The insertion sequence IS900 has been considered specific for Mycobacterium avium subsp. paratuberculosis (M. paratuberculosis) and has, therefore, been used as the target gene for diagnostic PCR of M. paratuberculosis. From a healthy dairy cow we have isolated and characterised a mycobacterium harbouring one copy of a sequence with 94% identity to IS900 at the nucleic acid level. The isolate was shown to be related to Mycobacterium cookii, as assessed by 16S rRNA sequencing. Strong amplifications were obtained with several PCR primers described for detection of IS900. This finding shows the need of alternative PCR systems based on other genes than IS900 to confirm the presence of M. paratuberculosis.
Mycobacterium avium subsp. paratuberculosis (M. paratuberculosis) is the causative agent of paratuberculosis, also known as Johne's disease, a severe chronic intestinal infection in cattle and other ruminants . M. paratuberculosis is also implicated as a possible cause of Crohn's disease, a gastrointestinal disease in humans . The standard diagnostic method for M. paratuberculosis is culture of the organism from tissue or faecal specimens. Identification of M. paratuberculosis is based on the slow growth rate and the dependence of exogenous mycobactin.
PCR methods based on IS900, an insertion sequence considered specific for M. paratuberculosis, have been developed and used to increase the specificity and the sensitivity of the laboratory diagnostics and to decrease the time needed for detection [3–6]. The insertion sequence IS900 is a 1451-bp segment that lacks inverted terminal repeats and does not generate direct repeats in target DNA . It belongs to the same family of insertion sequences as IS901, IS902, and IS1110, which have been described in M. avium subsp. avium, M. avium subsp. silvaticum, and M. avium subsp. avium, respectively.
PCR based on IS900 has been used to confirm the presence of M. paratuberculosis in milk, faecal specimens, and human intestinal tissue without primary culture or in cases when M. paratuberculosis has not been possible to cultivate or isolate . The IS900 PCR is also routinely used in control programmes for Johne's disease as the sole method to confirm the presence of M. paratuberculosis.
A definite method for identification of mycobacteria is sequence analysis of 16S rRNA or its gene and the phylogeny of mycobacteria based on this method has been extensively studied . Sequence analysis of the 16S rRNA gene has been particularly useful for identification of mycobacteria because of the difficulties with the conventional typing methods for these bacteria .
We have isolated a mycobacterium positive for IS900 by PCR, which by further investigation proved not to be M. paratuberculosis. The phenotypic features of the isolated organism were investigated and partial sequencing of the IS900-like fragment was performed. The nucleotide sequence of the 16S rRNA gene was determined and the phylogenetic relationship to other mycobacteria was established. Amplification with primers targeting other mycobacterial genes than IS900 was investigated as complementary tests to confirm or exclude the presence of M. paratuberculosis.
2Materials and methods
2.1Isolation and identification
A single colony, suspected to be M. paratuberculosis, was detected from a faecal culture of a healthy dairy cow in the Swedish Paratuberculosis Control programme. The isolate (strain 2333), detected on modified Löwenstein–Jensen medium with mycobactin, was picked for microscopic examination of acid-fast rods after Ziehl–Neelsen staining, PCR for IS900 with primers p36/p11  and subculture on media with and without mycobactin.
2.2Strains, growth conditions and extraction of genomic DNA
M. paratuberculosis, strain ATCC19698, was cultured on modified Löwenstein–Jensen medium with mycobactin at 37°C for 8 weeks. Strain 2333 was cultured on modified Löwenstein–Jensen medium at 37°C for 7 weeks and Mycobacterium cookii ATCC49103 at 30°C for 6 weeks. Genomic DNA was purified from the cultured mycobacteria as described previously .
Growth of strain 2333 and M. cookii was studied on modified Löwenstein–Jensen medium at 30°C, 37°C and 45°C for up to 10 weeks. Pigmentation with and without light exposure was studied after 2 and 4 weeks incubation of the cultures. Well-developed colonies of strain 2333 were subjected to the Accuprobe ‘M. avium complex culture identification test’ (M. avium subspp.- and Mycobacterium intracellulare-specific, GenProbe, San Diego, CA, USA).
