Molecular detection and characterization of the Mycobacterium tuberculosis complex subspecies responsible for bovine tuberculosis in Punjab, Pakistan

ABSTRACT Bovine tuberculosis (bTB), traditionally associated with Mycobacterium bovis, presents significant public health and economic challenges worldwide. This study investigated the causative agents of bTB in slaughtered cattle and buffalo in Lahore, Pakistan. Of the 3,581 animals screened, 34 were identified with gross TB-like lesions. The lesions were processed for culture, PCR, and Sanger sequencing to identify the causative agents of the disease. The results identified 10 Mycobacterium orygis and 8 Mycobacterium tuberculosis sensu stricto isolates. Whole-genome sequencing was performed on two M. orygis isolates, and the sequences were phylogenetically compared to 93 publicly available M. orygis sequences. The results also demonstrated that the JB21 and JB22 primers, which have been previously commonly applied to detect M. bovis in Pakistan, are unable to distinguish between M. tuberculosis complex subspecies. The identification of M. orygis and M. tuberculosis as causative agents of bTB in this slaughterhouse in Punjab may have important implications in identifying cases of zoonotic TB in humans and applying appropriate molecular tools to identify the prevalence of the disease. The data from this study align with recent findings suggesting M. orygis is the predominant cause of bTB in South Asia. IMPORTANCE The study findings hold significant relevance to the Journal of Clinical Microbiology, as they directly impact the field. The first-time identification of Mycobacterium orygis and Mycobacterium tuberculosis as the predominant causative agents of bovine tuberculosis in Lahore, Pakistan underscores the urgent need for enhanced diagnostic methods. The study emphasizes the importance of improved assays for the accurate detection and differentiation of Mycobacterium subspecies. Additionally, the research addresses zoonotic risk assessment and public health implications, advocating for a multidisciplinary approach that integrates clinical microbiology with veterinary and human health sectors. These insights contribute to clinical microbiology knowledge, shaping effective strategies for disease prevention, surveillance, and control. The study’s potential to advance the field makes it well suited for publication in the Microbiology Spectrum journal.

slaughterhouses (5), and constraints on animal trade (6).Mycobacterium bovis is the most commonly reported cause of bTB (7)(8)(9) and belongs to a related group of mycobac teria known to cause tuberculosis called the Mycobacterium tuberculosis complex (MTBC) (10).Mycobacterium orygis, a lesser-known MTBC subspecies, has also been reported as a cause of tuberculosis in cattle, particularly within South Asia (11)(12)(13).To detect bTB postmortem, histopathological techniques can be used to identify TB in granulomatous or fibrotic tissue lesions of animal tissue, but these are limited in their sensitivity and specificity (14).Alternatively, the detection of bTB using culture is a laborious process that takes weeks.Molecular assays have been proven to be a sensitive, reliable, and rapid method for the identification of bTB (15,16).They also offer the additional benefit of being able to differentiate between MTBC subspecies to determine the cause of the disease.Slaughterhouses may serve as sentinel sites for monitoring various diseases of animals and can also provide insight into the prevalence of bTB (17).In Pakistan, studies on the prevalence of bTB have varied by region, sample size, sampling strategy, and analysis, with estimates ranging from 14.00% (18) to 2.71% (19).Throughout Pakistan, M. bovis has previously been reported as the sole cause of zoonotic (20) and bovine TB (21).The majority of investigations in Pakistan used PCR with the JB21 and JB22 primers to detect zoonotic or bovine TB in human and animal samples (22)(23)(24).These primers were previously reported to amplify an M. bovis-specific 500-bp sequence (25).
In this study, we aimed to perform molecular assays to identify MTBC subspecies recovered from TB-like lesions in cattle and buffalo at a slaughterhouse in Lahore, Pakistan.We hypothesized that bTB would be present and would be caused by M. bovis.We applied molecular assays, whole-genome sequencing (WGS), and phylogenetic analysis to characterize the MTBC subspecies collected in our study.To our knowledge, this is the first report of whole genome sequences from cases of bTB in Pakistan.Identification of the etiological agents of bTB in the region may be important for understanding potential zoonotic transmission, estimating the prevalence of disease, and developing appropriate diagnostics for its detection.The results of this study may, therefore, better inform control policies for zoonotic and bovine TB in the region.

