6) d (ppm): 2.24 (s, 6 H), 2.43 (s, 3 H), 4.47 (s, br,
CH2 from diketo form), 6.80 (s, 1 H, CH of enol form), 7.10 (s, 1 H),
7.52 (s, 1 H); 13C NMR (50 MHz, DMSO-d6)
d
(ppm): 18.65, 19.35,
20.52, 51.46, 101.23, 130.31, 132.81, 133.27, 134.29, 135.50, 141.66,
163.43, 167.80, 195.23.
(Z)-2-hydroxy-4-oxo-4-(2,3,5,6-tetramethylphenyl)but-2-
enoic acid (7): C14H16O4, Mr ¼ 248.27, white powder (5.20 g, 78%),
Fig. 1. Structural similarity of phenyl-substituted aryldiketo acids and ciprofloxacin.
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
3
Fig. 2. Active sites of: a) HIV-1 integrase (PDB code 1K6Y); b) CrtM (PDB code 4F6V).
Fig. 3. Synthetic procedure for aryldiketo acids.
mp ¼ 156e158 ꢀC, dcc, (crystallized from AcOEt/PhMe), ESI-MS:
C
19H28O3, Mr ¼ 304.42, white powder, mp ¼ 126e128 ꢀC (crystal-
Calculated for C14H15O4 [MeH]- 247.0976, found 247.0976; IR (
n
,
lized from PhMe); ESI-MS: Calculated for
305.2111, found 305.2102; IR (
(200 MHz, CDCl3)
C
19H29O3 [MþH]þ
cmꢁ1): 2962 br (COOeH), 1743 (C¼O), 1599 (C¼O), 1261 (CeO); 1H
n
, cmꢁ1): 2989 br,1688,1224; 1H NMR
NMR (200 MHz, DMSO-d6)
d
(ppm): 2.04 (s, 6 H), 2.17 (s, 6 H), 6.21
d
(ppm): 1.23 (d, J ¼ 4.6 Hz, 12 H), 1.27 (d,
(s, 1 H, CH of enol form), 7.02 (s, 1 H); 13C NMR (50 MHz, DMSO-d6)
(ppm): 15.98, 19.29, 103.77, 129.19, 132.00, 133.87, 138.70, 163.77,
J ¼ 4.7 Hz, 6 H), 2.71 (hp, overlapped, 2 H), 2.79 (t, J ¼ 6.18, 12.92 Hz,
d
2 H), 2,90 (hp, overlapped, 1 H), 3.05 (t, J ¼ 6.74, 12.92 Hz, 2 H), 7.02
169.17, 196.89.
(s, 2 H); 13C NMR (50 MHz, CDCl3)
d (ppm): 24.13, 24.48, 27.78,
(Z)-2-hydroxy-4-oxo-4-(5,6,7,8-tetrahydronaphthalen-2-yl)
but-2-enoic acid (8): C14H14O4, Mr ¼ 246.26, white powder (0.84 g,
63%), mp ¼ 173e175 ꢀC (crystallized from AcOEt/PhMe), ESI-MS:
30.98, 34.47, 40.89, 121.24, 137.20, 143.91, 149.89, 179.07, 208.79.
4-([1,1′-biphenyl]-4-yl)-4-oxobutanoic acid (3a):
C16H14O3,
Mr ¼ 254.28, white powder, mp ¼ 157e159 ꢀC (crystallized from
Calculated for C14H15O4 [MþH]þ 247.0965, found 247.0963; IR (
n
,
PhMe); IR (
d6)
n
, cmꢁ1): 3034 br,1706,1263; 1H NMR (200 MHz, DMSO-
(ppm): 2.60 (t, J ¼ 6.18 Hz, 2 H), 3.28 (t, J ¼ 6.18 Hz, 2 H), 7.46
cmꢁ1): 3382 br, 1642, 1591, 1421; 1H NMR (200 MHz, CDCl3)
d
d
(ppm): 1.81 (m, 4 H), 2,82 (m. 4H), 7.12 (s, 1 H), 7.17 (d, J ¼ 8.42 Hz,
1 H), 7.69 (m, 2 H, overlapped signals of two o-H atoms); 13C NMR
(50 MHz, CDCl3) (ppm): 22.74, 22.88, 29.40, 29.86, 96.83, 125.15,
129.01, 129.92, 131.25, 138.17, 145.05, 164.18, 171.04, 189.54.
(m, 1 H), 7.52 (d, J ¼ 7.86 Hz, 2 H), 7.74 (d, J ¼ 7.86 Hz, 2 H), 7.82 (d,
J ¼ 8.42 Hz, 2 H), 8.02 (d, J ¼ 8.42 Hz, 2 H). 13C NMR (50 MHz, DMSO-
d
d6) d (ppm): 27.91, 33.16, 125.21, 126.98, 127.06, 128.66, 129.18,
135.31, 138.97, 144.61, 173.95, 198.13.
1-(4-ethyl-2,6-diisopropylphenyl)ethanone (5a):
C
17H26O,
(Z)-2-hydroxy-4-(4-methoxy-2,5-dimethylphenyl)-4-oxobut-
2-enoic acid (10): C13H14O5, Mr ¼ 250.25, pale yellow powder
(2.39 g, 85%), mp ¼ 167e169 ꢀC, dcc (crystallized from AcOEt/
PhMe), ESI-MS: Calculated for C13H13O5 [MeH]- 249.0768, found
Mr ¼ 246.39, white powder (3.12 g, 69%), mp ¼ 79e80 ꢀC (crys-
tallized from AcOEt/PhMe); ESI-MS: Calculated for C17H27
O
[MþH]þ 247.2056, found 247,2063; 1H NMR (200 MHz, CDCl3)
d
(ppm): 1.23e1.26 (m, overlapped doublets, 18 H), 2.50 (s, 3 H),
249.0766; IR (
(200 MHz, DMSO-d6)
3 H), 4.46 (s, br, CH2 of diketo form), 6.86 (s, 1 H, CH of enol form),
6.91 (s, 1 H), 7.63 (s, 1 H); 13C NMR (50 MHz, DMSO-d6)
(ppm):
n
, cmꢁ1): 2948 br, 1713, 1621, 1277; 1H NMR
2.73 (hp, J ¼ 6.74 Hz, 2 H), 2.89 (hp, J ¼ 6.74 Hz,1 H), 7.01 (s, 2 H); 13
C
d (ppm): 2.17 (s, 3 H), 2.53 (s, 3 H), 3.87 (s,
NMR (50 MHz, CDCl3)
d (ppm): 24.07, 24.47, 31.09, 34.06, 34.42,
121.17, 138.50, 143.25, 149.54, 209.59.
d
(Z)-2-hydroxy-4-oxo-4-(2,4,6-triisopropylphenyl)but-2-
15.39, 21.79, 51.19, 55.68, 100.99, 113.91, 123.57, 126.88, 132.13,
139.58, 160.71, 163.55, 166.51, 194.83.
enoic acid (9): C19H26O4, Mr ¼ 318.41, white powder (0.75 g, 56%),
mp ¼ 152e153 ꢀC (crystallized from AcOEt/PhMe), ESI-MS: Calcu-
(Z)-2-hydroxy-4-(4-hydroxy-3,5-dimethylphenyl)-4-oxobut-
2-enoic acid (11): C12H12O5, Mr ¼ 236.22, white powder (0.89 g,
70%), mp ¼ 191e193 ꢀC (crystallized from EtOH), ESI-MS: Calcu-
lated for C19H27O4 [MþH]þ 319.1904, found 319.1903; IR (
n
, cmꢁ1):
3024 br, 1712, 1626, 1277; 1H NMR (200 MHz, CDCl3)
d (ppm): 1.25
(overlapped d, 18 H), 2.88 (hp, 3 H), 3.91 (s, br, CH2 of diketo form),
lated for C12H13O5 [MþH]þ 237.0758, found 237.0756; IR (
n
, cmꢁ1):
6.56 (s), 6.65 (s, 1 H), 7.05 (two overlapped s, 2 H); 13C NMR
3315 br, 1692, 1634, 1261; 1H NMR (200 MHz, CD3OD)
d (ppm): 2.25
(50 MHz, CDCl3)
d
(ppm): 24.01, 24.49, 31.10, 31.19, 34.56, 53.36,
(s, 6 H), 7.02 (s, 1 H, CH of enol form), 7.65 (s, 2 H); 13C NMR
(50 MHz, CD3OD) (ppm): 16.67, 98.53, 125.87, 127.46, 130.20,
103.17, 104.69, 121.27, 121.36, 132.12, 133.61, 145.03, 145.36, 150.91,
151.28, 162.74, 163.30, 168.58, 171.82, 196.64, 199.45. Multiple
tautomeric forms are observed in solution.
d
160.87, 165.41, 169.08, 191.99.
1-(2,5-dicyclohexylphenyl)ethanone
(4a):
C20H28O,
4-oxo-4-(2,4,6-triisopropylphenyl)butanoic
acid
(2a):
Mr ¼ 284.44, white powder (2.12 g, 75%), mp ¼ 88e90 ꢀC, ESI-MS:
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
4
Calculated for C20H29O [MþH]þ 285.2213, found 285.2206; 1H NMR
189.11.
