MBE Advance Access originally published online on June 1, 2007
Molecular Biology and Evolution 2007 24(8):1861-1871; doi:10.1093/molbev/msm111
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
Contribution of Horizontally Acquired Genomic Islands to the Evolution of the Tubercle Bacilli
,
,1,2
* Genomics and Molecular Bioinformatics, Inserm U726 - Université Paris 7, Paris, France
Unit of Biodiversity of Emerging Pathogenic Bacteria, Institut Pasteur, Paris, France
Unit of Mycobacterial Genetics, Institut Pasteur, Paris, France
Centre National de la Recherche Scientifique, URA 2172, Paris, France
E-mail: neyrolle{at}pasteur.fr.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
The contribution of horizontal gene transfer (HGT) to the evolution of Mycobacterium tuberculosis — the main causal agent of tuberculosis in humans — and closely related members of the M. tuberculosis complex remains poorly understood. Using a combination of genome-wide parametric analyses, we have identified 48 M. tuberculosis chromosomal regions with atypical characteristics, potentially due to HGT. These specific regions account for 4.5% of the genome (199 kb) and include 256 genes. Many display features typical of the genomic islands found in other bacteria, including residual material from mobile genetic elements, flanking direct repeats, insertion in the vicinity of tRNA sequences, and genes with putative or documented virulence functions. Southern blotting analysis of nine of these 48 regions confirmed their presence in "Mycobacterium prototuberculosis," the ancestral species of the M. tuberculosis complex. Finally, our results strongly suggest that the ancestor of the tubercle bacilli was an environmental bacillus that exchanged genetic material with other bacterial species, including Proteobacteria in particular, present in its surroundings. This study describes a rational approach to searching for mycobacterial virulence genes, and highlights the importance of dissecting gene transfer networks to improve our understanding of mycobacterial pathogenicity and evolution.
Key Words: Horizontal gene transfer • Mycobacterium tuberculosis complex • Tuberculosis • Virulence
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Horizontal gene transfer (HGT) between unrelated species is thought to be one of the leading creative forces driving bacterial evolution (Ruiting and Reeves 1996
; Doolittle 1999a; Eisen 2000; Ochman, Lawrence and Groisman 2000; Dutta and Pan 2002). If a horizontally acquired gene is to be kept and expanded within a bacterial population, it must confer a selective advantage upon the host species and increase its fitness, for colonizing new environments or new hosts, for example. HGT has been demonstrated in many pathogenic and non pathogenic bacteria (Smith, Feng and Doolittle 1992; Syvanen 1994; Eisen 2000; Dutta and Pan 2002). Clusters of genes acquired as a single unit by horizontal transfer form so-called "genomic islands." These genomic islands typically contain intact mobile genetic elements or residual material from such elements. They are flanked by direct nucleotide repeats, and are sometimes inserted in the vicinity of tRNAs (Hacker and Kaper 2002; Canchaya et al. 2003; Hacker et al. 2004; Frost et al. 2005; Thomas and Nielsen 2005). Several types of genomic island have been described, including pathogenicity islands (PAIs), which contain virulence genes required for host colonization, and symbiosis islands (Dobrindt et al. 2004).
The tubercle bacilli comprise both the M. tuberculosis complex (MT-complex) — M. tuberculosis, M. bovis, M. microti, M. africanum, M. pinnipedii, and M. caprae species — and its ancestral species "Mycobacterium prototuberculosis" (Gutierrez et al. 2005). Very few of the genes of M. tuberculosis were initially thought to have been acquired by HGT (Cole et al. 1998). Several genes (Saves et al. 2001; Blanc-Potard and Lafay 2003; Kinsella et al. 2003), transposons and insertion sequences (IS) (Martin et al. 1990; Mariani et al. 1993; Picardeau, Bull and Vincent 1997) and plasmids (Le Dantec et al. 2001; Stinear et al. 2004) were thought to have been transferred to Mycobacteria by HGT. Further studies based on more sophisticated parametric methods showed that Mycobacteria have been subject to average levels of HGT (Garcia-Vallve, Romeu and Palau 2000; Ochman, Lawrence and Groisman 2000; Nakamura et al. 2004; Dufraigne et al. 2005). We recently showed that the Rv0986-8 virulence operon (Pethe et al. 2004; Rosas-Magallanes et al. 2006b) arose in the ancestor of the tubercle bacilli by HGT from a 
-proteobacterial species (Rosas-Magallanes et al. 2006a). We showed that this operon was specific to both the M. tuberculosis complex and to its ancestral species "M. prototuberculosis" (Gutierrez et al. 2005) but absent from all other mycobacterial genomes examined, including that of Mycobacterium marinum, the species most closely related to the MT-complex. This raised the question of the contribution of horizontal gene transfer to the evolution of the ancestor of the tubercle bacilli and the emergence of M. tuberculosis complex as major pathogens (Smith 2003).
Two in silico strategies are used to detect genes introduced by horizontal transfer: parametric methods and phylogenetic approaches. Phylogenetic methods detect potential HGT in the form of genes displaying phylogenetic incongruency (Lecointre et al. 1998; Maynard-Smith and Smith 1998; Doolittle 1999b; Wolf et al. 1999). The dependence of these methods on the availability of multiple homologous sequences limits their suitability for genome-wide HGT detection. In contrast, parametric methods detect potential HGT based solely on the genome sequence of an organism. Foreign DNA fragments are identified on the basis of differences in characteristics between these fragments and the rest of the genome (Lawrence and Ochman 1997; Nakamura et al. 2004). Various methods, mostly based on nucleotide content, codon usage and oligonucleotide frequencies, have been described (Karlin, Mrazek and Campbell 1998; Mrazek and Karlin 1999; Garcia-Vallvé, Romeu and Palau 2000; Karlin, 2001; Nakamura et al. 2004; Dufraigne et al. 2005). These methods sometimes give different results (Ragan 2001; Lawrence and Ochman 2002; Dufraigne et al. 2005). We therefore used a combination of parametric methods, to obtain a consensus for the detection of potential HGT (Azad and Lawrence 2005).
