Development of a cost-effective line probe assay for rapid detection and differentiation of Mycobacterium species

Line probe assay (LPA) is one of the most accurate diagnostic tools for detection of different Mycobacterium species. There are several commercial kits based on line probe assay for detection of Mycobacterium species. Because of their high cost, especially for underdeveloped and developing countries, and the discrepancy of NTM prevalence across geographic regions, it would be reasonable to consider the development of an in-house LPA. The aim of this study was to develop a LPA to detect and differentiate mycobacterial species, and also to evaluate the usefulness of PCR-LPA for direct application on clinical samples. One pair of biotinylated primers and fifteen designed DNA oligonucleotide probes were used based on multiple aligned ITS sequences. Specific binding of the PCR amplified products to the probes immobilized on a strip of nitrocellulose membrane was evaluated by hybridization method. Experiments were performed three times on separate days to evaluate the repeatability of the assay. Further, evaluation of the PCR-LPA was carried out directly on 9 clinical samples and cultivated isolates. All the fifteen probes used in this study were hybridized specifically to ITS sequences of the corresponding standard species. Results were reproducible for all of the strains in different days. Mycobacterium species of nine clinical specimens and their relevant cultivated isolates were correctly identified by PCR-LPA and confirmed by sequencing. In this study, we described a PCR-LPA which is readily applicable in clinical laboratory. This assay is a fast, cost-effective and highly specific method which requires no radioactive materials.


Introduction
Mycobacterium tuberculosis, the causative agent of 'tuberculosis', is an ancient human pathogen, which has re-emerged as a major health problem worldwide. According to the WHO reports, about one-third of world's population has been infected with M. tuberculosis. In 2017, 1.3 million people died from tuberculosis (TB) and about 54 million deaths recorded during 2000-2017 [1]. This disease is a life-threatening infection and can easily be transmitted through respiratory droplets. Therefore, rapid and accurate diagnosis of TB can result in the appropriate therapy and preventing its spread within the community.
In recent years, the prevalence of non-tuberculous mycobacteria (NTM) infections has increased significantly. Some infections of NTM, especially M. chelonae, M. avium complex (MAC) and M.
fortuitum, are rapidly increasing worldwide [2]. Unlike M. tuberculosis, NTMs are organisms that are often isolated from natural resources such as water and soil. Therefore, contamination with these organisms is almost inevitable especially in elderly people or immunodeficiency population [3].
Although, many of NTM cause the same pulmonary disease as TB, however the type of drugs, their doses and duration of treatment are different from those of TB [4]. For example, the presence of erm genes (erm 38, erm 39, and erm 41) in M. abscessus group, M. fortuitum and M. smegmatis induces resistance to macrolides and may lead to treatment failure [5][6][7]. Therefore, rapid identification and differentiation of Mycobacterium species are very necessary to ensure effective therapy of the disease. Traditional strategies based on culture methods and phenotypic characteristics like colonial morphology, growth rate, producing pigment and biochemical characteristics are very time-consuming, difficult to interpret and unreliable [8]. Several molecular techniques like PCR (polymerase chain reaction), RFLP (restriction fragment length polymorphism) and real-time PCR methods have also been developed as rapid tests for detection of Mycobacterium species. However, none of these techniques can simultaneously differentiate several species of Mycobacterium in the clinical laboratory. Therefore, a high discriminatory method is still needed for laboratory diagnosis and epidemiological studies. Line probe assay (LPA) is a diagnostic tool based on the detection of a specific RNA or DNA sequence by hybridizing the complementary strand of a nucleotide probe to a part of the target sequence [9]. This method has considerably improved the accuracy and rapidity of Mycobacterium spp. differentiation [10].
Among these, ITS target has been found to have an outstanding potential for this purpose, because of its variations in size and sequence from species to species and containing both variable and highly conserved regions. Previous studies have also indicated that ITS has a good reproducibility and high level of variation in mycobacteria [12,20].
There are several commercial kits based on line probe assay for detection of Mycobacterium species.
Because of their high cost, especially for underdeveloped and developing countries, and the discrepancy of NTM prevalence across geographic regions, it would be reasonable to consider the development of an in-house LPA for rapid identification and differentiation of Mycobacterium species. Therefore, the main aim of the present study was to design an in-house and cost-effective LPA-based method targeting ITS to identify prevalent NTM species. It was additionally evaluated for direct application to DNA from clinical specimens compared to that of extracting from bacterial colonies of the same specimen.

