Dodecanoic acid & palmitic acid disarms rifampicin resistance by putatively targeting mycobacterial efflux pump Rv1218c

The success rate of anti-tuberculosis (TB) treatment will reportedly decline from 83 per cent to 56 per cent if the infection is due to drug-resistant Mycobacterium tuberculosis (MTB)1. During rifampicin (RMP) resistance, the non-synonymous mutations that alter MTB RNA polymerase2 are common rather than deletions and insertion mutations3. Clinical isolates which are resistant to RMP have a higher chance to become isoniazid (INH) resistant and make RMP resistance a potential indicator of multi-drug resistance (MDR) TB4,5. Nearly 95 per cent of RMP resistance is due to spontaneous mutations in the 81 base pair regions of the rpoB gene6,7. The rest five per cent RMP resistance exhibits no evidence of any mutation and offers strong evidence of the role of other mechanisms, particularly chromosome-mediated drug efflux pumps (EPs) that export various molecules outside the bacterial cells8. Rv3065, belongs to the family of resistance-nodulation-division transporters and, is an important efflux of pyrrole and pyrazolone molecules. It is related to the susceptibility to RMP, INH, erythromycin, tetracycline, aminoglycosides and fluoroquinolones and has been reported in Indian MDR-TB isolates9. Knockout studies on Rv3065, Rv1258, Rv0849 and Rv1819c showed increased susceptibility to drugs in the pyrrole class compared to knockout mutant strains10. The efflux gene Rv2938 encodes an inner membrane protein of an EP system in MTB and has been reported to confer drug resistance in some MDR strains11. The Rv1819c gene codes for a putative ABC transporter that has a negatively charged cavity and specializes in transporting long-chain fatty acids, vitamin B12 and drugs such as bleomycin12,13. According to studies by Gopinath et al14, Rv1819c functions as a channel for the entry of vitamin B12, which plays a crucial role in the mycobacterial pathway. Being an ABC transporter, Rv1819c may also be involved in the transportation of lipids within the bacterium15. Rv2209 is required for virulence and is involved in the cell wall synthesis. The overexpression of Rv2477 and Rv2209 has been observed in response to stress from ofloxacin in MTB16. These findings highlight the importance of monitoring the emergence of drug resistance in MTB populations and the role that EP genes play in this process. Further research is needed to fully understand the mechanisms and interactions between these genes and other factors that contribute to drug resistance in MTB.

Certain EPs were reported to be specific to RMP mono-resistance17. However, the function of other active drug EPs which cause RMP resistance is yet to be explored. MDR pathogens are known to overexpress these EPs to facilitate a dip in the intracellular drug concentration. To date, five families of EPs were identified, including the ABC transporter family, which was identified to be one of the causes of RMP resistance10,18. Rv1218c EP is one such ABC transporter family pump of MTB that was reported to efflux RMP, novobiocins, pyrazolones, biarylpiperazines, bisanilino pyrimidines, pyrroles and pyridines19. Any loss in function of Rv1218c consequently lowered the minimum inhibitory concentration (MIC) of these drugs by increasing the drug bioavailability and intracellular accumulation by eight-fold10. Owing to the efflux-mediated drug resistance, more light has been shed on EP inhibitors (EPI) in the recent years to augment the anti-TB regimen20, with a primary interest in plant-based secondary metabolites. Verapamil is one such EPI that was approved by the Food and Drug Administration. However, it is incapable to inhibit a wide range of EPs and is not successful in phenotypic levels21. This study aimed to evaluate the efflux inhibiting potency of selected bioactive compounds from a plant-based database showing a positive combined inhibition with RMP against MDR-TB clinical isolates and in recombinant Mycobacterium smegmatis expressing Rv1218c. In silico and in vitro validation of efflux inhibitory activity by these molecules against the EP Rv1218c was done.

Material & Methods

This study was conducted at the department of Bacteriology, ICMR-National Institute for Research in Tuberculosis (NIRT), Chennai, India, from January 2018 to November 2020. The approval by the Institutional Ethics Committee was obtained before starting the study.