We further investigated this aspect of the evolutionary history of the tubercle bacilli by combining our consensus method with an analysis of genomic features and in vitro validation. Our results strongly suggest that the ancestor of the tubercle bacilli underwent episodes of substantial HGT, mostly from environmental Proteobacteria. We identified several DNA fragments limited to tubercle bacilli among Mycobacteria, with features typical of the genomic islands commonly found in other bacteria. Several of these genomic islands were found to contain genes with putative or documented virulence functions, and may therefore be seen as typical PAIs.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Sequences
The complete sequences of Mycobacterium tuberculosis H37Rv (NC_000962 [GenBank] .2) (Cole et al. 1998
; Camus et al. 2002), Mycobacterium tuberculosis CDC1551 (NC_002755
[GenBank]
.2) (Fleischmann et al. 2002), Mycobacterium bovis AF2122/97 (NC_002945
[GenBank]
.3) (Garnier et al. 2003), Mycobacterium ulcerans Agy99 (NC_008611
[GenBank]
.1) (Stinear et al. 2007), Mycobacterium avium subsp. paratuberculosis K-10 (NC_002944
[GenBank]
.2) (Li et al. 2005) and Mycobacterium leprae TN (NC_002677
[GenBank]
.1) (Cole et al. 2001) were retrieved from Genbank. The complete sequence of Mycobacterium smegmatis MC2 was obtained from the Institute for Genomic Research. Mycobacterium marinum sequence data was provided by the M. marinum Sequencing Group at the Sanger Institute. For convenience, Mycobacterium tuberculosis H37Rv is used as the reference, with all genes named as in this genome.
Bacterial Strains
DNA from representative strains of eight genetic groups of "Mycobacterium prototuberculosis" (A, B, C, D, F, G, H, I) (Gutierrez et al. 2005) was used to determine whether the sequences acquired by putative HGT were or not already present in the ancestral species of M. tuberculosis. Two strains belonging to the M. tuberculosis complex (M. tuberculosis Mt14323 and M. bovis BCG) were also studied, and M. tuberculosis H37Rv was used as a positive control.
Genomic Signatures
The characteristics of a given sequence can be analyzed with a four-letter word signature (Deschavanne et al. 1999; Dufraigne et al. 2005). This four-letter word signature is based on the frequencies of all tetranucleotides contained in the sequence, obtained with a very rapid algorithm derived from chaos game representation (CGR) (Jeffrey 1992; Deschavanne et al. 1999). Four-letter words were chosen because this length presents the best compromise between discriminatory power and length of windows (see below). Signatures were compared by measuring Euclidean distances between them, where i is one of the 256 possible four-letter words:
 |
Detection of Regions of Potential HGT
Horizontally acquired sequences are thought to have various atypical genomic characteristics, if they are recent enough to not have been ameliorated yet (Lawrence and Ochman 1997; Ochman, Elwyn and Moran 1999). We therefore used an approach for HGT detection based on the consensus of three parametric methods (Azad and Lawrence 2005). A gene was considered to have been horizontally transferred if detected by at least two of the following three methods:
(1) G+C content was calculated at the first and third codon positions of all genes. Genes with values higher or lower than the corresponding thresholds of the G+C content distribution (lower threshold = lower quartile – 0.25 x interquartile distance and upper threshold = upper quartile + 0.25 x interquartile distance) (Lawrence and Ochman 1998) are considered to have been potentially acquired by horizontal transfer;
(2) Factorial correspondence analysis of absolute codon frequencies for all genes, excluding the PE/PPE multigene families, was carried out (Medigue et al. 1991). Genes were clustered using a k-means method into four groups and genes belonging to neither the highly expressed gene group nor to the two housekeeping gene groups (a group of genes enriched in hydrophobic amino acid codons and the others) (Carbone, Zinovyev and Kepes 2003) were considered to be of foreign origin.
(3) The genome was scanned by calculating the signatures of 5 kb overlapping sliding windows (step = 0.5 kb). The Euclidean distance between each window signature and the signature of the whole genome sequence was calculated. Genes included in windows with a distance greater than the threshold value (threshold = 97.5% percentile) were considered to be of foreign origin (Dufraigne et al. 2005) (illustrations of HGT signatures are presented in Supplementary figure 1).
Adjacent atypical genes on the same strand were then concatenated into atypical regions.
Ortholog Determination
Orthologs in MT-complex genomes were defined as BlastP reciprocal best hits (E-value < 10–20) (Altschul et al. 1990). MT-complex genes were considered to have an M. marinum ortholog if their BlastP best hit (threshold: E-value < 10-20 and identity > 50% of protein length) against M. marinum predicted peptides displayed conserved synteny with adjacent orthologs. We checked that the absence of the corresponding gene was not due to a deletion in the M. marinum genome, by verifying the absence of these genes and their synteny in the following mycobacterial genomes: M. ulcerans Agy99, M. leprae TN, M. avium subsp. paratuberculosis K-10 and M. smegmatis MC2.
Southern Blot
The mycobacterial strains were grown on Lowenstein-Jensen medium at 37°C for 3 weeks. Mycobacterial chromosomal DNA was extracted and amplified. PCR was performed with Taq DNA polymerase (Qbiogene, Carlsbad, CA), according to the manufacturer's recommendations. PCR conditions were as follows: 5 min at 95°C, then 35 cycles of (1 min at 95°C, 1 min at 55°C, 1 min at 72°C) and 10 min at 72°C. Southern blotting was performed by a previously described method (van Embden et al. 1993). The membranes were sequentially probed with one of 13 probes specific to horizontal transfer sequences. The probes were labelled and detected with an enhanced chemiluminescence kit (ECL; Amersham International).