2.4Analysis of strain 2333 by fragment length polymorphism analysis (RFLP)
Strain 2333 and strain ATCC19698 were subjected to restriction RFLP. Preparation of genomic DNA and digestion by restriction endonucleases BstEII and PstI were performed according to Pavlı́k et al. . The resulting restriction fragments were separated in 0.8% w/v agarose by pulsed field electrophoresis (Chef-DR?, Bio-Rad, Hercules, CA, USA) at 5.3 V cm−1 with a linear ramping from 0.3 s to 10.0 s for 10 h, and transferred to Hybond-N+ membrane by Southern blotting. The labelling and detection system Gene Images™ Alk Phos Direct™ (Amersham Pharmacia Biotech, Uppsala, Sweden) was used in accordance with the instructions of the manufacturer. The 413-bp PCR product obtained by amplification of M. paratuberculosis DNA with primers p90 and p91 (Table 1) was labelled with alkaline phosphatase and used as a probe. Overnight hybridisation and primary washing were performed at 55°C.
Table 1. PCR primers used in this study
aThe sequence for primer sva-001 was as follows: 5′-RSPb-ACC TTG TTA CGA CTT CGT CCC AAT C-3′.
bRSP is reversed universal sequencing handle with the following sequence: 5′-CAC AGG AAA CAG CTA TGA CC-3′.
Strain 2333 was investigated by PCR with primers specific for IS900[3–6,12]. PCR was also performed with primers specific for the f57 gene and the p34 gene , and with the genus-specific primers 17A/17B . Amplifications were performed in 50 μl and were undertaken in a PTC-200 Thermo Cycler (MJ Research, Waltham, MA, USA). The PCR mixture comprised 67 mM Tris–HCl (pH 8.8), 2 mM MgCl2, 0.1 mM of each of the four dNTPs, 10 pmol of each oligonucleotide primer, 0.5 U of Taq DNA polymerase (Applied Biosystems), and 5 ng of DNA. Amplification with primers p90/p91, 150/921, s204/s749, and s347/s535 consisted of an initial denaturation at 94°C for 3 min followed by 35 cycles of denaturation at 94°C for 1 min, primer annealing at 63°C for 1 min, and extension at 72°C for 1 min. Amplification with primers p36/p11, MP3/MP4, and Mav17A/Mav17B was performed as mentioned above but with an annealing temperature of 55°C. The PCR products were analysed by electrophoresis in 2% agarose gels in 1×TBE buffer. As a marker BglI and HinfI cleaved pBR328 (Boehringer Mannheim, Germany) was used. The gels were stained for 20 min in ethidium bromide (1.5 μg ml−1), rinsed in distilled water and visualised by UV light transillumination.
Primer p90 and MP4 (Table 1) were used to amplify a 1043-bp region of the IS900-like sequence of strain 2333 under the PCR conditions mentioned for primers p90/p91 above. The 16S rRNA gene was amplified with primers 593 (F) and sva-001 (R) complementary to the universal region U1 and U8, respectively (Table 1). The resulting PCR products were sequenced as described below. The IS900-like sequence and the 16S rRNA sequence of strain 2333 have been deposited in the GenBank under accession No. AF455252 and AY065649, respectively.
2.6DNA sequence analysis
Double-stranded PCR products were sequenced using the Thermo sequenase fluorescently labelled primer cycle sequencing kit with 7-deaza-dGTP according to the protocol of the manufacturer (Amersham Pharmacia Biotech, Piscataway, NY, USA). The samples were loaded on ReadyMix gel on an ALFexpress DNA sequencer (Amersham Pharmacia Biotech) and subjected to electrophoretic separation, on-line detection, and computerised sequence evaluation. Fluorescently labelled IS900 primers were used for sequencing both strands of the p90/MP4 PCR product of strain 2333 (Table 1). Sequencing primers, originally designed for mycoplasma sequences, were used  to determine the 16S rRNA sequence of strain 2333.
