Novel C-3-(N-alkyl-aryl)-aminomethyl rifamycin SV derivatives exhibit activity against rifampicin-resistant Mycobacterium tuberculosis RpoBS522L strain and display a different binding mode at the RNAP β-subunit site compared to rifampicin

Article abstract of DOI:10.1016/j.ejmech.2021.113734

Antimicrobial resistance is a main concern in tuberculosis treatment and is often associated with the emergence of Mycobacterium tuberculosis strains resistant to rifampicin (RIF), which is one of the cornerstones of tuberculosis chemotherapy. In this study, aminoalkyl-aromatic ring tails were appended to the C3 position of rifamycin core to assess the role of C3 substitutions to the anti-mycobacterial activity of the rifamycin antibiotics. The typical hydrazone unit of RIF was replaced by an amino-alkyl linkage to connect the aromatic ring tails with the rifamycin naphthoquinone core. Eight novel C3-(N-alkyl-aryl)-aminoalkyl analogues of rifamycin SV were synthesised and screened in vitro against wild-type HR37Rv and “hypervirulent” HN-878 strains, and a panel of rifampicin-resistant M. tuberculosis clinical isolates carrying mutations at the 522, 531 and 455 positions of the rpoB gene (RpoB_S522L, RpoB_S531L and RpoB_H455D strains). The analogues exhibited anti-tubercular activity against H37Rv and HN-878 at submicromolar or nanomolar concentrations, and against clinical H37Rv isolates bearing the S522L mutations at low micromolar concentration. Benzylamine moiety-including analogue 8 was as active as rifampicin against HN-878 with a MIC₉₀ value of 0.02 μM, whereas 14 and 15, which included tryptamine and para-methyl-sulfonylbenzylamine C3-substituents, respectively, showed higher anti-tubercular activity (MIC₉₀ = 3 μM) compared to rifampicin against the S522L mutated H37Rv strain. Detailed in silico analysis of different RNAP molecular systems predicted a distinct, possibly novel, binding mode for the new rifamycin analogues. These were found to occupy a different space in the binding pockets of both wild type and mutated RNAP proteins compared to that of rifampicin. Moreover, the molecular modelling experiments investigated the ability of the novel analogues aromatic tails to establish key interactions at the RNAP binding site. These interesting findings might pave the way for generating rifamycin analogues that can overcome anti-microbial resistance in M. tuberculosis.

Full text of DOI:10.1016/j.ejmech.2021.113734

European Journal of Medicinal Chemistry 225 (2021) 113734
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Novel C-3-(N-alkyl-aryl)-aminomethyl rifamycin SV derivatives
exhibit activity against rifampicin-resistant Mycobacterium
tuberculosis RpoBS522L strain and display a different binding mode at
the RNAP b-subunit site compared to rifampicin
,
, *
Mire Zloh a b, Megha Gupta c, Tanya Parish d, Federico Brucoli e
a Faculty of Pharmacy, University Business Academy, Novi Sad, 2100, Serbia
b UCL School of Pharmacy, UCL, London, WC1N 1AX, UK
c Infectious Disease Research Institute, 1616 Eastlake Ave E, Suite 400, Seattle, WA, 98102, USA
d Center for Global Infectious Disease Research, Seattle Children's Research Institute, 307 Westlake Avenue North, Suite 500, Seattle, USA
e Leicester School of Pharmacy, De Montfort University, Leicester, LE1 9BH, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Antimicrobial resistance is a main concern in tuberculosis treatment and is often associated with the
emergence of Mycobacterium tuberculosis strains resistant to rifampicin (RIF), which is one of the cor-
nerstones of tuberculosis chemotherapy. In this study, aminoalkyl-aromatic ring tails were appended to
the C3 position of rifamycin core to assess the role of C3 substitutions to the anti-mycobacterial activity
of the rifamycin antibiotics. The typical hydrazone unit of RIF was replaced by an amino-alkyl linkage to
connect the aromatic ring tails with the rifamycin naphthoquinone core. Eight novel C3-(N-alkyl-aryl)-
aminoalkyl analogues of rifamycin SV were synthesised and screened in vitro against wild-type HR37Rv
and “hypervirulent” HN-878 strains, and a panel of rifampicin-resistant M. tuberculosis clinical isolates
carrying mutations at the 522, 531 and 455 positions of the rpoB gene (RpoBS522L, RpoBS531L and
RpoBH455D strains). The analogues exhibited anti-tubercular activity against H37Rv and HN-878 at
submicromolar or nanomolar concentrations, and against clinical H37Rv isolates bearing the S522L
mutations at low micromolar concentration. Benzylamine moiety-including analogue 8 was as active as
Received 12 April 2021
Received in revised form
26 July 2021
Accepted 28 July 2021
Available online 8 August 2021
Keywords:
Rifampicin analogues
Tuberculosis drug discovery
Reductive amination
Anti-microbial drug resistance
Medicinal chemistry
Molecular modelling
rifampicin against HN-878 with a MIC90 value of 0.02
and para-methyl-sulfonylbenzylamine C3-substituents, respectively, showed higher anti-tubercular ac-
tivity (MIC90 ¼ 3 M) compared to rifampicin against the S522L mutated H37Rv strain. Detailed in silico
mM, whereas 14 and 15, which included tryptamine
m
analysis of different RNAP molecular systems predicted a distinct, possibly novel, binding mode for the
new rifamycin analogues. These were found to occupy a different space in the binding pockets of both
wild type and mutated RNAP proteins compared to that of rifampicin. Moreover, the molecular modelling
experiments investigated the ability of the novel analogues aromatic tails to establish key interactions at
the RNAP binding site. These interesting findings might pave the way for generating rifamycin analogues
that can overcome anti-microbial resistance in M. tuberculosis.
© 2021 Elsevier Masson SAS. All rights reserved.
1. Introduction
discovery, showed enormous potential as broad-spectrum antimi-
crobials and anti-tuberculosis agents [3,4].
The rifamycins are ansa antibiotics (ansamycins) originally iso-
lated from fermentation cultures of Amycolatopsis rifamycinica
(previously mistakenly identified as Streptomyces mediterranei)
[1,2]. This family of bacterial secondary metabolites consists of
seven molecules, rifamycins A, B, C, D, E, S and SV, which, since their
Rifamycins B (1) and SV (2) were the first members of this family
to enter clinical trials as intravenous antibiotics, although these
compounds were found to be chemically unstable in vivo due their
benzoquinone core and showed some degrees of organ toxicity
after parental administration [5].
Several chemical modifications were subsequently carried out at
the rifamycin C-3 side chain to produce analogues with excellent
* Corresponding author.
E-mail address: federico.brucoli@dmu.ac.uk (F. Brucoli).
https://doi.org/10.1016/j.ejmech.2021.113734
0223-5234/© 2021 Elsevier Masson SAS. All rights reserved.

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
sterilizing activity against Mycobacterium tuberculosis (Mtb),
improved PK/ADME parameters and reduced toxicity. As part of this
drug development process, in the late 1960s the Lepetit SPA labo-
ethambutol and pyrazinamide, was marketed in Italy in 1969 and
approved by the FDA in USA in 1971 [1,4]. RIF is active against
actively growing and non-replicating (dormant) Mtb bacilli. How-
ever, this important anti-tubercular drug has several pitfalls,
including the selection of resistant mutants, if used in mono-
therapy, the occurrence of side effects, e.g., hepatotoxicity, and
cytochrome P450 induction activity, which might result in drug-
drug interaction issues.
ratories synthesised
a 3-(4-methyl-1-piperazinyl)-iminomethyl
derivative of rifamycin SV, rifampicin (3, RIF), as an orally available
drug that exhibited high bactericidal activity and enhanced intes-
tinal absorption properties (Fig. 1). RIF, which is a first-line anti-
tuberculosis drug and is used in combination with isoniazid,
Fig. 1. Structures of rifamycin derivatives developed to treat tuberculosis infections. Rifamycin B (1) and SV (2) were originally isolated from cultures of Amycolatopsis rifamycinica
and due to severe side effects were not developed as antibacterial dugs. Rifampicin (3, RIF), rifapentine (4, RPT) and rifabutin (5, RBT) are currently used to treat tuberculosis
infected patients. Early rifamycin analogue CGP-7040 (6) had longer half-life compared to RIF and showed superior activity against non-tuberculous bacteria (e.g., MAC), although its
drug development program was abandoned due to lack of financial incentives. Compound (7) is an experimental rifamycin SV derivative bearing a secondary amine unit that
showed interesting growth inhibitory properties in vitro against Mtb H37Rv.
2