(Z)-2-hydroxy-4-oxo-4-(4-(pyrrolidin-1-yl)phenyl)but-2-
(200 MHz, CD3OD)
d (ppm): 1.43 (m, 12 H), 1.88 (m, 8 H), 2.54 (m,
1 H), 2.60 (s, 3 H), 3.02 (m, 1 H), 7.16 (d, br, 1 H), 7.30 (d, br, 1 H), 7.32
(s, 1 H).
enoic acid (17):
C
14H15NO4, Mr
¼
261.27, red powder,
mp ¼ 160e162 ꢀC (crystallized from EtOH/PhMe), ESI-MS: Calcu-
(Z)-4-(2,5-dicyclohexylphenyl)-2-hydroxy-4-oxobut-2-enoic
acid (12): C22H28O4, Mr ¼ 356.46, white powder (1.30 g, 64%),
mp. ¼ 146e148 ꢀC, dcc (crystallized from AcOEt/hexane), ESI-MS:
lated for C14H16NO4 [MþH]þ 262.1074, found 262.1063; IR (
n,
cmꢁ1): 3155 br, 1716, 1599, 1272; 1H NMR (200 MHz, CDCl3/DMSO-
d6)
d
(ppm): 2.06 (qn, J ¼ 6.18 Hz, 4 H), 3.40 (t, J ¼ 6.18 Hz, 4 H), 6.56
Calculated for C22H29O4 [MþH]þ 357.2060, found 357.2062, IR (
n
,
(d, J ¼ 8.99 Hz, 2 H), 7.02 (s, 1 H), 7.89 (d, J ¼ 8.99 Hz, 2 H); 13C NMR
cmꢁ1): 2976 br, 1722, 1630, 1264; 1H NMR (200 MHz, CDCl3)
(50 MHz, CDCl3/DMSO-d6) d (ppm): 25.22, 47.54, 97.54, 111.25,
d
(ppm): 1.29e1.34 (m, 2 H), 1.36e1.45 (m, 8 H), 1.75 (d br, 2 H), 1.78
121.39, 130.52, 151.84, 164.61, 165.96, 189.61.
(m, 8 H), 2.51 (m, 1 H), 3.00 (m, 1 H), 3.94 (s, CH2 of diketo form),
(Z)-2-hydroxy-4-(4-morpholinophenyl)-4-oxobut-2-enoic
6.83 (s, 1 H, CH of enol form), 7.31 (s, 1 H), 7.33 (s, 2 H); 13C NMR
acid (18):
C
14H15NO5, Mr
¼
277.27, pale orange powder,
(50 MHz, CDCl3)
d
(ppm): 26.00, 26.13, 26.76, 26.83, 34.28, 34.68,
mp ¼ 164e165 ꢀC (crystallized from EtOH/PhMe), ESI-MS: Calcu-
40.10, 43.92, 101.38, 127.09, 127.35, 130.75, 134.50, 144.99, 145.62,
163.57, 170.60, 195.05.
lated for C14H16NO5 [MþH]þ 278.1023, found 278.1019. IR (
n
, cmꢁ1):
3320 br, 1713, 1604, 1268; 1H NMR (200 MHz, D2O)
d (ppm): 3.35 (t,
4-(2,5-dicyclohexylphenyl)-4-oxobutanoic
acid
(1a):
J ¼ 5.05 Hz, 4 H), 3.85 (t, J ¼ 5.05 Hz, 4 H), 3.92 (s, CH2 of diketo
C
22H30O3, Mr
¼
342.47, white powder (0.92 g, 71%),
form), 6.88 (d, J ¼ 8.99 Hz, 2 H), 7.01 (s, 1 H), 7.93 (d, J ¼ 8.99 Hz,
mp ¼ 122e124 ꢀC, dcc (crystallized from PhMe); ESI-MS: Calculated
2 H), 15.64 (sb, 1 H); 13C NMR (50 MHz, CDCl3)
d (ppm): 47.21, 53.19,
for
C
22H31O3 [MþH]þ 343.2268, found 343.2260; 1H NMR
66.57, 97.90, 113.39, 124.96, 130.37, 154.96, 163.35, 167.19, 190.04.
(Z)-4-(4-(1H-imidazol-1-yl)phenyl)-2-hydroxy-4-oxobut-2-
enoic acid (19): C13H10N2O4, Mr ¼ 258.23, pale yellow solid,
mp ¼ 184e186 ꢀC (crystallized from AcOEt/PhMe), ESI-MS: Calcu-
(200 MHz, DMSO-d6) d (ppm): 1.32 (m), 1.75 (m), 2.57 (m, 3 H), 2.93
(m, 1 H), 3.06 (t, J ¼ 5.62, 12.36 Hz), 3.21 (t, J ¼ 5.62, 12.36 Hz), 7.10
(d, J ¼ 7.86 Hz), 7.23 (s), 7.28 (s), 7.36 (d, J ¼ 8.42 Hz), 7.51 (d,
J ¼ 7.86 Hz), 7.90 (d, J ¼ 8.42 Hz), 12.16 (s, br). Two forms are visible
in solution, ratio 8:5 calculated according to the signals of aromatic
H-atoms. We assumed the presence of two conformers in solution.
(Z)-2-hydroxy-4-(2-methoxyphenyl)-4-oxobut-2-enoic acid
lated for C13H11N2O4 [MþH]þ 259.0713, found 259.0722; IR (
n,
cmꢁ1): 3452 br, 1710, 1608, 1277; 1H NMR (200 MHz, CDCl3)
d
(ppm): 7.10 (s, 1 H), 7.27 (m, 1 H), 7.38 (m, 1 H), 7.56 (d, J ¼ 8.99 Hz,
2 H), 7.99 (s,1 H), 8.14 (d, J ¼ 8.42 Hz, 2 H); 13C NMR (50 MHz, CDCl3)
(13):
C
11H12O4, Mr
¼
208.21, white powder (2.54 g, 70%),
d (ppm): 98.07, 117.76, 121.06, 130.09, 131.43, 133.68, 135.49, 141.40,
mp ¼ 129e131 ꢀC (crystallized from EtOH), ESI-MS: Calculated for
162.66, 169.81, 189.18.
C
11H11O5 [MþH]þ 223.0601, found 223.0600; IR (
n
, cmꢁ1): 2998 br,
d (ppm): 3.90 (s,
(Z)-2-hydroxy-4-(4-(4-methylpiperazin-1-yl)phenyl)-4-
oxobut-2-enoic acid (20): C15H18N2O4, Mr ¼ 290.31, pale yellow
powder, mp ¼ 162e163 ꢀC (crystallized from EtOH/PhMe); ESI-MS:
1711, 1634, 1259; 1H NMR (200 MHz, DMSO-d6)
3 H), 4.35 (s, br, CH2 of diketo form), 7.08 (t, J ¼ 7.30, 15.16 Hz, 1 H),
7.13 (s, 1 H, CH of enol form), 7.20 (d, J ¼ 7.86 Hz, 1 H), 7.59 (dt,
Calculated for C15H19N2O4 [MþH]þ 291.1339, found 291.1347. IR (
n,
J ¼ 7.30, 15.72 Hz, 1.68 Hz, 1 H), 7.78 (dd, J ¼ 7.86, 1.68 Hz, 1 H); 13
C
cmꢁ1):3420 (br), 1645, 1599, 1247; 1H NMR (200 MHz, DMSO-d6)
NMR (50 MHz, DMSO-d6)
d
(ppm): 53.87, 56.06, 102.76, 112.79,
d (ppm): 2.78 (s, 3 H), 3.33 (m, br, overlapped with residual water
120.88, 124.12, 130.13, 135.18, 158.91, 163.53, 170.14, 189.72.
(Z)-2-hydroxy-4-(3-methoxyphenyl)-4-oxobut-2-enoic acid
signal from solvent), 4.44 (s, CH2 of diketo form), 7.03 (s, 1 H, CH of
enol form), 7.10 (d, J ¼ 9.0 Hz, 2 H), 7.98 (d, J ¼ 9.0 Hz, 2 H), 11.42 (s
(14):
C
11H10O5, Mr
¼
222.19, white powder (2.64 g, 67%),
br, 1 H); 13C NMR (50 MHz, DMSO-d6)
d (ppm): 40.77, 41.86, 43.66,
mp ¼ 139e140 ꢀC (crystallized from EtOH), ESI-MS: Calculated for
51.62, 97.43, 114.03, 124.35, 130.27, 153.44, 163.53, 168.20, 189.64.
C
11H11O5 [MþH]þ 223.0601, found 223.0602; IR (
n
, cmꢁ1): 3256 br,
(Z)-4-(4-(dimethylamino)phenyl)-2-hydroxy-4-oxobut-2-
1687, 1626, 1258; 1H NMR (200 MHz, DMSO-d6)
d
(ppm): 3.83 (s,
enoic acid (21):
C
12H13NO4, Mr
¼
235.24, pale red solid,
3 H), 4.54 (s, br, CH2 of diketo form), 7.03 (s, 1 H, CH of enol form),
7.24 (dd, J1,3 ¼ 7.86 Hz, J1,2 ¼ 2.25 Hz, 1 H), 7.46 (t, J ¼ 7.86 Hz, 1 H),
7.48 (s, 1 H, overlapped with t at 7.46 ppm), 7.62 (d, J ¼ 7.86 Hz,1 H);
mp ¼ 155e157 ꢀC (crystallized from EtOH/PhMe); ESI-MS: Calcu-
lated for C12H14NO4 [MþH]þ 236.0917, found 236.0918. IR (
n
, cmꢁ1):
3382 br, 1719, 1602, 1274. 1H NMR (200 MHz, CD3OD)
d (ppm): 3.10
13C NMR (50 MHz, DMSO-d6)
d
(ppm): 53.14, 55.47, 98.15, 112.15,
(s, 6 H), 6.78 (d, J ¼ 8.99 Hz, 2 H), 7.01 (s, 1 H), 7.91 (d, J ¼ 8.99 Hz,
120.14, 120.42, 130.35, 136.29, 159.74, 163.34, 170.78, 189.88.
2 H); 13C NMR (50 MHz, CD3OD)
d
(ppm): 40.10, 98.19, 112.26,
(Z)-4-(3-bromophenyl)-2-hydroxy-4-oxobut-2-enoic
acid
123.02, 131.44, 155.98, 165.72, 168.36, 191.03.
