Flies were trapped from livestock (cattle, buffalo, sheep, goat, pigs) & poultry farm premises and Veterinary Clinical Complex (VCC) using manually prepared traditional traps. The unconventional fly-traps were made out of plastic bottles. Dried fish and chicken powder, in the ratio of 2:1, mixed with sugar syrup was used as bait. Sampling frequency for collection of flies was predetermined so as to collect pooled samples from farm premises. Traps were fixed at different places in the livestock/ poultry sheds where flies were abundant. For each food animal species, trapping was attempted six times, i.e. six pools of samples per species were collected (Table 1). Using no specific selection criteria, 40% of each catch was processed, provided that the numbers of flies caught are not ≤ 15, in such case, all the flies were processed for bacterial isolation.

Flies were first washed with 1x Phosphate Buffer Saline (PBS) to remove exterior debris and genus identification was done on the basis of wing pattern under 10x microscopic view (Sen and Fletcher, 1962). Identification at genus and species level was done by expert parasitologist. Identified flies were pooled and washing was carried out three times with 1× PBS to reduce surface contaminants. Washed pool samples were placed in a fresh tube and centrifuged for 10 min at 12000 rpm. Flies settled at the bottom of the centrifuge tubes were crushed by a sterile metal rod, and mixed with sterile PBS, again centrifuged for 10 min at 6000 rpm. The supernatant was collected and used for isolation of E. coli and Salmonella species. Fecal samples randomly collected from cattle, buffalo, sheep, goat, pigs and poultry using rectal/cloacae swabs were also processed for bacterial isolation.

For obtaining pure culture of E. coli and Salmonella, enrichment followed by selective plating on specific medium was performed. For E. coli, 1 ml supernatant was inoculated into 5 ml Enterobacteriaceae Enrichment Broth (EEB) with incubation at 37 °C for 24 h. Similarly, rectal/ cloacae swabs were also enriched in 10 ml of EEB for 24 h at 37 °C. A loop full of enriched culture was streaked on Eosin Methylene Blue (EMB) agar and plates were again incubated at 37 °C for 24 h. For isolation of Salmonella, pre-enrichment in Buffered Peptone Water (BPW), enrichment in EEB and selective plating on Xylose Lysine Deoxycholate (XLD) agar was done. Fecal swabs and 1 ml supernatant of pooled flies were first inoculated in 9 ml BPW and incubated at 37 °C for 24 h. Then, 1 ml of pre-enriched culture was inoculated with 5 ml EEB and incubated overnight at 37 °C. A loop full of the enriched culture was used for selective plating on the XLD agar plates incubated at 37 °C for 24 h. Three to five representative E. coli and Salmonella colonies were picked up, purified and confirmed by biochemical tests namely, catalase, oxidase, indole, methyl red, Voges Proskaur, H₂S production and citrate utilization.

Matrix-assisted laser desorption ionization time - of - flight (MALDI - TOF) mass spectrometry (MS) technique was used for confirmation of E. coli and Salmonella. Facilities available at Centre for Zoonoses, Department of Veterinary Public Health, Nagpur Veterinary College, MAFSU Nagpur were used for MALDI - TOF MS identification. Isolates were first grown on BHI agar and single colony from BHI agar plate was picked using sterile loop and placed in target plate. Colonies were dissolved in one to two micro liter of matrix solution. Matrix solution was composed of alpha-cyano-4-hydroxycinnamic acid saturated in 50% acetonitrile and 2.5% trifluoracetic acid. Target plates were allowed to dry at room temperature and placed into the plate chamber of the mass spectrometer. For calibration of the instrument known standard culture of E. coli was used during the assay. The further analysis of the spectra giving clear identification of the organisms was documented (Fig. 1, 2).

Antimicrobial susceptibility and resistance patterns of E. coli and Salmonella species isolated from flies and fecal samples were studied using Kirby-Bauer disc diffusion method and interpreted using criteria suggested by the Clinical and Laboratory Standards Institute (CLSI, 2017). AMR studies were performed using 12 different antimicrobials as depicted in Table 3. Overnight grown bacterial cultures in BHI broth were smeared on Mueller Hinton agar (MHA) plates and antimicrobial discs were placed at suitable distance. After drying, plates were incubated at 37 °C for 24 h. Zones of inhibition were measured and interpreted in accordance to the manufacturer’s instructions (HiMedia Laboratories, Mumbai).

