The insertion sequence (IS) 6110 has been described as a repetitive mobile element specific for the Mycobacterium tuberculosis complex (MTBC; Thierry et al., 1990). It has been used for years to diagnose and to characterise at the molecular level M. tuberculosis strains by the IS6110-restriction fragment length polymorphism (RFLP) technique (Thierry et al., 1990; Brisson-Noel et al., 1991; Otal et al., 1991). It is 1,355 bp and is flanked by inverted repeats (IR) of 28 bp (Thierry et al., 1990). When transposition occurs, it creates a duplication of 2–4 bp in the insertion site (Mendiola et al., 1992; Dale, 1995). IS6110 contains two overlapping reading frames called orfA and orfB. While the product orfAB is a transposase, the individual products orfA and orfB inhibit transposition, so the proportion of each product is important for the autoregulation of transposition (Sekine et al., 1997). Transcription of this element is regulated by a-1 frameshift due to a ribosomal slippery sequence and by a pseudoknot in the mRNA; these processes lead to a transposition based on a copy-out-paste-in mechanism (Gonzalo-Asensio et al., 2018).

The global M. tuberculosis population can be divided into several phylogeographic lineages (Ls) and families. The genetic diversity found in circulating strains plays an important role in disease outcome. The number of IS6110 copies in the genome is variable and seems to be related to the evolution of the strains. Modern Ls have more IS6110 copies than ancient Ls, and they are considered to be better adapted to high-density populations (Gonzalo-Asensio et al., 2018). It seems that if the IS6110 is inserted in a transcriptionally active region of the genome, the number of copies increases, so the variable number of copies present in different strains could be due to the genomic region in which they are inserted, although a large number of IS6110 copies may be deleterious for the strain because they can remove parts of the genome (Wall et al., 1999).

While IS6110 supposedly transposes randomly, there are some preferential location sites such as the CRISPR region, also known as the DR region in M. tuberculosis; it was probably the first IS copy because it is present in almost all MTBC isolates (Hermans et al., 1991). Other hot spots described are: the plcD gene (Vera-Cabrera et al., 2001), members of the PPE family genes (Sampson et al., 1999; Beggs et al., 2000), the intergenic region dnaA:dnaN (Supply et al., 2006), and other ISs (Fang and Forbes, 1997; Fang et al., 1999). In addition, characteristic locations of IS6110 in different Ls have been observed (Roychowdhury et al., 2015). The function of IS6110 remains unknown although some general effects have been described. The first is the disruption of the gene in which it has been inserted, with a possible deleterious effect (Sampson et al., 1999; Beggs et al., 2000; Warren et al., 2000; Yesilkaya et al., 2005). The second is the recombination and the deletion or inversion of the DNA in between, a phenomenon that can occur when there are two IS relatively close to each other (Sampson et al., 2003). A third effect is that IS6110 can act as a promoter of the downstream gene (Beggs et al., 2000; Safi et al., 2004; Soto et al., 2004).

For the MTBC, the rate of point mutations has been determined to be 1 × 10⁻⁹ events per site per generation. However, it has been demonstrated that the rate of point mutations for the IS6110 is about 7.9 × 10⁻⁵, higher compared with the rest of the genome, indicating that IS6110 is under selective selection (Tanaka, 2004). There are few studies focused on the sequence genetic variability of this element. Dale et al. (1997) amplified and sequenced IS6110 of four M. tuberculosis strains, each carrying one copy, and one of the copies carried by the Beijing W strain, which has 17 copies. They compared the sequences with the Bacillus Calmette and Guérin (BCG) copy published and found they are identical. Nevertheless, they only checked one copy of each strain. Conversely, Thabet et al. (2015) found several different haplotypes when they randomly cloned and sequenced only one of the IS6110 copies of some characteristic strains from their region.

In this work, we aimed to determine whether IS6110 is a conserved gene, as Dale et al. (1997) suggested or, on the contrary, it is relatively common to find this IS mutated, as Thabet et al. (2015) suggested. In addition to the clinical relevance, since it is widely used for TB diagnosis, we thought mutations of the IS6110 could affect its transposition ability. Therefore, we planned to find out the genetic variability of the sequence. For this purpose, we applied two different approaches. The first was the study of all the IS6110 copies present in the same strain to determine whether they had an identical sequence. For the second approach, a sample collection was used to study IS6110 copies in different MTBC families for which the location had been previously published (Reyes et al., 2012; Comín et al., 2021) to elucidate how conserved IS6110 is among families for the same location.

