Combination of Repurposed Drug Diosmin with Amoxicillin-Clavulanic acid Causes Synergistic Inhibition of Mycobacterial Growth

Effective therapeutic regimens for the treatment of tuberculosis (TB) are limited. They are comprised of multiple drugs that inhibit the essential cellular pathways in Mycobacterium tuberculosis (Mtb). The present study investigates an approach which enables a combination of Amoxicillin-Clavulanic acid (AMC) and a repurposed drug for its synergistic effect towards TB treatment. We identified Diosmin (DIO), by targeting the active site residues of L,D-transpeptidase (Ldt) enzymes involved in Mtb cell wall biosynthesis by using a structure-based drug design method. DIO is rapidly converted into aglycone form Diosmetin (DMT) after oral administration. Binding of DIO or DMT towards Ldt enzymes was studied using molecular docking and bioassay techniques. Combination of DIO (or DMT) and AMC exhibited higher mycobactericidal activity against Mycobacterium marinum as compared to individual drugs. Scanning electron microscopy study of M. marinum treated with AMC-DIO and AMC-DMT showed marked cellular leakage. M. marinum infected Drosophila melanogaster fly model showed an increased fly survival of ~60% upon treatment with a combination of AMC and DIO (or DMT). Finally, the enhanced in vitro antimicrobial activity of AMC-DIO was validated against Mtb H37Ra and a MDR clinical isolate. Our results demonstrate the potential for AMC and DIO (or DMT) as a synergistic combination for the treatment of TB.


Results and Discussion
Emerging antibiotic resistance is major concern in the current TB chemotherapy. The inhibition of both D,D-transpeptidase and Ldt enzymes prevents the formation of both classical and non-classical PG cross-linking, which is involved in the Mtb growth, survival and drug resistance. We repurposed an FDA approved drug (DIO) against key Mtb druggable targets Ldt Mt1 & Ldt Mt2 by using a structure-based drug design method. The anti-mycobacterial activity of the repurposed drug (DIO) was further validated by use of in vitro and in vivo techniques. Recent studies showed that the carbapenems such as MEM, ETP, and Tebipenem, bind to the key catalytic triad residues Cys226, His208, Ser209 of Ldt Mt1 and Cys354, His336, Ser337 of Ldt Mt2 14,[19][20][21][22] . X-ray crystallographic and kinetic studies revealed that Met175/303, Tyr190/318 His224/352, Gly225/353, and Asn228/356 residues of Ldt Mt1 /Ldt Mt2 form key non-covalent interactions with the inhibitor 9,14,19,22,23 . Further, these residues are well conserved in the MSA analysis, as shown in Fig. 1. Thus, the probability of a single compound to bind with the key catalytic residues Cys, His and Ser of both Ldt Mt1 and Ldt Mt2 enzymes were strongly predicted by MSA studies.

Multiple Sequence Alignment studies of the Ldt enzymes of
Drug repurposing by structure-based virtual screening against Ldt Mt1 and Ldt Mt2 . Crystal structure of both Mtb Ldt enzymes reveal that the key catalytic triad residues Cys226, His208, Ser209 of Ldt Mt1 and Cys354, His336, Ser337 of Ldt Mt2 are directly involved in the Mtb non-classical PG cross-linkage 19,22 . We selected Ldt Mt1 & Ldt Mt2 molecular targets for our computer-aided drug repurposing study. The FDA approved drugs (1556) from the DrugBank database were used for the high-throughput virtual screening (VS) by molecular docking technique against the active sites of Ldt Mt1 and Ldt Mt2 enzymes. Best 10% of the Glide high scoring drugs common to both VS hits against Ldt Mt1 and Ldt Mt2 were selected, which include Acarbose and DIO ( Fig. 2a and Table 1). Acarbose, was obtained as the best VS hit with a high docking score of −10.01 and −13.16 kcal/mol against Ldt Mt1 and Ldt Mt2 respectively, as shown in Table 1. Acarbose is used for the treatment of Type 2 diabetes mellitus and has a very low bioavailability. An earlier report suggested that Acarbose binds to trehalose synthase enzyme of M. smegmatis 29 . DIO was the second highest ranked VS hit with a docking score of −9.37 kcal/mol for Ldt Mt1 and −10.55 kcal/mol for Ldt Mt2 enzymes, shown in Table 1. DIO is a bioflavonoid compound used to treat venous disorders and also prescribed as a dietary supplement. It also possesses anti-inflammatory and anti-oxidant activities [30][31][32] . Since Acarbose has very low bioavailability (~2%) 33 , we selected DIO for our in vitro and in vivo validation studies.