Pre-aligned mycobacterial 16S rRNA sequences were retrieved from the Ribosomal Database Project (RDP, URL: http://rdp.cme.msu.edu/html/). A few sequences (Mycobacterium spp., strains IMVS B76676 and IWGMT 90236, and Mycobacterium sydneyiensis) retrieved from GenBank and the sequence of Mycobacterium sp. strain 2333 were then manually aligned with the RDP sequences. The final alignment comprised 1350 nucleotide positions. Phylogenetic calculations were performed using the neighbour-joining method, implemented in the phylogenetic program package PHYLIP .
3Results and discussion
3.1Isolation of strain 2333
The primary growth of the isolate appeared as a yellow colony on one slope of mycobactin-containing Löwenstein–Jensen medium on the final reading at 4 months. The colony analysed by PCR with primers p36/p11 was IS900 positive and the isolate was, therefore, first identified as M. paratuberculosis. After Ziehl–Neelsen staining of the colony, acid-fast rods could be seen. Most rods were comparatively long and slender and some were slightly bent. This appearance is not typical for M. paratuberculosis. At subculture, growth appeared after 4 weeks and the strain grew equally well on media with and without mycobactin. These findings indicated the isolate not to be M. paratuberculosis. Further, the Accuprobe test for the M. avium complex was negative, providing another indication that the isolate was not M. paratuberculosis.
On subculture, strain 2333 had a strong yellow to orange pigmentation that was scotochromogenic. Growth appeared more vigorous at 37°C than at 30°C. No growth was observed at 45°C. Visible colonies appeared after approximately 3 weeks. Strain 2333 differed from M. paratuberculosis in growth rate, pigmentation of colonies, and microscopic morphology. It was similar to M. cookii, except for temperature optimum, where strain 2333 grew best at 37°C while M. cookii grew best at 30°C.
3.3Characterisation of the IS900-like element
Amplification of genomic DNA of strain 2333 with the IS900-specific primers p90/p91, p36/p11, 150/921, s204/s749, s347/s535, and MP3/MP4 yielded one single PCR product, with the respective primer pairs (Fig. 1 and Table 1). The resulting amplicons were of the same sizes as the corresponding PCR products from amplified M. paratuberculosis DNA. The result indicated strain 2333 to harbour the IS900 sequence and would, therefore, if only based on the PCR results, be judged as a M. paratuberculosis isolate. Amplification with primers for the f57 sequence, which to our knowledge is specific for M. paratuberculosis, did not yield any amplicons with DNA from strain 2333 but produced a 439-bp PCR product with M. paratuberculosis DNA (Fig. 2). Likewise, amplification with primers myc3 and myc1, targeting the us-p34 sequence in the M. tuberculosis complex and the M. avium complex , yielded a 257-bp product with M. paratuberculosis control DNA but not with DNA from strain 2333 (Fig. 2). The results of PCR based on the f57 sequence and the us-p34 sequence showed strain 2333 not to be M. paratuberculosis, any subspecies of M. avium, or M. intracellulare. With the genus-specific primers 17A/17B  a weak banding pattern was obtained for strain 2333 whereas one distinct band was observed for M. paratuberculosis (Fig. 2). The result showed genetic differences between the strains and gave further evidence for strain 2333 not to be M. paratuberculosis.
The number of IS900-like fragments in strain 2333 was determined by digestion with BstEII and PstI followed by Southern blotting with an IS900-specific probe. One single band was obtained with BstEII and PstI, respectively, confirming the presence of a single IS900-like fragment. The control DNA of M. paratuberculosis showed the banding pattern described by Pavlı́k et al. , verifying the specificity of the probe and the stringency of the hybridisation conditions used.