DNA probes were generated using the following primers: 0301Fd, 5'-CTGATCGACAAGTCGGCGC-3', 0301Rv, 5'-CCCGACAGTTCGGCTGTC-3', 0323cFd, 5'-CGGACCGACCGAAACCTC-3', 0323cRv, 5'-GATCGGCGATCGGTTGGG-3', 0656cFd, 5'-CACCGAGGCCTGGAACTC-3', 0656cRv, 5'-CTAGTCGTCGGCGCTGAC-3', 1044Fd, 5'-CGCGATGCGCGAGACATC-3', 1044Rv, 5'-CCGTCGTGACGGGTATTCC-3', 1053cFd, 5'-GGATTCGCACAAGGTCTG-3', 1053cRv, 5'-CTATTGATGCCGACGGCTG-3', 2307BFd, 5'-GAAGTCCCTACTGGCCCG-3', 2307BRv, 5'-CGCTGGTACCCGCACTTG-3', 2804cFd, 5'-CGGGCCGGTCAACATCTG-3', 2804cRv, 5'-CCTCGATGCCCTCATCCG-3', 2819cFd, 5'-CTACGTGCCGGGTTCGAC-3', 2819cRv, 5'-GATTGAGGTTCCCGGCGC-3', 2956Fd, 5'-CGAACCGCTATCCGGACC-3', 2956Rv, 5'-GTTGATTTGCCCCCGGCG-3', 3108Fd, 5'-CCCAATGCGGCGAGTACC-3', 3108Rv, 5'-GCTCAGCACGTCGGACTTC-3', 3123Fd, 5'-CCTGGTCGCAGTGGAGTTG-3', 3123Rv, 5'-CACGATCGGGGTGGCTTC-3', 3178Fd, 5'-GCCGGATTCCGCAAACCG-3', 3178Rv, 5'-GCTCGGTCCAGCTTTGGC-3', 3189Fd, 5'-GAGCCGGGAGTCTGGTAC-3', 3189Rv, 5'-GGGCTCGATGTTGGGCAG-3'.
Potential Origin of Genomic Islands
We generated and maintained our own database containing the genomic signatures of 65,105 species, strains, organelles, viruses and plasmids. For each genomic island, we calculated the signature of adjacent atypical genes — genes detected by the signature method alone or by the consensus. We used the signature of these regions of HGT to search the signature database for the 35 signatures closest in terms of Euclidean distance. As genomic signature is species-specific, the species/strains with the closest signatures could be considered as potential donors of the genomic islands only if the distances obtained were below the average thresholds used for HGT detection (Dufraigne et al. 2005).
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Detection of Atypical Regions
We analyzed the genomes of three members of the MT-complex — M. tuberculosis H37Rv, M. tuberculosis CDC1551, and M. bovis AF2122/97 — with a consensus of parametric methods, to detect potential horizontal gene transfers. In each genome, we identified 160 to 200 regions of potential HGT, from 300 bp to 10 kb in length and containing 1 to 10 genes (table 1). The amount of atypical DNA detected accounted for about 230 kb per genome, corresponding to 5% of the size of the entire genome. The G+C content of the detected regions generally differed little from that of the whole genome. The difference in the number of detected genes acquired by HGT between the genome of M. tuberculosis CDC1551 and the other two genomes could be confusing knowing that they share 99% of identity at the nucleotide level (Fleischmann et al. 2002
). The ORF detection and annotation where not conducted the same way, therefore M. tuberculosis CDC1551 has 54 detected HGT with no homolog in M. tuberculosis H37Rv and M. bovis AF2122/97, even though sequences are present at the nucleotide level.
|
Genomic Features of MT-Complex Specific Horizontal Transfer Regions
We investigated the evolutionary history of the MT-complex, focusing exclusively on potential horizontal transfer regions specific to the complex — common to all three MT-complex genomes but absent from other Mycobacteria. We found that 153 of the potential HGT in M. tuberculosis H37Rv had orthologs in the other two MT-complex genomes. Of these genes, 76 had no syntenic homolog in M. marinum, M. ulcerans, M. leprae, M. avium or M. smegmatis. Assuming that adjacent detected genes were transferred during a single HGT event, the 76 genes could be grouped into 56 chromosomal regions each corresponding to a potential horizontal transfer event and specific to the MT-complex genomes.
The horizontal transfer of genomic material generally involves a genomic region including several genes. The genes transferred in a single horizontal transfer event may be subject to different selective pressures. Some adapt to the genomic characteristics of the host, whereas others retain their foreign characteristics, and the remainder develop characteristics intermediate between these two extremes (Lawrence and Ochman, 1997; Nakamura et al. 2004). Thus, our consensus method is unlikely to detect all the genes transferred in a single horizontal transfer event. We therefore extended the 56 regions, adding adjacent genes conserving the local synteny, with no match in M. marinum, M. ulcerans, M. leprae, M. avium or M. smegmatis, to generate 48 genomic islands (table 2). These genomic islands accounted for 4.5% of the M. tuberculosis H37Rv genome (199 kb) and included 256 genes. One of these regions, corresponding to the Rv0986-8 virulence operon, has already been reported to have been acquired through horizontal transfer (Rosas-Magallanes et al. 2006a). Furthermore, 45% of these regions were found to contain features typical of genomic islands, such as IS, repeats (direct repeats or genes of the PPE or PE_PGRS families, which are highly conserved and repeated in Mycobacteria), transposases, phage genes and tRNA (supplementary table). The genomic features of the Rv1041c-Rv1055 and Rv3173c-Rv3191c islands are illustrated, as an example, in figure 1. Based on the pattern of detection of the genes, we were able to identify three types of gene: genes that have adapted completely to the host (not detected by any of the parametric methods), genes that have partly adapted to the host (detected by only one method) and genes that have retained their foreign characteristics (detected by the consensus method).