We aimed to determine the genetic variability of the IS6110 sequence. With this purpose, a large number of copies were analysed. The first approach of the study focused on the comparison of all the IS copies present in a strain to determine whether different copies coexist in the same genome. Initially, we analysed the strains whose genomes were available in the databases. Furthermore, we studied several strains whose IS6110 locations were known as result of our previous studies. For the second approach, we studied IS6110 in specific locations defined for different families, including Haarlem, LAM, S, and L6 strains (Reyes et al., 2012; Comín et al., 2021) to determine its inter- and intra-strain variability using a sample collection. In addition, CAS_DELHI and T strains were included to study the IS copy located in the DR region. As a result, 158 IS6110 sequences from 55 different strains were analysed (Figure 1). Twenty-one of the copies, belonging to 16 strains, had a mutation (13.3%). Among the high copy number strains, seven out of the 113 copies studied (6.2%) were mutated while 14 out of the 45 copies (31.1%) showed a mutation in the LCN strains.

This work deepens the knowledge of the IS6110 sequence in the MTBC. IS6110 has been used for a long time for diagnosis and genotyping, but very little is known about its genetic variability, mainly due to the difficulties in determining its complete sequence for each copy location in the genome. The limited data that have been published regarding this issue have shown dissimilar results (Dale et al., 1997; Thabet et al., 2015).

We have analysed the sequences of both orfA and orfB genes that constitute the transposase in several strains for which we knew its exact location in the genome using specific primers for the amplification of each copy. So far, the only data reported on the sequence of IS6110 were obtained using internal primers that interchangeably amplify all the IS copies of the genome (Thabet et al., 2015), or the number of analysed copies was really low (Dale et al., 1997). In addition, as only one copy of each strain was analysed, the intra-strain variability was not studied. Therefore, the novelty of this work, as far as we know, is that for the first time the IS6110 sequence corresponding to each location has been identified and sequenced, and likewise, the sequences of the different copies in the same strain or among families have been compared. Furthermore, IS6110 cannot be studied directly using the most frequent whole genome sequencing (WGS) technologies, as is the case of Illumina, due to repetitive sequences are not correctly aligned; thus, this study complements this handicap.

We studied 158 IS6110 sequences from 55 different strains belonging to L2 (Beijing), L3 (CAS_DELHI), L4 (Haarlem, LAM, T, S, and X), L6, and the animal branch, including strains with a high and a LCN of IS6110. There were variations of the wild-type sequence in 21 (13.3%) of the copies belonging to 16 strains (29%). We observed a higher percentage of mutated copies among the LCN strains (31.1 vs. 6.2% in high copy number), which could be related with a less transposition ability. To support this idea, a higher number of copies with a widely strain spectrum should be studied.

We found 10 different mutations (Figure 2). Five of them are located in the orfB of the transposase vs. two in the orfA, findings that seem to indicate that orfB is more susceptible to mutation. A possible explanation could be that the orfB is three times longer than orfA. One mutation is in the overlapping region affecting both ORFs. In addition, two other mutations are in the IR, which we do not expect to be involved in the transposase function as they are outside the coding region. Thabet et al. (2015) also found more polymorphic sites for orfB than for orfA. They suggested that non-synonymous mutations tend to be eliminated because the ancestral IS6110 copy inherited by the MTBC would be functionally optimal. However, we found five out of eight mutations in the coding orfA and orfB that are non-synonymous. The fact that some of them are conserved in different strains supports the idea that those non-synonymous mutations are neutral in these cases, as Thabet et al. (2015) suggested. They described different haplotypes for orfA and orfB, reporting the wild type as the most frequent, similarly to our findings. On the other hand, our results differed from Thabet et al. (2015) in other aspects, including the description of some haplotypes with 1–3 SNPs. This caught our attention because all the variant copies we found have just one SNP. The high variability found by these authors could be partially caused by the different procedures applied to obtain the IS6110 DNA sequence. We used specific primers for each location, what allowed us to sequence directly, for each copy, the DNA obtained from PCR. In contrast, other authors used IS internal primers to amplify, clone and sequence one of the copies of each strain without knowing what copy had been amplified or if the sequence obtained was the sum of several. One of the mutations they found (haplotype 20 of orfB) is Gly215Ser (orfB), the same we found in the X family. Since their collection did not include any X strains, we conclude that this mutation is not restricted to the X family. None of the other mutations obtained matched their results. Our results completely differed from Dale et al. (1997), who did not find any mutation when comparing five IS copies of different M. tuberculosis strains, so they affirmed that the IS has been totally conserved among the MTBC. Surely, the low number of copies they studied caused these differences.