(15): C10H7BrO4, Mr ¼ 271.06, white powder (3.35 g, 68%),
The CAS numbers for already reported compounds are: 113416-
61-8 for 1; 1224740-28-6 for 2; 1224740-52-6 for 3; 1224740-57-1
for 4; 149281-53-8 for 5; 1799810-91-5 for 6; 1799810-93-7 for 7;
1799811-01-0 for 8; 1207167-72-3 for 9; 1799811-06-5 for 10;
760207-27-0 for 13; 868616-56-2 for 14; 260558-95-0 for 15;
1224740-53-7 for 16; 1799811-04-3 for 17; 927904-77-6 for 18;
1224739-86-9 for 19; 1146317-98-7 for 22; 1207167-69-8 for 23;
1207167-70-1 for 24; 1207167-71-2 for 25; 160428-91-1 for 26;
36330-85-5 for 3a; 2234-14-2 for 5a.
mp ¼ 143e144 ꢀC (crystallized from EtOH/PhMe), ESI-MS: Calcu-
lated for C10H6BrO4 [MeH]- 268.9455, found 268.9453; IR (
n
, cmꢁ1):
3524, 1628, 1301, 1237; 1H NMR (200 MHz, DMSO-d6)
d (ppm): 4.57
(s, br, CH2 of diketo form), 7.08 (s, 1 H, CH of enol form), 7.49 (t,
J ¼ 7.86 Hz, 1 H), 7.85 (dd, J1,2 ¼ 7.86 Hz, J1,3 ¼ 2.25 Hz, 1 H), 8.03 (d,
J ¼ 8.42 Hz, 1 H), 8.14 (s, 1 H); 13C NMR (50 MHz, DMSO-d6)
d (ppm):
49.24, 97.64, 122.56, 127.02, 130.31, 131.33, 136.54, 136.97, 163.18,
170.43, 188.79.
(Z)-2-hydroxy-4-oxo-4-(3-(trifluoromethyl)phenyl)but-2-
enoic acid (16): C11H7F3O4, Mr ¼ 260.17, white powder (0.62 g,
61%), mp ¼ 135e137 ꢀC (crystallized from AcOEt/PhMe), ESI-MS:
The full characterization of compounds 8, 10 and 17 is, for the
first time, given in this manuscript.
Calculated for C11H8F3O4 [MþH]þ 261.0369, found 261.0373; IR (
n
,
2.1.1. Antibacterial activity
cmꢁ1): 3439 br, 1673, 1622, 1228; 1H NMR (200 MHz, CDCl3)
Bacterial Strains. Antimicrobial susceptibility testing control
organisms Escherichia coli NCTC 12241, Staphylococcus aureus (NCTC
12981 and NCTC 13373), Enterococcus faecalis NCTC 12697 and Ba-
cillus subtilis NCTC 10400 were obtained from the National
Collection of Type Cultures (Public Health England, UK).
Tetracycline-resistant MRSA isolate XU212 (expressing a TetK efflux
d
(ppm): 7.10 (s, 1 H, CH of enol form), 7.20 (s, 1 H, CH of second enol
form), 7.67 (dt, J ¼ 7.86 Hz, 1 H), 7.87 (dd, J ¼ 7.86 Hz, 1 H), 8.24 (s br,
1 H), 8.18 (d, J ¼ 7.86 Hz, 1 H); 13C NMR (50 MHz, CDCl3)
d (ppm):
96.07, 98.13, 124.83, 124.91, 129.79, 129.91, 130.25, 130.32, 130.54,
130.61, 131.10, 134.30, 135.72, 161.91, 162.53, 170.39, 174.75, 185.76,
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
5
pump) was obtained from Dr. E. Udo [11]. Erythromycin-resistant
2.1.4. Molecular docking
S. aureus RN4220 was provided by Dr. Jonathan Cove [12].
EMRSA-15 [13] was from the culture collection of Dr. Paul Staple-
ton. Norfloxacin-resistant S. aureus 1199B, which over-expresses
the NorA efflux pump, was kindly donated by Professor Glenn W.
Kaatz [14].
In order to elucidate potential bacterial proteins targeted by
ADKs, docking studies were performed. Targets were chosen ac-
cording to literature data, and Pharmmaper [32] searches for the
two most potent derivatives. The following proteins were selected
as putative targets from S. aureus: DNA gyrase and dehy-
drosqualene synthase (CrtM). Furthermore, additional bacterial
targets were selected from E. coli (undecaprenyl diphosphate syn-
Minimum inhibitory concentration (MIC) determination.
MICs were determined by the microdilution method [15] in 96-well
microtitre plates in a total volume of 200
m
L per well. The growth
thase (UPSS), methionine aminopeptidase (MetAP) and
b-lacta-
medium used was cation-adjusted (20 mg/l Ca2þ and 10 mg/L
Mg2þ) Mueller-Hinton broth (Oxoid, UK) with a bacterial inoculum
of 5 ꢂ 105 colony-forming units per milliliter. Plates were incubated
at 37 ꢀC and read by visual inspection after 16 h. The MIC was
recorded as the lowest concentration that inhibited visible growth
within the time period. All compounds were prepared in DMSO and
diluted such that the solvent was less than 1% in well of the
microtitre plate. DMSO concentrations 5% or less has previously
been demonstrated not to affect growth of the organisms assessed
(unpublished observation). Compounds were evaluated within the
mase), K. pneumoniae ( -lactamase) and S. enterica (tryptophan
b
synthase). These targets are also present in the strains that were
used in the in vitro testing and were considered for molecular
docking studies without homology modeling. The information of
Protein Data Bank (PDB) codes used for retrieving the experimental
structures is given in Table 6. After removal of ligand and water
molecules hydrogen atoms were added, and proteins were
neutralized through the addition of sodium or chloride ions and
embedded in the water box. Solvated proteins were minimized by
keeping the protein backbone, DNA backbone (for 2XCT) and metal
ions fixed to preserve the experimental structure. The 10 ps con-
jugate gradient minimization was performed in NAMD 2.12 pro-
gram, using CHARM force field. AutoDock Vina software was used
for docking. Exhaustiveness was set to 15, and only the most
favorable binding mode was calculated for each molecule. Raising
the exhaustiveness to 250 gave the same results for docking against
the 4H3A as a target.
The.sdf database consisting of ADKs in monoanionic and dia-
nionic form as well, and compounds 1a-5a were prepared, starting
from the best-ranking conformation generated with OMEGA 2.5.1.4
software, and further optimized using semiempirical PM7 method
implemented in MOPAC 2016. Vega ZZ software was used as a GUI
for all calculations. Images of docking solutions were prepared
using BIOVIA Discovery Studio 2016.
range 0.12e128 mg/ml and MICs were determined at least twice.
MIC values were converted into molar units (
values were input for SAR study.
mmol/L), and such
2.1.2. Molecular similarity
Structural similarity between compounds studied and fluo-
roquinolone antibiotics was calculated by overlapping conforma-
tional assembly of each ADK with the structure of norfloxacin as a
template molecule, using Openeye ROCS program [16]. Degree of
overlap was quantified using Tanimoto scoring function [17].
2.1.3. 3D QSAR study
In order to correlate antibacterial activity with the structure of
compounds, we made 3D QSAR model using Pentacle 1.06 software.
As a dependent variable, we have chosen potencies toward
norfloxacin-resistant S. aureus 1199B, expressed as elog(MIC). For
each ADK, a set of conformers was generated in OMEGA 2.4.3 [18]
using a MMFF94s force field [19]. In order to simulate appropriate
ionization state under physiological conditions, compounds were
modeled as carboxylates (monoanionic form). Structures were
additionally optimized using semiempirical PM6 method [20]
implemented in MOPAC 2016 [21].
Molecular interaction fields (MIFs) [22] were calculated using
four GRID probes: DRY which represents hydrophobic interactions,
O is sp2 carbonyl probe which maps H-bond donor features of
molecules; N1 is neutral flat eNH probing H-bond acceptors in
molecules, and TIP probe introduces the shape of molecules in
model building [23]. Pentacle software computes alignment-
independent descriptors (GRIND-2 descriptors) where each vari-
able represents the product of the most favorable interaction en-
ergies (IE) of two GRID probes in a given distance range. The
majority of 3D-QSAR programs like CoMFA [24] and CoMSIA [25]
require spatial alignment of structures before MIF computation,
and since there are no strict criteria for alignment, results are
strongly affected by the quality of the alignment. Alignment-
independent methods like GRIND overcome the problem with
alignment and this method is widely used in SAR research [26e29].
Important positions around molecules (hot spots) are extracted
from MIFs using AMANDA discretization algorithm [30]. Encoding
of the filtered MIFs into GRIND variables was performed by the
maximal auto- and cross-correlation (MACC2) algorithm. The
weight of GRIND correlograms that encode DRY probe is increased
by changing its cut-off value from ꢁ2.8 (default) to ꢁ2. All other
parameters for MIF computation were retained at default values.
GRIND variables are correlated with antibacterial activity using
partial least square (PLS) method [31].
2.1.5. HSA binding experiments
Fatty acid-free human serum albumin (HSA, <0.007% fatty acids,
Mw ¼ 66478 Da) was purchased from Sigma, as well as potassium
dihydrogen phosphate, disodium hydrogen phosphate, sodium
chloride and potassium chloride used for PBS preparation. A stock
solution of HSA (c ¼ 1.96 ꢂ 10ꢁ4 M) was prepared by dissolving
accurately weighted mass of commercially available lyophilized
HSA in freshly prepared 30 mM PBS (pH 7.38). A stock solution of
2,5-di-cyclohexyl ADK (c ¼ 1.63 ꢂ 10ꢁ3 M) was prepared by dis-
solving a substance in DMSO, because of its low solubility in the
buffer. For HSA-ADK interaction studies, HSA solution was freshly
prepared from the stock by dilution with a buffer (HSA concen-
tration was kept constant, c ¼ 5 ꢂ 10ꢁ7 M), and titrated with small
volume increments of ADK stock solution, to avoid large sample
dilution. ADK was added in 0.5, 1, 1.5, 2, 3, 4, 6 and 8 molar
equivalents. After each aliquot addition, the system was stirred and
left to equilibrate for 20 min before UV/Vis absorption and fluo-
rescence emission spectra recording. At the end of the titration (8
equivalents of substance added) the final DMSO volume did not
exceed 2%. It was shown that addition of 15% of DMSO did not
induce structural changes in bovine serum albumin (BSA), a protein
structurally similar to HSA [33]. Therefore, it is unlikely that the
conformation of HSA was changed with the level of DMSO used in
this study. Millipore water was used for buffer solution preparation.