|
|
Putative Functions of Transferred Genes
A putative function has been attributed to about 50% of the genes included in the genomic islands (table 3A) (Cole et al. 1998
; Camus et al. 2002). Genes of unknown function, and genes implicated in genetic mobility such as IS and transposases, repeated sequences, phage-related proteins (integrases and excisionases), were found to be overrepresented in our genomic islands. In total, 21 genes, including the three genes of the Rv0986-8 operon (Pethe et al. 2004; Rosas-Magallanes et al. 2006b), have previously been reported to be involved in M. tuberculosis virulence, by in vitro and in vivo screens of mycobacterial mutant libraries (table 3B). Other genes, such as Rv2303c, may be involved in M. tuberculosis antibiotic resistance, which has yet to be investigated. As (Sassetti and Rubin 2003) made a complete virulence gene screen, we could compare the proportion of virulence genes in the detected islands, 8.2%, to their proportion in the whole genome, 6.5%. It seems that virulence genes are slightly overrepresented in the genomic islands.
|
|
Rearrangements in Genomic Islands
Twenty-four out of the 231 genes (integrases/transposases and PE-PGRS genes excluded) in our 48 genomic islands have paralogs. Twenty-three of them can be attributed to posterior rearrangements of the genomic island, i.e. native genes were duplicated and their copy was integrated into the horizontal transfer detected. For the last gene, Rv0628c, as it constitutes a genomic island by itself, it cannot be concluded if the gene was acquired by horizontal transfer and then duplicated or if it is a duplicated native gene that derived in genomic characteristics. This gene is around 65% identical with its paralog Rv0874, which has been detected by one method only, instead of two for Rv0628c. Furthermore, the paralog Rv0874 has no ortholog in other Mycobacteria (M. marinum, M. ulcerans, M. avium, M. smegmatis, M. leprae). Therefore we rather consider it as a horizontal transfer that was duplicated later on, one copy being ameliorated faster than the other.
By analyzing the genomic features of the 10 largest (> 6 kb) genomic islands in the three MT-complex genomes, we were able to show that four of these islands had undergone genomic rearrangements, such as gene conversions, insertions or deletions of 0.5 to 3 kb. Two examples — the Rv3108-Rv3126c and Rv3173c-Rv3191c islands — are presented in figure 2. In the first, Rv3117, Rv3118 and Rv3119 are present as multiple copies in the MT-complex genomes, whereas M. bovis has one fewer copies of Rv3118 and Rv3119 and lacks Rv3120 and Rv3121. Furthermore, although the other Mycobacteria have multiple copies of Rv3117 and Rv3118, they nonetheless have one copy less than M. tuberculosis strains. We therefore suggest that a duplication event and its subsequent transposition led to the ancestor of the MT-complex gaining three genes (Rv3117-9) within the Rv3108-Rv3126c island. Then, solely in M. bovis AF2122/97, two genes (Rv3118-9) of the island were deleted, with this deletion also affecting the two adjacent genes, Rv3120 and Rv3121. Assuming parsimony, an insertion event that only affected M. tuberculosis H37Rv seems to have led to a gain of four genes within the Rv3173c-Rv3191c island. However, as these four genes are transposases for IS6110, a mobile element, evolution may have followed two different courses: either these genes were lost in CDC1551 and M. bovis AF2122/97 or they were transposed into this genomic island solely in M. tuberculosis H37Rv.
|
Presence of Horizontal Transfer Islands in the "M. prototuberculosis" Genetic Groups
We also used Southern blotting to investigate the presence of orthologs of the genes of the inferred genomic islands in 8 "M. prototuberculosis" genetic groups (Gutierrez et al. 2005
), the M. bovis BCG Pasteur strain, and a clinical strain (Mt14323) of M. tuberculosis. The reference M. tuberculosis strain H37Rv was used as a positive control of hybridization (table 4). All the tested genes were present in "M. prototuberculosis" indicating that all horizontal transfer events occurred in the progenitor species of the M. tuberculosis complex. However the presence of the genes tested seems to be not constant in all the genetic groups of "M. prototuberculosis". For example, we detected Rv0301 only in groups F and H, and we failed to detect Rv1044-Rv1053c in groups A to D. Two different evolutionary scenarios can thus be proposed: either all these genes were acquired by the progenitor of "M. prototuberculosis" and then lost in some of its groups, or they were independently acquired by the different groups of the species "M. prototuberculosis" and then shuttled to the last common ancestor of the M. tuberculosis complex by intra-species recombination. Multiple events of intra-species recombination have been already demonstrated for housekeeping genes in "M. prototuberculosis" (Gutierrez et al. 2005). Moreover, in relation with the previous analysis with the members of the M. tuberculosis complex, some genomic islands seem to have undergone different rearrangements. Three types of islands can be described: islands with no apparent rearrangement, such as Rv2302-Rv2312 (gene Rv2307B in table 4); partially deleted islands, such as Rv2801c-Rv2827c (genes Rv2804c and Rv2819c) and mostly deleted islands, such as Rv1041c-Rv1055, corresponding to genes Rv1044 and Rv1053.