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
Other rifamycins currently used in the clinic as anti-tuberculosis
drugs include rifapentine (4, RPT) and rifabutin (5, RBT). RPT (4) is
an N-amino-N0-cyclopentanyl-piperazine derivative of rifamycin SV
developed at Lepetit SPA in the late 1960s as a long-acting version
of RIF (3) and approved by the FDA in 1998 [6]. RBT (5) is a
spiropiperidyl-rifamycin discovered by the Achifar drug company
in the mid-1970s that gained FDA approval in 1992 to treat Myco-
bacterium avium complex (MAC) disease in AIDS patients. RBT is
also used to treat tuberculosis [7]. RIF, RPT and RBT are on the list of
WHO essential medicines [8]. A number of earlier ryfamycins were
also investigated to treat TB, including rifalazil (KRM-1648) and 3-
(2,4,6-trimethylbenzylpiperazinyl)rifamycin SV (6, CGP-7040),
although their development was terminated due to either
adverse side effects in patients or shortage of drug discovery funds
(Fig. 1) [9e11].
protocol was devised to incorporate at the C-3 position of rifamycin
SV a small set of amino-alkyl aromatic/heteroaromatic rings that
contained H-bond-accepting and -donating groups. An
aminomethyl-piperazinyl-ethanol moiety was also included at the
end of the C-3 alkyl tail of the macrolide to furnish a RIF-analogue
(16) with increased hydrophilicity and a lower LogP value (4.1)
compared to that of parent compound 3 (LogPRIF ¼ 4.7) (physico-
chemical properties were predicted using ACD Labs ACD/Phys
Chem suite). The cleavable hydrazone unit of RIF was substituted
with a more stable, protonable, secondary amino linker, which
according to previous work [23], might increase intermolecular H-
bond interactions and enable rifamycin analogues to resist myco-
bacterial inactivation via hydrolysis of their C-3 tails. As a defence
mechanism against rifampicin, RIF-resistant M. tuberculosis strains
use their enzymatic armoury to hydrolyse the hydrazone unit of the
ansa antibiotic [29] and it is anticipated that the aminoalkyl linker
of the novel rifamycin analogues might inhibit this inactivation
process. The effectiveness of the rifamycin SV derivatives against
selected RR-TB strains was measured using liquid, whole-cell
phenotypic assays to determine minimum inhibitory concentra-
tions (MIC90). Interactions of the new compounds with RIF RNAP
binding site were extensively investigated using molecular docking
tools.
RIF (3) inhibits mycobacterial transcription by binding to the b-
subunit of DNA-dependent RNA polymerase (RNAP). RNAP reads
DNA sequences and catalyses the polymerisation of complemen-
tary RNA chains using nucleotide building-blocks, thus being ulti-
mately responsible for the transcription and expression of
mycobacterial genes [12]. The increase of resistance to RIF and RFB
in Mtb is a result of mutations in the 81-bp area of the rpoB gene,
termed RIF resistance-determining region (RRDR), which encodes
for the b-subunit (RpoB protein) of RNA polymerase [13].
Globally, in 2019 3.3% of new TB cases and 18% of previously
treated cases were either multidrug- or rifampicin-resistant
tuberculosis (MDR/RR-TB), with an estimated 465 000 incident
cases of RR-TB [14]. The emergence of multi-drug resistant strains
and the lack of new drugs are two main factors that contribute to
the re-emergence of tuberculosis.
1.1. Chemistry
The reductive amination reaction of 3-formylrifamycin 7 with
eight primary alkyl(aromatic) amines was carried out in a parallel
synthetic fashion using sodium triacetoxyborohydride [NaB-
H(OAc)3] as a mild and selective reducing agent (Scheme 1). Pre-
vious reductive amination attempts on the same substrate using
sodium cyanoborohydride (NaBH3CN) resulted in lower-yielding,
slower and less clean reactions compared to NaBH(OAc)3 and
therefore NaBH3CN was not further pursued as the reducing agent.
The reactions proceeded smoothly, and the title compounds
(8e16) were produced in medium to high yield. The presence of a
protonated nitrogen atom at the C-38 position was corroborated by
NMR analysis, as can be noted in the proton spectrum of 8 (Fig. 2a).
This confirmed previous findings that C-3 amino linked rifamycin
conjugates undergo intramolecular zwitterionisation as a result of
proton transfer between the acidic naphthoquinone C-8-OH and C-
38-nitrogen in protic solvents or in the presence of water [30,31].
The structure assignment of carbon and proton nuclei of de-
rivative 8 are illustrated in the 1He13C HSQC spectrum of Fig. 2b,
Table S1 and Fig. S1 (HMBC spectrum).
The RNAP core enzyme (400 kDa) consists of the five subunits:
a
-dimer (
a2), b subunit, b’ subunit and u, which form a holoenzyme
and initiate transcription from promoters. RIF binds to the RNAP
b
-
subunit, which is located near the DNA/RNA channel, by forming
hydrogen bond interactions between its four hydroxyl groups at C-
1, C-8, C-21, C-23 and acetoxy carbonyl oxygen at C-25, and key
polymerase amino acid residues [15]. The C3-hydrophobic chain of
RIF do not appear to be involved in crucial interactions at the RNAP
active site and might modulate the antibiotic activity of rifamycin
derivatives by improving their bacterial membrane permeation
abilities [15]. RIF inhibits RNAP by steric occlusion blocking the exit
of the growing RNA chain [16].
Mutations at positions 526 and 531 of the RpoB protein are
generally associated with high-level of RIF resistance with high RIF
MICs, whereas isolates bearing mutations at positions 516, 521, 522
and 533 exhibit moderate resistance levels to the rifamycin anti-
biotics RIF and RFB [17,18].
Reactions of 3-formylrifamycin SV with primary amines and
amino acids [19] has been attempted in the past and several de-
rivatives, which were modified at their C3 and C4 positions and
contained amines, ketone, hydrazine and heterocyclic moieties and
sulphonium and phosphonium ylides and oximes, have been pre-
pared [20e27]. However, a focussed library of rifamycin SV de-
rivatives bearing amino alkyl-heterocyclic chains at the C3 position
has not been previously explored or screened for Mtb growth in-
hibition properties. In earlier work, benzyl moieties have been
appended to either piperazinyl unit directly linked to the C-3 po-
sition of the ansa macrolide core (6) [9,28], or piperidyl ring
attached via methylamino bridge to the rifamycin SV C-3 position
[23], leading to analogues (e.g., 7) with anti-tubercular activities
against non-RR-TB Mtb strains comparable to that of RIF.
1.2. Strains RRDR mutations analysis and antitubercular activity
evaluation
The rpoB gene of the 65 RR-isolates examined in this study was
successfully sequenced from codon 507 through 533 to identify
clinically relevant mutations (Table 1). The most frequent mutation
(41%) in the RpoB RRDR was found be at the 531 position, with 18
isolates bearing the S531L mutations (serine replaced by leucine).
The second most frequent mutation (40%) was at position 526, with
12 isolates carrying the H526D (histidine replaced by aspartic acid).
A relevant number of isolates also possessed the S522L (9%)
mutation.
The novel rifamycin derivatives were screened for whole-cell
growth inhibition of M. tuberculosis H37Rv and HN878 wild-type
strains, and RR-resistant strains RpoBS522L RpoBS531L and
RpoBH455D (HN-0258218-RM1), which was isolated as a sponta-
neous rifampicin resistant mutant from HN-878,32 and minimum
inhibitory concentrations (MICs) were determined at five days
(Table 2). The “hypervirulent” Mtb HN878 was selected as it is a
Here, we sought to investigate the contributions of the hydra-
zone linkage and methyl-piperazinyl tail of RIF to the anti-
tubercular activity of the parent compound against wild-type and
RR-TB Mtb strains. To this end, a robust reductive amination
3

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
Scheme 1. Synthesis of rifamycin analogues 8e16. Reagents and conditions: amines (1 equiv.), NaBH(OAc)3 (1.4 equiv.), dichloroethane (DCE, 5 mL), room temperature, 3 h, 55e83%
yield.
particularly insidious strain due to its ability to grow fast and
drastically reduce the survival rate of immune-competent infected
mice [33]. RIF-resistant strains with S531L and S522L mutations
were chosen as they are generally found in isolates from a high
proportion of MDR-TB patients [17].
analogues 14 and 15, which contained tryptamine and
methylsulfonyl-benzylamine residues, respectively, exhibited cidal
activity against Mtb RpoBS522L at a concentration as low as
3.1e3.2
>8 M).
mM, whereas RIF was not effective against this strain (MIC
m
The new rifamycin analogues 8e16 were active against Mtb
H37Rv, HN-878 and RpoBS522 RIF-resistant strains, but not against
the RpoBS531L RR-strain. Mtb H37Rv was found to be the most
susceptible strain to the anti-tubercular activity of the compounds.
Analogues 8,10, 11, 12, 13, 15, 14 and 16 exhibited significant growth
inhibitory properties against H37Rv with MIC values ranging from
1.3. Molecular modelling studies
Docking studies were carried out to evaluate the binding mode
of the RIF-analogues and investigate the interactions of the com-
pounds’ functional groups, including their amino-methyl aromatic
tails or aminomethyl-piperazinyl-ethanol moiety (16), with the
0.034 to 0.097
compared to RIF (MIC ¼ 0.0045
at a concentration of 0.12 0.046
m
M, albeit with a 7- to 20-fold reduced potency
M). Compounds 9 was only active
M, indicating that hydrogen
m
amino acid residues in the binding pockets of the b-subunits of
m
wild-type (PDB ID 5UHC) [34], and S531L- and S522L-mutated
RNAP enzymes. The optimized 3D structures of rifampicin ana-
logues were docked into the RNAP binding site defined by the
location of rifampicin in the complex. The resulting docking scores
of the zwitterionic analogues (Table 3) indicated favourable in-
teractions between the analogues and the amino acid residues of
the RNAP. Although there is no direct correlation between the
docking score and MICs of the majority of tested molecules, the
molecular docking has revealed possible a novel mode of binding
for the rifampicin analogues.
bond acceptor groups, e.g., methyl carboxylate unit, are not well
tolerated in the RNAPH37Rv binding pocket.
Benzylamino residue-containing analogue 8 was as effective as
RIF (3) in killing the hypervirulent strain HN-878 with a MIC90
value of 0.026 0.035
14 were less active against HN-878 compared to 8 with MIC90
ranging from 0.20 to 0.53 M. Compounds 15 and 16 arrested the
growth of HN-878 at concentrations of 0.096 0.064 and
0.070 0.042 M, respectively. Compounds 8, 10 and 11 were
mM. Rifamycin derivatives 9, 10, 11, 12, 13 and
m
m
active, albeit at higher concentrations (9e16 mM), against the RR
Molecular docking of the analogues into the rifampicin binding
pocket of the wild type RNAP crystal structure (PDB ID 5UHC)
indicated a possibility of three distinct modes of their interaction
with RNAP (Figs. S2aed). Most of the analogues occupied the same
strain (RpoBH455D) isolated from HN-878.
The title compounds 8e16 inhibited the growth of Mtb
RpoBS522L with MIC values ranging from 3 to 12 M. Notably,
m
4