Fluorescence measurements were performed on spectrofluo-
rometer Fluoromax-4 Jobin Yvon (Horiba Scientific, Japan), equip-
ped with Peltier element for temperature control and magnetic
stirrer for cuvette, using quartz cell with 1 cm path length and 4 ml
volume. Before the fluorescence spectra recording, diluted HSA
solution was filtrated using 0.22
mm pore size filters. An excitation
wavelength was 280 nm, with 5 nm slits; emission spectra were
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
6
recorded in 300e450 nm wavelength range, with 5 nm slits, and
0.1 s integration time. Background PBS signal was subtracted from
each spectrum.
UV/Vis spectra were recorded on a GBC Cintra6 UV/Vis spec-
trophotometer, using quartz cell with 1 cm path length and 4 ml
volume. All spectra were recorded against the corresponding blank
(PBS) in 250e450 nm wavelength range at room temperature.
field interactions analysis indicated the importance of DRY-DRYand
DRY-N1 variables for the activity of these compounds. Furthermore,
we postulated that the keto-enol moiety is responsible for the
antibacterial activity.
Therefore, we have designed a series of new congeners by sys-
tematic introduction of bulky hydrophobic substituents at different
positions in the aromatic ring. An additional set of congeners with
small hydrophobic or small polar substituents on the aromatic ring
and congeners without keto-enol moiety were selected for syn-
thesis and testing, for the comparison.
As anticipated, several compounds (Table 1) were found to have
activity against MDR Gram-positive (Gþ) organisms, Enterococcus
faecalis, Bacillus subtilis and Staphylococcus aureus. As predicted by
rational design, compound 12 exhibited notable activity against
E. faecalis, and S. aureus, including both methicillin-susceptible and
-resistant isolates. Compound 12 was 10 times more potent toward
SA 1199B strain than compound 25, the most potent ADK from
previously studied set of ADK, [6], and also 10 times more potent
than norfloxacin against the same bacterial strain. None of the
compounds were active against the Gram-negative (Ge) organism
Escherichia coli (MICs >128).
2.1.6. Acidity constant (pKa value) determination
Acidity constant (pKa2 value) of compound 12 was determined
using spectrophotometric titration at t ¼ 20 0.5 ꢀC. Working so-
lution (c12 ¼ 5 ꢂ 10ꢁ5 M) was prepared by dissolving an accurately
weighed substance in 2.50 ml of ethanol, 10 mM phosphate buffer
(pH ¼ 10.44, I ¼ 0.1 M (NaCl)) was added up to 100.00 ml. In-
crements of concentrated or 1 M H3PO4 were added stepwise to
reach pH~4 when precipitation occurred (total volume change at
the end of the titration was lower than 2%). pH Value was contin-
uously measured during titration and UV/Vis spectra recorded after
solution pH value equilibration. UV/Vis spectra were recorded on
GBC Cintra6 UV/Vis spectrophotometer, using quartz cell with 1 cm
path length and 4 ml volume. All spectra were recorded against the
corresponding blank in 210e600 nm wavelength range. Three sets
of experiments were performed. Acidity constant value (Ka2) was
determined according to transformed form of classical spectro-
photometric equation: [34].
ADKs are good chelators of Mg2þ [39] and keto-enolic moiety is
involved in interaction with a metal ion. Since prenyl-transferases
with the Mg2þ/Asp motif in the active site, are identified as
possible targets for ADK in a bacterial cell, we wanted to check
whether ADK metal complexation ability influences their potency
toward MDR bacterial strains. Antibacterial activities of two sets of
compounds, comprising three phenyl-substituted aroylpropanoic
acids (1a-3a) and two acetophenones (4a, 5a) (Fig. 4), were tested
(Table 1). Potencies of compounds lacking diketo moiety (1a-5a)
are notably lower compared to the ADKs with the same phenyl
substituent pattern. This result is in line with a hypothesis that
metal complexation ability is important for the antibacterial ac-
tivity of ADKs.
ꢀ
1
Ka2
ꢁ
A ¼ AA2ꢁ
þ
½H3Oþ ðAHA ꢁ AÞ
(1)
where AHA-, AA2-, and A represent absorbances of monoanionic
(HAe), dianionic (A2ꢁ) forms of 12 and their mixture at specified
wavelengths, respectively. Equation (1) gave linear dependence
where the spectrum of only one “pure” form (HAꢁ) was needed for
Ka2 determination. Ka2 Value was calculated by linear regression
analysis from the slope of corresponding fitting line, (results are
shown as an average value 95% confidence interval).
The most potent compounds were also tested on human MRC-
5 cell lines, and showed notable cytotoxicity: IC90 (compound
pH Values were measured using CRISON pH-Burette 24 2S
equipped with CRISON 50 29 micro-combined pH electrode (CRI-
SON INSTRUMENTS, S.A. Spain). The electrode was calibrated by
standard CRISON buffer solutions (pH 4.01, 7.00, and 9.21).
9) ~ 90 mM, IC90 (compound 12) ~ 30 mM. It is important to note that
the most potent compounds show selectivity toward bacterial vs.
human cells in some extent, but further structural modifications
would be necessary to increase the selectivity.
2.1.7. Shake-flask logD determination
3.2. Molecular similarity between ADK and norfloxacin
Standard shake-flask method was used for logD (pH 7.4, 10 mM
phosphate buffer) determination [35,36]. Stock solution of com-
pound 12 (c12 ¼ 5.0 ꢂ 10ꢁ3 M) was prepared in octanol saturated
with buffer. Octanol/buffer volume ratio was set according to
estimated logD (pH 7.4) value; MarwinSketch (www.chemaxon.
com) was used for logD value estimation.
In our previous study of ADKs' antibacterial activities, structure
similarity between ADK and norfloxacin was examined. It was
previously shown in a qualitative manner that the better overlap of
structures corresponds to the more potent derivatives [6]. A
quantitative evaluation of molecular similarity was carried out for
the set of molecules in this study. It was noticed that the most
potent ADK derivatives show the best shape complementarity with
norfloxacin, as seen from ROCS Overlap Function. However, phar-
macophoric similarity between the most potent ADK and nor-
floxacin was the smallest, as seen from the Color Tanimoto scores
(Table 2). One may assume that the mechanism of antibacterial
action of ADK with bulkier phenyl substituents is different from the
norfloxacin (inhibition of topoisomerase II and IV and prevention of
cell division).
The concentration of compound 12 in aqueous phase was
determined spectrophotometrically using
a calibration curve
method. Due to a low octanol volume used, concentration of 12 in
octanol phase was determined as a difference between total and
aqueous phase compound 12 concentration. Three octanol/buffer
volume ratios were used for the experiment and the result is given
as a mean 95% confidence interval.
3. Results
3.1. Antibacterial activity
3.3. QSAR
The antibiotic activity of ADK against Gram-positive MDR
strains was reported for the first time in our previous study, and
two congeners had a comparable activity with norfloxacin. The
analogues with the 2,4-di-iPr and 2,4,6-tri-Et substitutions on the
aromatic ring exerted broad spectrum activity [6]. The molecular
3.3.1. PCA model
Principal component analysis (PCA) is performed on the whole
set of GRIND descriptors. 5 PC model explained 84.45% of X sum of
squares (SSXacc), and 76.43% of X variance (VarXacc). In order to
reduce number of variables, one cycle of fractional factorial design
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
7
Table 1
Structure and antibacterial activities of ADKs (compounds 1e26 including previously reported activities for compounds 1 and 22e26), compounds without diketo moiety (1a-
5a, see Fig. 4) and antibiotics (norfloxacin, ciprofloxacin, levofloxacin, moxifloxacin) against several bacterial strains. Minimal inhibition concentrations (MIC) which
inhibited bacterial growth are given in mM.
Comp. R-
E.faecalis
NCTC 12697
B.subtilis NCTC 10400 S. aureus NCTC 12981 (MSSA) SA-NCTC
13373 (MRSA)
SA- 1199B SA
SA
SA XU212
RN4220 EMRSA-15
1
2
H
2-Me
n.d.
e
n.d.
e
n.d.
e
n.d.
e
>2660
e
n.d.
e
n.d.
e
n.d.
e
3
3-Me
e
e
e
e
e
e
e
e
4
5
6
7
8
9
4-nBu
129
e
e
258
e
258
e
258
e
258
e
258
e
258
e
2,4-di-Me
2,4,5-tri-Me
2,3,5,6-tetra-Me
e
e
e
e
e
e
e
e
e
516
260
e
e
e
e
e
e
e
e
b-tetralinyl
e
e
e
520
25
e
e
520
25
e
e
2,4,6-tri-iPr
4-OMe ꢁ2,5-di-Me
4-OH-3,5-di-Me
2,5-di-cyclohexyl
2-OMe
3-OMe
3-Br
3-CF3
4-pyrrolidinyl
4-morpholinyl
4-imidazolyl
4-N-Me piperazinyl
4-N,N-diMe
4-t-Bu
2,4-di-iPr
2,5-di-iPr
2,4,6-tri-Et
4-Ph
2,5-di- cyclohexyl AP n.d.
25
e
13
e
e
50
e
50
e
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
1a
2a
3a
4a
5a
e
e
e
e
e
e
e
e
e
e
6
e
125
e
11
e
11
e
11
e
11
e
22
e
11
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
496
e
496
e
e
248
e
124
e
e
e
496
e
e
e
e
e
e
e
e
e
e
e
e
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
94
e
n.d.
n.d.
n.d.
n.d.
n.d.
94
e
1030
232
232
116
477
47
e
e
e
e
232
464
232
e
232
232
232
e
464
464
232
e
47
e
94
e
47
e
2,4,6-tri-iPr AP
4-Ph AP
2,5-di-Ch A
2,4,6-tri-iPr A
n.d.
n.d.
n.d.
n.d.
6.3
1.5
1.4
n.d.
e
e
e
e
e
e
450
e
450
e
450
e
225
e
225
e
225
e
Norfloxacin [37] [14]
Ciprofloxacin [37,38]
Levofloxacin [37] [38]
Moxifloxacin [38]
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
106
24.2
5.53
0.62
e
3,3
n.d.
n.d.
n.d.
e
0.76
0.69
0.31
n.d.
n.d.
n.d.
n.d. e not determined; e inactive (MIC > 128 mg/mL); AP- aroylpropanoic acids; - A-ketone (Fig. 4).
Fig. 4. Structural modifications of ADK lacking diketo (keto-enolic) moiety.