DIO is composed of a flavanoid moiety which is linked to a carbohydrate moiety via O-glycosidic bond (Supplementary Figure S1). Molecular docking studies revealed that DIO interacts with the Ldt Mt1 residues-His208, His224, Val222, Asn228, Thr193, Thr194, and Ser223 via hydrogen bonding, thiol group of Cys226 forms a π-bond with the flavonoid moiety of DIO; His224 and Trp213 are making π-π stacking interactions with DIO, as shown in Table 1 and Fig. 2b,c. Further, the interaction of DIO towards Ldt Mt2 showed that Asp304, Gln327, www.nature.com/scientificreports www.nature.com/scientificreports/ Tyr330, Ser331, Tyr308, and Thr320 residues are making hydrogen bonding interactions with DIO. Further, π-π stacking interactions were observed with Met303, His352, Tyr330 and Tyr318 residues, as shown in Table 1 and Fig. 2e,f. The interactions of DIO towards both the Ldt enzymes consisted of reported key residues His224, Cys226, Asn228 of Ldt Mt1 and Ser331, Tyr308, Thr320, His352, Met303, Tyr318 of Ldt Mt2 9,14,19,22,23 required for the binding of carbapenem antibiotics, shown in Fig. 2 and Table 1. Thus, we observed that DIO adopts a similar binding mode as that of carbapenem with the active sites of both Ldt enzymes, using molecular docking studies.
It is well known that DIO is hydrolyzed into its aglycone form Diosmetin (DMT) by intestinal microflora enzymes 34,35 . Further, it was reported that DMT showed good synergistic inhibitory activity when used in combination with Erythromycin against pyruvate kinase of MRSA 36 . We analyzed the binding mode of DMT towards both the Ldt Mt1 and Ldt Mt2 enzymes with respect to DIO for understanding its molecular mechanism of interactions. The GLIDE molecular docking and electrostatic potential surface mapping studies of these complexes showed that the flavonoid moiety of DIO is occupied in the active site tunnel of Ldts (Fig. 2d,g), and its carbohydrate moiety lies outside the active site tunnel. This implies that the flavonoid moiety of DIO is contributing towards the inhibition of both the Ldt enzymes. Further, the binding mode of DMT at the active sites of Ldt Mt1 and Ldt Mt2 revealed a similar orientation and position as flavonoid moiety of DIO, with a Glide docking score of −6.48 kcal/mol and −6.23 kcal/mol respectively, as shown in Supplementary Fig. S2a-d.