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
Fig. 2. a) 1H NMR spectrum of 8 within the 5.9e9.5 ppm range recorded in DMSO‑d6 at 600 MHz highlighting the presence of the protonated amino group at C-38. b) 1He 13C-HSQC
spectrum of 8 in DMSO‑d6. Spectral region at dH ꢀ0.50 e 7.50.
space as rifampicin, with their cores overlapping with rifampicin
core and different position of their tails (Fig. 3a). Consequently,
these analogues form interactions with the key residues in WT
RNAP similar to those observed for rifampicin [34]. In particular,
hydrogen bonds were observed between those analogues and
amino acids Q435, Q438, R454 and S456. Analogues 15 has a pos-
sibility to form additional hydrogen bond interactions with nucle-
otide and G491 (Fig. 3a).
Analogues 12 and 13 partially occupied the cavity with their
cores occupying the similar space as the rifampicin tail (Fig. S2b).
These analogues do not form the hydrogen bonds with the same
residues despite being in close proximity, but their binding poses
are stabilised by forming hydrogen bonds with R173 and F439 and
via interactions with hydrophobic residue I497. While the most
favourable pose of 16 partially occupied the RNAP binding site
(Fig. S2c), the second pose occupied a similar space as rifampicin
(Fig. S2d). Furthermore, some of the less favourable docking poses
of other analogues were adopting conformations that are less
deviating from the rifampicin binding mode.
There is a lack of direct correlation between the docking scores
and experimentally observed activities (Tables 1 and 3), that may
be due not only to different space of the binding site the ligands
occupy when compared to rifampicin, but also due to their multiple
conformations that can be formed within the binding site (Fig. S4).
5

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
While rifampicin forms fewer conformations with one preferred, its
analogues can form between 7 and 13 conformations, and some are
unlikely to be bioactive. As the binding scores can be very similar,
the probability of having both bioactive and inactive conformations
are high, which could be one of the reasons for the lower activities
of rifampicin analogues. The higher number of possible binding
modes most likely results from introduction of rotatable bonds in
the tails, thus increasing the flexibility of analogues. The number of
torsion angles that can be changed increased from 5 in rifampicin to
either 7, 8 or 9. Additionally, the higher predicted logP and lower
predicted solubility of analogues, which are reflected in lower
predicted drug likeness, when compared to rifampicin, may also
contribute to lower biological activity (Table S2). Additionally, these
difference in molecular properties may also affect the permeability
of the analogues as demonstrated for a different set of rifampicin
analogues [30].
Notably, rifampicin and analogues did not fully occupy the
available space inside the binding site, and one of the pockets at the
bottom of the cleft may be explored by the analogues (Fig. S2b).
Since the protein, DNA and RNA conformations from the crystal
structure may lead to biased docking results, a short molecular
dynamics simulation was conducted to relax the 3D structure of the
whole system. The docking of all analogues against the 3D structure
extracted from the final frame of the trajectory has resulted in the
docking poses that had considerable higher docking score. The
careful analysis of the results indicated that the shape of the
binding cavity was changed due to the change of RNA chain
conformation and shrinking the space available for ligand binding.
This resulted in rifampicin and all analogues interacting mainly
with the RNA chain and therefore providing potentially misleading
information.
Therefore, the binding of analogues with the protein only was
explored by considering the 3D structure of the beta subunit on its
own, an approach that was previously reported elsewhere [35].
Consequently, the structure of the beta subunit and rifampicin
complex was extracted from the crystal structure file [34] and
subjected to a 10 ns molecular dynamics simulation of a partially
restrained protein chain that was more than 10 Å away from the
rifampicin as a preparation of the protein structure for the docking.
Similar to results of the docking against the whole RNAP, different
binding modes were observed. The first pose of the cores of 9 and
11 occupied the similar space as rifampicin (Fig. S3a), while most
analogues had their tails occupying previously unexplored space of
the binding site (pocket marked with a red circle) that can be seen
on the bottom left of Fig. S2b. Their cores (except 16) did not occupy
the similar space as the rifampicin and have formed set of in-
teractions with a new residue within previously unoccupied space
of the RNAP binding site, namely T488 (Fig. 4a).
Finally, the reasons behind a diminished anti-tubercular activity
of the analogues due to S531L mutation were explored in silico by
replacing the serine side chain with a leucine residue at 456 posi-
tion in the PDB file with RNAP crystal structure (PDB ID 5UHC) [34].
The 50 ns molecular dynamics simulation was conducted on the
modified protein and the final frame of the trajectory was used as a
target in the additional docking experiments. Albeit the docking
scores are lower when compared to those obtained for the wild
type protein, these cannot provide an explanation for the complete
loss activity of these analogues and rifampicin. However, the shape
of the binding site in the wild type protein appears considerably
deeper than the putative binding site in the mutated protein
(Fig. 5a-b). Interestingly, it can be observed that introduction of the
leucine instead of serine changes the conformation on surrounding
hydrophobic residues (V176, L458 and I497 in Fig. 5b). These
changes not only lead to modifications to the volume, but also have
a profound effect on the hydrophobicity of the cleft. This, in turn,
6

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
Table 2
Minimum inhibitory concentrations (MICs) of 8e16 in M. tuberculosis H37Rv, HN-878 and rifampicin resistant strains RpoBS522L, RpoBS531L and RpoBH455D isolated from HN-
878.
Cmp ID
Mycobacterium tuberculosis strains MIC90 (m
M)a
H37Rv-LPb
HN-878c
RpoBS522L H37Rvd
RpoBS531L
RpoBH455D
e
f
8
9
10
11
12
13
14
15
0.097 0.041
0.12 0.046
0.090 0.028
0.082 0.021
0.075 0.021
0.066 0.019
0.039 0.060
0.039 0.046
0.034 0.028
0.0045
0.026 0.035
0.53 0.064
0.45 0.140
0.20 0.057
0.31 0.140
0.22 0.045
0.39 0.010
0.096 0.064
0.070 0.042
0.02
12
>20
>20
>20
>20
>20
>20
>20
>20
>20
>8.0
0.55
9.9
>20
16
7.7
>20
8
6.7
5.8
3.1
3.2
4.7
>8.0
0.29
14
>20
>20
>20
>20
>20
>8.0
0.62
16
RIF (3)
Isoniazid
0.28
0.29
a
MIC90 was defined as the concentration required to inhibit growth of M. tuberculosis in liquid medium by 90% after 5 days.
H37Rv (ATCC 25618 Wild-type).
HN-878 wild-type.
(RIF-R1) Rifampicin-resistant strain (RpoBS522L H37Rv-LP).
(RIF-R2) Rifampicin-resistant strain (RpoBS531L).
b
c
d
e
f
(HN-0258218-RM1) Rifampicin-resistant strain (RpoBH455D) isolated from HN-878. The screening was conducted in triplicate.
Table 3
Docking scores (kcal/mol) of zwitterionic 8e16 against M. tuberculosis H37Rv RpoB and rifampicin resistant strains RpoBS522L and RpoBS531L as targets. The docking scores were
predicted using LeDock software. The docking score in brackets correspond to second pose.
Cmp RpoB H37Rv whole
ID
RpoB H37Rv whole system (10 ns RpoB H37Rv b-subunit (10 ns RpoBS531L H37Rv whole system (50 RpoBS522L H37Rv b-subunit (150 ns
system[a]
simulated)
simulated)
ns simulated)
simulated)
8
9
ꢀ7.9
ꢀ8.5
ꢀ9.2
ꢀ8.4
ꢀ8.8
ꢀ8.7
ꢀ8.7
ꢀ9.5
ꢀ9.5
ꢀ9.8
ꢀ8.7
ꢀ6.5
ꢀ6.9
ꢀ6.4
ꢀ6.0
ꢀ6.7
ꢀ6.6
ꢀ7.4
ꢀ7.1
ꢀ6.5
ꢀ6.1
ꢀ7.2
ꢀ8.3
ꢀ7.2
ꢀ7.4
ꢀ7.5
ꢀ7.4
ꢀ8.6
ꢀ8.2
ꢀ7.9
ꢀ7.4
ꢀ6.8
ꢀ8.7
ꢀ7.6
10
11
12
13
14
15
16
ꢀ7.9
ꢀ6.8
ꢀ8.1
ꢀ7.1
ꢀ7.8
ꢀ7.0
ꢀ7.9
ꢀ6.9
ꢀ8.7
ꢀ7.7 (ꢀ7.7)
ꢀ7.8 (ꢀ7.3)
ꢀ7.4
ꢀ8.5
ꢀ7.0 (ꢀ6.8)
RIF ꢀ7.1
ꢀ6.5
(3)
[a] wild type protein PDB ID 5UHC (system was not simulated prior to docking). [b] Binding score for the second pose.
has potentially a significant effect on ligand binding and can pro-
vide a structural information regarding the lack of activity against
this particularly resistant strain.
scores somewhat correlated with the compounds’ MIC90 values (3.1
and 3.2 M, respectively), which was one of the lowest of the series
against RpoBS522L H37Rv.
m
Moreover, the change of the binding site shape in the protein
with S522L mutation may explain the loss of anti-tubercular ac-
tivity of rifampicin (3), the different range of activities displayed by
most analogues and the increased efficacy of 15 (Fig. 5c). As the
S522 (position 447 in 5UHC) is not part of the binding site, a longer
molecular dynamics simulation of the mutated beta subunit was
conducted, which revealed that the larger leucine side chain dis-
placed nearby binding site residues (Fig. 5d). In particular, R454 and
R613 were displaced when compared to their locations in the
crystal structure [34], thus preventing formation of favourable in-
teractions and resulting in the loss of activity of 3. This is indicated
to some extent by lower binding scores obtained for rifampicin
docked against the binding site of wild type beta subunit
(ꢀ8.7 kcal/mol) and S522L mutated beta protein (ꢀ7.4 kcal/mol)
(Table 3). On the contrary, most of analogues retained their activity,
which may be a result of their potentially different binding modes,
due to their C3-aromatic tails, when compared to rifampicin.
Interestingly, in addition to previously explored binding site space,
tails of two most potent analogues were found to bind into a pre-
viously unoccupied space in the second most favourable poses. As
illustrated in Fig. 6a, the aromatic tails of 14 and 15 nicely fitted into
the space of the S522L mutated beta protein showing a favourable
docking scores of ꢀ8.6 and ꢀ8.2 kcal/mol, respectively. These
These computational investigations indicate a possible rationale
for the loss of rifampicin activity against resistant strains based on
the changes of the binding site space as a result of amino acid
mutations near the active site. However, further extended molec-
ular dynamics simulations or crystallographic studies would be
needed to confirm the exact nature of the protein structure changes
resulting from the mutations of the residues that are not part of the
binding site. Importantly, insights into putative mechanism of ac-
tion of the rifampicin analogues were revealed and opened op-
portunities for optimising the activity of RNAP inhibitors. These
studies provide a basis for further computer aided molecular design
by targeting space previously unoccupied by rifampicin.
2. Conclusions
Mutations in M. tuberculosis (Mtb) drug targets are one of the
main hurdles preventing effective treatment and management of
the tuberculosis disease. Mutations in the rpoB gene encoding for
the b-subunit of DNA-dependent RNA polymerase (RNAP) lead to
an increased resistance in Mtb to rifamycin antibiotics, such as
rifampicin (3), a first-line drug widely used for tuberculosis treat-
ment. Extensive SAR studies previously demonstrated that the
naphtohydroquinone core and hydroxyl groups attached to the
7

