A total of 60 tissue samples from 57 individual African buffaloes from the Hluhluwe-iMfolozi Park (HiP), South Africa (SA), were selected in the present study and included pooled head lymph nodes (n = 5), pooled thoracic lymph nodes (n = 7), lung tissue (n = 7), tonsil (n = 5), tracheobronchial lymp nodes (n = 12), mediastinal lymph nodes (n = 8), parotid (n = 4), retropharyngeal lymph nodes (n = 6), prescapular lymph nodes (n = 3), two mandibular lymph nodes, and one subiliac lymph node. All buffalo samples had previously been confirmed as infected using the aforementioned culture methods, and the presence of the M. bovis region of difference 4 (RD4) signature was confirmed from culture-positive crude DNA extracts22. Spacer oligonucleotide typing hybridization assay (spoligotyping) revealed that all samples fit one of two profiles, SB0130 (n=39) and SB1474 (n=21) (Suppl. Table 1).
ONT targeted amplicon sequencing
All samples showing amplification of >1 target gene were included for amplicon sequencing, whereas four samples (2 tonsils, one lung, and one parotid) for which none of the targets could be visualized after electrophoresis were excluded from downstream analysis. The hsp65, rpoB, and 16S rRNA gene regions were successfully amplified from 55 samples (Figure 1). One pooled head lymph nodes sample did not amplify the rpoB target. The number of reads per specimen with Q12 or greater ranged from 63,705 to 897,809 (Mean (M) = 320,985 reads, Standard Deviation (SD) = 154,129), and the total number of bases sequenced ranged from 38,921,972 to 554,606,081 (M = 199,128,119 bases, SD = 95,521,803). The N50 read length (median read length of the longest contigs) varied from 558 to 781 bases. The mean read length for the samples ranged from 576 to 658 bases. The mean read quality, as indicated by the Phred score, ranged between 14.2 and 14.5. A total of 200,000 reads with >Q12 were randomly selected from each sample for downstream analysis (Figure 2).
Since amplification of the three main targets occurred in a single tube, PCR products amplified were not normalized, and therefore shorter sequences were predominant. The distribution of reads for each target varied, with hsp65 ranging from 31-79%, rpoB 2-30%, and 16S rRNA 1-8% of the selected reads (Table 1, Figure 3a).
Table 1. Distribution of read counts for the seven African buffalo tissue samples (identification number in first row) where the presence of non-tuberculous mycobacteria or heterogeneous mycobacterial communities was identified. Reference-free sorting and assembly of consensus sequences were performed on a maximum of 200,000 randomly selected reads (>Q12) for each sample. For tissue samples with fewer reads, the entire available dataset was included. Reads were generated using Oxford Nanopore Technologies (ONT) on targeted amplified housekeeping genes shown in the first column. Reads that could not be assigned to any of the three targets were grouped as unclassified and excluded from downstream analysis. NB17057 parotid; NB17074 and KS19028 pooled head lymph nodes; NB17050 retropharyngeal lymph node; NB18206 subiliac lymph node; NB18261 mediastinal lymph node; NB18258 lung.
|
NB17057
|
NB17074
|
NB17050
|
NB18206
|
NB18261
|
NB18258
|
KS19028
|
hsp65
|
38771
|
153928
|
101257
|
90572
|
119897
|
62228
|
159430
|
rpoB
|
1644
|
NA#
|
49733
|
49978
|
56224
|
53955
|
10135
|
16S rRNA
|
4791
|
9471
|
2876
|
3010
|
4725
|
2186
|
9757
|
unclassified
|
18499
|
36601
|
46134
|
33795
|
19154
|
81631
|
20678
|
Included for analysis
|
63705
|
200000
|
200000
|
177355
|
200000
|
200000
|
200000
|
Total number reads
|
63705
|
204856
|
230365
|
177355
|
351470
|
271981
|
249582
|
# For sample NB17074 no amplification was observed for rpoB.
Mycobacteriome composition
Evidence of Mycobacterium spp. DNA presence was detected in 93.3% (56/60) of the DNA samples extracted directly from tissue homogenates. Among the 56 samples subjected to amplicon generation and subsequent deep sequencing for the three targets, the rpoB PCR exhibited the highest sensitivity for MTBC detection (98.2%), followed by hsp65 (94.6%), and 16S rRNA (92.8%). The shorter target PCR (hsp65) demonstrated the highest efficacy in detecting NTM and heterogeneous mycobacterial populations compared to alternative targets (Table 2), as well as closely related microorganisms such as Streptomyces sp. (Figure 3b).