To effectively block the PG biosynthesis of Mtb, simultaneous inhibition is essential for both the classical and non-classical PG cross linkages mediated by the D,D-transpeptidase and Ldt enzymes (Ldt Mt1 and Ldt Mt2 ). The β-lactam antibiotic, AMX target D,D-transpeptidase enzyme. Glide molecular docking of AMX at the active site of D,D-transpeptidase (PDB ID:5CXW 37 ) showed binding energy of −5.64 kcal/mol, and the molecular interactions of the complex (AMX-D,D transpeptidase) is shown in Supplementary Fig. S3a,b. These in silico results broadly reveal the mode of binding of DIO/DMT with Mtb Ldt enzymes as described above, and AMX with D,D-transpeptidase enzyme active sites, leading to a potential synergistic Mtb growth inhibition.  (Fig. 3a,b). However, Ldt Mt2 -MEM was found to be unstable throughout MD simulation, as indicated by the high backbone fluctuations with an average root mean square deviation (RMSD) of 2.34 Å (Fig. 3b). Whereas the average RMSD  Table S1). Our in silico results thus broadly indicate that DIO has a better binding affinity towards both the Ldt enzymes in comparison to that of MEM. Ldt Mt1 and Ldt Mt2 with DIO was studied using a bioassay method. The mycobacterial Ldt enzymes were allowed to bind to the Ni-NTA column and DIO (or DMT) was allowed to interact with the column bound Ldt enzymes. The complexes of Ldt-DIO (or Ldt-DMT) were eluted, and DIO (or DMT) was separated from the Ldt proteins by ultrafiltration. Moldovan et al. reported a sensitive spectrophotometric method for the determination of DIO with an absorbance maxima at 263.5 nm 38 . We also measured the optical density (OD) at 263.5 nm of the filtrate. DIO was found to be interacting with both the Ldt enzymes and the OD values of DIO in Ldt Mt2 -DIO fractions was 0.245 and for Ldt Mt1 -DIO fractions was 0.15, shown in Fig. 4a. This result suggests that DIO has stronger binding affinity towards Ldt Mt2 in comparison to that of Ldt Mt1 enzyme. Our molecular docking results also showed a similar trend of binding affinity when DIO was docked at the active site of Ldt Mt2 and Ldt Mt1 , as described earlier (Table 1). Similarly, we analysed the binding of DMT towards both the mycobacterial Ldt enzymes. The absorption peak at 344 nm 39 indicated the presence of DMT in the eluted fractions. The OD values of DMT in the eluted fraction of both Ldt Mt1 and Ldt Mt2 enzymes are almost similar, as shown in Fig. 4b. Thus, DMT has similar binding affinity towards both the enzymes. Hence, our experimental binding studies are in consonance with the computational molecular docking simulations, as described above.

In vitro anti-mycobacterial activity studies of different drugs towards M. marinum. The in-vitro
anti-mycobacterial activity of AMC, DIO, DMT alone and their combinations were tested against M. marinum, as described in the materials and methods section. No significant growth inhibition was observed when M. marinum was cultured in the presence of increasing concentrations of 100, 250 or 500 µg/ml DIO or DMT. Minimum bactericidal concentration (MBC) of AMC against M. marinum obtained in our study was 8 µg/ml. Next, increasing amount of AMC (0.5 µg/ml to 4 µg/ml) was added to the M. marinum cultures along with a fixed concentration of DIO or DMT (500 µg/ml), which showed a gradual decrease in the colony forming unit (CFU), as compared to AMC alone (Fig. 4c). The MBC of AMC was reduced to 4 µg/ml from 8 µg/ml concentration. Further, AMC-DMT  www.nature.com/scientificreports www.nature.com/scientificreports/ combination of drugs exhibited higher anti-mycobacterial activity as compared to AMC-DIO, as indicated by the reduction in the CFU, as shown in Fig. 4c. Recently, Sridharan et al. predicted DIO to have a higher binding affinity towards the active site of mycobacterial panthothenate kinase by molecular docking studies 40 . However, this study also showed no anti-mycobacterial activity for DIO alone, and which is in good agreement with our present study. Further, the present results indicate that AMC has a synergistic anti-mycobacterial activity in combination with DIO or DMT against M. marinum.

M. marinum treated with AMC-DIO (or DMT) combination of drugs severely altered its cell surface morphology.
It is well known that PG layer of mycobacteria provides structural integrity and is essential for its growth and survival. Schoonmaker et al., demonstrated that the loss of Ldt Mt1 and Ldt Mt2 enzymes alter the cell surface morphology, shape and size of tubercle bacilli 12 . Based on these observations, we also hypothesised that the treatment of M. marinum with AMC-DIO or AMC-DMT combination of drugs would inhibit the PG cell wall cross-linking and cause changes in their cellular surface morphology. Scanning Electron Microscope (SEM) imaging study was performed for M. marinum treated with AMC, DIO or DMT individually and its combinations, AMC-DIO and AMC-DMT.