Bacterial cultures and strains used: Middlebrook 7H9 medium supplemented with oleic acid (Ol), albumin, dextrose and catalase (OADC) (7H9) (BD Biosciences, catal) was used to grow by eight-fold and M. smegmatis while Luria-Bertani (LB) medium (Himedia, India) was used for Escherichia coli DH5α. Two MDR-TB clinical isolates (NIRT-001 and NIRT-002) from the ICMR-NIRT repository were used in the study. These isolates were well characterized by phenotypical, bacteriological and genetic methods such as line probe assay. The vector system pMV261 and E. coli DH5α competent cells (Novagen) were used as per the manufacturer’s instructions for the molecular techniques.

Statistical analysis: All experiments were carried out in triplicate. Statistical analysis was carried out using GraphPad Prism 5 software for Windows version 5.03. (www.graphpad.com). Significant differences between treated and untreated control groups were found using the unpaired t test. P<0.05 was considered significant.

In silico homology modelling and protein structure prediction: Based on the protein sequence of the EP gene Rv1218c, the 3D protein structure was predicted using I-TASSER, which is a hierarchical approach to predict the protein structure based on a known functional template22. The model with low confirmation was selected as a final structure for screening the lead molecules.

Ligand preparation: A total of 25,000 biologically active small molecules were retrieved from Duke’s database (www.phytochem.nal.usda.gov). These were geometrically minimized for liquid simulations using the optimized potentials. PDBQT file was created in MOL2 format, followed by the addition of hydrogen atoms and AM1-BCC charges by applying the AMBER force field that increases the accuracy of the ligand molecules23.

Protein preparation: Protein preparation and refinement were performed using Gasteiger charges, in which the following reactions occur, namely removal of water molecules, addition of hydrogen atoms, bond orders assignment hydroxyl group orientation and optimization. Finally, using default constraint 0.3 Å RMSD, energy minimization was carried out. The modelled protein structure of the EP was used as the receptor and the active sites for the proteins were predicted using the online free software SCFBio, (IIT, Delhi). The centre of the receptor grid was generated for the active site residues.

Screening of compounds against Rv1218c using mastero: The Mastero v9.1(22) software was used to perform the docking analysis, using the Rv1218c receptor and compounds optimized with an OPLS_2005 force field. Multiple conformations of the ligands were generated, and compounds with a fitness score higher than 1.0 were selected for docking with Rv1218c.

Analysis of Rv1218c transcriptional levels in multi-drug resistance-tuberculosis (MDR-TB) clinical isolates: The transcription level of Rv1218c was analyzed by quantitative real-time-PCR method by using H37Rv as a reference strain. MTB H37Rv and MDR clinical isolates were grown to the mid-log phase in the presence and absence of sub-inhibitory concentration of RMP (1/4 the MIC). Since RMP at its MIC will lead to inhibition of MTB growth, the test strains were exposed to sub-inhibitory MIC to ensure its survival and induction of the EPs. RNA was extracted using the TRIZOL reagent and DNA contamination was removed using RNase-free DNase. Verso cDNA synthesis kit (Thermo Scientific, Lithuania, USA) was used to reverse transcribe to total RNA cDNA using the commercially available kit (. The real-time PCR reactions were performed using Thermoscientific ABI-7500 one-step reverse transcription (RT)–PCR in a 20 µl total reaction volume of consisting of 10 pmol of each primer and SYBR green. Protocol: 55°C for 10 min, followed by PCR activation at 95°C for 30 sec and 45 cycles of amplifications at room temperature (RT). The experiment was performed in triplicates and the housekeeping gene 16s rRNA was used as the internal control.

Results

In silico screening for efflux pump inhibitors against Rv1218c was carried out. The compounds from Duke’s database were prepared for docking and induced fit docking was performed in the search for active EP inhibitors against Rv1218c. The top ten compounds were selected based on the Lipinski’s rule of five.