Multidrug resistant strains were studied for ESBL production using double disk synergy test (DDST). The procedure involved initial screening followed ESBL confirmation. Cefotaxime (30 µg) and ceftazidime (30 µg) discs were used for initial screening, and cefotaxime/ clavulanic acid (30/10 µg) and ceftazidime/clavulanic acid (30/10 µg) discs for ESBL confirmation. BHI inoculated overnight grown cultures of E. coli and Salmonella were streaked on the MHA plates. Plates were incubated at 37 °C for 24 h and difference in the zone of inhibition was recorded. Isolates showing difference in zone of inhibition of ≥5 mm of cephalosporin discs and cephalosporin plus clavulanic acid containing disc were considered as potential ESBL producers.

The isolates showing resistance to carbapenem discs (ertapenem, doripenem, meropenem and imipenem) during initial phenotypic screening were suspected as carbapenemase (KPC) positive. MIC was estimated using ertapenem/ertapenem plus boronic acid Ezy MIC strips (HiMedia). Ten randomly selected KPC positive isolates were cultured overnight in BHI broth and smeared on the MHA plates. The Ezy MIC strips were placed on the MHA plate carefully and incubated at 37 °C for 24 h. The ratio of ertapenem/ ertapenem + boronic acid (ETP/ETP+) more than 8 was considered positive for KPC.

Improved AmpC detection Ezy MIC strips (HiMedia) were used and all the isolates that were commonly resistant to amoxicillin-clavulanic acid, aztreonam and piperacillintazobactum were screened. The Ezy MIC strips containing MIX+ (ceftazidime, cefotaxime, cloxacillin & clavulanic acid- 0.032 - 4) and MIX (ceftazidime, cefotaxime & cloxacillin- 0.125 - 16) were placed on MHA plate and incubated at 37 °C for 24 h. If the ratio of the value of MIX and MIX in combination with clavulanic acid (MIX+) found more than 8 or no zone for MIX and zone obtained in MIX+, it was considered as ESBL+AmpC positive strain. When ratio of the value obtained for MIX and MIX+ was ≤ 8, it was considered as AmpC positive (AmpC present, but ESBL absent).

For the determination of MIC of suspected MBL (Metallo- β Lactamase) producing isolates, the common isolates showing ESBL and AmpC production were selected. Isolates were cultured in BHI broth and streaked on MHA plates as described above. The MBL plus ESBL Detection Ezy MIC™ strips containing ESBL+ (ceftazidime, cefotaxime, EDTA & clavulanic acid- 0.032 - 4) and ESBL (ceftazidime, cefotaxime & EDTA- 0.125 - 16) were placed on the agar plate and incubated. When the ratio of the value obtained for ESBL and ESBL+ found more than 8 or no zone obtained for ESBL and zone obtained for ESBL+, it was considered as MBL+ESBL positive strain.

When the ratio of the value obtained for ESBL and ESBL+ was ≤ 8, it was considered as MBL positive strain (ESBL is not present along with MBL).

Isolates resistant to ESBL, KPC, MBL and colistin were further screened for estimation of MIC for colistin using Ezy MIC strips (HiMedia). Thirty isolates were cultured in BHI broth and smeared on the MHA plates. The Ezy MIC strips were placed on MHA agar plates and incubated as described previously. Zone inhibition reading at the value of ≥ 2 µg/ml was considered as resistant, and a lesser value was termed as susceptible.

Phenotypically confirmed colistin resistant isolates of E. coli and Salmonella were screened by multiplex PCR targeting mobile colistin resistance genes viz. mcr1 to mcr5 as per the method described previously (Rebelo et al., 2018). Briefly, bacterial DNA was extracted by boiling and snap chilling method and 2 µl of supernatant was used as the DNA template for PCR. Multiplex PCR was performed in 25 μl volume containing 12.5 μl 2x PCR master mix (HiMedia), 0.5 μl of each forward and reverse primers (mcr1 - mcr5), 2 μl DNA template and 5.5 μl nuclease free water to make final volume of 25 μl. Cycling conditions were set as: initial denaturation (94 ºC/15 min^{- 1} cycle) followed by 35 cycles of denaturation (94 ºC/30 sec), annealing (58 ºC/90 sec), and extension (72 ºC/60 sec). Final extension was attained at 72 ºC/10 min and holding at 4 ºC. Five microliter amplified product was separated by electrophoresis in 1.5 % agarose gel dissolved in 0.5 × TBE stained by ethidium bromide. Oligonucleotide sequences used are depicted in Table 2.