(FFD) is performed and results are shown in Table 3.
Inspection of PCA scores plot (PC2 vs PC1) reveals clustering of
compounds into two groups (Fig. 5).
donors. First cluster, located at the far right of the graph, contains
molecules with two H-bond donor (HBD) groups, second cluster
groups molecules with one HBD, while compounds 1a and 4a are
outliers since they have no HBD groups.
Compounds are clustered according to number of H-bond
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
8
Table 2
Tanimoto and ROCS overlap scoring functions for overlaps of ADK with norfloxacin template.
Comp.
Tanimoto
Combo
Shape
Tanimoto
Color
Tanimoto
Overlap
Comp.
Tanimoto
Combo
Shape
Tanimoto
Color
Tanimoto
Overlap
1
2
3
4
5
6
7
8
0.940
0.909
0.958
0.904
0.923
0.918
0.839
0.971
0.793
0.844
0.953
0.805
0.891
0.635
0.603
0.654
0.615
0.616
0.612
0.564
0.699
0.734
0.564
0.693
0.672
0.608
0.305
0.306
0.305
0.289
0.306
0.306
0.275
0.273
0.059
0.280
0.260
0.133
0.283
644.132
641.157
672.227
691.311
665.772
679.306
661.592
757.711
860.108
661.289
733.759
876.715
662.311
14
15
16
17
18
19
20
21
1a
2a
3a
4a
5a
0.931
0.937
0.998
1.047
0.993
0.974
1.053
0.948
0.846
0.886
1.064
0.976
0.735
0.652
0.657
0.694
0.670
0.690
0.667
0.662
0.635
0.703
0.599
0.708
0.795
0.569
0.279
0.280
0.304
0.376
0.303
0.307
0.391
0.313
0.143
0.287
0.355
0.181
0.166
689.710
670.207
757.203
758.256
790.267
766.072
787.194
694.573
878.132
740.503
783.744
866.392
655.108
9
10
11
12
13
Table 3
obtained using three CV procedures is shown (Table 4).
Statistics of the PCA model generated after 1 FFD cycle (out of 490 variables, 272
were retained).
The graphs of calculated vs experimental (Fig. 6a) and predicted
vs experimental potency (Fig. 6b) for 2LV model with LOO CV are
shown.
Component
SSX
SSXacc
VarX
VarXacc
The potency of all compounds is predicted within 2 SDEP limit.
PLS coefficients plot (Fig. 7) shows which descriptors have
positive or negative impact on potency.
Variables with the highest influence on potency are labeled in
Fig. 7, and their association with structural elements of compounds
is described in Table 5.
1
2
3
4
5
48.11
14.87
9.65
8.22
3.6
48.11
62.98
72.63
80.85
84.45
44.29
12.82
8.47
8.09
2.76
44.29
57.11
65.57
73.67
76.43
The information obtained from each block of variables are
described in following lines:
ꢃ DRY-DRY correlograms define optimal distance between two
hydrophobic groups. While small distance variables have
negative impact on potency, two DRY hot spots at distances from
10 to 16 Å positively contribute to antibacterial activity;
ꢃ N1-N1 block of variables is directly correlated with biological
activity of compounds, with optimal node-node distance of
11.60e12.00 Å;
ꢃ DRY-N1 cross-correlograms define optimal distance between
HBA group and non-polar moiety (16.40e16.80 Å). Variables
with smaller node-node distances (11.20e11.60 Å) are inversely
correlated with potency;
ꢃ Variable block O-N1 describes optimal distance between HBD
and HBA groups. All variables up to 16.40 Å distance are directly
proportional to potency, while larger distances seemed
inversely correlated.
ꢃ O-TIP correlograms provide the same information as O-N1
variables.
Fig. 5. PCA scores plot of PC2 vs PC1, for all compounds included in 3D QSAR model.
The most relevant variables for antibacterial activity of studied
compounds are shown in Fig. 8.
In order to rationalize which structural features of studied
compounds contribute to the overall potency, we made partial least
square (PLS) model using pMIC (elogMIC) values against SA1199B
strain as dependent variable, and the whole set of GRIND de-
scriptors as independent ones. The resulting 2 latent variable (LV)
The SAR study on a series of ADK congeners with notable ac-
tivity against MDR Gram-positive bacteria revealed that two hy-
drophobic regions at distances from 10 to 16 Å positively influence
antibacterial activity. The presence of two HBA groups increases
potency, with optimal distance of 11.60e12.00 Å. The optimal
model had r2
¼
0.88, q2 (leave-one-out, LOO cross-
validation) ¼ 0.43 and standard deviation of error of prediction
(SDEP) ¼ 0.62. The predictivity of model is highly improved by
applying one cycle of FFD (2LV, r2 ¼ 0.90, q2 (LOO) ¼ 0.72 and
SDEP ¼ 0.43), which reduced the number of variables from 490 to
272. Besides leave-one-out, leave-two-out (LTO) and random group
(RG) cross-validation (CV) procedures are also performed. Splitting
of this relatively small set of compounds (16 comp.) to training and
test is not done, since random group validation essentially provides
the same information on model predictivity. Statistics of models
Table 4
Statistics of 2LV PLS model after application of different CV procedures.
Component
SSXacc
SDEC
SDEP
r2
q2
acc
acc
2 LV (LOO)
2 LV (LTO)
2 LV (RG)
57.51
57.87
60.33
0.26
0.26
0.26
0.43
0.46
0.53
0.90
0.90
0.90
0.72
0.70
0.59
FFD-LV ¼ 2, 20% of dummy variables, 20 randomization cycles, 3 random groups.
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
9
Fig. 6. The graphical representation of: a) calculated vs experimental pMIC, and b) predicted vs experimental pMIC.
to decreased potency of compounds. The presence of HBA group at
the distance up to 16 Å from HBD positively influences potency,
while larger distances are negatively correlated. Steric hot spots,
represented through TIP probe, have similar optimal distance from
HBD probe (11e14 Å), and nodes at distance larger than 16 Å have
negative impact on model. While these ligand based SAR obser-
vations provide some guidance for design of novel analogues with
improved activity, the structure based design would be needed to
improve the selectivity of ADKs towards bacterial targets.
3.4. Molecular docking
Molecular docking was carried out to elucidate potential bind-
ing modes with bacterial enzymes and propose a putative target for
most active molecules (Table 6). Our previous reported studies
indicate that ADK can exist in monoanionic and dianionic forms at
physiologically relevant pH, as well as confirmed by our judicious
determination and analysis of pKa values for compound 12 reported
further down in the manuscript. Therefore, we have carried out
docking using two different forms for all ADK compounds in this
study. The docking results suggest that most likely targets in
S. aureus are dehydrosqualene synthase and DNA gyrase with
binding energies over ꢁ9 kcal/mol for monoanionic form and even
over ꢁ10 kcal/mol for dianionic forms. It is apparent that com-
pound 12 in these two targets have a potential to interact with
Fig. 7. PLS coefficients plot (2 LV model) showing variables with positive values
(positive impact on potency) or negative values (negative impact on potency). More
intensive variables means higher influence of corresponding structural features on the
potency of compounds. The most important variables are labeled by numbers, and
discussed in the text.
distance between hydrophobic moiety and HBA group is found to
be 16.40e16.80 Å, and smaller DRY e N1 node distances are linked
Table 5
Highlights of variables having highest influence on potency of compounds toward SA1199B.
Probe
block
Variable Distance Impact Structural elements, comments
No
(Å)
DRY-
10
4.00
e4.40
10.40
e10.80
11.60
e12.00
11.20
e11.60
16.40
e16.80
17.20
e17.60
12.00
e12.40
16.40
e
þ
þ
e
þ
e
þ
e
Aromatic ring and another hydrophobic moiety at short distance. More pronounced in less potent compounds
DRY
DRY-
DRY
N1-N1 127
DRY-N1 273
DRY-N1 286
26
The most potent compounds (12 and moxifloxacin) have the highest variable values (VV). Optimal distance between two
hydrophobic groups.
Only present in the most potent compounds. Represents distance between carboxylate and keto or methoxy group oxygen at given
distance.
More pronounced in less potent compounds. The smallest VV is for moxifloxacin, differentiating it from norfloxacin, ciprofloxacin
and levofloxacin. Compounds 9 and 12 also have small VV.
Optimal distance between HBA and hydrophobic group. More pronounced in more potent compounds, with highest VV for
moxifloxacin.
Large distance between HBA (carboxylate) and protonated nitrogen atom of norfloxacin, ciprofloxacin and 19. Not present in most
potent compounds.
O-N1
O-TIP
O-TIP
386
422
433
More pronounced in more potent compounds. Complementary to O-N1 variable block at smaller node-node distances.
HBD group relatively far from steric hot spot. Only expressed in less potent compounds. Provides the same information as variable
386.
e16.80
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
10
Table 6
The docking scores of the best-docked compounds into the several bacterial target proteins.