The compounds raffinose pentahydrate (RP), palmitic acid (PA), tetra decanoic acid (TD), magnesium stearate (MS), 3, 4, 5-trihyhydroxy benzoic acid (TH), dodecanoic acid (DA), flavone (FL), rhamnose monohydrate (RM), Ol and stearic acids (SA) were selected for in vitro validation based on parameters including the glide score, glide energy, ADMET (absorption, distribution, metabolism, elimination and toxicity) properties and interactions with the target protein (Table I). The 2D and 3D interactions of the identified EPs with the protein (Rv1218c) are given in Figure 1.

Close

Fig. 1

Docking pose of identified efflux pump inhibitors interacting with RV1218C visualized using Maestro v9.1 for two-dimensional and Pymol viewer version 1.3 for three-dimensional images.

Export to PPT

A significant differential expression of Rv1218c was observed only in the MDR-TB clinical isolates when exposed to sub-inhibitory levels of RMP (¼ MIC). However, such a difference was not observed in the RMP-sensitive MTB H37Rv (both RMP treated and untreated) (Fig. 2). Hence, increased expression of Rv1218c can be considered as one of the RMP-resistant mechanisms of the strains used in the study.

Close

Fig. 2

Analysis of mRNA transcriptional levels of Rv1218c in MDR-TB clinical isolate and H37Rv. Rv1218c: RMP Untreated versus treated P=0.035, H37Rv: RMP Untreated vs. treated P=0.287. MDR-TB, multi-drug resistance-tuberculosis; RMP, rifampicin.

Export to PPT

The MICs of RMP against MDR-TB and rMSM were determined in the range of 128 to 0.003 µg/ml. Similarly, the MICs for the shortlisted molecules were also determined. The MIC of RMP for the MDR-TB clinical isolates (n=2) and rMSM was found to be >128 µg/ml, 32 µg/ml, and 8 µg/ml, respectively. The compounds OI and SA showed minimal inhibition even at higher concentrations with MDR-TB clinical isolates. MIC of TD, MS and RP was 1000 µg/ml each. MICs of the molecules RM, PA, TH and FL were 700, 100, 100 and 100 µg/ml, respectively, while the DA was highly potent even at a lower concentration of 10 µg/ml. The MIC of the molecules was comparatively higher in rMSM (Table II).

The RMP-MIC lowering the potential of the selected molecules was determined using the checkerboard method in microtitre plates. As OI and SA exhibited poor inhibition activity (MIC >1500 µg/ml against MDR-TB clinical isolates and rMSM), except for these two, the remaining 8 molecules were included in the checkerboard assay. The MIC of RMP was reduced by more than 4-8 fold in the presence of identified active compounds (Table II).

New EP inhibitors counteracted Rv1218c and retained intracellular EtBr accumulation: The EP inhibitory efficiency of the identified putative EP inhibitors RP, PA, TD, MS, TH, DA, FL and RM was tested against rMSM in the presence of EtBr. Bacterial cells unaccompanied by identified active compounds and the known efflux inhibitor piperine were considered as controls. The compound MS, RP, TH, RM, PA and DA showed increased efflux inhibition as compared to piperine (P<0.001). Even though the molecules FL and TA reduced the MIC of RMP, it showed minimal accumulation of EtBr (Fig. 3). The expansion of the fluorescence intensity of EtBr in the presence of most of the inhibitors used confirms that the mechanism of inhibition through the EP inhibition in MDR clinical isolates.

Close

Fig. 3

Effect of shortlisted efflux pump inhibitors on the accumulation of ethidium bromide. The efflux inhibitors were used at ½ MIC and ethidium bromide at 0.5 µg/ml concentration. The rMSM without efflux pump inhibitor serves as a control. MIC, minimum inhibitory concentration; rMSM, recombinant Mycobacterium smegmatis mc² 155.