Study area
In the district of Lahore, which is the capital of Punjab, there are three public and five private sector red meat slaughterhouses.The single largest public-sector slaughterhouse in the city of Lahore (31°32′59″N 74°20′37″E) was selected for the present study to accommodate logistics and accessibility issues.The slaughterhouse is adjacent to one of the biggest live animal markets (locally called maweshi mandi) in Punjab where animal traders (locally called Beopari) routinely bring animals for sale from different neighbor ing cities of Lahore.The majority of beef contractors/traders from the slaughterhouse purchase animals from this adjacent live animal market.This slaughterhouse is the major catchment area for the animals brought for slaughter in Lahore.

Study population and design
A cross-sectional study was conducted for 5 months (November 2021-March 2022) at the selected slaughterhouse.The study population comprised 3,581 animals (441 cattle and 3,140 buffalo) slaughtered at the site within the study period.Animals slaughtered in slaughterhouses belonging to the corporate sector were excluded from the target population as they prefer to slaughter young male animals from feedlot fattening farms.In the lairage, animals were subjected to antemortem examination, and information about the species, age, and gender of the animals was recorded.All animals were declared fit for slaughter at antemortem.Beef contractors/traders from the slaughter house were also interviewed individually about the point of purchase of animals to assess the spatial distribution of slaughtered animals.Geographical information was only available for 772 animals.

Sample collection and processing
Initial postmortem inspection of carcasses resulted in the collection of specimens, mostly lymph nodes (pharyngeal, retropharyngeal, and mediastinal), lungs, and liver, presenting any kind of superficial visible lesions or abscesses.All collected samples were then examined thoroughly for the presence of granulomatous gross visible lesions attribut able to tuberculosis and declared clinically suggestive if present (9,26).Tissues with TB-like lesions were sealed in sterile pre-labeled plastic containers and were transported at refrigerated conditions to the laboratory at the Department of Epidemiology and Public Health, University of Veterinary and Animal Sciences, Lahore where these samples were stored at 4°C (9).
In the laboratory, TB-like lesions from necropsy samples from each carcass were triturated and divided into two aliquots.One aliquot was transported to the Provincial TB Reference Laboratory, Lahore for a Mycobacterial Growth Indicator Tube (MGIT) culture test using the BACTEC MGIT 960 instrument.One sample (bR3) was cultured on Lowenstein-Jensen (LJ) media when MGIT stock was low.This reference labora tory primarily handles human tuberculosis samples under the Primary and Secondary Healthcare Department, Ministry of Health Punjab, Pakistan.If the sample was culturepositive, DNA was extracted by boiling.The second aliquot was suspended in lysis buffer and placed for overnight incubation in a dry hot bath.The DNA extraction was performed using phenol-chloroform-isoamyl alcohol (PCI) and the ethanol precipitation method (Supplementary Methods), and the eluted DNA was stored at −20°C (27).