Out of 2187 flies trapped, all were identified as house fly (Musca domestica) and only one blow fly (Calliphora erythrocephala) could be trapped at VCC of the institute. E. coli was isolated from all the flies and fecal samples (100%). However, isolation rate of Salmonella species was comparatively less. It could be isolated from 21 (58.33%) out of 36 pools of flies and 15 (20.83%) fecal samples. Thus, total 109 E. coli and 36 Salmonellae species were confirmed. Similarly, 35 randomly selected isolates (15 E. coli and 20 Salmonellae) were also confirmed by MALDITOF MS identification system. All the 35 isolates were confirmed as E. coli and Salmonella enterica by MALDI -TOF MS system. Livestock, poultry farm and rural environment provides ideal environment for breeding and propagation of house flies and therefore Muscidae must be the most abundant species near human and animal dwellings. Abundance of house flies and blow flies in the animal farms were previously reported from Thailand (Fukuda et al., 2018). E. coli, Salmonella, Shigella and helminth eggs were recently identified in Musca domestica collected from cattle byres, poultry farms and human houses (Issa, 2019). Musca domestica exposed to the chickens challenged with Salmonella Enteritidis get colonized with strain within 24 - 48 h and were able to transmit Salmonellae to naive chicken population (Holt et al., 2007). Foodborne pathogens including bacteria, gastrointestinal parasites and viruses have been isolated from exterior body surface and gut of filth flies collected from human houses, livestock and poultry farms (Hald et al., 2004; Lindeberg et al., 2018). A study from Tamil Nadu, have established link between fly densities and diarrheal episodes in rural and urban households and 99.9% flies were from Muscidae family. Rotavirus, Salmonella, Shigella and E. coli were isolated from them (Collinet-Adler et al., 2015).

All the 145 isolates (109 E. coli and 36 Salmonella) were examined for antimicrobial resistance pattern by disc diffusion method (Table 3). Almost all the E. coli isolates expressed resistance to multiple antibiotics. Most of the E. coli strains irrespective of species/source of isolation has shown resistance to amoxicillin-clavulanic acid (100%), cefotaxime (93.57%), aztreonam (59.63%), cefpodoxime (58.71%) and imipenem (48.62%). Overall colistin resistance was recorded very low (12.84%) and only 14 isolates expressed resistance to colistin by disc diffusion method. Species/source wise prevalence of colistin resistance was recorded as: cattle-buffalo (66.6%), sheep (5.5%), goat (22.77%), poultry (11.11%), pigs (11.1%) and VCC isolates (5.26%). E. coli isolates were multi-drug resistant strains and out of 109 isolates, 92 isolates (84.40%) expressed resistance to at least 3 out of 12 antibiotics. Multiple antibiotic resistance (MAR) index for the individual strains vary from 0.25 to 0.75. AMR pattern of E. coli and Salmonella species isolated from flies and fecal samples of food animals is presented in Table 4 and 5. Overall, Salmonella isolates showed 100% resistance to Ampicillin-clavulanic acid followed by cefotaxime (91.66%), cefpodoxime (94.44%), imipenem (91.66%), colistin (61.11%) and piperacillin-tazobactam (52.77%).

Resistance to ertapenem, meropenem and doripenem was very low. Colistin resistant strains of Salmonella were isolated from flies as well as fecal samples and its prevalence was high (61.11%) as compared to colistin resistant E. coli (12.84%). Highest number of colistin resistant strains were recovered from pig farm (90%) and VCC (80%) premise. Resistance to multiple antimicrobials was also recorded in the Enterobacteriaceae including E. coli isolated from flies collected at dairy and poultry sites in Portugal (Barreiro et al., 2013). Detection of AMR strains of E. coli and Salmonella isolated from flies and fecal samples of farm animals is worrisome as flies can strains of E. coli and Salmonella isolated from flies and fecal samples of farm animals is worrisome as flies can carry antimicrobial resistant microbes mechanically from animals to humans. AMR in E. coli isolated from flies (Sobur et al., 2019) and food animals (Malhotra-Kumar et al., 2016) were studied earlier. In contrast to present findings ertapenem, imipenem and meropenem susceptible E. coli were found in healthy cattle from Spain (Hernández et al., 2017). In another cross sectional study on houseflies collected from educational hospitals, vegetable centre and livestock slaughter house, E. coli resistant to gentamicin, cefotaxine, ciprofloxacin and other routine drugs were detected (Nazari et al., 2017). Salmonella isolates under study showed very high resistance to ampicillinclavulanic acid, cefotaxime, cefpodoxime, imipenem and colistin. However, resistance to ertapenem, meropenem and doripenem was very low. A recent study on AMR diversity of Salmonella Typhimurium exhibited high degree of AMR diversity in pig isolates as compared to chicken and cattle origin isolates (Mellor et al., 2019). Isolation and characterization of E. coli and Salmonella from flies for AMR is not extensively investigated and there are no supporting reports from India with reference to AMR in bacteria associated with flies trapped at livestock and poultry farms. Present findings affirm that Musca domestica flies prevalent at livestock and poultry farms harbor multi-drug resistant strains of E. coli and Salmonella.