Materials And Methods
Strains and clinical specimens: In this study, 25 bacterial species were used including 15 Mycobacterium and 10 non-mycobacterium species (Table 1). Moreover, nine clinical specimens and their relevant isolates from culture medium were collected from Pasteur Institute of Iran, to evaluate the results of our developed LPA in a pilot study. The clinical samples were positive for acid-fast bacilli and the cultivated isolates were negative for niacin test, and (i.e., the patients were suspected of NTM infection). These samples were evaluated in parallel by PCR-LPA and sequencing.
Probe and primer design: The 55 sequences of ITS gene of mycobacteria were obtained from NCBI GenBank database (Table 2). Multiple alignment analysis of available sequences were performed using BioEdit software (Ibis Biosciences, Carlsbad, CA, USA) and species-specific probes were selected from the polymorphic and conserved regions of mycobacterial sequences to ensure probe specificity.
Afterwards, Gene Runner software (Hastings Software, Colorado, USA) was used to compute the probes parameters. Then, selected oligonucleotide probes were analyzed with BLAST search (Basic Local Alignment Search Tool) at nucleotide collection database to determine the specificity of the selected regions. A set of primers Sp1 (5'-ACC TCC TTT CTA AGG AGCACC-3') and Sp2 (5'-GAT GCT CGC AAC CAC TAT YCA-3'), targeting ITS region of Mycobacterium were used from the previous study [12]. Primer Sp2 was labeled with biotin at the 5' end and a poly-(T) tail was added at the 3' end of oligonucleotide probes.
DNA extraction and PCR amplification: The extraction of DNA from samples was carried out with boiling method [21]. Briefly, after two times of washing with TE buffer (10 mM Tris-HCl pH 7.6, 1 mM EDTA), 50 µl TE buffer was added to the samples and then boiled for 30 min in a water bath.
Afterward, the microtubes were centrifuged at 10000 rpm for 10 minutes. The supernatant solution was used as template for the PCR reaction. PCR was performed in a total volume of 25 µl containing 1 µl each of forward and reverse primers (10 pmol), 12.5 µl PCR Master Mix (Parstous biotechnology, Iran), and 1 or 10 µl DNA samples (1µl for cultivated samples and 10 µl for clinical samples). PCR was performed by a Mastercycler Gradient (ASTEC, Japan). The reaction scheme was as follows: an initial denaturation step at 95°C for 5 min, a total of 40 cycles with the following steps: denaturation at 95°C for 30 s, annealing at 60°C for 30 s, extension at 72°C for 30 s, and a final extension step at 72°C for 10 min. Analysis of the PCR products was performed by electrophoresis on 2% agarose gel stained with Green viewer™. The predicted sizes of the amplicons (varied from 200-300 bp) were visualized by UV-transilluminator. For each PCR run, sterile water and 1 ng of M. tuberculosis H37Rv genomic DNA were used instead of template as negative and positive control, respectively. Line probe assay: the LPA was modified from a previous publication [22]. Briefly, 10 µl of 20 pmol/µl species-specific probes in TE buffer were immobilized onto a nitrocellulose membrane (pore size 0.

Results
Designed specific probes: The ITS sequences of 55 mycobacteria were evaluated to select the best conserved region of each species. It was essential to design more than one probe for some of the species like M. fortuitum and M. kansasii. Some other species like M. ulceranse and M. marinum was not differentiable by ITS target. So, one genus-specific and fifteen species-specific probes for 12 Mycobacterium species were designed after consideration of length and melting temperature.
According to table 3, the lengths of designing probes were19-23 bp, which was suitable for hybridization procedure. The optimal melting temperatures (Tm) were in range 60.6-62.5 °C (Tm difference ≤ 1.9 °C) and the optimal free energy values (ΔG) for hairpin and self-dimer were lower than -2.2 and -5.3 kcal/mol, respectively (Table 3). Then, the selected probes were checked with the nucleotide collection database and it was confirmed that there was no homology among the designed probes and non-related bacterial sequences.
PCR-LPA: The ITS fragments were amplified successfully for all Mycobacterium species by PCR with the SP1 and SP2-biotinylated primers. The PCR reaction was performed at annealing temperature of 59°C and produced a single band ranging in length approximately from 220 to 300 bp for different Mycobacterium species (Figure1). In this study, the rapidly growing Mycobacterium species like M. fortuitum show larger amplicons (300 bp) than slowly growing species. The genus-specific primers (Sp1 and Sp2) were shown to be specific for mycobacteria as were unable to amplify any DNA fragment from bacteria other than mycobacterium species. All immobilized oligonucleotide probes on the nitrocellulose membrane strips were successfully hybridized with their relevant biotinylated PCR products from Mycobacterium species. The specificity of LPA assay was 100% using DNA extracted from mycobacterial and non-mycobacterial reference strains (Table 1  comparable. The results of PCR-LPA and sequence analysis were also comparable except for the two mixed infection specimens in which the sequencing results were not interpretable. For this reason, these two specimens were amplified again, and then PCR products were purified using a gel extraction kit (GeNet Bio, Korea) and the purified PCR products were sent for re-sequencing. The results showed that the two species extracted from sample 7 belong to different types of M. fortuitum and sample 9 contained M. fortuitum and M. simiae, which was similar to PCR-LPA result.