M. marinum with or without DIO (or DMT) drug treatment revealed a smooth cell surface, shown in Fig. 4d(i and iii). However, AMC treated M. marinum exhibited slightly corrugated cell surface with dents, as shown in Fig. 4d(ii). M. marinum cultured in the presence of DMT also exhibited slight changes in the cellular morphology ( Fig. 4d(iv)). M. marinum on treatment with AMC-DIO or AMC-DMT showed dramatic changes in their cellular surface morphology consisting of unrecognizable large amorphous mass of cell debris (Fig. 4d(v  and vi)). In addition, degraded cells appeared shrunk and corrugated with evident atrophy, and also bulges were present all over the mycobacterial cell surface. Also, Mtb cell wall has both DD-transpepetidase and Ldt enzymes, and which play a key role in its PG cross-linking. Hence, M. marinum on treatment with AMC-DIO or www.nature.com/scientificreports www.nature.com/scientificreports/ AMC-DMT combinations showed severe morphological changes, suggesting its potential role in the inhibition of PG cross-linking enzymes.
The WHO recommends AMC along with the MEM as an add-on therapeutic option for the treatment of drug resistant TB, since clavulanic acid is only available in combination with AMX 41 . However, the major drawback of MEM is its intravenous administration, thrice a day, and which further requires prolonged hospitalization of the patient. The present work has the potential significance in which the combination of drugs, AMC-DIO (or AMC-DMT) showed significant in vitro anti-mycobacterial activity, and the drugs can be administered orally for an effective TB therapy.  (Fig. 5b). It was reported earlier that in mammalian system, DIO gets converted into its aglycone form DMT by intestinal micro flora 34,35 . Thus, using mass spectrometry (MS) analysis, we confirmed that the conversion of DIO to DMT occurs in D. melanogaster flies after oral administration of DIO, as shown in Fig. 5a. Treatment of infected flies with different drug concentrations of either AMC or DIO individually does not show any marked fly survival (Fig. 5c,d). However, the survivability of the flies was significantly improved on treatment with increasing concentration of AMC and DIO combination. Maximum fly survival of 66.67% was observed at the higher concentration of AMC (1 mg/ml) and DIO (2 mg/ml) treated flies, revealing a synergistic effect of the combination of drugs (Fig. 5e).

Synergistic effects of AMC
Further, we analysed the efficacy of DMT (2 mg/ml) at the same concentration of DIO in combination with AMC (1 mg/ml) in the infected flies. AMC-DMT also exhibited a synergistic activity with a percentage survival of 60.0 of the infected flies, as shown in Fig. 5f. Thus, the present MS analysis and fly survival studies indicated that DIO after oral administration is converted into DMT. Also the therapeutic effect of either AMC-DIO or AMC-DMT is similar in M. marinum infected D. melanogaster fly model. Additionally, we checked the efficacy of AMC-DIO along with the standard TB drug, RIF. When infected flies were fed with a lower concentration of RIF (0.25 mg/ml), no significant fly survival was observed. Interestingly, a better fly survival of 83.33% was observed when infected flies were treated with a combination of AMC (1 mg/ml), DIO (2 mg/ml) and RIF (0.25 mg/ml), as shown in Fig. 5g.
Additionally, the bacterial load in the infected flies on 3 rd , 5 th , 7 th and 9 th day of drug treatment were determined by AFB (Acid Fast Bacilli) staining technique, as shown in Fig. 6. The number of bacteria present in 100 fields was counted and graded according to WHO/IUATLD Quantification scale (Supplementary Table S1). Progressive increase in the severity of M. marinum infection was seen in the flies treated with AMC, DIO or DMT individually, shown in Fig. 6a(ii,iii and v). The corresponding AFB smear grading of infected flies treated with these drugs was escalated to 3+ within five days of treatment, as shown in Fig. 6b. However, a reduction in the bacterial load was observed on treatment with AMC-DIO or AMC-DMT. The flies on the 9 th day of treatment with AMC-DIO or AMC-DMT were completely free from M. marinum infection as indicated by the absence of bacilli in the smears (Fig. 6a(iv and vi) and Fig. 6b). These in vivo observations strongly support our earlier fly survival studies and revealed a synergistic anti-mycobacterial activity of the combination of drugs. Further, we checked the bacterial load in the infected flies treated with RIF alone and in combination with AMC-DIO. In RIF treated fly group after 9 days of treatment, only a few bacilli were present in the smear samples ( Fig. 6a(vii)). However, the survival of infected flies was only ~20% as shown in Fig. 5g. Further, RIF-AMC-DIO was very effective and no bacilli were seen in the fly smear from 7 th day onwards, as presented in Fig. 6a,b(viii). In summary, treatment of M. marinum infected flies with the individual drugs did not show any marked improvement in their survival, while AMC-DIO or AMC-DMT or AMC-DIO-RIF showed better survival of infected flies.