Table 2. Synopsis of the reference-free sorting and assembly outcomes of consensus sequences for the three Mycobacterium spp. target genes (hsp65, rpoB, 16S rRNA) included in the culture-independent next-generation sequencing approach. The composition of the mycobacteriome varied depending on the three independent targets. Notably, the shorter target PCR (hsp65) demonstrated the ability to identify the highest proportion of mycobacteria, with amplification and Mycobacterium spp. sequences detected in 93.3% of the samples tested (56/60). The rpoB PCR displayed the highest sensitivity for MTBC detection, identifying MTBC in 91.7% of all samples included (55/60).
|
hsp65
|
rpoB
|
16S rRNA
|
Mycobacterium tuberculosis complex (MTBC)
|
49
|
54
|
52
|
Non-tuberculous mycobacteria (NTM)
|
3
|
0
|
2
|
MTBC + NTM
|
4
|
1
|
0
|
Total
|
56
|
55#
|
54§
|
# One sample could not be amplified using the rpoB target
§ For two samples (NB18154 and KS19028), the 16S rRNA consensus sequences were classified as organisms not belonging to the Mycobacterium genus. Further analysis using NCBI Basic Local Alignment Search Tool (BLAST) identified the presence of Niallia sp. in NB18154 and Streptomyces sp. in KS19028.
Based on previous molecular analyses, which involved RD-PCRs and spoligotyping used to confirm the presence of M. bovis within the selected samples, our goal was to assess the ability of the novel culture-independent methodology to detect and speciate MTBC members. Using a subset of eleven tissue DNA samples, the gyrA and gyrB genes were amplified and sequenced on a single Flongle flow cell (ONT). The number of reads with Q12 or greater ranged from 118 to 1,918 (M = 1,157 reads, SD = 489), and the total number of bases sequenced ranged from 20,411 to 328,224 (M = 200,948 bases, SD = 85,764). The mean read quality, as indicated by the Phred score, was consistently above 13 (11/11). All reads with >Q12 quality scores were selected from each sample for downstream analysis.
The three additional targets employed for confirming the presence of M. bovis DNA (gyrA, gyrB1, and gyrB2) were successfully amplified in all 11 selected samples. Specifically, the gyrA gene target amplicon sequences confirmed the presence of MTBC DNA in approximately three-quarters of the samples (8 out of 11), while gyrB1 and gyrB2 confirmed the RD and spoligotyping speciation results obtained from culture-derived DNA in all samples. For two samples (NB17057 and KS19028), however, the total number of reads (656 and 118) and the number of consensus sequences identified as MTBC in one or more targets (92 and 43) were lower compared to the corresponding mean values generated for the other samples (total number of reads 10,149).
In addition to the detection of MTBC DNA, tNGS of the hsp65 gene successfully identified diverse mycobacterial species in different tissue specimens, including two pooled head lymph nodes, one lung, one mediastinal lymph node, one parotid, one retropharyngeal lymph node, and one subiliac lymph node. Mycobacterium avium complex (MAC) was detected in four samples (7.1%), M. smegmatis in two (3.6%), M. komaniense and an unclassified Mycobacterium sp. in one sample each (1.8%). Additionally, DNA of mycobacteria’s closely related organisms such as Streptomyces sp. were amplified and further identified using the hsp65 target gene (Figure 3b). In the DNA sample NB17074, originating from lung tissue, M. smegmatis was identified using hsp65 and 16S rRNA targets, but no amplification was observed using rpoB PCR. Conversely, in one retropharyngeal lymph node (sample NB17057), M. komaniense was detected using hsp65 and 16S rRNA PCRs, while MTBC DNA was detected using rpoB.
At least two distinct strains of MAC were detected, exhibiting 15 genomic modifications, including single nucleotide variations and indels. These strains displayed the highest hsp65 sequence homology to M. avium and M. colombiense. The MAC sequences were exclusively detected in tissue samples exhibiting heterogeneous bacterial populations (4/4), including MTBC in three samples, and a combination of Streptomyces sp. and Mycobacterial sp. in one sample (Table 3). The samples originated from various locations including lung tissue, mediastinal lymph nodes, head lymph nodes, and subiliac lymph nodes. Finally, the relative abundance of reads classified as MAC varied between 0.3 and 17% for the hsp65 target in the above-mentioned samples (Figure 3b).
Table 3. Percentage of hsp65 PCR target reads (%) indicating the presence of non-tuberculous mycobacteria or heterogeneous mycobacterial communities in each African buffalo tissue sample where the DNA of these organisms was detected. NB17057 parotid; NB17074 and KS19028 pooled head lymph nodes; NB17050 retropharyngeal lymph node; NB18206 subiliac lymph node; NB18261 mediastinal lymph node; NB18258 lung.
|
NB17057
|
NB17074
|
NB17050
|
NB18206
|
NB18261
|
NB18258
|
KS19028
|
M. komaniense
|
100
|
0
|
0
|
0
|
0
|
0
|
0
|
M. smegmatis
|
0
|
100
|
0.86
|
0
|
0
|
0
|
0
|
MTBC
|
0
|
0
|
99.14
|
99.68
|
95.7
|
83.01
|
0
|
MAC
|
0
|
0
|
0
|
0.32
|
4.3
|
16.99
|
10.9
|
Mycobacterium sp.
|
0
|
0
|
0
|
0
|
0
|
0
|
6.13
|
Streptomyces sp.
|
0
|
0
|
0
|
0
|
0
|
0
|
82.97
|