Out of 109 E. coli, 23 (21.10%) were ESBL positive strains. Among these, 8 (34.78%) isolates were from flies and 15 (65.21%) were from fecal samples. The distribution of ESBL positive E. coli revealed maximum isolates from broiler (6/23), followed by goat (4/23), sheep (4/23), cattle (4/23), pig (3/23) and VCC (2/23) origin samples. Similarly, prevalence of ESBL positive Salmonella was 19.44% and out of 36, seven strains were ESBL positive. Only one isolate was from fly source and other six strains were from fecal samples. Isolates resistant to amoxicillinclavulanic acid, aztreonam and piperacillin- tazobactam during initial screening, were suspected as AmpC positive strains, however, all the isolates were negative for AmpC by Ezy MIC strips. Similarly, eight common ESBL and amoxicillin-clavulanic acid resistant isolates were screened for MBL type of resistance using Ezy MIC strips. Two isolates were positive for MBL, ESBL and AmpC production, while six were positive for MBL only. Further, 10 common isolates resistant to doripenem, meropenem, ertapenem and imipenem were screened for KPC using Ezy MIC strips and only one goat fecal origin E. coli was carbapenemase positive. Resistance to ESBLs, AmpC cephalosporinases, and carbapenem antibiotics are of greater public health significance. Thus, detection of ESBL, MBL, AmpC, KPC and colistin resistance in E. coli and Salmonella is highly significant in the recent context. ESBL positive E. coli strain was also detected in two fly pools collected from the poultry farms in the Netherlands (Blaak et al., 2014). In contrast to our findings, AmpC producing E. coli were highly prevalent in the broiler birds and farmers working at Dutch farms (Dierikx et al., 2013).

Initial phenotypic screening by the disc diffusion method revealed 12.84% E. coli and 61.11% Salmonella resistant to colistin. Out of these, 30 isolates (13 E. coli and 17 Salmonella) were examined for MIC of colistin. Overall, 16 (53.33%) isolates found positive for colistin at a MIC level of ≥ 2 mg/L. Of these, 10 were E. coli and 6 were Salmonella. Maximum E. coli isolates showing MIC ≥ 2 mg/L were of goat origin (three), followed by cattle and pig (one each). Colistin positive E. coli with MIC ≥ 2 mg/L could be isolated from two fly pools trapped at poultry farm and one each from cattle farm, goat farm and VCC. Out of six positive Salmonella isolates showing MIC of ≥ 2 mg/L, two each, were from flies collected at sheep farm, pig farm and VCC (Fig. 3, 4, 5).

Multiplex PCR was performed on all the 34 colistin positive isolates as per the standard protocol, however, we could not detected mcr genes in any of the phenotypically colistin resistant E. coli and Salmonella isolates. Use of colistin (polymyxin E) has been re-introduces in human medicine and now it is a last resort antibiotic to treat infections from multi-drug resistant bacteria. As compared to our study, less prevalence of colistin resistant E. coli isolated from livestock and food was noted in Germany (Irrgang et al., 2016).

Similar to the present observations high frequency of colistin resistance in pigs (23.75%) and poultry (7.92%) was recorded in a study from China (Huang et al., 2017). Sporadic reports are available on detection of mcr1 and mcr3 genes in E. coli isolated from flies (Fukuda et al., 2018; Sobur et al., 2019). Other than India, prevalence of mcr genes in E. coli and Salmonella isolated from food animals has been reported globally (El Garch et al., 2018). As colistin is used as last resort antibiotic, development of resistance against this important antibiotic is great cause of concern.