The isolation and molecular characterization of Mycobacterium avium subsp. paratuberculosis in Shandong province, China

1038 individual bovine serum samples were collected between October 2014 and August 2015 from 19 different dairy farms located in nine distinct geographic regions of Shandong province. The herds were selected at random and varied between 300–800 animals. Individual cows showing diarrhoea were selected in addition to some animals showing no clinical signs. Approximately 10 % of the animals were selected all of them being adult cows over 24 months of age. Milk, faeces and tissues were sampled from the herd showing the highest frequency of positive sera. All the faecal samples were obtained via per-rectal drags, and were kept at 4 °C up to 48 h prior to processing for culture and DNA extraction. Milk samples were collected in 50-ml tubes after four initial strippings were discarded. Ileum and mesenteric lymph nodes were collected from two euthanized cows with positive antibody for Map. All protocols and procedures were performed according to the Chinese Regulations of Laboratory Animals—The Guidelines for the Care of Laboratory Animals (Ministry of Science and Technology of People’s Republic of China) and Laboratory Animal Requirements of Environment and Housing Facilities (GB 14925–2010, National Laboratory Animal Standardization Technical Committee). The license number associated with their research protocol was 20110611–01 and the animal study proposal was approved by The Laboratory Animal Ethical Committee of China Agricultural University.

In order to identify the herd with highest frequency of positive sera the 1038 sera were tested for the presence of antibodies to Map using a commercial antibody test kit (IDEXX). Sera with S/P ratios ≤0.45 and ≥0.55 were considered negative and positive, respectively. Intermediate S/P values were considered “suspect” and regarded as negative for the data analysis.

10 mg feces without any treatment was taken to spread evenly on to the surface of a clear glass slide with 20 ml water drop. For milk samples, there was a little difference. Centrifugation was needed to get more bacterium. After centrifugation at a speed of 12000 rpm for 5 min, the bottom precipitate was collected as the samples for smear, and the same procedure was used as that of faecal samples. Tissues were used as frozen sections. Slides were stained using the Ziehl–Neelsen method and examined by light microscopy for the presence of acid-fast bacilli. Map presence was confirmed by PCR amplification of IS900 as described previously [33].

A modification of the double centrifugation methods was used to cultivate Map [34]. Briefly, two grams of faeces were added to 35 ml of sterile distilled water. Samples tubes were vortexed for 1 min and then allowed to stand undisturbed for 45 min. Five ml of the supernatant were added to 25 ml of 0.9 % hexadecylpyridinium chloride (HPC) and then vortexed for 2 min and allowed to stand undisturbed for 12 h at room temperature. Tubes were centrifuged at 1500g for 30 min, the supernatant was discarded and the pellet was resuspended in 1 ml of antibiotic mixture (50 µg/ml amphotericin B, and 100 µg/ml nalidixic acid, and 100 µg/ml vancomycin). This was mixed well and incubated for 12 h at 37 °C. Herrold’s Egg Yolk Agar with mycobactin J and ANV (BD, USA) were inoculated with 0.2 ml of the suspension and incubated at 37 °C for 6 weeks.

Colonies that appeared within 1 week were discarded and slower-growing colonies with typical colonial morphology were checked for acid fast bacteria, and then detected by PCR based on IS900 as mentioned above. In order to generate pure and sufficient biomass, a single colony of each confirmed Map isolate was subcultured into 7H9 broth supplemented with OADC and Mycobactin J and incubated at 37 °C. The culture was reconfirmed as Map by PCR and the purity checked by light microscopy with Ziehl–Neelsen method staining. The strain type was identified by IS1311 PCR and restriction endonuclease analysis (REA) using the primers and method described previously [33]. The identification as M. avium subsp. paratuberculosis was confirmed by PCR using IS900 and REA using AlwI digestion. The IS1311 PCR-REA was used to distinguish C strains from S strains using HinfI and MseI (details shown in Table 1). PCR primers are detailed in Table 2.