PCR analysis
A PCR that has been previously used to identify M. bovis was first performed at the University of Veterinary and Animal Sciences, Lahore.This PCR used a set of primers (JB21 and JB22) to generate a 500-bp fragment reported to be specific to M. bovis (25) (Supplementary Methods; Table S1).An annealing temperature of 55°C was used as opposed to the 68°C annealing temperature from reference (25).During PCR optimiza tion, no amplicon was generated using 68°C, so a gradient PCR was performed using a 0.5°C difference interval.This found that the product could be generated using 55°C.
DNA from all samples was also sent to the Research Institute at the McGill University Health Centre in Montreal for additional analysis of the MTBC subspecies collected.Samples were analyzed using two previously described PCR assays, which differentiate between M. tuberculosis sensu stricto, M. bovis/M.bovis BCG, and M. orygis (28).The first assay is a three-primer PCR, which detects the presence or absence of region of difference (RD) 9.A ~200 bp fragment is amplified if RD9 is present and a ~400 bp fragment is amplified if it is absent.M. tuberculosis sensu stricto has RD9 and M. bovis/M.bovis BCG and M. orygis do not.The second assay is a six-primer PCR, which detects differences in the deletion size of RD12 (Supplementary Methods; Table S1).RD12 is found in M. tuberculosis sensu stricto and absent in M. bovis/M.bovisBCG and M. orygis with the deletion size being larger in M. orygis.Using this PCR, a ~400 bp fragment is amplified if RD12 is present, a ~600 bp fragment is amplified if the M. bovis/M.bovis BCG size deletion is present, and a ~250 bp fragment is amplified if the M. orygis size deletion is present.Samples that did not produce an amplicon on either the RD9 or RD12 assay were subject to PCR of the 16S or hsp65 regions followed by Sanger sequencing to confirm their identity (Table S1) (29,30).The samples were initially tested using primers targeting the hsp65 sequence.If the sample was negative, the PCR was repeated using primers to detect the 16S sequence.

Whole-genome sequencing
Samples prepared by the PCI and ethanol precipitation method and which had been identified as an MTBC subspecies by PCR were submitted for confirmation by WGS.These samples were chosen due to the preparation method of the DNA samples.Ideally, DNA would have been prepared from the positive cultures using a method compatible for WGS, but this was not available.The Illumina S4 reagent kit was used to prepare paired-end sequencing libraries, and shotgun sequencing was performed on the Illumina NovaSeq 6000.Sequences were analyzed using the validated SNP (vSNP) pipeline (available at https://github.com/USDA-VS/vSNP)developed by the United States Department of Agriculture-Veterinary Services to align sequences to M. tuberculosis H37Rv (NC_000962.3) and assign an MTBC subspecies identity as has been previously described (28).Integrative Genomics Viewer was used to visualize sequences (31).Sequences were deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under the BioProject accession number PRJNA958208.

Phylogenetic tree assembly
Newly collected sequences were phylogenetically compared with other publicly available M. orygis sequences available on the SRA database (Table S2).A phylogenetic tree file was generated by vSNP and rooted to M. tuberculosis H37Rv (NC_000962.3).The tree was visualized and annotated using the interactive tree of life (version 6) (32).

Spatial distribution analysis
QGIS version 3.28.2Bonn (available at http://qgis.org/) was used to generate the map for the spatial distribution of slaughtered animals.

Examination of animals for gross TB-like lesions
Throughout the study period, a total of 400 carcasses were selected based on the presence of any type of lesion (granulomas, necrosis, abscess, fibrosis, and cysts) on their tissues (Fig. 2).Upon thorough examination of those 400 carcasses for lesions clinically suggestive of tuberculosis based on the presence of gross TB-like lesions or tubercles, only 34 samples were identified (Fig. 2; Table 1).Of these samples, 30 were from buffalo (28 females and 2 males) and four were from cattle carcasses (all female).TB-like lesions were observed on the lungs (n = 28), lymph nodes (n = 13), and liver (n = 3).

Culture of samples
Culture test was performed on samples from 16 of the 34 animals with TB-like lesions that were submitted to the Provincial TB Reference Laboratory.Due to a power outage, the lab was unable to perform the culture test for the remaining 18 samples.Of the 16 cultured samples, 12 of these were positive on MGIT or LJ (Table 1).