Discussion
Due to increasing incidence of Mycobacterium infection, rapid diagnosis and differentiation of the Mycobacterium species is crucial for successful treatment. Traditional methods usually take about 4-10 weeks (or more) to give the results. Therefore, molecular methods can be an appropriate alternative for rapid diagnosis and differentiation of Mycobacterium species. Although, Mycobacterium genome is more than 4000 bp in length, a few genes can be the candidate target for molecular diagnosis of these bacteria. Some of the most prevalent NTM species lack a unique sequence within the candidate genome region to be differentiated from each other, due to the high genomic similarity among the Mycobacterium species [23].
Most of the PCR assays which are used for detecting mycobacteria, identify a single or limited number of Mycobacteriums species [24]. Therefore, these methods seem to be inefficient for identifying Mycobacteria to the species level. Restriction fragment length polymorphism (RFLP) is known as a reliable method for differentiation of mycobacteria [25], but interpretation of its results is actually quite difficult, because of low distance between band sizes in some species and the emergence of new patterns that have not been reported before [26]. Furthermore, RFLP method is unreliable in detection of mixed cultures and is not applicable to clinical samples. Real time PCR technique is also too costly and requires considerable expertise. Although, sequencing is known as the gold standard method in detection of mycobacteria, it is unable to identify mixed culture samples or co-infection with more than one NTM species in a patient as it requires a high quality and pure PCR product.
Generally, the advantages and disadvantages of different molecular techniques are summarized in Due to disadvantages of these diagnostic techniques, use of a molecular method based on reverse hybridization can be useful in detecting and differentiating of Mycobacterium species. The PCR-LPA is one of the reverse hybridization techniques which is able to differentiate a large number of Mycobacterium species from each other. LPA is faster than traditional methods, more accurate than chromatographic tests, and less expensive than sequencing for diagnosis of mycobacteria [12].
However, LPA assay is a PCR-based method. The low amount of DNA and high amount of inhibitors in the clinical specimen can lead to false-negative results, especially in smear-negative specimens [27].
Therefore, cultivation of bacteria from clinical specimens and testing on colonies is the most appropriate way for NTM identification. However, a major drawback with mycobacterial culture is the overgrowth of one NTM species that may cause to miss the growth of another one in the cases of coinfection with more than one NTM species.
In the present study, ITS showed a strong potential for discriminating closely related Mycobacterium species. The sequence analysis of ITS in different Mycobacterium species revealed that the 16S-23S spacer sizes of slowly growing Mycobacteria are shorter than those of rapidly growing species (Figure1). Moreover, the differences in the ITS length of various Mycobacterium species have made it easy to detect the mixed culture or co-infections by electrophoresis of PCR products. We found the presence of more than one species in 2 samples according to the electrophoresis and sequencing results. The rest of the samples (1,2,3,4,5,6,8) were showed only one species (Figure3). Therefore, this advantage of ITS target can be very useful for diagnostic purposes.
To date, several lung infections have been reported with more than one species of NTM or more than one genotype of a specific species. However, information on such infections is very limited. The However, PCR-LPA analysis indicated that the discrimination between M. marinum and M. ulcerans which have a high degree of sequence homology is impossible. Interestingly, PCR-LPA was able to discriminate between M. chelonae and M. abscessus with two nucleotide differences between their ITS sequences, and between M. avium and M. intracellulare which cannot be distinguished by chromatographic methods [12].
In our pilot study on nine clinical samples and their cultured isolates, M. fortuitum was the most common mycobacterial species in the sample collection region (Tehran, Iran). In consistent with our results, Irandoost and colleagues reported the same species as the most prevalent one in the same region [29]. Our finding showed that this method is capable of detecting mycobacterium species in clinical samples as accurate as culture or sequencing methods (Figure 4). Although, the results of LPA performed on clinical samples showed band intensity weaker than those of cultured isolates, we had no problem for interpretation of the results.
In the present study, we described a kind of PCR-LPA which is readily applicable for respiratory specimens in clinical laboratories. This is the first study which compared the results of PCR-LPA in clinical and cultivated samples from the same patients. Our PCR-LPA based on the ITS target can differentiate 12 mycobacterial species and genotype by the use of 15 probes. The speed of the assay is similar to commercial kits with high specificity and required no radioactive material. Moreover, the cost of this test is very low compared to the similar available kits. Each test only has a cost of 6-7€ (Table 4). We routinely performed the amplification by PCR in less than 3 h, followed by a line probe assay in 4 h.