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
Fig. 3. aec. Binding modes of the new rifamycin analogues within the binding pocket of the whole system of wild type Mtb RNAP (3D structure extracted from PDB ID: 5UHC)[34] as
a target: a) 15 (pink sticks) and 12 (light blue sticks). Amino acid residues of the RNAP binding are shown as thin sticks coloured according to CPK scheme, while the carbon atoms of
the nucleotide are coloured in cyan. Key residues that form interactions with the analogues are labelled, while the interactions are shown as dotted lines.
aliphatic ansa bridge of rifampicin are essential pharmacophoric
characteristics enabling tight contacts with key amino acids in the
activity [36].
Here, in an effort to explore the contribution of C-3 tails to the
RNAP
b
-subunit binding site. On the other hand, the 4-methyl-1-
anti-tubercular activities and ligand binding properties of rifamycin
analogues, the piperazinyl-iminomethyl unit of rifampicin (3) was
replaced with eight amino-alkyl-aromatic rings. The latter were
connected to the naphtoquinone chromophore with a stable sec-
ondary amino linker, which in turn replaced RIF's cleavable
hydrazone linkage. The resulting novel C3-(N-alkyl-aryl)-amino-
methyl rifamycin analogues were screened against M. tuberculosis
H37Rv and HN-878 strains, and clinical isolates bearing 522, 531
and 455 mutations. Interestingly, it was found that benzylamino-
including analogue 8 was as effective as RIF (3) in inhibiting the
piperazinyl-iminomethyl unit of rifampicin might have a less
prevalent role in establishing interactions with the binding pocket,
although several authors reported different, sometimes contrast-
ing, viewpoints on the ability of the rifampicin tail to form in-
teractions with key residue E445 in the RNAP site [20,23,27]. It was
therefore postulated that structurally diverse C3-RIF tails did not
affect ligand-protein interactions but might improve the DM/PK
profiles of the molecules and alter permeability of bacterial cell
walls instead, leading to RIF variants with different bactericidal
8

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
growth of the hypervirulent HN-878 strain with a MIC90 value of
0.02 mM. The RpoBH455D HN-878 strain bearing a mutation outside
the hotspot region of rpoB gene was also sensitive to compound 8.
Moreover, tryptamine- and methyl-sulfonyl-benzyl-containing
analogues 14 and 15, respectively, were found to be active against
RpoBS522L H37Rv (MIC90 ¼ 3
mM), whilst RIF was not effective
against this strain. The S522L mutation has a high incidence in Mtb
clinical isolates, and the scaffold of rifamycin derivatives active
against strains bearing this mutation might be used for the devel-
opment of more effective anti-tubercular agents.
In silico docking studies provided a possible rationale for the
anti-tubercular activity of the novel analogues, which might have
alternative modes of ligand interactions with the RNAP subunit.
Moreover, the docking results showed that the analogues’ aromatic
tails might play a more prominent role in establishing molecular
interactions within the binding site compared to RIF piperazine
unit. Our molecular modelling experiments revealed novel binding
Fig. 4. Second favourable binding mode of 15 within the binding pocket of the beta
subunit system of wild type Mtb RNAP (3D structure extracted from the molecular
dynamics simulation trajectory). Carbon atoms of the analogue are coloured in pink
and the amino acid residues of the RNAP binding are shown as thin sticks coloured
according to CPK scheme. Key residues are labelled, while the interactions are shown
as dotted lines.
pockets within the b-subunit that can be chemically explored by ad
hoc designed rifampicin analogues. For example, the aromatic tails
of some analogues, including 14 and 15, were found to be posi-
tioned in a space of RpoBS531L binding site that was not occupied by
Fig. 5. aed. Comparison of the rifampicin binding sites of the Mtb RNAP found in the: a) crystal structure of the wild type protein (PDB entry 5UHC), b) final frame of the trajectory
obtained by molecular dynamics simulation of the protein with S531L mutation and c) final frame of the trajectory obtained by molecular dynamics simulation of the protein with
S522L mutation. Binding sites are shown as spatial distribution of hydrophilicity and hydrophobicity mapped onto a surface. Most of the pocket residues are shown in thin stick
representation, with S522, L522, S531 and L531 residues shown as thick sticks and selected hydrophobic residues shown as medium sticks in the mutated protein; d) the overlay of
the residues of the binding site extracted from the crystal structure [34] (green sticks) and S522L structure (grey sticks), where the thickness of the sticks indicates the residues that
affected the shape of the binding site in the mutated protein.
9

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
3. Experimental
3.1. General chemistry information
1H and 13C Nuclear Magnetic Resonance (NMR) spectroscopy
analyses were carried out using a JEOL JNM-ECZR 600 MHz
(equipped with a ROYAL probe) or Bruker Avance 400 MHz NMR
spectrometers. Solvent signals for hydrogen and carbon NMR were
used as the internal reference. Chemical shifts (dH) are quoted in
parts per million and are relative to the solvents residual peaks in
the 1H and 13C NMR spectra: CDCl3 (7.26 and 77.0 ppm), MeOD-d4
(3.31 and 49.1 ppm) and DMSO‑d6 (2.50 and 39.52 ppm). Coupling
constants (J) are given in Hertz (Hz) and the signal multiplicity is
described as singlet (s), doublet (d), doublet of doublets (dd), triplet
of doublets (td), triplet (t), quartet (q) and multiplet (m). Chemicals
were purchased from Acros Organic, Alfa Aesar, Fisher Scientific,
Sigma Aldrich and VWR. The deuterated solvents (CDCl3, DMSO‑d6
and MeOD-d4) used for NMR spectroscopy experiments were pur-
chased from Cambridge Isotope Laboratories Inc. Thin Layer Chro-
matography (TLC) was performed using aluminium backed
20 ꢁ 20 cm silica gel 60 F254, which were purchased from Merck for
viewing colourless spots under 254 nm wavelength ultraviolet
light. Flash column chromatography purifications of the in-
termediates and final products were conducted in a glass column
using irregular, 60 Å pore size silica gel, 63e200 mm, 70e230 mesh.
LC-MS analysis was conducted on a Thermo Fisher e Agilent 6100
series Quadrupole LC-MS system with a G4220A 1290 binary
pump/DAD. The column used was an Agilent Zorbax SB-C19
2.1 ꢁ 50 mm 1.8
mm (400 bar). Parallel synthesis was carried out
using Radleys Carousel 12 Plus reaction station.
3.2. Reductive amination procedure
The aromatic primary amine (1 equiv.) was added to a solution
of rifaldehyde 7 (0.1 mmol, 72.6 mg) in 1,2-dichloroethane (10 mL)
and subsequently treated with sodium triacetoxyborohydride (1.4
equiv., 29.7 mg). The mixture was stirred at room temperature for
2 h under nitrogen atmosphere. After HPLC analysis showed con-
sumption of the starting material, the reaction was quenched with
10% NaHCO3 (10 mL) and the compounds extracted with EtOAc
(3 ꢁ 15 mL). The organic phase was dried with MgSO4 and the
solvent was evaporated under reduced pressure to yield the free
base as a crude reddish residue, which was purified by column
chromatography using a Hexane:EtOAc/1:9 solvent system.
8. A red solid (45 mg, 55%) Rf ¼ 0.34 (EtOAc - MeOH 100:1 v/v);
MS m/z 817.4 (Mþ1); 1H NMR (600 MHz, DMSO‑d6) dH 9.27 (s, 1H,
NH-amide), 8.93 (s, 1H, NH-amine), 8.55 (s, 1H, NH-amine),
7.49e7.48 (m, 2H, H-41/45), 7.37e7.36 (m, 3H, H-42/44, H-43),
6.40e6.36 (m, 1H, H-18), 6.27e6.26 (m, 1H, H-17), 6.26e6.24 (m,
1H, H-29), 6.02 (dd, J ¼ 15.9, 7.5 Hz, 1H, H-19), 5.06 (d, J ¼ 11.1 Hz,
1H, H-25), 4.90 (dd, J ¼ 12.8, 8.5 Hz,1H, H-28), 4.25e4.23 (m, 1H, H-
38), 4.19 (s, 2H, H-39), 3.92 (d, J ¼ 9.0 Hz, 1H, C-23-OH), 3.57e3.55
(m, 2H, H-21, H-38), 3.22 (d, J ¼ 8.7 Hz, 1H, H-27), 2.87 (s, 3H, H-37),
2.81e2.77 (m, 1H, H-23), 2.25e2.21 (m, 1H, H-20), 1.97 (s, 3H, H-
36),1.94 (s, 3H, H-30),1.91 (s, 3H, H-14),1.65 (s, 3H, H-13),1.61e1.60
(m, 1H, H-22), 1.20e1.17 (m, 1H, H-24), 0.96e0.93 (m, 1H, H-26),
0.89 (d, J ¼ 7.0 Hz, 3H, H-32), 0.78 (d, J ¼ 7.0 Hz, 3H, H-31), 0.39 (d,
J ¼ 6.9 Hz, 3H, H-33), ꢀ0.38 (d, J ¼ 6.8 Hz, 3H, H-34); 13C NMR
(151 MHz, DMSO‑d6) dC 185.3 (C-11), 180.1 (C-8), 172.0 (C-6), 169.9
(C-15), 169.4 (C-35), 151.4 (C-1), 145.0 (C-4), 142.1 (C-29), 140.4 (C-
19), 137.2 (C-40), 135.3 (C-17), 131.0 (C-16), 129.8 (C-41/45), 128.7
(C-42/44), 126.3 (C-18), 119.0 (C-2), 118.2 (C-28), 118.0 (C-10), 117.6
Fig. 6. a-b. The second favourable docking poses of a) 14 and b) 15 inside the binding
site found in the simulated structure of S522L mutated Mtb RNAP. The red circle in-
dicates unoccupied spaces of the wild type Mtb RNAP as an additional pocket to
accommodate the tail of rifampicin analogues. The surface represents the binding site
coloured according to its hydrophobicity.
the rifampicin piperazine moiety. Also, it was noted that the
binding site of the RpoBS531L mutant strain was narrower compared
to the one of the H37Rv RpoB protein, this probably occurring as
direct result of the substitution of a serine with a leucine residue. In
summary, this study offers an insight in the design of more active
class of RIF-based anti-TB agents and highlights the role of selected
C-3 appended aromatic tails in increasing the molecular in-
teractions of rifamycin analogues with amino acid residues of the
RNAP
b subunit. These results might also serve as a basis for further
computational studies to explain the effects of other mutations on
rifampicin derivatives activities.
10