In vitro anti-mycobacterial activity of AMC-DIo combination of drugs against Mtb and MDR clinical isolate. Anti-mycobacterial activity of various concentrations of AMC, DIO and their combinations
were tested against Mtb strain H37Ra using an automated MGIT 960 system. AMC (upto 20 μg/ml) and DIO (upto 500 μg/ml), when tested individually against Mtb, did not show any inhibitory activity. However, the combination of AMC (12 μg/ml) with 500 μg/ml of DIO was found to have a synergistic anti-mycobacterial effect. Increasing concentrations of AMC, 12 μg/ml to 20 μg/ml, when combined with 500 μg/ml of DIO, showed a gradual reduction in the mycobacterial growth (Fig. 7a). We observed a complete inhibition of mycobacterial growth at 20 μg/ml concentration of AMC in combination with 500 μg/ml of DIO, as shown in Fig. 7a.
Resistance to different antibiotics is of major concern in the current TB treatment regimen. Thus, we tested the in vitro efficacy of AMC and DIO both individually and in combination at different concentrations against MDR-Mtb clinical isolate. We observed that the individual drugs (AMC or DIO) were not able to inhibit the growth of MDR clinical isolate even at high concentrations (24 μg/ml of AMC; 500 μg/ml of DIO). However, AMC at concentrations above16 μg/ml, in combination with 500 μg/ml of DIO were able to show a concentration dependent reduction of mycobacterial growth. A complete growth inhibition of MDR clinical isolate was observed for AMC 24 μg/ml concentration in combination with 500 μg/ml of DIO, shown in Fig. 7b. Thus, the present in vitro efficacy studies clearly demonstrated a synergistic inhibitory activity of the AMC-DIO combination of drugs against a MDR clinical isolate of Mtb.

Conclusions
Using computer-aided drug repurposing method we identified DIO by targeting both Ldt Mt1 and Ldt Mt2 enzymes of Mtb. After oral administration, DIO rapidly gets converted into its aglycone form, DMT. In the present study, we proved that both DIO and DMT interact with Ldt Mt1 and Ldt Mt2 enzymes using molecular docking technique and in vitro bioassay. A synergistic anti-mycobacterial activity of AMC and DIO or DMT combination www.nature.com/scientificreports www.nature.com/scientificreports/ was observed against M. marinum and Mtb H37Ra. Further, the in vivo efficacy of AMC-DIO or AMC-DMT combination of drugs was validated in M. marinum infected D. melanogaster fly model. Multi-drug resistance is a major challenge in the current TB combination therapy, leading to poor treatment outcomes. We demonstrated the synergistic anti-tubercular activity of AMC-DIO combination against a MDR-Mtb clinical isolate. The carbapenems are effective in killing both MDR and XDR strains of Mtb. However, they are not hydrolytically stable; this limits its antibiotic administration to a controlled intravenous infusion and also requires prolonged hospitalization. Both DIO and AMC are available as oral formulation, which can be a safe and inexpensive alternative for TB treatment. Further, DIO is a flavonoid compound with proven anti-oxidant, anti-inflammatory, chemo-preventive properties and it also modulates the host immune system. These pharmacological activities of DIO will thereby decrease the destruction of tissues and internal organs during Mtb infection. In conclusion, we have successfully repurposed DIO as an anti-mycobacterial agent in combination with AMC and which can be a potential alternative for the treatment of TB.