Export to PPT

Dodecanoic acid and palmitic acid are cytologically safe: The lysis of RBC was more than 80 per cent at the 1× concentration for TA and MS whereas < 5 per cent RBC lysis was seen in the presence of RM, TH and PA while the integrity of RBCs was completely secured in the presence of DA and FL even at 2× MIC concentration. The molecules RM and RP had normal ranges of 2.7 and 2.9 per cent, respectively, at a 2× MIC. However, due to their increased individual MIC with MDR clinical isolates, these compounds were not pursued further (Fig. 4). Although MS showed a high efflux inhibition, it indeed showed higher haemolytic activity in a concentration-dependent manner. The toxicity of molecules TH, DA, PA and FL was further tested by MTT assay using PBMCs. DA and PA showed the least toxicity even at higher concentrations (8× MIC) (Fig. 5). Both FL and TH showed only <20 per cent of cell viability even at 1× MIC.

Close

Fig. 4

Haemolytic activity of shortlisted active compounds. Dodecanoic acid, Flavone caused no haemolysis at its 2× MIC. RBCs treated with distilled water and PBS was considered as positive and negative control, respectively. PBS, phosphate buffered saline; RBCs, red blood cells.

Export to PPT

Close

Fig. 5

Cytotoxicity of shortlisted efflux pump inhibitors on the PBMCs. Palmitic acid and Dodenoic acid showed more than 90 per cent cell viability at 2× MIC. Untreated PBMCs were used as a control.

Export to PPT

The cytotoxicity of short-listed EPIs along with Piperine was further tested on Vero (kidney), THP-1 monocyte (macrophage) and HepG2 liver cells belonging to different cell origins. As predicted, EPIs which showed less toxicity in PBMCs (PA and DA) showed promising results with >85 per cent cell viability in all three cell lines. In THP-1, both DA and PA showed 80 per cent of cell viability in 1× of MIC. In Vero and HepG2 cells, both EPIs (DA and PA) showed 85 and 90 per cent of cell viability, respectively (Fig. 6). The in vitro and ex vivo study outcomes align to each other supporting that these molecules DA and PA as a possibly reliable EPIs.

Close

Fig. 6

Ex vivo cytotoxicity of dodecanoic acid, palmitic acid and Piperine on different cell lines.

Export to PPT

Time kill kinetics: The potentiality of the identified molecules on the time-dependent killing activity of RMP against the MDR clinical isolate (MIC >128 µg/ml) was assessed. There was a significant five log₁₀ reduction in the CFU (P<0.001) for all identified EP inhibitors (data shown for only PA and DA), at the earliest of 48 h in combination with RMP, when compared with Piperine (Fig. 7).

Close

Fig. 7

Representative graph showing time-dependent killing by RMP and reduction in viable MTB MDR-TB clinical strain in the presence of DA and PA. The actual RMP MIC of 128 µg/ml was drastically reduced to 0.0078 µg/ml and 0.25 µg/ml, respectively by PA and DA, which efficiently killed MTB MDR strain within 48 h of RMP exposure (P<0.001).