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
(C-9), 113.8 (C-3), 108.8 (C-12), 100.4 (C-7), 98.6 (C-5), 77.0 (C-27),
75.7 (C-23), 71.2 (C-25), 70.4 (C-21), 55.6 (C-37), 42.1 (C-38), 41.7
(C-39), 40.2 (C-26), 38.1 (C-24), 37.8 (C-20), 32.6 (C-22), 22.1 (C-13),
20.7 (C-36), 19.9 (C-30), 18.2 (C-31), 11.1 (C-32), 9.0 (C-34), 8.6 (C-
33), 7.4 (C-14); HRMS: found 816.3838, calculated for C45H56N2O12
816.3833.
117.9, 114.5, 108.8, 102.3, 76.3, 75.2, 73.1, 73.0, 55.6, 43.0, 40.0, 38.1,
32.5, 22.0, 20.6, 20.0, 19.8, 18.1, 11.0, 8.9, 8.6, 7.3. HRMS: found
817.3781, calculated for C44H55N3O12 817.3786.
13. A dark red solid (68 mg, 81%) Rf ¼ 0.31 (EtOAc - MeOH 100:1
v/v); MS m/z 847.3 (Mþ1); 1H NMR (600 MHz, DMSO‑d6) dH 9.28 (s,
1H), 8.40 (s, 1H), 8.27 (s, 1H), 7.02e7.00 (m, 2H), 6.71e6.67 (m, 2H),
6.33 (d, J ¼ 11.0 Hz, 1H), 6.27e6.24 (m, 1H), 6.19e6.15 (m, 1H),
5.96e5.91 (m, 1H), 5.06 (d, J ¼ 10.7 Hz, 1H), 4.91e4.87 (m, 1H),
4.34e4.30 (m, 1H), 4.12e4.09 (m, 1H), 4.03 (q, J ¼ 7.1 Hz, 1H), 3.90
(d, J ¼ 9.3 Hz, 1H), 3.73e3.67 (m, 1H), 3.09 (d, J ¼ 10.3 Hz, 1H), 2.93
(m, 1H), 2.87 (s, 3H), 2.82e2.79 (m, 1H), 2.24e2.19 (m, 1H), 1.97 (s,
3H), 1.91 (s, 3H), 1.65 (m, 1H), 1.27e1.23 (m, 1H), 0.92e0.90 (m, 1H),
0.86 (dd, J ¼ 10.0, 6.9 Hz, 3H), 0.81 (d, J ¼ 6.9 Hz, 3H), 0.47 (d,
J ¼ 6.2 Hz, 3H), ꢀ0.33 (d, J ¼ 6.9 Hz, 3H); 13C NMR (151 MHz,
DMSO‑d6) dC 186.6, 186.0, 172.0, 170.2, 169.5, 148.9, 144.9, 142.9,
139.0, 137.4, 136.5, 135.5, 131.7, 131.1, 130.1, 129.0, 125.9, 117.6, 117.4,
115.9, 115.1, 114.0, 108.7, 100.9, 99.5, 76.8, 74.9, 73.4, 72.9, 56.6, 47.8,
41.9, 41.2, 40.6, 38.4, 37.5, 36.5, 32.3, 22.6, 20.4, 19.2, 17.8, 11.5, 9.1,
8.9, 8.0; HRMS: found 846.3939, calculated for C46H58N2O13
846.3936.
14. A purple-red solid (61 mg, 71%) Rf ¼ 0.25 (EtOAc - MeOH
100:1 v/v); MS m/z 870.5 (Mþ1); 1H NMR (600 MHz, DMSO‑d6) dH
10.93 (s, 1H), 8.55 (bs, 1H), 7.53 (q, J ¼ 3.9 Hz, 1H), 7.35 (dd, J ¼ 8.1,
2.9 Hz, 1H), 7.24e7.19 (m, 1H), 7.10e7.07 (m, 1H), 7.02e6.96 (m, 1H),
6.34 (d, J ¼ 10.3 Hz, 1H), 6.27e6.24 (m, 1H), 6.10e6.05 (m, 1H),
5.93e5.89 (m, 1H), 5.05e4.99 (m, 1H), 4.92 (d, J ¼ 12.1 Hz, 1H), 4.39
(d, J ¼ 12.0 Hz, 1H), 4.16 (d, J ¼ 11.0 Hz, 1H), 3.92 (d, J ¼ 8.6 Hz, 1H),
3.75 (d, J ¼ 12.1 Hz, 1H), 3.69 (d, J ¼ 9.0 Hz, 1H), 3.24 (d, J ¼ 7.9 Hz,
1H), 3.10e3.05 (m, 1H), 2.91 (s, 3H), 2.83e2.80 (m, 1H), 2.29e2.24
(m, 1H), 1.99 (s, 3H), 1.92 (s, 3H), 1.91 (s, 3H), 1.65 (s, 3H), 1.25e1.23
(m,1H),1.01e0.98 (m,1H), 0.91 (d, J ¼ 7.2 Hz,1H), 0.81 (d, J ¼ 6.9 Hz,
3H), 0.30 (d, J ¼ 6.9 Hz, 3H), ꢀ0.32 (d, J ¼ 6.9 Hz, 3H); 13C NMR
(151 MHz, DMSO‑d6) dC 185.3, 184.9, 172.0, 170.3, 169.4, 149.2, 148.9,
144.4, 142.4, 137.8, 136.3, 132.1, 131.6, 131.4, 131.3, 126.7, 126.1, 123.2,
121.2, 118.5, 118.1, 115.2, 114.6, 111.5, 108.2, 101.3, 97.9, 76.1, 75.7,
73.2, 72.8, 55.7, 47.3, 43.0, 40.7, 40.0, 38.1, 37.8, 37.0, 32.6, 29.5, 22.0,
20.0, 19.8, 18.2, 11.1, 9.0, 8.4, 7.4; HRMS: found 869.4091 calculated
for C48H59N3O12 869.4099.
9. A red solid (59 mg, 65%) Rf ¼ 0.25 (EtOAc - MeOH 100:1 v/v);
MS m/z 875.5 (Mþ1); 1H NMR (400 MHz, DMSO‑d6) dH 9.22 (s, 1H),
9.09 (s, 1H), 8.54 (s, 1H), 7.91 (d, J ¼ 8.4 Hz, 2H), 7.64 (d, J ¼ 8.1 Hz,
2H), 6.52 (d, J ¼ 3.1 Hz, 1H), 6.28e6.24 (m. 1H), 6.18 (d, J ¼ 12.4 Hz,
1H), 5.96 (q, J ¼ 7.5 Hz, 1H), 5.05 (d, J ¼ 11.5 Hz, 1H), 4.91 (dd,
J ¼ 12.9, 8.1 Hz, 1H), 4.33 (bs, 2H), 4.23e4.19 (m, 1H), 4.00 (d,
J ¼ 8.3 Hz, 1H), 3.91 (s, 3H), 3.58e3.55 (m, 1H), 3.24e3.22 (m, 1H),
2.95 (s, 1H), 2.88 (s, 3H), 2.79 (t, J ¼ 7.4 Hz, 1H), 2.22e2.20 (m, 1H),
1.97 (s, 3H), 1.91 (s, 3H), 1.65 (s, 3H), 1.62 (s, 3H), 1.23 (bs, 1H), 0.88
(d, J ¼ 7.1 Hz, 3H), 0.84 (d, J ¼ 7.4 Hz, 3H), 0.42 (d, J ¼ 7.1 Hz,
3H), ꢀ0.37 (d, J ¼ 6.8 Hz, 3H); HRMS: found 874.3888, calculated for
C
47H58N2O14 874.3880.
10. A red solid (60 mg, 70%) Rf ¼ 0.29 (EtOAc - MeOH 100:1 v/v);
MS m/z 807.1 (Mþ1); 1H NMR (600 MHz, DMSO‑d6) dH 9.27 (s, 1H),
9.01 (s, 1H), 8.57 (s, 1H), 7.67 (d, J ¼ 1.4 Hz, 1H), 6.60 (d, J ¼ 3.4 Hz,
1H), 6.53e6.49 (m, 1H), 6.46 (q, J ¼ 1.6 Hz, 1H), 6.29 (d, J ¼ 10.3 Hz,
1H), 6.25 (d, J ¼ 12.4 Hz, 1H), 6.06 (q, J ¼ 7.8 Hz, 1H), 5.07 (d,
J ¼ 11.7 Hz, 1H), 4.90 (dd, J ¼ 13.1, 8.3 Hz, 1H), 4.36e4.32 (m, 1H),
4.24e4.20 (m, 2H), 3.95 (d, J ¼ 8.3 Hz, 1H), 3.69e3.64 (m, 1H), 3.23
(d, J ¼ 9.0 Hz, 1H), 2.94 (s, 1H), 2.88 (s, 3H), 2.81 (t, J ¼ 7.9 Hz, 1H),
2.27 (q, J ¼ 7.8 Hz, 1H), 1.98 (d, J ¼ 4.1 Hz, 3H), 1.93 (s, 3H), 1.91 (d,
J ¼ 3.4 Hz, 3H), 1.65 (s, 3H), 1.61 (s, 1H), 1.23 (s, 1H), 0.91 (d,