Materials and Methods
Computational Methods. Primary sequence analysis, 3D structural studies, molecular modelling, docking and MD simulation studies were performed on a Linux workstation using Schrödinger, AMBER and PyMol molecular modeling packages. were retrieved from the UniProt sequence database and multiple sequence analysis (MSA) was performed using Clustal-X program 43 . Crystal structure of Ldt Mt1 (PDB ID: 4JMN) 19 and Ldt Mt2 (PDB ID: 4GSR) 22 were retrieved from Protein Data Bank (PDB) and the structures were optimized using the Protein Preparation Wizard module of Schrödinger 44 . Initially, the 3D structures of both Ldts were pre-processed by assigning the bond orders, adding hydrogen and deleting water molecules and finally the 3D structure minimization was done using OPLS2005 (Optimized Potential for Liquid Simulations) force field.
Structure-based virtual screening using FDA approved drugs against Ldt Mt1 and Ldt Mt2 . The chemical structures of all the FDA approved drugs (1556) was collected from the DrugBank database 45 for our drug repurposing study. The geometry of these structures were optimized by employing MMFF94 (Merk Molecular Force Field) using the LigPrep wizard of Schrödinger 46 . Using GLIDE (Grid-based Ligand Docking with Energetics) module of Schrödinger v9.2 47 , molecular docking of these FDA approved drugs (1556) with the active sites of both Mtb Ldts was computed. Initially, centre of the grid box for molecular docking was selected by enclosing the experimentally reported active site residues for both Ldt Mt1 and Ldt Mt2 enzymes 19,22 . GLIDE docking program has been validated and successfully implemented in our previous VS as well as the host-pathogen interaction studies [48][49][50] . Molecular docking was performed in two stages: The first stage includes SP (standard-precision) docking followed by XP (extra-precision) docking with higher accuracy and less false positives. The GLIDE XP binding free energy is calculated using the following equation: Where E coul , E vdW , and E bind represent the electrostatic potential energy, van der Waals energy, and binding interaction energy respectively. E penalty is the strain energy from protein, ligand or both, loss of entropy of the ligand and protein and desolvation of the ligand or protein 46 .

Molecular dynamics simulation and binding free energy calculations of Ldt-drug complexes.
Molecular dynamics (MD) simulations were performed using AMBER12 51 program for 20 ns time interval for Initially, the atom types and the AM1-BCC charges for the ligands-MEM and DIO was assigned using Antechamber module and general amber force field (GAFF). The atom types of the proteins are assigned using the Leap program embedded in AMBER12, by employing the ff99SB force field. The uneven charge of these proteins was neutralized by adding sodium ions followed by solvation by immersing in a rectangular periodic box containing TIP3P water molecules. The prepared systems were then subjected to two step energy minimization employing steepest descent and conjugate gradient methods. These molecular systems were then subjected to initial equilibration, where it was heated from 0 to 300 K for 20 ps time interval with a cut-off of 10.0 Å for long range non-bonded interactions. The entire equilibration was performed using Langevin dynamics temperature control method and harmonic restraints was used with a force field constant of 10 kcal/mol Å 2 to all the protein atoms. The MD production run without any restraints for 20 ns time interval was performed. Trajectory coordinates were recorded at every 250 ps of MD simulation and was analysed using PTRAJ module in AMBER12. Using MM-GBSA 52 module, the binding free energy (ΔG bind ) was calculated for Ldt Mt2   www.nature.com/scientificreports www.nature.com/scientificreports/ containing 50 mM NaH 2 PO 4 , 300 mM NaCl and 20 mM imidazole (pH 8.0). DIO/DMT solution (50 µg/ml) was added into the Ni-NTA columns and subsequently washed with wash buffer. The proteins (Ldt Mt1 and Ldt Mt2 ) were then eluted using the same buffer with 250 mM concentration of imidazole. The eluted protein fractions were incubated at 60 °C for 10 mins and the proteins and DIO (Mw: 609.181 Da) or DMT (301.071 Da) in the respective elutes were separated by ultrafiltration using Amicon Ultra-4 10 K centrifugal filter devices (Millipore). The optical density of the respective filtrates at 263.5 nm and 344 nm was determined for DIO and DMT respectively using Nanodrop Microvolume UV-Vis Spectrophotometer (Thermo Scientific). E.coli BL21 alone was used as the control for the experiment.