|
Putative Origin of Transferred Genes
We used the genomic signature, previously shown to be species-specific, to assign a putative origin to the 48 foreign genomic islands specific to tubercle bacilli. Thus, if the genomic islands have retained their foreign genomic characteristics, then it should be possible to search for species with a similar signature in our bank as the candidate species of origin of these horizontal transfers. We identified plausible donors (i.e. with a similar enough signature) for only 18 of the 48 genomic islands (see Methods section). Three major groups of donors were identified (figure 3): 1) Actinobacteria, including the genus Mycobacterium and the order Bifidobacteriales; 2) Proteobacteria, including the orders Burkholderiales, Pseudomonadales, Rhizobiales and Sphingomonadales and 3) viruses. Most of the possible donor species identified were Actinobacteria and the closest possible donors, in terms of similarity of signature, were Proteobacteria. The genomic islands detected can be clustered into four groups according to donor spectrum (table 5). The potential donors of the islands in the first group were mostly Proteobacteria, and those of the second group were mostly Mycobacteria and Proteobacteria. The third group was unusual in that the potential donors always included a bacteriophage. Finally, all the potential donors of the islands of the fourth group were Actinobacteria.
|
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
We investigated the contribution of HGT to the evolution of the tubercle bacilli, using parametric methods to detect horizontal transfer, this approach being more sophisticated than the comparisons of base composition generally used in the M. tuberculosis strain H37Rv genome description (Cole et al. 1998
). We found that the level of HGT in the genomes of the MT-complex was about average with respect to other studied species (Ochman, Lawrence and Groisman 2000; Garcia-Vallve et al. 2003; Nakamura et al. 2004; Dufraigne et al. 2005). The consensus method used here resulted in the detection of lower (Nakamura et al. 2004) or similar (Ochman, Lawrence and Groisman 2000; Garcia-Vallve et al. 2003; Dufraigne et al. 2005) levels of HGT to those reported previously. The regions of HGT detected in MT-complex genomes had a high G+C content, similar to that of the host MT-complex genomes. This feature probably hindered previous attempts to identify regions of HGT in MT-complex genomes (Cole et al. 1998; Fleischmann et al. 2002; Garnier et al. 2003). The selection of specific HGT of the MT-complex species occurred in 4 steps: on a first step an average of 6.2% of genes was detected as atypical in the 3 genomes studied (table 1). From this set of genes 153 were common to the 3 genomes of the MT-complex. Out of them, 76 were absent from the other sequenced mycobacterial genomes and were retained as MT-complex specific. In a last step, we extended these horizontal transfer regions with adjacent genes conserving synteny within the MT–complex genomes and with no matches in the other sequenced mycobacterial genomes. Those horizontal transfer regions correspond to 48 genomic islands MT-complex specific. The cumulative length of these 48 regions accounted for about 4.5% of the genomes of the MT-complex, indicating that these genomes underwent a substantial exchange of genetic material with other species in their environment during the course of speciation. If we add the regions of HGT identified here to the acquired genomic islands common to all Mycobacteria sequenced to date or specific to only one species, MT-complex genomes are among the top 25% of species in terms of HGT levels. About half the MT-complex-specific regions identified presented features typical of transferred regions (figure 1, table 3A). Indeed, we observed IS, repeats and tRNA at the boundaries of these regions, IS and repeated genes within these regions and the gene content of these regions differed from that of the rest of the genomes, with many "unknown" and virulence genes (table 3A&B). The combination of these genomic features with the detection of atypical genes and their presence in only MT-complex members of Mycobacteria supports the acquired nature of these regions.
Each of these specific regions is thought to have resulted from a single transfer event. However, not all the genes included in these genomic islands were identified as HGT. Indeed, genes in these regions undergo different selective pressures (Lawrence and Ochman 1997; Ochman, Lawrence and Groisman 2000). Some are subject to the same selective pressures as the host genome and gradually acquire the same characteristics, to the point that they become undetectable and present all the features of a host gene. Some genes are partially, but not entirely modified by selection pressure, rendering them more similar to the host genome but detectable by only one method. Finally, some genes conserve their original foreign characteristics and are readily detected by the consensus method. It is unclear whether these genes have retained their properties because they cannot be changed, as they are indispensable to their new host, or whether they are essentially useless and likely to be deleted in the near future.
Further support for the recent transfer of these regions is provided by their ongoing rearrangement. Tubercle bacilli are thought to have arisen about a million years ago (Gutierrez et al. 2005). If one cause of this speciation event was a substantial invasion of the genome by foreign sequences, then the detected regions may be considered young and still under host selection pressure. This selection pressure involves the deletion of unused parts of sequences or of sequences involved in the process of transfer itself (such as IS or repeats). These rearrangements may result in only one isolated gene from the original cluster remaining in the host genome (table 2). We observed a large proportion of rearranged regions in MT-complex genomes (figure 2), a collection of bacilli of the "M. prototuberculosis" species, and a clinical strain (table 4). All these findings are consistent with a recent speciation event in the MT-complex ancestor species, linked to an invasion of the genome by foreign sequences. This hypothesis has been proposed in previous studies (Brosch et al. 2002; Gutierrez et al. 2005; Rosas-Magallanes et al. 2006a) and is supported by the results presented here.
What role did the transferred genes play? Several M. tuberculosis genes belonging to the genomic islands identified in this study have been previously identified as virulence genes in various genetic screening systems (table 3B) (Sassetti and Rubin 2003; Pethe et al. 2004; Stewart et al. 2005; Rosas-Magallanes et al. 2006b). The Rv0986-8 operon, for instance, was characterized as playing a key role in the binding of M. tuberculosis to eukaryotic cells and trafficking within these cells in vitro (Pethe et al. 2004; Rosas-Magallanes et al. 2006b). This example illustrates the role played by the acquired genes in the development of tubercle bacilli pathogenicity in mammals. These findings strongly suggest a role for HGT in the emergence of the members of the MT-complex as major pathogens.