J ¼ 6.9 Hz, 3H), 0.86 (d, J ¼ 6.9 Hz, 3H), 0.46 (d, J ¼ 6.9 Hz, 3H), ꢀ0.36
(d, J ¼ 6.9 Hz, 3H); 13C NMR (151 MHz, DMSO‑d6) dC 186.9, 184.0,
174.1, 171.4, 167.4, 154.4, 148.7, 145.3, 144.8, 143.1, 140.2, 139.2, 137.7,
131.7, 130.9, 130.1, 128.0, 126.4, 126.2, 123.0, 120.5, 116.6, 115.1, 112.5,
111.0, 108.8, 98.6, 77.9, 76.2, 73.1, 73.0, 55.6, 49.4, 42.7, 40.9, 40.2,
37.9, 37.4, 33.8, 22.0, 20.7, 19.8, 18.1, 11.1, 9.0, 8.7, 7.3; HRMS: found
806.3630, calculated for C43H54N2O13 806.3626.
11. A red solid (72 mg, 78%) Rf ¼ 0.33 (EtOAc - MeOH 100:1 v/v);
MS m/z 823.4 (Mþ1); 1H NMR (600 MHz, DMSO‑d6) dH 9.23 (s, 1H),
7.62e7.58 (m, 1H), 7.26 (d, J ¼ 12.1 Hz, 1H), 7.21 (d, J ¼ 11.4 Hz, 1H),
6.49 (d, J ¼ 23.8 Hz, 1H), 6.45 (dd, J ¼ 15.5, 11.4 Hz, 0H), 6.28 (d,
J ¼ 12.7 Hz,1H), 6.24 (d, J ¼ 12.7 Hz,1H), 6.04 (q, J ¼ 7.8 Hz,1H), 5.07
(d, J ¼ 11.0 Hz, 1H), 4.91 (dd, J ¼ 12.7, 8.3 Hz, 0H), 4.45e4.37 (m, 1H),
4.25e4.20 (m, 1H), 4.03 (q, J ¼ 7.1 Hz, 1H), 3.94 (d, J ¼ 9.0 Hz, 0H),
3.63 (d, J ¼ 12.1 Hz, 1H), 3.25e3.23 (m, 1H), 2.89 (d, J ¼ 4.8 Hz, 3H),
2.82e2.79 (m, 0H), 2.27e2.22 (m, 1H), 1.98 (d, J ¼ 6.4 Hz, 3H), 1.95
(s, 3H), 1.91e1.91 (m, 3H), 1.64 (d, J ¼ 3.4 Hz, 3H), 1.63 (s, 1H),
1.25e1.22 (m, 1H), 0.91 (d, J ¼ 6.5 Hz, 3H), 0.83 (d, J ¼ 6.5 Hz, 3H),
0.42 (d, J ¼ 6.9 Hz, 3H), ꢀ0.36 (d, J ¼ 6.9 Hz, 3H); 13C NMR (151 MHz,
DMSO‑d6) dC 185.1, 184.5, 172.3, 171.0, 169.6, 167.6, 149.0, 145.2,
143.3, 137.9, 135.0, 133.0, 132.3, 131.6, 128.6, 127.6, 117.7, 117.4, 116.9,
116.4, 113.8, 109.1, 108.4, 98.8, 76.6, 76.0, 73.1, 72.9, 60.0, 55.9, 42.5,
40.9, 40.3, 38.1, 37.2, 22.3, 20.9, 20.3, 18.4, 11.3, 8.9, 8.5, 7.6; HRMS:
found 822.3392, calculated for C43H54N2O12S 822.3397.
12. A red solid (65 mg, 79%) Rf ¼ 0.25 (EtOAc - MeOH 100:1 v/v);
MS m/z 818.3 (Mþ1); 1H NMR (600 MHz, DMSO‑d6) dH 9.24 (s, 1H),
8.74 (s, 1H), 8.61 (d, J ¼ 5.5 Hz, 1H), 8.56 (d, J ¼ 5.5 Hz, 2H),
7.48e7.51 (2H), 6.51 (s, 2H), 6.27 (s, 1H), 6.24 (s, 1H), 6.02 (q,
J ¼ 7.8 Hz, 1H), 5.09 (d, J ¼ 3.8 Hz, 1H), 5.06 (d, J ¼ 11.0 Hz, 1H), 4.97
(q, J ¼ 5.5 Hz, 1H), 4.26 (s, 1H), 3.93 (d, J ¼ 9.0 Hz, 1H), 3.58 (d,
J ¼ 12.1 Hz, 2H), 3.23 (d, J ¼ 9.0 Hz,1H), 3.14 (d, J ¼ 10.3 Hz,1H), 3.08
(d, J ¼ 10.0 Hz, 1H), 2.88 (s, 3H), 2.82e2.78 (m, 1H), 2.25e2.20 (m,
1H), 1.99 (s, 3H), 1.98 (s, 3H), 1.91 (s, 3H), 1.65 (s, 3H), 1.23 (s, 1H),
0.90 (s, 3H), 0.88 (d, J ¼ 3.8 Hz, 3H), 0.76 (d, J ¼ 6.9 Hz, 3H), 0.42 (d,
J ¼ 6.9 Hz, 3H), ꢀ0.36 (d, J ¼ 6.9 Hz, 3H); 13C NMR (151 MHz,
DMSO‑d6) dC 186.5, 183.4, 171.8, 170.8, 169.3, 150.0, 144.0, 143.0,
139.8,139.0,138.8,131.9,131.7,130.9,128.1,126.3,125.9,124.2,119.4,
15. A red solid (65 mg, 73%) Rf ¼ 0.23 (EtOAc - MeOH 100:1 v/v);
MS m/z 895.4 (Mþ1); 1H NMR (600 MHz, DMSO‑d6) dH 0.24 (s, 1H),
8.74 (s, 1H), 8.56 (s, 1H), 8.00 (d, J ¼ 8.6 Hz, 2H), 7.71 (d, J ¼ 8.6 Hz,
2H), 6.28 (d, J ¼ 14.5 Hz, 1H), 6.19e6.13 (m, 1H), 5.93 (dd, J ¼ 16.2,
5.9 Hz, 1H), 5.07 (dd, J ¼ 10.8, 4.6 Hz, 1H), 5.03e5.01 (m, 1H), 4.97
(dd, J ¼ 11.0, 4.1 Hz, 1H), 4.90e4.86 (m, 1H), 4.17 (s, 1H), 4.05e4.02
(m, 2H), 3.90 (d, J ¼ 10.0 Hz, 1H), 3.73 (d, J ¼ 6.5 Hz, 1H), 3.59e3.55
(m, 1H), 3.23 (d, J ¼ 5.5 Hz, 1H), 3.21 (s, 3H), 2.87 (s, 3H), 2.80 (d,
J ¼ 5.2 Hz, 1H), 2.25e2.18 (m, 1H), 1.99 (s, 3H), 1.94 (d, J ¼ 2.8 Hz,
3H), 1.91 (s, 3H), 1.65 (d, J ¼ 5.9 Hz, 3H), 1.63e1.61 (m, 1H), 1.23 (s,
1H), 0.96 (d, J ¼ 3.8 Hz, 1H), 0.90 (d, J ¼ 7.2 Hz, 3H), 0.80 (dd, J ¼ 6.7,
1.9 Hz, 3H), 0.45 (d, J ¼ 6.9 Hz, 3H), ꢀ0.36 (d, J ¼ 6.9 Hz, 3H). 13
C
NMR (151 MHz, DMSO‑d6) dC 185.0, 184.6, 172.0, 171.1, 168.5, 167.4,
148.9,147.0,146.7, 143.8,140.4,138.4,137.4, 133.1, 132.7, 129.7,129.6,
127.2,126.0,117.6,117.1,115.8,114.9,107.0,101.1, 77.8, 76.1, 73.7, 72.8,
59.7, 56.8, 47.1, 43.4, 41.7, 40.0, 38.3, 33.1, 22.0, 20.6, 19.9, 18.1, 10.9,
8.8, 8.3, 7.4; HRMS: found 894.3602 calculated for C46H58N2O14
S
894.3609.
16. A dark orange solid (71 mg, 83%) Rf ¼ 0.32 (EtOAc - MeOH
100:1 v/v); MS m/z 856.0 (Mþ1); 1H NMR (600 MHz, DMSO‑d6) dH
15.60 (s, 1H), 12.51 (s, 1H), 9.49 (s, 1H), 8.80 (s, 1H), 7.11 (m, 1H), 6.24
(d, J ¼ 10.7 Hz, 1H), 6.20 (d, J ¼ 12.8 Hz, 1H), 5.90 (dd, J ¼ 15.9,
6.7 Hz, 1H), 5.35 (bs, 1H), 5.06 (d, J ¼ 10.7 Hz, 1H), 5.03 (s, 1H), 4.92
(dd, J ¼ 12.7, 8.1 Hz, 1H), 4.18 (d, J ¼ 8.8 Hz, 1H), 3.74e3.72 (m, 1H),
3.23 (d, J ¼ 9.1 Hz, 1H), 2.89 (s, 3H), 2.82 (s, 1H), 2.20e2.16 (m, 1H),
1.96 (s, 3H), 1.90 (s, 3H), 1.63 (s, 3H), 1.58 (d, J ¼ 7.2 Hz, 1H),
1.34e1.30 (m, 1H), 1.05e1.00 (m, 1H), 0.88 (d, J ¼ 6.8 Hz, 3H), 0.83
11