In vitro anti-mycobacterial activity assays. scanning electron microscopy study. Two days old M. marinum cultures were treated with AMC (8 µg/ml; AMX and clavulanic acid in the ratio 4:1), DIO (500 µg/ml), DMT (500 µg/ml) and their combinations for 48 h. Samples were fixed using 2.5% (v/v) glutaraldehyde and dehydrated in graded ethanol series 55,56 . Samples were gold sputter coated and imaged to study mycobacterial morphology.
Mass spectrometry study. Drosophila melanogaster flies were raised in standard banana agar medium and allowed to feed DIO (2 mg/ml) for 5 days. Flies were then homogenized with a pestle in 100 μl of sterile double deionized water and extracts were used for MS studies. The Q-TOF LC-MS/MS analysis were carried out using an Agilent 1290 Infinity Ultra High Performance Liquid Chromatographic system coupled to an Agilent 6540 UHD Accurate Mass Q-TOF mass spectrometer equipped with Dual AJS (ESI) source. The samples were infused to the mass spectrometer through a reversed-phase column (Agilent Poroshell-120, 2.1 × 150 mm, 2.7 µm). The mobile phase (A: water with 0.1% formic acid; B: acetonitrile with 0.1% formic acid) flow rate was maintained at 0.2 ml/ min. A linear gradient of 5-95% mobile phase B in 20 min was used to separate the molecules. The mass spectra were acquired between 100 and 1000 m/z and the dry gas flow and dry gas temperature were set at 6 ml/min and 320 °C, respectively. Targeted MS/MS data of DIO (theoretical monoisotopic protonated mass, 609.181 Da), and DMT (theoretical monoisotopic protonated mass, 301.071 Da) were collected by setting the collision energies at 20 V and 25 V, respectively. All the MS data were acquired in positive ionization mode using Agilent Mass Hunter data acquisition software, Vers. B.05.01 and the results were analyzed using Agilent Mass Hunter qualitative analysis software, Vers. B.07.01 57 . Solvents (acetonitrile, water and formic acid) used for MS study was of HPLC grade and procured from Merck.

In vivo anti-mycobacterial activity studies in M. marinum infected D. melanogaster fly model.
In vivo experiments were carried out using 5 to 7 days old male D. melanogaster fly (Flybase Stock 2:FBst0000002). Flies were fed and raised at 26 °C on banana agar media with 60% humidity 58 . One day prior to the infection, the flies were starved for 6 h. A Whatman filter paper disc of 2.3 cm diameter was placed on top of the banana medium in glass vials and 60 µl of cyclophosphomide (stock concentration: 500 µg/ml, dissolved in double distilled water) was added to the filter paper disc. Flies were then transferred to these glass vials and allowed to feed. After 24 h, these flies were injected with M. marinum (OD 600 = 3.0) at the junction area between the ventral and dorsal cuticles of the fly, using needle of 31 Gauge thickness. Flies were then returned to the standard fly culture vials containing respective drugs, AMC (AMX and clavulanic acid in the ratio 4:1), DIO, DMT, and RIF; the combination of drugs, AMC-DIO, AMC-DMT and AMC-DIO-RIF respectively. A Whatman filter paper disc was placed on the banana agar surface of the fly culture vials and 60 µl of different concentrations of specific drugs, AMC (0.25, 0.5 and 1 mg/ml), DIO (0.25, 0.50, 1 and 2 mg/ml), DMT (2 mg/ml) and RIF (0.25 mg/ml) and their combinations were added to the filter paper disc. The flies were transferred to fresh media and fly survival