The use of the signature species-specificity (Deschavanne et al. 1999; Deschavanne et al. 2000; Dufraigne et al. 2005) made it possible to suggest a potential origin for horizontal transfers retaining their original characteristics. We first checked the similarity of the signature of genes included in the same inferred transferred region. The results obtained were consistent with the hypothesis that these genes/regions were transferred together, in the same HGT event (Supplementary figure 2). The species specificity of the signature can therefore be used to back-search for the species of origin of the transferred genes, or at least to provide some clue to the genus or family of the donor species, as signature conserves phylogenic relationships (Pride et al. 2003; Chapus et al. 2005). By comparing the signature of the detected HGT with a bank of species signatures, we were able to identify potential donors for 18 of the 48 inferred regions (figure 3). As previously shown (Dufraigne et al. 2005), the spectrum of donor species was MT-complex specific. This spectrum was dominated by Actinobacteria, including Mycobacteria, and Proteobacteria, represented by several families and including many pathogenic and soil bacteria. This provides a first clue to the origin of the virulence properties gained by the tubercle bacilli. We identified no very closely related donors for the recently transferred regions, confirming that these regions are under strong selective pressure. Alternatively, our bank may not contain the actual donor species, and may include only related species, as shown by similarity in signatures, but not a close relationship. Our bank of signatures based on GenBank represents only a small set (about 65,000 species) of existing prokaryotic species (at least millions of species) and this fact may explain why, in some cases, mycobacterium species are proposed as donors by default. A more detailed analysis (table 5) led to the classification of regions for which we had some indication as to the donor into four groups based on the potential origin of the HGT. This classification provides an indication of the minimum number of transfer events that took place during the MT-complex species emergence in addition to other events involved in speciation (Brosch et al. 2001; Fitzgerald and Musser 2001; Springer et al. 2004; Arnold 2007). It therefore reflects the complexity of the birth of a new species. In our analysis of the events for which it was possible to suggest an origin (table 5), one of the groups was found to contain only Actinobacteria as donor species. Two groups had Actinobacteria species as major donors but secondary sources were also identified: Proteobacteria for one group and Viruses for the other. For the fourth group, only Proteobacteria were identified as potential donor species. This result is consistent with the conclusions of previous work in MT-complex genomes attributing the origin of a horizontal transfer region associated with virulence to a -Proteobacterium (Rosas-Magallanes et al. 2006a).
In conclusion, the results reported here support the hypothesis that the tubercle bacilli emerged as major human and mammalian pathogens following a substantial invasion of the genome of the ancestral environmental mycobacterial species by foreign DNA from soil bacteria (Rhizobiales) and bacteria capable of colonizing animals, such as -Proteobacteria, which include a wide range of pathogens (Pseudomonadales, Burkholderiales). These results are consistent with those of a recent study dealing with HGT in M. avium sp. paratuberculosis (Marri et al. 2006). Future functional studies, using M. tuberculosis mutant strains, should investigate the precise role of the genomic islands identified here in Mycobacteria pathogenicity. It would also be interesting to evaluate the occurrence of these genomic islands in various present-day M. tuberculosis genotypes (Filliol et al. 2006) and in various groups of the pathogenic ancestral species "M. prototuberculosis."
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
M. marinum sequence data were produced by the Mycobacterium marinum Sequencing Group at the Sanger Institute and can be obtained from
ftp://ftp.sanger.ac.uk/pub/M_marinum. Preliminary M. smegmatis sequence data were obtained from the Institute for Genomic Research website at http://www.tigr.org.
We thank Marie Gon
alves for excellent technical assistance. This work was supported by the Pasteur Institute. V. R. M. holds fellowship from Conacyt and Fondation pour la Recherche Médicale.
|
Footnotes |
|---|
1 These authors share senior authorship.
2 Present address: O. Neyrolles Institut Pasteur, Unit of Mycobacterial Genetics 28 rue du Dr Roux, 75015 Paris, France Tel: +33 (0)1 45 68 88 40, Fax: +33 (0)1 45 68 88 43.
Jennifer Wernegreen, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgementsReferences |
|---|
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J Mol Biol. (1990) 215(3):403–410.[CrossRef][Web of Science][Medline]
Arnold C. Molecular evolution of Mycobacterium tuberculosis. Clinical Microbiology and Infection (2007) 13:120–128.[Web of Science][Medline]
Azad RK, Lawrence JG. Use of artificial genomes in assessing methods for atypical gene detection. PLoS Comput Biol. (2005) 1(6):e56.[CrossRef][Medline]
Blanc-Potard AB, Lafay B. MgtC as a horizontally-acquired virulence factor of intracellular bacterial pathogens: evidence from molecular phylogeny and comparative genomics. J Mol Evol. (2003) 57(4):479–486.[CrossRef][Web of Science][Medline]
Brosch R, Gordon SV, Marmiesse M, et al. A new evolutionary scenario for the Mycobacterium tuberculosis complex. P Natl Acad Sci USA (2002) 99(6):3684–3689.
Brosch R, Pym AS, Gordon SV, Cole ST. The evolution of mycobacterial pathogenicity: clues from comparative genomics. Trends Microbiol. (2001) 9(9):452–458.[CrossRef][Web of Science][Medline]
Camus JC, Pryor MJ, Médigue C, Cole ST. Re-annotation of the genome sequence of Mycobacterium tuberculosis H37Rv. Microbiology (2002) 148:2967–2973.
Canchaya C, Fournous G, Chibani-Chennoufi S, Dilmann ML, Brussow H. Phage as agents of lateral gene transfer. Curr Opin Microbiol. (2003) 6:417–424.[CrossRef][Web of Science][Medline]
Carbone A, Zinovyev A, Kepes F. Codon adaptation index as a measure of dominating codon bias. Bioinformatics (2003) 19(16):2005–2015.