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
(d, J ¼ 6.8 Hz, 3H), 0.43 (d, J ¼ 6.8 Hz, 3H), ꢀ0.25 (d, J ¼ 6.8 Hz, 3H).
simulations for 10 ns. Temperature was kept constant at 300 K with
the Nose-Hoover thermostats and pressure was maintained with
the MartynaeTobiaseKlein barostats within the NPT ensemble
[47]. The cut off value of 9.0 Å was set in calculations of van der
Waals and short-range coulombic interactions. The equation of
motion was solved with the RESPA integrator, with an inner time
step of 2.0 fs and an outer time step of 6.0 fs. [48]. The results were
saved as trajectory by storing coordinates and the energies to disk
at every 5 ps. The 10 ns MD simulation was carried out to relax the
protein and use the final frame of the trajectory as the target
structure of the wild type protein for further docking.
In the truncated systems, the large conformational change was
observed for the protein segment of residues between I1041and
E1153 that resulted in formation of undesired intermolecular pro-
tein interactions and unrealistic binding pocket. Therefore, posi-
tional restraints with a force constant of 50 kcal mol-1 Åꢀ2 were
applied for all atoms that were away more than 10 Å from bound
rifampicin. The 10 ns MD simulation was repeated to relax the
protein and use the final frame of the trajectory as the target
structure of the wild-type protein for further docking.
HRMS: found 854.4318 calculated for C44H62N4O13 854.4313.
4. Microbiology
4.1. Strains and culture
M. tuberculosis strains were cultured in Middlebrook 7H9 me-
dium plus OADC (oleic acid, albumen, dextrose, catalase) supple-
ment and 0.05% w/v Tween 80 (7H9-OADC-Tw). Rifampicin
resistant strains from three different backgrounds were used. RIF-
R2 (RpoBS531L) was obtained from ATCC (ATCC 35838). RIF-R1
(RpoBS522L) was isolated as a spontaneous rifampicin resistant
mutant from H37Rv-LP (ATCC 25618). HN-0258218-RM1
(RpoBH455D) was isolated as a spontaneous rifampicin resistant
mutant from HN-878 [32]. Strain HN878 (NR-13647) was obtained
through BEI Resources, NIAID, NIH. The rpoB gene was sequenced in
the resistant isolates to identify mutations.
4.2. Determination of minimum inhibitory concentrations
The final frames of the resulting trajectories were also used to
generate 3D structures of the full system with a S531L mutation and
MICs were determined as previously described using OD590 as
measurement of growth [37,38]. Briefly, compounds were tested as
10-point 2-fold serial dilutions. Bacterial growth was measured
after 5 days in 7H9-OADC-Tw. Growth curves were generated using
the Levenberg-Marquardt algorithm. MIC90 was defined as the
concentration required to inhibit growth by 90% as compared to
controls.
only
b subunit with a S522L mutation. Maestro function “Mutate
residue” was used to change the serine residues (residue numbers
447 and 456 in the 5UHC entry) into leucine. The modified protein
systems were subjected to the protocol for building solvated sys-
tems and prepared for the molecular dynamics simulation using
the same procedure and settings as for the systems containing wild
type structure. Additional 50 ns of the full system and 150 ns of the
truncated system production runs NPT simulations were conducted
to explore possible effects of mutation on the interactions of resi-
dues from the binding site with the rifampicin and analogues. The
final frame of the extended simulation trajectories were extracted
and used as a target in the molecular docking.
The 3D structures of rifampicin analogues were initially built by
modifying the previously determined X-ray structure of zwitter-
ionic analog of rifampicin [49]. These structures were further fully
optimized using MOPAC software and PM7 Hamiltonian [50], and
saved as mol2. The protein structures extracted from crystal
structure (PDB ID 5UHC) [34] and final frames of the trajectories
were used as targets in the docking with the binding sites posi-
tioned in the geometrical centers of a bound rifampicin. The
docking of ligands in zwitterionic form was conducted using
LeDock software. The center of the 20 Å ꢁ 20 Å ꢁ 20 Å box was
positioned on the rifampicin ligand and 20 poses were generated
for each analogue. The additional Root Mean Square Deviation
(RMSD) cutoff of 1 Å was set to redundancy of poses. The reliability
of the docking simulations was demonstrated by redocking rifam-
picin into the binding site and reproducing the binding modes of
the ligand crystal structure within RMSD of 0.7 Å. The largest
change of the molecule was observed for piperazine moiety as it is
highly dependent on the RNA conformation that changes during
the molecular dynamics simulation.
4.3. Molecular modelling
The molecular modelling work was performed using Desmond
[39] with Maestro [40] as graphical user interface (GUI) ver 2020-1,
AutoDock Vina ver. 1.1.2 [41] with VegaZZ ver. 3.2.1.33 [42,43] as
GUI, LeDock [44], Avogadro ver 1.2.0 [45] and ChemBioOffice ver.
16.0.1.4. All the analysis and image preparations were conducted
using Maestro and Discovery Studio Visualizer ver. 20.1.0.19295
[46].
The crystal structure of the wild-type M. tuberculosis RNAP in a
complex with rifampicin (PDB ID: 5UHC) [34] was used as a target
protein to investigate possible interactions of all ligands with the
wild type and mutated protein RpoBS522L. Initially, the three
dimensional (3D) structures of all analogues were prepared using
ChemBioOffice and saved in a. mol2 format, while the 3D structure
of the wildtype RNAP was prepared using Protein Preparation
Wizard implemented in Maestro by adding hydrogen atoms and
setting protonation states of all ionizable groups for pH 7.
The system comprising protein in complex with rifampicin, DNA
and RNA was used for further evaluations of the analogues binding
in the pocket located on the beta subunit in the wild type and
mutated protein. The System Builder Tool was utilized to prepare a
cubic periodic solvated system using single point charge (SPC)
model of water molecules, with box size 15 Å larger than a mo-
lecular system in all directions. Adequate number of Naþ and Clꢀ
ions were added to neutralize the system and mimic conditions
with 0.05 M NaCl concentration. Additionally, following previously
reported study structure where it was demonstrated that each
Funding
No funding was received for this work.
subunit is relatively independent and that only
b subunit be
simulated [35], the subunit in complex with rifampicin was
b
extracted. The truncated system was prepared in the same way as
the full system described as above.
Declaration of competing interest
Initially, the default Desmond “Molecular Dynamics” protocol
and OPLS2005 all atoms force field for that include minimization,
equilibration and production run steps was used to conduct
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
12