PCR results
In the initial conventional PCR screening conducted in Lahore using JB21 and JB22 primers, five were positive (Table 2; Fig. S1).Further analysis of the samples using the RD9 and RD12 assays indicated that 18 of the 34 samples were positive for MTBC subspecies: 8 of which were identified as M. tuberculosis sensu stricto and 10 of which were identified as M. orygis.No samples were identified as M. bovis.The 16 remaining samples produced no band on either the RD9 or RD12 assays.These samples were subjected to PCR of the 16S or hsp65 sequences followed by Sanger sequencing to confirm their identity: eight produced no band on either PCR indicating a lack of bacterial DNA present, six were identified by 16S PCR and Sanger sequencing as non-mycobacterial species, and two were identified as MTBC following hsp65 PCR and Sanger sequencing.The RD9 and RD12 assays were re-attempted twice on the two samples identified as MTBC by hsp65 PCR but returned negative results each time.These samples were, therefore, classified as MTBC (Table 2).The complete workflow description of how samples were selected and screened is given in Fig. S2.

Validation of JB21 and JB22 primers
Given that no M. bovis was identified by the RD9 or RD12 assays, this prompted investigation of the JB21 and JB22 primers, which have been previously reported to amplify an M. bovis-specific fragment and have repeatedly been used to identify M. bovis in studies within Pakistan (20)(21)(22)(23)(24)(33)(34)(35)(36)(37)(38)(39).A PCR was performed with the JB primers using DNA from various confirmed mycobacteria including several M. tuberculosis lineages (three to four isolates per lineage) including H37Rv, M. bovis AF2122/97, five M. bovis BCG strains, M. orygis, Mycobacterium caprae, Mycobacterium microti, Mycobacterium africanum, and eight non-tuberculosis mycobacteria.The expected 500-bp band was total animals in the study.The color of each district in map "a" shows the number of animals screened from that location.The specific number collected from each district is listed below the district name.Samples from the Lahore peripheries are listed within the Lahore district.Map "a" also identifies the locations where samples with visible granulomas came from (n = 13).If more than one sample came from that location, the number is listed on top of the green symbol.Map "b" shows the distribution of eight MTBC subspecies with location information available, which were identified in carcasses with visible granulomas: four were M. orygis, three were M. tuberculosis, and one MTBC.
produced for the M. bovis and M. bovis BCG strains.However, amplification was also observed with DNA from several M. tuberculosis strains, M. orygis, M. caprae, and M. africanum (Fig. 3A).Further analysis of the amplicon using the Basic Local Alignment Search Tool (BLAST) from NCBI and visualization via Geneious Prime determined that the JB primers amplify a 492-bp fragment of a helicase gene: Mb2049c in M. bovis AF2122/97 or RJtmp002096 in M. orygis 51145.Specifically, the primers amplify nucleoti des 2,253,648-2,254,140 in M. bovis and 2,249,132-2,249,624 in M. orygis.This sequence was not found in M. tuberculosis H37Rv but was identified in other M. tuberculosis isolates (Fig. 3B).This analysis confirmed that the JB primers are not a reliable method for distinguishing M. bovis from other MTBC subspecies.