As far as we know, this is the first time that all the IS copies of a strain have been studied. In some cases, we observed that it exists variability among the IS6110 copies of the same strain, finding wild-type and mutated copies at the same time. In addition, we did not find copies with different mutations in the same strain.

Our expertise in genotyping M. tuberculosis by IS6110-RFLP supports the fact, previously described (Gonzalo-Asensio et al., 2018), that modern lineages (L2, L3, and L4) have accumulated more copies of IS6110 in their genomes than ancient lineages (L1, L5, and L6). Modern lineages are better adapted to high-density populations, so it is possible that the number of ISs is related to this success (Gonzalo-Asensio et al., 2018). This eventuality agrees with the obtained results because the two most successful strains in our population, used in this work, are high copy number strains (MtZ and GC1237) and, curiously, all their copies (12 and 18, respectively) are wild type. Remarkably, all the strains of the X family (L4) studied, such as CDC1551; share a characteristic variant in at least one of their IS6110 copies. According to these results, it is possible that this mutation could affect the transposition ability of IS6110, reducing its transposition frequency. Thereby, there would be a greater probability that the wild-type copies transpose, leading to a higher copy number of IS6110 in the genome. This phenomenon could partially explain why the X family strains, belonging to L4, are LCN strains despite being a successful clinical family. Nevertheless, the mutation found in the X family strains did not prevent the transposition, as there are more copies with the same SNP in the same strain. Dale et al. (1997) suggested that the differences observed in the copy number are not related to structural changes in the transposase, but rather to the genomic region in which it is inserted. In view of our results, focusing especially on the X strains, we think that both facts are not mutually exclusive and more than one factor must be affecting its transposition.

It is notable that the Haarlem and X family strains studied have an IS6110 inserted in Rv0403c at the same point and direction (483,296, −), in addition to an identical 3 bp created duplication. This finding suggests a common origin for both families, supporting all the phylogenetic trees in which the X and Haarlem families are closely related (Stucki et al., 2016). All the X strains studied in this work have a mutated copy located in Rv0403c, unlike the Haarlem strains, which lack a mutation. The evolutionary process that is inferred by these results is that the common ancestor of these two families had the IS6110 copies located in the DR region and in the Rv0403c gene. Subsequently, the mutation in the copy located in Rv0403c of the X strains occurred after the phylogenetic separation between the Haarlem and X families, and this mutated copy was the one that transposed and led to the rest of the copies in this family, because all IS6110 copies have the same mutation, except the one located in the DR region. The exception is HCU 3717, which has its two DR copies also mutated, suggesting a different evolutionary process.

Researchers demonstrated that the synthesis of the transposase is regulated post-transcriptionally by the ribosome (Gonzalo-Asensio et al., 2018). orfA finishes on Leu92 (UUA), but the ribosome can make a −1 frameshift and starts orfB with Lys1 (AAA). Considering this, we conclude that the mutation in the IS sequences of the X family strains has changed glycine 215 for serine (orfB). Glycine is the smallest amino acid and the only one that is not chiral. Furthermore, it could provide more flexibility to the protein region. On the contrary, serine is a polar amino acid, uncharged, and is bigger than glycine, a factor that diminishes the flexibility of the protein chain. We believe it is possible that this amino acid change may affect the protein structure but, according to our results, it does not prevent the transposition ability.

Regarding M. africanum, in both NCBI strains one of its IS6110 copies is mutated. While the mutation found in the GM041182 strain is synonymous, strain 25 has a gap in the sequence, which completely alters the reading frame and probably disrupts the transposase. Among the LCN strains we sequenced, isolated in our region, M. africanum HMS 2382 has a non-synonymous mutation in one of its three IS6110 copies. It changes aspartic acid for glycine, an important change because aspartic acid has negative charge and is larger than glycine. Although the Haarlem family is a completely different lineage (L4) than M. africanum (L6), we found that two Haarlem strains (HMS 18009 and HMS 18046) share the same mutation as HMS 2382, in the same nucleotide and in the same IS copy (DR region). In addition, IS6110 has been inserted at exactly the same point in the three strains. As both L6 and L4 are separate phylogenetic branches, it is not likely that they acquired the mutation from a common ancestor; therefore, the mutation could be random. There is one mutation in the IR of the IS6110 for HMS 1693 and HMS 14017 (L6). These two strains seem to be from the same cluster according to their RFLP patterns, a finding that would be supported by this shared mutation. As it is out of the coding region, we do not know if it affects the transposase somehow.