Target
PDB Compounds docked as monoanions
Compounds docked as dianions
code
Compounds with the most favorable binding
Binding energy range Compounds with the most favorable binding Binding energy range
energies (values given in brackets in kcal/mol)
for all compounds
(kcal/mol)
energies (values given in brackets)
for all compounds
(kcal/mol)
S. aureus CrtM
4F6V 23, 12, 1a
(e9.9, ꢁ9.7, ꢁ9.7)
2XCT 12, 1a, 4a
(e9.6, ꢁ8.8, ꢁ8.8)
1C21 12, 1a, 4a
(e8.5, ꢁ7.5, ꢁ7.1)
4H3A 12, 1a, 26
(e7.7, ꢁ7.4, ꢁ7.3)
4LEN 16, 8, 22
(e6.0, ꢁ5.8, ꢁ5.7)
1RCJ 12, 1a, 4a
(e8.3, ꢁ8.2, ꢁ7.9)
1QOP 1a, 12, 26
(e8.1, ꢁ8.0, ꢁ7.8)
ꢁ9.9 to ꢁ7.2
ꢁ9.6 to ꢁ6.4
ꢁ8.5 to ꢁ5.5
ꢁ7.7 to ꢁ5.5
ꢁ6.0 to ꢁ4.6
ꢁ8.3 to ꢁ6.2
ꢁ8.1 to ꢁ5.6
12, 23, 26
ꢁ10.6 to ꢁ7.8
ꢁ9.6 to ꢁ7.1
ꢁ8.1 to ꢁ5.6
ꢁ8.0 to ꢁ6.0
ꢁ6.1 to ꢁ4.9
ꢁ8.4 to ꢁ6.7
ꢁ8.3 to ꢁ6.0
(e10.6, ꢁ9.9, ꢁ9.9)
S. aureus DNA
gyrase
E. coli MetAP
12, 23, 20
(e9.6, ꢁ9.2, ꢁ8.7)
12, 24, 8
(e8.1, ꢁ6.8, ꢁ6.7)
E. coli UPPS
12, 26, 4
(e8.0, ꢁ7.6, ꢁ7.4)
8, 22, 16
E. coli serine
b
-lactamase
K. pneumoniae
-lactamase
S. enterica
tryptophan
synthase
(e6.1, ꢁ5.9, ꢁ5.9)
12, 8, 7
b
(e8.4, ꢁ8.0, ꢁ7.8)
12, 26, 8
(e8.3, ꢁ8.0, ꢁ7.7)
UPPS- Undecaprenyl diphosphate synthase; MetAP - Methionine aminopeptidase. CrtM - Dehydrosqualene synthase.
positively charged ions (Mn2þ and Mg2þ), confirming our hypoth-
esis that ADK e metal ion interactions may play an important role
for their activity. Albeit the binding mode for two different ionic
forms appear different in the docking poses in the CrtM binding site
(Fig. 9), their importance for the inhibitory activity for this enzyme
may be similar. Furthermore, our most active compound may form
interactions with other important enzymes in the bacterial life
cycle, including UPPS interacting in a similar binding mode as the
ADK with a large hydrophobic moiety reported in Ref. [8]. This
provides an opportunity for designing a molecule that is multi-
targeting and could effectively inhibit resistant pathways, and
offer a more effective approach against infectious diseases [40].
emission wavelengths, respectively [42]. Since absorbancies at
both, excitation and emission, wavelengths did not exceed 0.05 in
all measurements, we used raw data for further calculations.
Fluorescence quenching data were processed using Stern-
Volmer (S-V) equation (3):
F0
F
¼ 1 þ Ksv½Qꢄ ¼ 1 þ Kqt0½Qꢄ
(3)
where F0 and F represents HSA fluorescence intensities in absence
(F0) and presence of quencher (F), Ksv and Kq are S-V's quenching
constant and the quenching rate constant of protein, respectively;
t0 is the average fluorescence lifetime (7.09 ns for HSA) [43] and [Q]
is the concentration of the quencher (comp. 12). The plot of F0/F vs.
[Q] at three temperatures is shown in Fig. 11, and results of linear
regression analysis are shown in Table 7.
3.5. Physico-chemical profiling of compound 12
3.5.1. HSA binding
Linearity of S-V plots indicates single quenching mechanism for
all compound/HSA ratios.
Human serum albumin (HSA) is the major soluble and the most
versatile transport protein in blood circulation with plenty of
physiological functions. HSA contributes significantly to colloidal
osmotic pressure and aids the transport, distribution and meta-
bolism of many endogenous and exogenous ligands, with particular
affinity to hydrophobic, negatively charged molecules [41]. There-
fore, knowledge of binding properties of potential drug molecule to
HSA is crucial in drug design and development. HSA binding of 12
was investigated using fluorescence spectroscopy and molar ratio
method for stability constants determination.
The changes in HSA emission spectrum at T ¼ 303 K upon
addition of compound 12 are given in Fig. 10 (changes in HSA
emission spectra upon addition of 12, recorded at 293 and 310 K,
are given in Supplementary material, Fig. S26). Decrease of HSA
fluorescence intensity (fluorescence quenching) is observed during
titration at all temperatures, with the blue shift of emission max-
ima. Large blue shift indicates the formation of HSA-ADK complex,
leading to more hydrophobic microenvironment around fluores-
cent residue Trp214.
Fluorescence quenching mechanism is classified as static or
dynamic. One way to resolve the quenching mechanism is the
evaluation of binding constants' temperature dependence. Dy-
namic quenching is highly dependent upon diffusion. Higher
temperatures result in faster diffusion and hence larger values for
biomolecular quenching constant. On the other hand, higher tem-
peratures will typically result in the dissociation of weakly bound
complexes, and therefore decrease the biomolecular quenching
constant in static process. Ksv Value decreases as temperature in-
creases (Table 7), indicating static quenching mechanism. Another
confirmation of static quenching mechanism is the value of Kq,
which is significantly higher than the maximum scatter collision
quenching constant value (2 ꢂ 1010 Mꢁ1
s
ꢁ1) [44]. The quenching
process was further analyzed using modified S-V equation (4):
F0
1
1
1
¼
þ
(4)
DF
faKa ½Qꢄ fa
With the increasing amount of quencher, results may deviate
from the linearity due to instrumental inner filter effect. This effect
can be removed using Lakowicz equation (2):
where F is the difference in fluorescence intensity of HSA in the
D
absence (F0) and in the presence of the quencher at concentration
[Q], Ka represents the effective quenching constant for the acces-
sible fluorophores, and fa is the fraction of accessible fluorophore.
The results of linear fit are shown in Fig. 12 and in Table 8.
Results show that the fluorophore (Trp214) is fully accessible to
quencher (fa ~1).
þA
Fcorr ¼ Fobs ꢂ 10A
(2)
ex em
2
where Fcorr and Fobs are corrected and observed fluorescence in-
tensities, and Aex and Aem are absorbancies at the excitation and
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
11
Fig. 8. Variables having highest impact on antibacterial activity of compounds represented using compound 12 as potent (red lines), and norfloxacin as less potent compound (blue
lines). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
3.5.2. Analysis of binding equilibria e binding constant and the
number of binding sites
3.5.3. Thermodynamic binding parameters
The enthalpy ( H) and entropy change (DS) during the binding
D
When small molecules bind independently to a set of equivalent
sites on the protein, the equilibrium between free and bound
molecules, for the static quenching process, is given by Equation
(5):
of small molecule to protein can be determined measuring the
binding constants at several temperatures, and following the Van't
Hoff Equation (6):
D
H
DS
R
lnK ¼ ꢁ
þ
(6)
RT
F0 ꢁ F
log
¼ log Kb þ n log½Qꢄ
(5)
where R is the universal gas constant, T is the temperature (in K),
and K is the S-V's binding constant at the corresponding temper-
ature. Results are shown in Fig. 14 and Table 10.
According to Ross' view [45] the signs and magnitudes of ther-
modynamic parameters for protein reactions can account for the
main forces contributing to protein stability. From the thermody-
F
where F0 and F have the same meaning as in previous equations; Kb
is the binding constant, and n is the number of binding sites. Linear
dependence is shown in Fig. 13, and the results of linear regression
analysis are given in Table 9.
Compound 12 binds to one independent binding site (n ~ 1,
Table 9). As can be seen in Table 9, logKb value decreases with the
increase of temperature, which is an indication of static quenching
mechanism.
namic stand-point,
action, H<0 and S<0 reflects the van der Waals force or hydrogen
H<0 and S>0 suggesting an electrostatic
DH>0 and DS>0 implies a hydrophobic inter-
D
D
bond formation and
force.
D
D
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
12
Fig. 9. The orientation of the most active compound 12 in S. aureus CrtM (left) and DNA Gyrase (right) binding sites. Active sites are represented as surfaces coloured according to
hydrophobicity and protein residues that are 5 Å away from the ligand. The docked poses of monoanionic form (thick sticks with carbon atoms in green color) and dianionic form
(thick sticks with carbon atoms in cyan color) are overlaid. Hydrogen atoms are not shown for clarity. (For interpretation of the references to colour in this figure legend, the reader
is referred to the web version of this article.)
Fig. 10. Changes in HSA fluorescence emission spectra (c ¼ 0.5
m
M) upon addition of
comp. 12 (c12 ¼ 0.0; 0.25; 0.5; 0.75; 1; 1.5; 2.0; 3.0; 4.0
pH ¼ 7.38.
m
M); T ¼ 303 K; 30 mM PBS,
Fig. 11. S-V plot for binding of comp. 12 (0.5e8.0 eq.) to HSA (c ¼ 0.5
temperatures (293, 303 and 310 K); pH ¼ 7.38 (30 mM PBS).
mM) at three
The negative enthalpy and positive entropy change indicate that
the electrostatic forces are dominant for binding of 12 to HSA.
Entropic term (TDS) is the main contributor to overall free energy of
Table 7
S-V constants at three temperatures; t0 ¼ 7.09 ns. Ksv is obtained as a slope, while
the intercept value was fixed as 1 (equation (3)).
binding.
T (K)
Ksv ꢂ 106 (Mꢁ1
)
Kq ꢂ 1014 (Mꢁ1 sꢁ1
)
r2
We might draw the conclusion that compound 12 strongly binds
to HSA and quenches its fluorescence intensity through static
quenching mechanism. Trp214 as a main fluorophore of HSA is fully
accessible for quenching with 12. Although HSA possesses many
binding sites, 12 binds to one independent binding site. Electro-
static forces are the principal forces responsible for the formation of
12eHSA complex, with the entropic term being the main contrib-
utor to the free energy of binding.
293
303
310
1.72 0.04
1.60 0.03
1.48 0.03
2.42
2.26
2.09
0.997
0.998
0.998
with HSA as a drug carrier, a strategy often employed in active
therapeutics development [47].
The binding constant of 12 to HSA is relatively high, but still
around one order of magnitude lower than the binding constant of
myristic acid [46]. Therefore, this molecule can be efficiently stored
and transported in the body. Potential problems with pharmaco-
kinetics of this molecule may be overcome by making conjugates
3.5.4. Acidity constants of 12
Aryldiketo acids in aqueous solution act as diprotic acids. pKa
Value of carboxylic group is near 2 (pKa1), [48]; the second proton
dissociates from enolic eOH group (pKa2 value around 7). Spec-
trophotometric titration was used to experimentally determine pKa
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
13
Fig. 14. The plot of lnK vs. 1/T for the interaction of 12 with HSA.