Whole genome sequencing
A total of 10 DNA samples that were identified as MTBC subspecies by PCR and prepared using the PCI and ethanol precipitation method were submitted for whole-genome sequencing.Of these, three failed library preparation, and five had an inadequate mean depth of coverage and percent genome coverage for analysis.This was likely caused by the low concentration of DNA present in the sample and poor sample purity.The two sequences produced were from samples DR2 and DR20, which had been identi fied as M. orygis by PCR.These sequences have been deposited under the BioProject accession number PRJNA958208.WGS and analysis by the vSNP pipeline confirmed both samples were M. orygis.The two sequences were 47 single nucleotide polymorphisms (SNPs) apart, indicating they were likely independently acquired infections (40).Notably, however, the sequence qualities for both samples were still low with a mean depth of coverage of 10× for DR2 and 8× for DR20 and genome coverage of 97.87% for DR2 and 97.73% for DR20.As an added measure, the sequences were checked for the presence of two SNPs known to be specific to M. orygis: Rv0444c g698c and a C to T mutation found at nucleotide 2,850,607 (Fig. S3 and S4).Both DR2 and DR20 had these two M. orygisspecific SNPs.The two M. orygis sequences were then phylogenetically compared with 93 additional M. orygis sequences available on the SRA database, which demonstrated their relatedness to previously published strains (Fig. 3; Table S2) .M. orygis sequences included in the tree were isolated from other animals including antelope, cattle, deer, and bison, as well as humans.
Given the combined results of all PCR assays and Sanger sequencing, 20 samples with TB-like lesions were positive for an MTBC subspecies.Of those, 10 were identified as M. orygis (50%) and 8 were identified as M. tuberculosis (40%).The final two samples (10%) were classified as MTBC because they were identified as MTBC by Sanger sequencing of hsp65, but no bands were produced using the RD9 or RD12 PCRs.The MTBC samples were collected from animals brought to the slaughterhouse from five different districts, indicating that these infections occurred independently (Fig. 1).The proportion of

DISCUSSION
The World Organization for Animal Health, the World Health Organization, and the Food and Agriculture Organization have long defined M. bovis as the causal pathogen of bovine and zoonotic tuberculosis (9).Pakistan is located in South Asia, and bTB is frequently reported from various regions of the country.Previous studies from Pakistan reported M. bovis as the sole cause of zoonotic and bTB (21,41).The majority of these studies concentrated on herd testing and risk factor analysis at the animal and/or herd level (42,43).This was the first surveillance-based study at a slaughterhouse in Lahore, Pakistan, which aimed to employ PCR assays and sequencing to identify the MTBC subspecies causing bTB in the area.A total of 20 MTBC samples were identified from 34 TB-like lesions: 10 M. orygis, 8 M. tuberculosis sensu stricto, and 2 MTBC (Table 2).Two M. orygis whole genome sequences were generated from this study and a phyloge netic comparison confirmed their relatedness with other previously deposited M. orygis genomes (Fig. 4).Since our data suggested erroneous detection of M. bovis with the JB21 and JB22 primers, we analyzed a panel of confirmed MTBC isolates with these primers.Our results indicated that amplification of the 500-bp product does occur in other MTBC subspecies other than M. bovis including M. tuberculosis and M. orygis (Fig. 3A).After performing a BLAST search of the amplicon, it was determined that this identical sequence is found in a variety of MTBC genomes (Fig. 3B).Taken together, we suggest that M. orygis was not previously identified in Pakistan because the majority of studies used the JB primers and reported positive samples as M. bovis.As a result, the prevalence and identity of the MTBC subspecies contributing to disease burden remained uncertain in Pakistan.The cause of bovine and zoonotic TB should be further investigated on a larger scale using diagnostic tools that are capable of reliably differentiating between MTBC subspecies.
To our knowledge, this study is the first of its kind reporting M. orygis from animals from Pakistan and the first study from Pakistan, which has deposited whole genome sequences from bTB cases.The findings of this study are concordant with previous reports that detected M. orygis in animals and humans within South Asia.Previously, M. orygis was reported in cattle and humans from South Asia (44) and in emigrants originat ing from South Asia (45)(46)(47).In Bangladesh, M. orygis was reported in cattle and mon keys, suggesting that M. orygis may have been introduced to Bangladesh with the import of cattle from Pakistan (12).Finally, a screening of 940 cultures in India and 715 publicly available genomes from South Asia identified several M. orygis isolates; there was a  complete absence of M. bovis (28).Taken together, evidence from the literature along with the data from our study proves the endemicity of M. orygis across South Asia.
Another interesting finding from this study is the presence of M. tuberculosis sensu stricto in 8 (40%) of the 20 MTBC isolates identified in domestic bovines, highlighting the possibility for either reverse zoonosis or animal-to-animal transmission of M. tuberculosis.This requires further investigation as M. tuberculosis has not previously been reported as a prevalent cause of bTB in animals from Pakistan.However, M. tuberculosis has been reported in captive antelopes (48) and wild animals (49) in Pakistan.The isolation of M. tuberculosis from cattle has been reported in South India, and the cattle isolates were found to be phylogenetically related to isolates from farmers (50).
This study has several limitations.The first is that it relied on the presence of visible TB-like lesions or granulomas on the organ surfaces to identify samples for analysis.A postmortem examination is not a reliable method to identify all cases of bTB (51).Second, two isolates could only be identified as MTBC (Table 2) due to repeated failure of molecular assays to differentiate the samples.This was likely a result of low DNA concentrations and the quality of the samples.Another limitation was the disproportion ate distribution of buffaloes and female animals relative to cattle and male animals.At the slaughterhouse selected for this study, the proportion of buffalo (89.25%) brought in for slaughtering was higher as compared to cattle (10.72%), resulting in the segrega tion of fewer tissues with TB-like lesions from cattle carcasses.The central region of the Punjab province has a higher buffalo population density due to different cultural preferences (52), which could result in the pooling of a higher number of buffalo in the slaughterhouse.Male animals are also usually slaughtered at a few months of age or raised as sacrificial animals for a religious occasion (Eid-Ul-Adha) (53).These factors might have caused an increased number of female animals in the slaughterhouse and, therefore, more females in our study.Finally, during this study, screening was performed on animals at a single slaughterhouse over a relatively short period of time, limiting the potential generalizability of the results to the rest of the country and highlighting a need for larger geographically representative surveys of bovine tuberculosis in the country.
We conclude that bTB was present in bovines in this slaughterhouse in Lahore and was caused by M. orygis and M. tuberculosis sensu stricto.Our results are in agreement with recent reports indicating that M. orygis is the predominant cause of bTB in South Asia.Detection of M. tuberculosis in bovines at our study site suggests that reverse zoonosis may have occurred, but transmission between animals is another possibility that warrants future exploration.A high proportion of M. tuberculosis in animals may have important implications in understanding the prevalence and risk of zoonotic TB in humans.Future studies identifying bovine or zoonotic TB in Pakistan should use molecular methods capable of reliably differentiating members of the MTBC.Collectively, our findings stress the importance of a One Health approach to manage tuberculosis in Pakistan.administration, Resources, Supervision, Validation, Visualization, Writing -review and editing