Chapus C, Dufraigne C, Edwards S, Giron A, Fertil B, Deschavanne P. Exploration of phylogenetic data using a global sequence analysis method. BMC Evol Biol. (2005) 5:63–83.[CrossRef][Medline]
Cole ST, Brosch R, Parkhill J. (42 co-authors). Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature (1998) 393(6685):537–544.[CrossRef][Medline]
Cole ST, Eiglmeier K, Parkhill J. (44 co-authors). Massive gene decay in the leprosy bacillus. Nature (2001) 409:1007–1011.[CrossRef][Medline]
Deschavanne P, Giron A, Vilain J, Dufraigne C, Fertil B. Genomic signature is preserved in short DNA fragments. In: BIBE 2000 IEEE International Symposium on bio-informatics & biomedical engineering (2000) Washington, D.C. USA.
Deschavanne P, Giron A, Vilain J, Fagot G, Fertil B. Genomic signature: characterization and classification of species assessed by Chaos Game Representation of sequences. Mol Biol and Evol. (1999) 16:1391–1399.
Dobrindt U, Hochhut B, Hentschel U, Hacker J. Genomic islands in pathogenic and environmental microorganisms. Nat Rev Microbiol. (2004) 2(5):414–424.[CrossRef][Web of Science][Medline]
Doolittle WF. Lateral genomics. Trends Cell Biol. (1999a) 9:M5–8.[CrossRef][Web of Science][Medline]
Doolittle WF. Phylogenetic classification and the universal tree. Science (1999b) 284:2124–2129.[CrossRef][Web of Science][Medline]
Dufraigne C, Fertil B, Lespinats S, Giron A, Deschavanne P. Detection and characterization of horizontal transfers in prokaryotes using genomic signature. Nucleic Acids Res. (2005) 33(1):e6.
Dutta C, Pan A. Horizontal gene transfer and bacterial diversity. J Biosciences (2002) 27:27–33.[CrossRef]
Eisen JA. Horizontal gene transfer among microbial genomes: new insights from complete genome analysis. Current Opinion in Genetics & Development (2000) 10:606–611.[CrossRef][Web of Science][Medline]
Filliol I, Motiwala AS, Cavatore M, et al, (25 co-authors). Global phylogeny of Mycobacterium tuberculosis based on single nucleotide polymorphism (SNP) analysis: insights into tuberculosis evolution, phylogenetic accuracy of other DNA fingerprinting systems, and recommendations for a minimal standard SNP set. J Bacteriol. (2006) 188(2):759–772.
Fitzgerald JR, Musser JM. Evolutionary genomics of pathogenic bacteria. Trends Microbiol. (2001) 9(11):547–53.[CrossRef][Web of Science][Medline]
Fleischmann RD, Alland D, Eisen JA, et al, (26 co-authors). Whole-genome comparison of Mycobacterium tuberculosis clinical and laboratory strains. J Bacteriol. (2002) 184(19):5479–5490.
Frost LS, Leplae R, Summers AO, Toussaint A. Mobile genetic elements: the agents of open source evolution. Nature Rev Microbiol. (2005) 3:722–732.[CrossRef]
Garcia-Vallve S, Guzman E, Montero MA, Romeu A. HGT-DB: a database of putative horizontally transferred genes in prokaryotic complete genomes. Nucleic Acids Res. (2003) 31(1):187–189.
Garcia-Vallvé S, Romeu A, Palau J. Horizontal gene transfer in bacterial and archeal complete genomes. Genome Res. (2000) 10:1719–1725.
Garnier T, Eiglmeier K, Camus JC, et al, (19 co-authors). The complete genome sequence of Mycobacterium bovis. P Natl Acad Sci USA (2003) 100(13):7877–7882.
Gutierrez MC, Brisse S, Brosch R, Fabre M, Omais B, Marmiesse M, Supply P, Vincent V. Ancient origin and gene mosaicism of the progenitor of Mycobacterium tuberculosis. PLoS Pathog. (2005) 1(1):e5.[CrossRef][Medline]
Hacker J, Hochhut B, Middendorf B, Schneider G, Buchrieser C, Gottschalk G, Dobrindt U. Pathogenomics of mobile genetic elements of toxigenic bacteria. Int J Med Microbiol. (2004) 293:453–461.[CrossRef][Web of Science][Medline]
Hacker J, Kaper JB. Pathogenicity islands and the evolution of pathogenic microbes (2002) Berlin: Springer-Verlag.
Jeffrey HJ. Chaos game visualization of sequences. Comp Graph. (1992) 16(1):25–33.[CrossRef]
Karlin S. Detecting anomalous gene clusters and pathogenicity islands in diverse bacterial genomes. Trends in Microbiol. (2001) 9(7):335–343.[CrossRef][Web of Science][Medline]
Karlin S, Mrazek J, Campbell AM. Codon usages in different gene classes of the Escherichia coli genome. Mol. Microbiol. (1998) 29(6):1341–1355.[CrossRef][Web of Science][Medline]
Kinsella RJ, Fitzpatrick DA, Creevey CJ, McInerney JO. Fatty acid biosynthesis in Mycobacterium tuberculosis: lateral gene transfer, adaptive evolution, and gene duplication. P Natl Acad Sci USA (2003) 100(18):10320–10325.
Lawrence JG, Ochman H. Amelioration of bacterial genomes: rates of change and exchange. J of Mol Evol. (1997) 44:383–397.[CrossRef]
Lawrence JG, Ochman H. Molecular archaeology of the Escherichia coli genome. P Natl Acad Sci USA (1998) 95:9413–9417.
Lawrence J, Ochman H. Reconciling the many faces of lateral gene transfer. Trends Microbiol. (2002) 10:1–4.[CrossRef][Web of Science][Medline]
Le Dantec C, Winter N, Gicquel B, Vincent V, Picardeau M. Genomic sequence and transcriptional analysis of a 23-kilobase mycobacterial linear plasmid: evidence for horizontal transfer and identification of plasmid maintenance systems. J Bacteriol. (2001) 183(7):2157–2164.