Among the findings in the other MTBC families studied is the mutation in the IS6110 copy of the DR region in HMS 18014 (T family), which changes threonine to isoleucine. Both amino acids have a similar size, but they are in different classification groups. Threonine has a polar lateral chain without a charge, whereas isoleucine has a hydrophobic lateral chain, so the interactions will be different. Notably, the three Haarlem strains studies have a different mutation in the copy located in Rv0963c, one with a mutation in the IR of the IS (HMS 18037) and two with a synonymous mutation in the coding region of the transposase (HSJ 234 and HMS 18031). These mutations seem to be random as they are not the same or conserved among strains. The chance of three different mutations in the same location could be because a high number of these copies were sequenced. HMS 18025 (LAM) has a mutation in the copy inserted in lpqQ:Rv0836c, with a change from glycine to arginine. This is an important structural change because arginine is larger than glycine and has a positive charge. More studies are needed to determine how these mutations affect the transposition ability.

In total, we studied 38 IS copies located in the DR region. We successfully amplified this copy in 30 strains of our collection and examined eight more from the NCBI strains. Among them, we have found three different variants in five strains, described above. This is the copy in which we have found the most mutations, a phenomenon that can be explained because this is the region for which more IS copies have been studied, even though the percentage of mutation is constant with regard to the total mutations found. As almost all MTBC strains have an IS copy in this region, it would be reasonable to think that this copy could have been the first IS6110 that started to transpose and thus has increased its number within the genome (McEvoy et al., 2007). If this is true, we have to assume that mutations in this copy occurred after the transposition, as the other studied copies do not have any mutations. Again, the exception is HCU 3717 (X family), in which all its IS6110 copies, including the two of the DR region, seems to be mutated (in the absence of knowing the genotype of the copy located in ppe46).

Ates et al. (2018a) showed that modern Beijing families have different deletions involving the ppe38 locus, suggesting that these changes may contribute to a higher growth rate in vivo and lung inflammation due to the block of PE_PGRS secretion, rendering some Beijing strains hypervirulent. The most virulent deletion described in their work was the one that disrupts the ppe38 and ppe71 genes with the loss of the esx genes between ppe38 and ppe71. This is the deletion we found in the Beijing GC1237 strain (Figure 3). We found that this IS copy has no direct repeats, which suggests that there has been a recombination event between two IS6110 copies with the removal of the genes between them (Brosch et al., 1999; Sampson et al., 2003). Regarding the MtZ strain, there is an IS6110 within the ppe38 gene and another one in the ppe71 gene. We do not know how the insertion affects both genes. Ates et al. (2018b) demonstrated that a single copy of these ppe genes is enough to support PE_PGRS secretion; therefore, MtZ could have this secretion affected. Several experiments are ongoing to determine whether the secretion of PE_PGRS in the MtZ strain is altered. McEvoy et al. (2009) found IS6110 insertion events in the ppe38 locus in other MTBC families. As we also found this in the MtZ strain, we conclude that it is not something specific to the Beijing family. As we have stated previously, none of the IS6110 copies amplified that are inserted in ppe genes have a mutation.

In summary, we have studied the genetic variability of IS6110. We have analysed the DNA sequence of the different IS6110 copies in the same strain of M. tuberculosis as well as in different MTBC strains. We have observed mutation in 13.3% of the copies studied. Gly215Ser (orfB) mutation seems to be characteristic of the X family. In general, the high copy number strains analysed carry wild-type copies whereas several LCN strains, such as the X family and L6, have one or more of their copies mutated. Some strains share the location as well as mutations in these copies, allowing us in these cases to establish the transposition events that have occurred over time and indicating a close relationship in the evolution of these strains. The detailed study of each of the copies has also allowed us to provide information regarding the evolution of some MTBC families. Many publications have studied the variability among strains of the MTBC caused by changes in the number and location of IS6110. To these, we add the changes produced over time in the sequence itself as one more factor that will probably affect its evolution.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: https://www.ncbi.nlm.nih.gov/; MZ574181–MZ574188.

SS and IO: conceptualised the work. JC: carried out the laboratory experiments and curated the data. JC, IO, and SS: wrote the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by the Carlos III Health Institute in the context of a Grant (FIS18/0336), and JC was awarded a scholarship by the Government of Aragon/European Social Fund, “Building Europe from Aragon.”

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.