Fig. 12. Modified S-V plot for binding of 12 (0.5e8.0 eq.) to HSA (c ¼ 0.5
temperatures (293, 303 and 310 K); pH ¼ 7.38 (30 mM PBS).
mM) at three
Table 10
Thermodynamic parameters of compound's 12 binding to HSA.
T
(K)
lnK
D
G
D
H
DS
Table 8
(Mꢁ1
)
(kJmolꢁ1
)
(kJmolꢁ1
)
(Jmolꢁ1Kꢁ1
)
Effective quenching constants at three temperatures.
293
303
310
14.36
14.29
14.21
34.97
35.99
36.62
ꢁ6.64
96.76
T (K)
Ka ꢂ 106 (Mꢁ1
)
fa
r2
293
303
310
1.06 0.02
1.02 0.03
0.95 0.03
1.14 0.03
1.15 0.05
1.17 0.06
0.998
0.996
0.994
Fig. 15. Absorption spectra of 12 used for pKa2 value determination (fully deproto-
nated, A2ꢁ, form is shown as blue line, HAꢁ form shown as red line). (For interpretation
of the references to colour in this figure legend, the reader is referred to the web
version of this article.)
Fig. 13. Log-log plot for determination of binding constant (logKb), and the number of
binding sites (n) for binding of 12 to HSA at three temperatures.
process going on in the solution within studied pH range. Spectral
Table 9
Binding constant (logKb) and the number of binding sites (n) for binding of 12 to HSA
at three temperatures.
data are processed according to equation (1) and pKa2
value
7.56 0.04 is obtained. In solutions with pH < 4, where 12 is pre-
sent in H2A and/or HAꢁ form, solubility is lower, thus compound 12
precipitates. Therefore, the pKa1 value of 12 could not be deter-
mined under these experimental conditions. According to pKa2
value, pH-dependent distribution of HAꢁ and A2ꢁ was calculated:
at blood pH value (7.4), compound 12 exists as a mixture of mon-
oanionic (59% HAe) and dianionic form (41% A2ꢁ).
T (K)
logKb (Mꢁ1
)
n
r2
293
303
310
7.02 0.09
6.84 0.12
6.75 0.15
1.14 0.02
1.11 0.02
1.10 0.03
0.999
0.998
0.996
values of 12. Absorption spectra changes of 12 with pH change are
shown in Fig. 15. The existence of clearly visible isosbestic point at
278.3 nm proves that the dissociation of eOH group is the only
3.5.5. logD7.4 determination of compound 12
The logD7.4 value (logD ¼ 2.6
0.1) was determined using
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
14
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
[7] J.A. Grobler, K. Stillmock, B. Hu, M. Witmer, P. Felock, A.S. Espeseth, A. Wolfe,
shake-flask method.
M. Egbertson, M. Bourgeois, J. Melamed, J.S. Wai, S. Young, J. Vacca,
D.J. Hazuda, Diketo acid inhibitor mechanism and HIV-1 integrase: implica-
tions for metal binding in the active site of phosphotransferase enzymes, Proc.
Natl. Acad. Sci. 99 (2002) 6661e6666, https://doi.org/10.1073/
pnas.092056199.
4. Conclusions
The aim of this work was to design and synthesize a series of
novel analogues of aryl diketo acids (ADKs) that have antimicrobial
activity against multidrug resistant Gram positive bacterial strains.
Among them, compound 12 having 2,5 di-cyclohexyl substitution
at the aryl moiety, has exhibited a promising activity against a
number of multidrug resistant strains and ten folds higher activity
than the standard antibiotic norfloxacin. Through molecular dock-
ing approach and synthesis of active molecules analogues, we had
demonstrated that diketo moiety is responsible for antimicrobial
activity of this series, most likely through sequestering divalent
ions in the active sites of several target proteins. This may provide a
potential to simultaneously affect different biochemical pathways
in bacteria and thus minimize the development of resistance
against this class of molecules. Albeit the most active compound
has high affinity for serum albumin, this class of molecules may
serve as a good structural source for further lead optimization
through conjugation to serum albumin or a development of novel
antimicrobials through further rational modifications to allow tar-
geting of relevant enzymes in Gram negative bacteria.
[8] Y. Zhang, Fu-Yang Lin, K. Li, W. Zhu, Y. Liu, R. Cao, R. Pang, E. Lee, J. Axelson,
M. Hensler, K. Wang, K.J. Molohon, Y. Wang, D.A. Mitchell, V. Nizet, E. Oldfield,
HIV-1 integrase inhibitor-inspired antibacterials targeting isoprenoid
biosynthesis, ACS Med. Chem. Lett.
10.1021/ml300038t.
3 (2012) 402e406, https://doi.org/
[9] O.A. Sof’ina, N.M. Igidov, E.N. Koz’minykh, N.N. Trapeznikova, Y.S. Kasatkina,
V.O. Koz’minykh, Reactions of acylpyruvic acids and 2,3-dihydrofuran-2,3-
diones with 2,3-diaminopyridine, Russ. J. Org. Chem. 37 (2001) 1017e1025,
https://doi.org/10.1023/A:1012438902959.
ꢁ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
[10] I.N. Cvijetic, T.Z. Verbic, B.J. Drakulic, D.M. Stankovic, I.O. Juranic,
ꢀ
D.D. Manojlovic, M. Zloh, Redox properties of alkyl-substituted 4-aryl-2,4-
dioxobutanoic acids, J. Serb. Chem. Soc. (2017), https://doi.org/10.2298/
JSC161118021C.
[11] S. Gibbons, E.E. Udo, The effect of reserpine, a modulator of multidrug efflux
pumps, on the in vitro activity of tetracycline against clinical isolates of
methicillin resistant Staphylococcus aureus (MRSA) possessing the tet(K)
determinant, Phyther. Res. 14 (2000) 139e140, https://doi.org/10.1002/(SICI)
1099-1573(200003)14:2<139::AID-PTR608>3.0.CO;2e8.
[12] J.I. Ross, A.M. Farrell, E.A. Eady, J.H. Cove, W.J. Cunliffe, Characterisation and
molecular cloning of the novel macrolide-streptogramin B resistance deter-
minant from Staphylococcus epidermidis, J. Antimicrob. Chemother. 24
(1989) 851e862, https://doi.org/10.1093/jac/24.6.851.
[13] J.F. Richardson, S. Reith, Characterization of a strain of methicillin-resistant
Staphylococcus aureus (EMRSA-15) by conventional and molecular
methods, J. Hosp. Infect. 25 (1993) 45e52, https://doi.org/10.1016/0195-
6701(93)90007-M.
[14] G.W. Kaatz, S.M. Seo, C.A. Ruble, Efflux-mediated fluoroquinolone resistance
in Staphylococcus aureus, Antimicrob. Agents Chemother. 37 (1993)
1086e1094.
Conflict of interest
The authors confirm that this article content has no conflict of
interest.
[15] E.C. for A.S.T. (EUCAST) of the E.S. of C.M. and I.D. (ESCMID), Determination of
minimum inhibitory concentrations (MICs) of antibacterial agents by broth
dilution, Clin. Microbiol. Infect.
j.1469-0691.2003.00790.x.
9 (2003) ixexv, https://doi.org/10.1046/
Acknowledgements
[16] T.S. Rush, J.A. Grant, L. Mosyak, A. Nicholls, A shape-based 3-D Scaffold
Hopping method and its application to a bacterial ProteinꢁProtein interac-
tion, J. Med. Chem. 48 (2005) 1489e1495, https://doi.org/10.1021/
jm040163o.
Ministry of Education, Science and Technological Development
of the Republic of Serbia supported this work, Grant No. 172035.
University of Hertfordshire is acknowledged for its support to the
project.
ꢀ
ꢀ
[17] D. Bajusz, A. Racz, K. Heberger, Why is Tanimoto index an appropriate choice
for fingerprint-based similarity calculations? J. Cheminform 7 (2015) 1e13,
https://doi.org/10.1186/s13321-015-0069-3.
[18] P.C.D. Hawkins, A.G. Skillman, G.L. Warren, B.A. Ellingson, M.T. Stahl,
Conformer generation with OMEGA: algorithm and validation using high
quality structures from the protein databank and Cambridge structural
database, J. Chem. Inf. Model 50 (2010) 572e584, https://doi.org/10.1021/
ci100031x.
[19] T.A. Halgren, MMFF VI. MMFF94s option for energy minimization studies,
J. Comput. Chem. 20 (1999) 720e729, https://doi.org/10.1002/(SICI)1096-
987X(199905)20:7<720::AID-JCC7>3.0.CO;2-X.
[20] J.J.P. Stewart, Optimization of parameters for semiempirical methods V:
modification of NDDO approximations and application to 70 elements, J. Mol.
Model 13 (2007) 1173e1213, https://doi.org/10.1007/s00894-007-0233-4.
[21] J.J.P. Stewart, MOPAC: a semiempirical molecular orbital program, J. Comput.
Aided. Mol. Des. 4 (1990) 1e105, https://doi.org/10.1007/BF00128336.
Appendix A. Supplementary data
Supplementary data associated with this article can be found in
the online version, at https://doi.org/10.1016/j.ejmech.2017.10.045.
These data include MOL files and InChiKeys of the most important
compounds described in this article.
References
[22] P.J. Goodford,
A computational procedure for determining energetically
[1] E. Oldfield, X. Feng, Resistance-resistant antibiotics, Trends Pharmacol. Sci. 35
(2014) 664e674, https://doi.org/10.1016/j.tips.2014.10.007.
favorable binding sites on biologically important macromolecules, J. Med.
Chem. 28 (1985) 849e857, https://doi.org/10.1021/jm00145a002.
[23] F. Fontaine, M. Pastor, F. Sanz, Incorporating molecular shape into the
alignment-free GRid-INdependent descriptors, J. Med. Chem. 47 (2004)
2805e2815, https://doi.org/10.1021/jm0311240.
[24] M. Clark, R.D. Cramer, The probability of chance correlation using partial least
squares (PLS), Quant. Struct. Relat. 12 (1993) 137e145, https://doi.org/
10.1002/qsar.19930120205.