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
formylrifamycin SV derivatives obtained via ‘click chemistry’approach, Eur.
J. Med. Chem. 84 (2014) 651e676.
Acknowledgment
[24] R. Cricchio, G. Tamborini, P. Bravo, G. Gaudiano, The reaction of 3-
formylrifamycin SV with sulphonium and phosphonium ylides, Farmaco.
[Sci. ] 29 (1974) 358e365.
Kyle Krieger and Renee Allen are thanked for technical
assistance.
[25] M. Taguchi, N. Aikawa, G. Tsukamoto, Reaction of 3-formylrifamycin S with
secondary amines, Chem. Pharm. Bull. 32 (1984) 4388e4395.
Appendix A. Supplementary data
ꢀ
[26] K. Bujnowski, L. Synoradzki, T.A. Zevaco, E. Dinjus, E. Augustynowicz-Kopec,
A. Napiorkowska, Rifamycin antibioticsdnew compounds and synthetic
methods. Part 4: study of the reaction of 3-formylrifamycin SV with secondary
amines and ketones, Tetrahedron 71 (2015) 158e169.
NMR spectral data for 8e16. Molecular docking scores for
rifampicin analogues against M. tuberculosis H37Rv RpoB and im-
ages showing the binding modes of selected analogues within the
binding pocket of the b-subunit of wild-type M. tuberculosis RNAP.
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.ejmech.2021.113734.
[27] S.K. Gill, H. Xu, P.D. Kirchhoff, T. Cierpicki, A.J. Turbiak, B. Wan, N. Zhang,
K. Peng, S.G. Franzblau, G.A. Garcia, Structure-based design of novel benzox-
azinorifamycins with potent binding affinity to wild-type and rifampin-
resistant mutant Mycobacterium tuberculosis RNA polymerases, J. Med.
Chem. 55 (2012) 3814e3826.
[28] J.M. Dickinson, D.A. Mitchison, In vitro activity of new rifamycins against
rifampicin-resistant M. tuberculosis and MAIS-complex mycobacteria, Tu-
bercle 68 (1987) 177e182.
References
[29] H.G. Floss, T. Yu, Rifamycin mode of action, resistance, and biosynthesis,
Chem. Rev. 105 (2005) 621e632.
[30] K. Pyta, A. Janas, N. Skrzypczak, W. Schilf, B. Wicher, M. Gdaniec, F. Bartl,
P. Przybylski, Specific interactions between rifamycin antibiotics and water
influencing ability to overcome natural cell barriers and the range of anti-
bacterial potency, ACS Infect. Dis. 5 (2019) 1754e1763.
[31] K. Pyta, P. Przybylski, F. Bartl, Regioselective long-range proton transfer in
new rifamycin antibiotics: a process in which crown ethers act as stronger
Brønsted bases than amines, ChemPhysChem 16 (2015) 938e942.
[32] T.R. Ioerger, T. O'Malley, R. Liao, K.M. Guinn, M.J. Hickey, N. Mohaideen,
K.C. Murphy, H.I. Boshoff, V. Mizrahi, E.J. Rubin, Identification of new drug
targets and resistance mechanisms in Mycobacterium tuberculosis, PLoS One
8 (2013), e75245.
[1] P. Sensi, P. Margalith, M.T. Timbal, Rifomycin, a new antibiotic; preliminary
report, Farmaco. [Sci. ] 14 (1959) 146.
[2] T.D. McHugh, Tuberculosis: Diagnosis and Treatment; Cabi, vol. 21, 2013.
[3] W. Rifampin Lester, A semisynthetic derivative of rifamycin-A prototype for
the future, Annu. Rev. Microbiol. 26 (1972) 85e102.
[4] P. Sensi, History of the development of rifampin, Rev. Infect. Dis. 5 (1983)
S402eS406.
[5] N. Maggi, C.R. Pasqualucci, R. Ballotta, P. Sensi, Rifampicin: a new orally active
rifamycin, Chemotherapy 11 (1966) 285e292.
[6] V. Arioli, M. Berti, G. Carniti, E. Randisi, E. Rossi, R. Scotti, Antibacterial activity
of DL 473, a new semisynthetic rifamycin derivative, J. Antibiot. 34 (1981)
1026e1032.
[7] R.J. O'Brien, M.A. Lyle, D.E. Snider Jr., Rifabutin (ansamycin LM 427): a new
rifamycin-S derivative for the treatment of mycobacterial diseases, Rev. Infect.
Dis. 9 (1987) 519e530.
[8] World Health Organization No title, World Health Organization Model List of
Essential Medicines: 21st List 2019, 2019.
[9] N. Lounis, G. Roscigno, In vitro and in vivo activities of new rifamycin de-
rivatives against mycobacterial infections, Curr. Pharmaceut. Des. 10 (2004)
3229e3238.
[33] C. Manca, L. Tsenova, A. Bergtold, S. Freeman, M. Tovey, J.M. Musser, C.E. Barry,
V.H. Freedman, G. Kaplan, Virulence of a Mycobacterium tuberculosis clinical
isolate in mice is determined by failure to induce Th1 type immunity and is
associated with induction of IFN-a/b, Proc. Natl. Acad. Sci. U.S.A. 98 (2001)
5752e5757.
[34] W. Lin, S. Mandal, D. Degen, Y. Liu, Y.W. Ebright, S. Li, Y. Feng, Y. Zhang,
S. Mandal, Y. Jiang, Structural basis of Mycobacterium tuberculosis tran-
scription and transcription inhibition, Mol. Cell. 66 (2017) 169e179, e8.
[35] Q. Zhang, S. Tan, T. Xiao, H. Liu, S.J.A. Shah, H. Liu, Probing the molecular
mechanism of rifampin resistance caused by the point mutations S456L and
D441V on Mycobacterium tuberculosis RNA polymerase through Gaussian
accelerated molecular dynamics simulation, Antimicrob. Agents Chemother.
(2020) 64.
[36] I. Artsimovitch, M.N. Vassylyeva, D. Svetlov, V. Svetlov, A. Perederina,
N. Igarashi, N. Matsugaki, S. Wakatsuki, T.H. Tahirov, D.G. Vassylyev, Allosteric
modulation of the RNA polymerase catalytic reaction is an essential compo-
nent of transcription control by rifamycins, Cell 122 (2005) 351e363.
[37] T.R. Ioerger, Y. Feng, K. Ganesula, X. Chen, K.M. Dobos, S. Fortune,
W.R. Jacobs Jr., V. Mizrahi, T. Parish, E. Rubin, C. Sassetti, J.C. Sacchettini,
Variation among genome sequences of H37Rv strains of Mycobacterium
tuberculosis from multiple laboratories, J. Bacteriol. 192 (14) (2010)
3645e3653.
[10] P.A. Aristoff, G.A. Garcia, P.D. Kirchhoff, H.H. Showalter, Rifamycinseobstacles
and opportunities, Tuberculosis 90 (2010) 94e118.
[11] W.A. Vischer, P. Imhof, S. Hauffe, P. Degen, Pharmokinetics of new long-acting
rifamycin-derivatives in man, Bull. Int. Union Tuberc. 61 (1986) 8.
[12] S. Ramaswamy, J.M. Musser, Molecular genetic basis of antimicrobial agent
resistance in Mycobacterium tuberculosis: 1998 update, Tuber. Lung Dis. 79
(1998) 3e29.
[13] A.N. Unissa, L.E. Hanna, Molecular mechanisms of action, resistance, detection
to the first-line anti tuberculosis drugs: rifampicin and pyrazinamide in the
post whole genome sequencing era, Tuberculosis 105 (2017) 96e107.
[14] World Health Organization, Global Tuberculosis Report 2019, World Health
Organization, Geneva (Switzerland), 2020, 2020.
[15] E.A. Campbell, N. Korzheva, A. Mustaev, K. Murakami, S. Nair, A. Goldfarb,
S.A. Darst, Structural mechanism for rifampicin inhibition of bacterial RNA
polymerase, Cell 104 (2001) 901e912.
[16] A. Feklistov, V. Mekler, Q. Jiang, L.F. Westblade, H. Irschik, R. Jansen,
A. Mustaev, S.A. Darst, R.H. Ebright, Rifamycins do not function by allosteric
modulation of binding of Mg2þ to the RNA polymerase active center, Proc.
Natl. Acad. Sci. U.S.A. 105 (2008) 14820e14825.
[17] Z.L. Berrada, S.G. Lin, T.C. Rodwell, D. Nguyen, G.F. Schecter, L. Pham,
J.M. Janda, W. Elmaraachli, A. Catanzaro, E. Desmond, Rifabutin and rifampin
resistance levels and associated rpoB mutations in clinical isolates of Myco-
bacterium tuberculosis complex, Diagn. Microbiol. Infect. Dis. 85 (2016)
177e181.
[18] F.B. Jamieson, J.L. Guthrie, A. Neemuchwala, O. Lastovetska, R.G. Melano,
C. Mehaffy, Profiling of rpoB mutations and MICs for rifampin and rifabutin in
Mycobacterium tuberculosis, J. Clin. Microbiol. 52 (2014) 2157.
[19] D. Czerwonka, J. Domagalska, K. Pyta, M.M. Kubicka, P. Pecyna, M. Gajecka,
P. Przybylski, Structureeactivity relationship studies of new rifamycins con-
taining l-amino acid esters as inhibitors of bacterial RNA polymerases, Eur. J.
Med. Chem. 116 (2016) 216e221.
[38] J. Ollinger, M.A. Bailey, G.C. Moraski, A. Casey, S. Florio, T. Alling, M.J. Miller,
T. Parish, A Dual read-out assay to evaluate the potency of compounds active
against Mycobacterium tuberculosis, PLoS One 8 (2013), e60531.
[39] K.J. Bowers, D.E. Chow, H. Xu, R.O. Dror, M.P. Eastwood, B.A. Gregersen,
J.L. Klepeis, I. Kolossvary, M.A. Moraes, F.D. Sacerdoti, In scalable algorithms
for molecular dynamics simulations on commodity clusters, in: SC'06: Pro-
ceedings of the 2006 ACM/IEEE Conference on Supercomputing, IEEE, 2006,
p. 43.
€
[40] D. E. Shaw Research, New York, NY, Schrodinger Release 2020-1: Desmond
Molecular Dynamics System, Maestro-Desmond Interoperability Tools,
€
Schrodinger, New York, NY, 2020, 2020.
[41] O. Trott, A.J. Olson, AutoDock Vina, Improving the speed and accuracy of
docking with
a new scoring function, efficient optimization, and multi-
threading, J. Comput. Chem. 31 (2010) 455e461.
[42] A. Pedretti, L. Villa, G. Vistoli, VEGAean open platform to develop chemo-bio-
informatics applications, using plug-in architecture and script programming,
J. Comput. Aided Mol. Des. 18 (2004) 167e173.
[43] A. Pedretti, L. Villa, G. Vistoli, VEGA: a versatile program to convert, handle
and visualize molecular structure on Windows-based PCs, J. Mol. Graph.
Model. 21 (2002) 47e49.
[20] K. Pyta, A. Janas, M. Szukowska, P. Pecyna, M. Jaworska, M. Gajecka, F. Bartl,
P. Przybylski, Synthesis, docking and antibacterial studies of more potent
amine and hydrazone rifamycin congeners than rifampicin, Eur. J. Med. Chem.
167 (2019) 96e104.
[44] N. Zhang, H. Zhao, Enriching screening libraries with bioactive fragment
space, Bioorg. Med. Chem. Lett 26 (2016) 3594e3597.
[45] M.D. Hanwell, D.E. Curtis, D.C. Lonie, T. Vandermeersch, E. Zurek,
G.R. Hutchison, Avogadro: an advanced semantic chemical editor, visualiza-
tion, and analysis platform, J. Cheminf. 4 (2012) 1e17.
ꢀ
[21] K. Pyta, P. Przybylski, B. Wicher, M. Gdaniec, J. Stefanska, Intramolecular
proton transfer impact on antibacterial properties of ansamycin antibiotic
rifampicin and its new amino analogues, Org. Biomol. Chem. 10 (2012)
2385e2388.
ꢁ
[46] Dassault Systemes BIOVIA, Discovery Studio Modeling Environment, Release,
ꢁ
Dassault Systemes, San Diego, 2017, 2016.
ꢀ
[22] K. Pyta, P. Przybylski, K. Klich, J. Stefanska, A new model of binding of
[47] W.G. Hoover, Canonical dynamics: equilibrium phase-space distributions,
Phys. Rev. A 31 (1985) 1695.
[48] D.D. Humphreys, R.A. Friesner, B.J. Berne, A multiple-time-step molecular
dynamics algorithm for macromolecules, J. Phys. Chem. 98 (1994)
rifampicin and its amino analogues as zwitterions to bacterial RNA poly-
merase, Org. Biomol. Chem. 10 (2012) 8283e8297.
[23] K. Pyta, K. Klich, J. Domagalska, P. Przybylski, Structure and evaluation of
antibacterial and antitubercular properties of new basic and heterocyclic 3-
13

M. Zloh, M. Gupta, T. Parish et al.
European Journal of Medicinal Chemistry 225 (2021) 113734
6885e6892.
[50] J.J. Stewart, Optimization of parameters for semiempirical methods VI: more
modifications to the NDDO approximations and re-optimization of parame-
ters, J. Mol. Model. 19 (1) (2013) 1e32, https://doi.org/10.1007/s00894-012-
1667-x.
[49] B. Wicher, K. Pyta, P. Przybylski, E. Tykarska, M. Gdaniec, Redetermination of
rifampicin pentahydrate revealing a zwitterionic form of the antibiotic, Acta
Crystallogr.
C
68 (5) (2012) o209eo212, https://doi.org/10.1107/
S0108270112015296.
14