Export to PPT

Discussion

RMP and INH are known to be two drugs which are more efficient in terms of efficacy and toxicity to treat drug-sensitive TB infections32. However, MDR-TB creates an unenviable dispute in treating and controlling TB. Since 1980, when the influence of EPs on drug resistance was reported, EPs are being viewed as a scope for managing bacterial drug resistance possibly with the first-line anti-TB drugs themselves33,34. The expression of EPs increases in the presence of anti-TB drugs. The clinical TB isolates (RMP MIC >128 µg/ml) used in this study were also found to overexpress Rv1218c, an EP of the ABC transporter family in the presence of RMP (Fig. 2). This portrays the significance of these EPs in the pumping-out of the drug molecules19 and as potential targets to handle the drug resistance menace. Hence, there has been a pursuit of natural molecules with EPI activity in the recent past18. Using in silico methods, we selected ten molecules from the pool of 25000 plant-based molecules. Based on the outcome of in vitro experiments for anti-TB activity and RMP-MIC lowering activity, eight molecules were chosen for toxicity testing based on which these were further narrowed down to two molecules, namely DA and PA. DA and PA potentially decreased the MIC of RMP by 8 to 1000 folds in the MDR-TB clinical isolates notably a drop in MIC from >128 µg/ml to 0.0078 µg/ml in one of the studied isolates (Table II). The time-kill kinetic study showed that both DA and PA can reduce the time to kill these MDR isolates of RMP by 48 h which otherwise could not be achieved even after 240 h in the presence of RMP alone (Fig. 7). DA showed more than 80 per cent cell viability even at 10-fold higher concentrations of its MIC. PA was found to be non-toxic at 2× of its MIC while being toxic at a higher concentration of 1 mg/ml (Fig. 5). Though the other molecules like TH and FL showed significant RMP MIC lowering activity and EP inhibition, these were more toxic to PBMCs. The non-toxic characteristics of DA and PA were found to be consistent across different cell lines belonging to different organ sources (Figs 5-6). However, both TH and EL were non-lytic to RBCs (Figs 4 and 5). Interestingly, even though TA and FL efficiently minimized the MIC of RMP and showed significant binding with Rv1218c (in silico), these did not cause accumulation of EtBr, unlike DA and PA. This could be due to a combined inhibitory activity mediated by efflux inhibition but by targeting other EPs that could be present in the respective clinical strains. Since the EtBr accumulation assay was carried out in recombinant M. smegmatis transformed with the Rv1218c, only the Rv1218c EP inhibitors showed the accumulation. Interestingly, a time-dependent fluorescence quenching was noted which could be due to EtBr decay35. Hence, to summarize, both DA and PA were able to reduce the MIC of RMP by specifically binding with Rv1218c in the limited evidence provided. This emphasizes the need for further in-depth studies (Fig. 3).

PA is a saturated fatty acid that is abundant in the human physiological system and can be consumed as a part of a regular diet. Formerly, PA was putatively contemplated as a detrimental molecule, but later scientific insights contradicted this view36. PA is also plentiful in the natural food products such as milk, coconut oil and palm oil37. PA by itself was found to inhibit MTB in this study. Previously, PA was reported to bind with the Rv3249c and Rv1816 transmembrane proteins of MTB that further hindered the MTB Mmpl protein expression38. DA, otherwise called lauric acid, is present abundantly in coconut oil along with PA39. DA has been shown to possess anti-microbial properties against various microbial pathogens against a broad range of bacteria, fungi and viruses. Besides a previous report that showed the anti-TB activity of DA40, this study has demonstrated the RMP-MIC lowering capacity of DA to the scientific community. Coherently, we have also previously reported the anti-TB activity of a herbal plant, Lantana camara that was known to contain DA according to the Duke’s Database41. DA exhibits rapid non-ionic passive membrane transport across lipid bilayer42. This quality could be advantageous as MTB is an intracellular pathogen. However, DA gets quickly cleared from the bloodstream and easily digested and metabolized in the liver when administered through intravenous and oral routes, respectively39. This may be surmounted by using an efficient and targeted drug delivery system. Interestingly, PA in combination with stearic acid itself could be used for the construction of such solid-lipid nanodelivery systems43. We also reported the drug efflux-inhibitory anti-TB activity of other molecules such as FL and TH that was dropped due to toxicity. Using advanced chemical technologies, non-toxic potent analogues of these compounds can be generated.

Several plant-based compounds, including the secondary metabolites, were reported previously to encompass the EPI activity against various pathogens18,43. However, authors believe that this is the first report to indicate that both DA and PA were able to bring down the MICs of RMP in more than one MDR-TB isolate, while both were found to putatively bind with RV1218C. Even though this supports the high likelihood of re-establishing RMP into the MDR regimen, these compounds should be validated in the supplementary number of drug-resistant clinical isolates of MTB, including bedaquiline resistance. Furthermore, the specificity of DA and PA by gene knock-out studies also need confirmation. The smaller number of strains included in this study is, however, the major limitation. Further validations to understand other potential targets of these EPIs, their chemical as well as functional interaction with other anti-TB drugs, in vivo safety and in vivo efficacy testing need to be undertaken before these molecules could be considered for therapeutic purposes.