[25] G. Klebe, U. Abraham, T. Mietzner, Molecular similarity indices in a compar-
ative analysis (CoMSIA) of drug molecules to correlate and predict their bio-
logical activity, J. Med. Chem. 37 (1994) 4130e4146, https://doi.org/10.1021/
jm00050a010.
[26] A.A. Alizadeh, M. Hamzeh-Mivehroud, B. Sokouti, S. Dastmalchi, An
alignment-independent 3D-QSAR study on series of hydroxamic acid-based
[2] F. Rossi, L. Diaz, A. Wollam, D. Panesso, Y. Zhou, S. Rincon, A. Narechania,
G. Xing, T.S.R. Di Gioia, A. Doi, T.T. Tran, J. Reyes, J.M. Munita, L.P. Carvajal,
~
A. Hernandez-Roldan, D. Brandao, I.M. van der Heijden, B.E. Murray,
P.J. Planet, G.M. Weinstock, C.A. Arias, Transferable vancomycin resistance in a
community-associated MRSA lineage, N. Engl. J. Med. 370 (2014) 1524e1531,
https://doi.org/10.1056/NEJMoa1303359.
[3] W. Zhu, Y. Zhang, W. Sinko, M.E. Hensler, J. Olson, K.J. Molohon, S. Lindert,
R. Cao, K. Li, K. Wang, Y. Wang, Y.-L. Liu, A. Sankovsky, C.A.F. de Oliveira, D.
a Mitchell, V. Nizet, J.A. McCammon, E. Oldfield, Antibacterial drug leads
targeting isoprenoid biosynthesis, Proc. Natl. Acad. Sci. U. S. A. 110 (2013)
123e128, https://doi.org/10.1073/pnas.1219899110.
[4] G.Y. Liu, A. Essex, J.T. Buchanan, V. Datta, H.M. Hoffman, J.F. Bastian, J. Fierer,
V. Nizet, Staphylococcus aureus golden pigment impairs neutrophil killing and
promotes virulence through its antioxidant activity, J. Exp. Med. 202 (2005)
209e215, https://doi.org/10.1084/jem.20050846.
[5] C.-I. Liu, G.Y. Liu, Y. Song, F. Yin, M.E. Hensler, W.-Y. Jeng, V. Nizet, A.H.-
J. Wang, E. Oldfield, A cholesterol biosynthesis inhibitor blocks Staphylococcus
aureus virulence, Sci. 319 (2008) 1391e1394, https://doi.org/10.1126/
science.1153018.
tumor necrosis factor-
a converting enzyme inhibitors, J. Chemom. (2016)
537e547, https://doi.org/10.1002/cem.2817.
ꢁ
ꢁ
ꢁ
ꢀ
ꢀ
ꢀ
ꢀ
[27] I.N. Cvijetic, Z.P. Zizak, T.P. Stanojkovic, Z.D. Juranic, N. Terzic, I.M. Opsenica,
ꢀ
ꢀ
D.M. Opsenica, I.O. Juranic, B.J. Drakulic, An alignment independent 3D QSAR
study of the antiproliferative activity of 1,2,4,5-tetraoxanes, Eur. J. Med. Chem.
45 (2010) 4570e4577, https://doi.org/10.1016/j.ejmech.2010.07.019.
ꢁ
ꢁ
ꢁ
[28] L.M. Kabeya, C.H.T.P. da Silva, A. Kanashiro, J.M. Campos, A.E.C.S. Azzolini,
A.C.M. Polizello, M.T. Pupo, Y.M. Lucisano-Valim, Inhibition of immune
ꢀ
ꢁ
ꢀ
ꢀ
[6] B.J. Drakulic, M. Stavri, S. Gibbons, Z.S. Zizak, T.Z. Verbic, I.O. Juranic, M. Zloh,
Aryldiketo acids have antibacterial activity against MDR Staphylococcus
aureus strains: structural insights based on similarity and molecular interac-
tion fields, ChemMedChem. 4 (2009) 1971e1975, https://doi.org/10.1002/
cmdc.200900273.
complex-mediated neutrophil oxidative metabolism:
model for 3-phenylcoumarin derivatives using GRIND-based 3D-QSAR and
2D-QSAR procedures, Eur. J. Med. Chem. 43 (2008) 996e1007, https://doi.org/
a pharmacophore
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045

ꢀ
I.N. Cvijetic et al. / European Journal of Medicinal Chemistry xxx (2017) 1e15
15
10.1016/j.ejmech.2007.07.003.
https://doi.org/10.1093/jac/dkg044.
[39] T.Z. Verbic, B.J. Drakulic, M. Zloh, I.O. Juranic, The effect of phenyl substituents
on C-13 NMR shifts and metal ions binding to 4-phenyl-2,4-dioxobutanoic
acid derivatives, Lett. Org. Chem.
10.2174/157017808786857589.
[40] N.M. Vandevelde, P.M. Tulkens, F. Van Bambeke, Modulating antibiotic activity
towards respiratory bacterial pathogens by co-medications: a multi-target
approach, Drug Discov. Today 21 (2016) 1114e1129, https://doi.org/
10.1016/j.drudis.2016.04.001.
ꢁ
ꢀ
ꢀ
ꢀ
[29] I. Jabeen, P. Wetwitayaklung, P. Chiba, M. Pastor, G.F. Ecker, 2D- and 3D-QSAR
studies of a series of benzopyranes and benzopyrano[3,4b][1,4]-oxazines as
inhibitors of the multidrug transporter P-glycoprotein, J. Comput. Aided. Mol.
Des. 27 (2013) 161e171, https://doi.org/10.1007/s10822-013-9635-9.
5 (2008) 692e699, https://doi.org/
ꢀ
ꢀ
[30] A. Duran, G.C. Martínez, M. Pastor, Development and validation of AMANDA, a
new algorithm for selecting highly relevant regions in molecular interaction
fields, J. Chem. Inf. Model 48 (2008) 1813e1823, https://doi.org/10.1021/
ci800037t.
[31] M. Baroni, G. Costantino, G. Cruciani, D. Riganelli, R. Valigi, S. Clementi,
Generating optimal linear PLS estimations (GOLPE): an advanced chemo-
metric tool for handling 3D-QSAR problems, Quant. Struct. Relat. 12 (1993)
9e20, https://doi.org/10.1002/qsar.19930120103.
[41] X.M. He, D.C. Carter, Atomic structure and chemistry of human serum albu-
min, Nature 358 (1992) 209e215.
[42] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, third ed., Springer, US,
2006 https://doi.org/10.1007/978-0-387-46312-4.
[32] X. Liu, S. Ouyang, B. Yu, Y. Liu, K. Huang, J. Gong, S. Zheng, Z. Li, H. Li, H. Jiang,
PharmMapper server: a web server for potential drug target identification
using pharmacophore mapping approach, Nucleic Acids Res. 38 (2010)
W609eW614, https://doi.org/10.1093/nar/gkq300.
[33] B. Ojha, G. Das, The interaction of 5-(Alkoxy)naphthalen-1-amine with bovine
serum albumin and its effect on the conformation of protein, J. Phys. Chem. B
114 (2010) 3979e3986, https://doi.org/10.1021/jp907576r.
[43] M. Amiri, K. Jankeje, J.R. Albani, Origin of fluorescence lifetimes in human
serum albumin. Studies on native and denatured protein, J. Fluoresc. 20
(2010) 651e656, https://doi.org/10.1007/s10895-010-0597-1.
[44] M.R. Eftink, C.A. Ghiron, Fluorescence quenching studies with proteins, Anal.
Biochem. 114 (1981) 199e227, https://doi.org/10.1016/0003-2697(81)90474-
7.
[45] P.D. Ross, S. Subramanian, Thermodynamics of protein association reactions:
forces contributing to stability, Biochemistry 20 (1981) 3096e3102, https://
doi.org/10.1021/bi00514a017.
[34] E.P. Serjeant, A. Albert, The Determination of Ionization Constants: a Labo-
ratory Manual, second ed., Chapman and Hall, London, 1971.
[35] U. EPA, Product Properties Test Guidelines OPPTS 830.7550 Partition Coeffi-
cient (n-Octanol/Water), Shake Flask Method, Washington, 1996.
[36] A. Andres, M. Roses, C. Rafols, E. Bosch, S. Espinosa, V. Segarra, J.M. Huerta,
Setup and validation of shake-flask procedures for the determination of
partition coefficients (log D) from low drug amounts, Eur. J. Pharm. Sci. 76
(2015) 181e191, https://doi.org/10.1016/j.ejps.2015.05.008.
[37] EUCAST, Determination of minimum inhibitory concentrations (MICs) of
antibacterial agents by broth dilution, Clin. Microbiol. Infect. 9 (2003) ixexv,
https://doi.org/10.1046/j.1469-0691.2003.00790.x.
[46] A.O. Pedersen, R. Brodersen, Myristic acid binding to human serum albumin
investigated by dialytic exchange rate, J. Biol. Chem. 263 (1988)
10236e10239.
[47] B. Elsadek, F. Kratz, Impact of albumin on drug delivery d new applications
on the horizon, J. Control. Release 157 (2012) 4e28, https://doi.org/10.1016/
j.jconrel.2011.09.069.
ꢁ
ꢀ
ꢀ
ꢀ
ꢀ
[48] T.Z. Verbic, B.J. Drakulic, M.F. Zloh, J.R. Pecelj, G.V. Popovic, I.O. Juranic, An
LFER study of the protolytic equilibria of 4-aryl-2,4-dioxobutanoic acids in
aqueous solutions, J. Serb. Chem. Soc. 72 (2007) 1201e1216, https://doi.org/
10.2298/JSC0712201V.
[38] S. Gibbons, M. Oluwatuyi, G.W. Kaatz, A novel inhibitor of multidrug efflux
pumps in Staphylococcus aureus, J. Antimicrob. Chemother. 51 (2003) 13e17,
ꢀ
Please cite this article in press as: I.N. Cvijetic, et al., Design, synthesis and biological evaluation of novel aryldiketo acids with enhanced
antibacterial activity against multidrug resistant bacterial strains, European Journal of Medicinal Chemistry (2017), https://doi.org/10.1016/
j.ejmech.2017.10.045