FIG 1
FIG 1 Spatial distribution of 772 screened bovines in nine districts of Punjab Province, Pakistan.Geographic information was only available for 772 of the 3,581

FIG 2
FIG 2 Sample selection and characteristics.

FIG 4 A
FIG 4 A phylogenetic tree of the two new M. orygis sequences from Pakistan and 93 additional publicly available M. orygis genomes.The inner band shows the country of origin listed on the NCBI Biosample information for each sequence.The outer band shows the host species.The two sequences from this study are highlighted in green.The reference sequence M. orygis 51145 and the root of the tree M. tuberculosis H37Rv are labeled.

TABLE 1
Characteristics of bTB samples collected from November 2021 to March 2022 from the slaughterhouse in Lahore, Pakistan a NA, location not available.b Gender of sampled animals: F, female and M, male.c Types of tissues collected: LNs, lymph nodes.d Culture results: P, positive by MGIT; P*, positive by LJ; N, negative; and NA, culture results not available.

TABLE 2
PCR identification of bTB samples collected from November 2021 to March 2022 from the slaughterhouse in Lahore, Pakistan bR, culture positive and DNA preparation was done using the boiling method and DR, DNA extracted directly from tissue lesions using PCI and ethanol precipitation method.
a b N, negative results for PCR with JB21 and JB22 primers.c N, negative for RD9 PCR.d N, negative for RD12 PCR.e N, negative for Sanger sequencing.