Massive parallel sequencing using the Solexa-technology (Illumina) was performed at the Chinese Academy of Sciences. Two isolates were sequenced using the Illumina HiSeq 2500 platform using one flow cell lane with 125-cycle paired end chemistry. Short reads were assembled using SOAPdenovo (http://​www.​soap.​genomics.​org.​cn), a genome assembler developed specifically for next-generation short-read sequences [35]. The SOAP GapCloser was also used to close gaps. Reads were mapped by using BWA [36], and SNPs and indels were called and filtered by SAM tools [37]. A rooted tree was computed using MEGA. Phylogenetic tree was calculated using the maximum likelihood method by whole genome alignment and performing 1000 bootstraps replicates. Several references are selected for comparation, include type C, type B, and type S (Table 3).

Statistical comparisons between milk detection and fecal detection were determined using the Chi square test with p < 0.05 denoting statistical significance. SPSS software (Chicago, IL, USA) was used.

Screening of 1038 serum samples from 19 dairy farms by the Mycobacterium paratuberculosis antibody test kit detected 121 (11.7 %) animals positive for Map antibody, and 11 (57.9 %) dairy farms were positive for Map antibody. Of 11 dairy farms with signs of diarrhoea, 7 (63.6 %) were positive and of 8 dairy farms with no signs of diarrhea, 3 (37.5 %) were positive. The highest positive rate within these herds was 78.9 % and this herd was selected for isolation and identification of Map (Table 4).

Nineteen faecal and milk samples were sampled for Ziehl–Neelsen staining to detect individual cattle positive for acid fast bacilli. Of these 13 (68.4 %), 6 (31.6 %) faecal samples and milk samples respectively were positive for acid-fast bacilli (Fig. 1a, b). We euthanized two cattle with diarrhoea, emaciation and reduced milk production. Postmortem examination showed gross lesions in the mesenteric lymph nodes showing 5–6 fold enlargement compared with normal mesenteric lymph nodes (date not shown). However, there were no obvious gross lesions in the gastro-intestinal tract. Histopathological examination of mesenteric lymph nodes showed the characteristic diffuse granulomas with epithelioid cells, multinucleated giant cells and macrophages (Fig. 2a, b) compared with normal mesenteric lymph nodes(Fig. 2c, d). However, no granulomas were observed in the ileum (date not shown). Ziehl–Neelsen stained impression smears of the mesenteric lymph nodes showed large numbers of acid-fast bacilli (Fig. 1c). Large numbers of clustered acid-fast bacilli were observed in the smears taken from a purified colony (Fig. 1d).

The PCR products are shown Fig. 3. Two isolates were both identified as positive using the IS900 PCR. We obtained the target fragment, whose size was 413 bp (Fig. 3a) which was digested using restriction endonuclease AlwI which resulted in two bands, one at 147 bp, and other at 266 bp (Fig. 3b). For further identification of strain type, IS1311 PCR was conducted and followed by digestion by two restriction endonucleases (HinfI/MseI). The results showed that both isolates were identified as positive with the target band of 608 bp (Fig. 3c). Restriction endonuclease analysis of the IS1311 PCR reaction products confirmed the presence of bands of 67, 218, 285 and 323 bp (Fig. 3d), which indicated that both isolates were type C.

Although we grouped the two isolates, it was not enough to deeper understand the genomic differences between our isolates and reference strains. Here Next-generation sequencing technology (Illumina, HiSeq 2500) was used to sequence the whole genome of the two isolates from cattles in China. As a result, we successfully got two draft sequences named 2015WD-1 and 2015WD-2, the accession numbers are LKUS00000000 and LKUT00000000, respectively. The results showed that the genome size of 2015WD-1 and 2015WD-2 are respectively 475,0273 and 472,7050 bp with a same G + C content of 69.3 %. The numbers of single nucleotide polymorphisms (SNPs) in our Map isolates against Map K-10 are respectively 292 and 296. Also, 84 and 96 indels are found in our isolates against Map K-10. More information between our isolates and reference strains are presented in Table 5. In the phylogenetic tree (Fig. 4) based on the nucleotide sequences, we can see our two isolates are most closely relate to type C representative strains Map K-10 and Map ATCC19698, and farthest from type S representive strains Map JQ5 and Map s397.