Sequence data of the amplified IS900-like sequence were generated from the nucleotide positions corresponding to 28–1071 of the IS900 nucleotide sequence (GenBank accession No. X16293). There were 58 nucleotide differences between the IS900 sequences of M. paratuberculosis and the IS900-like sequence of strain 2333, giving an identity at the nucleic acid level of 94.4%. Within the first 450 bases, where most PCR primers are positioned, only six bases differed from IS900. One mismatch was found in the sequence for primer p91 and MP3, respectively, and two mismatches were found in the s749 primer sequence. None of these mismatches prevented amplification.
In a previous study of IS900-like sequences in mycobacteria , the resulting IS900 PCR products were subjected to restriction endonuclease analysis to identify false positive results. The restriction enzyme HaeIII was used to examine the PCR product obtained by the truncated version of primers 150/921, and AlwI and MseI were used to assess the PCR product generated by primers p90/p91. DNA of strain 2333 amplified with the original primers 150/921 and p90/p91 yielded PCR products where the restriction sites for HaeIII, AlwI, and MseI were identical to the restrictions sites in amplified DNA of M. paratuberculosis. For that reason, restriction endonuclease analysis could not be used to solve the problem with false positives. Sequence analysis of the PCR product remains the definite method to confirm the identity of the IS900 sequence.
A preliminary phylogenetic analysis of Mycobacterium sp. strain 2333 and 70 representative mycobacterial species and strains retrieved from RDP showed in which phylogenetic cluster this strain grouped. A number of species were selected from this group together with a few species and strains selected from the list obtained by similarity searches with the 16S rRNA sequence of strain 2333 in GenBank. The final phylogenetic analysis was performed with these species and other suitable representatives (Fig. 3). It was found that strain 2333 grouped with M. cookii and Mycobacterium sp. strain IMVS B76676. This cluster formed a sister lineage to another cluster composed of Mycobacterium branderi and Mycobacterium celatum. The similarity values for strain 2333 to M. cookii and Mycobacterium sp. strain IMVS B76676 were 98.3% and 98.8%, respectively and to M. branderi and M. celatum 97.2% and 97.8.0%, respectively. The cluster comprising these two sister lineages will be referred to as the M. cookii cluster. The branching order within the M. cookii cluster was not affected by removal of gaps and ambiguously aligned positions. The similarity value to M. paratuberculosis was only 96.4%.
A long or extended helix 18 is characteristic for slow-growing mycobacteria and strain 2333 possessed an extended helix 18. The sequence of the hypervariable region B (nucleotide positions 433–503 according to Escherichia coli numbering), which comprises helix 18, is very informative for mycobacteria  and it was found to be unique for strain 2333. Hypervariable region A (positions 129–266), which comprises helix 10, was also found to be unique to strain 2333. In total, there were 51 nucleotide differences in the 16S rRNA genes of M. paratuberculosis and strain 2333.
In a recent study  it was shown that a number of field isolates of mycobacteria possess IS900-like sequences. These isolates had significant 16S rRNA sequence similarities with M. intracellulare, Mycobacterium scrofulaceum, and Mycobacterium sp. strain IWGMT 90236. As can be seen from Fig. 3, strain 2333 is not very closely related to these taxa and the sequence similarity was 96.9% to all of them. Thus, phylogenetic data clearly show that strain 2333 should not be classified as M. paratuberculosis and this strain probably represents a new mycobacterial species.
The finding of an insertion sequence very similar to IS900 in a Mycobacterium sp. unrelated to M. paratuberculosis has consequences for the reliability of the PCR methods used for detection of M. paratuberculosis. The results of clinical specimens tested positive for IS900 by direct PCR should be confirmed either by isolation of M. paratuberculosis or by a PCR method targeting another gene in M. paratuberculosis.
We thank Anna-Lena Andersson for technical assistance and Malin Heldtander Königsson for valuable scientific discussions. This study was financially supported by grants from the Swedish Council for Forestry and Agricultural Research.