Lecointre G, Rachdi L, Darlu P, Denamur E. Escherichia coli molecular phylogeny using the incongruence length difference test. Mol Biol Evol. (1998) 15:1685–1695.[Abstract]
Li L, Bannantine JP, Zhang Q, Amonsin A, May BJ, Alt D, Banerji N, Kanjilal S, Kapur V. The complete genome sequence of Mycobacterium avium subspecies paratuberculosis. P Natl Acad Sci USA (2005) 102:12344–12349.
Mariani F, Piccolella E, Colizzi V, Rappuoli R, Gross R. Characterization of an IS-like element from Mycobacterium tuberculosis. J Gen Microbiol. (1993) 139(8):1767–1772.
Marri PR, Bannantine JP, Paustian ML, Golding GB. Lateral gene transfer in Mycobacterium avium subspecies paratuberculosis. Can J Microbiol. (2006) 52:560–569.[CrossRef][Web of Science][Medline]
Martin C, Timm J, Rauzier J, Gomez-Lus R, Davies J, Gicquel B. Transposition of an antibiotic resistance element in mycobacteria. Nature (1990) 345(6277):739–743.[CrossRef][Medline]
Maynard-Smith J, Smith NH. Detecting Recombination from gene trees. Mol Biol Evol. (1998) 15:590–599.[Abstract]
Medigue C, Rouxel T, Vigier P, Henaut Danchin A. Evidence for horizontal gene transfer in Escherichia coli speciation. J Mol Biol. (1991) 222:851–856.[CrossRef][Web of Science][Medline]
Mrazek J, Karlin S. Detecting alien genes in bacterial genomes. Ann NY Acad Sci. (1999) 870:314–329.[CrossRef][Web of Science][Medline]
Nakamura Y, Itoh T, Matsuda H, Gojobori T. Biased biological functions of horizontally transferred genes in prokaryotic genomes. Nature Genetics (2004) 36:760–766.[CrossRef][Web of Science][Medline]
Ochman H, Elwyn S, Moran NA. Calibrating bacterial evolution. P Natl Acad Sci USA (1999) 96:12638–12643.
Ochman H, Lawrence JG, Groisman EA. Lateral gene transfer and the nature of bacterial innovation. Nature (2000) 405(6784):299–304.[CrossRef][Medline]
Pethe K, Swenson DL, Alonso S, Anderson J, Wang C, Russell DG. Isolation of Mycobacterium tuberculosis mutants defective in the arrest of phagosome maturation. P Natl Acad Sci USA (2004) 101(37):13642–13647.
Picardeau M, Bull TJ, Vincent V. Identification and characterization of IS-like elements in Mycobacterium gordonae. FEMS Microbiol Lett. (1997) 154(1):95–102.[CrossRef][Web of Science][Medline]
Pride DT, Meinersmann RJ, Wassenaar TM, Blaser MJ. Evolutionary implications of microbial genome tetranucleotide frequency biases. Genome Res. (2003) 13(2):145–158.
Ragan MA. On surrogate methods for detecting lateral gene transfer. FEMS Microbiology letters (2001) 201:187–191.[CrossRef][Web of Science][Medline]
Rosas-Magallanes V, Deschavanne P, Quintana-Murci L, Brosch R, Gicquel B, Neyrolles O. Horizontal transfer of a virulence operon to the ancestor of Mycobacterium tuberculosis. Mol Biol Evol. (2006a) 23:1129–1135.
Rosas-Magallanes V, Stadthagen-Gomez G, Rauzier J, et al, (11 co-authors). Signature-tagged Transposon Mutagenesis Identifies Novel Mycobacterium tuberculosis Genes involved in the Parasitism of Human Macrophages. Infect Immun. (2006b) 75(1):504–507.[Web of Science][Medline]
Ruiting L, Reeves PR. Gene transfer is a major factor in bacterial evolution. Mol Biol Evol. (1996) 13:47–55.[Medline]
Sassetti CM, Rubin EJ. Genetic requirements for mycobacterial survival during infection. P Natl Acad Sci USA (2003) 100(22):12989–12994.
Saves I, Westrelin F, Daffe M, Masson JM. Identification of the first eubacterial endonuclease coded by an intein allele in the pps1 gene of mycobacteria. Nucleic Acids Res. (2001) 29(21):4310–4318.
Smith I. Mycobacterium tuberculosis pathogenesis and molecular determinants of virulence. Clin Microbiol Rev. (2003) 16(3):463–496.
Smith MW, Feng DF, Doolittle RF. Evolution by acquisition: the case for horizontal gene transfers. Trends in Biol Sci. (1992) 17:489–493.[CrossRef]
Springer B, Sander P, Sedlacek L, Hardt WD, Mizrahi V, Schar P, Bottger EC. Lack of mismatch correction facilitates genome evolution in mycobacteria. Mol Microbiol. (2004) 53(6):1601–1609.[CrossRef][Web of Science][Medline]
Stewart GR, Patel J, Robertson BD, Rae A, Young DB. Mycobacterial mutants with defective control of phagosomal acidification. PLoS Pathog. (2005) 1(3):e33.[CrossRef]
Stinear TP, Mve-Obiang A, Small PL, et al, (15 co-authors). Giant plasmid-encoded polyketide synthases produce the macrolide toxin of Mycobacterium ulcerans. P Natl.Acad Sci USA (2004) 101(5):1345–1349.[CrossRef]
Stinear TP, Seemann T, Pidot S, et al, (23 co-authors). Reductive evolution and niche adaptation inferred from the genome of Mycobacterium ulcerans, the causative agent of Buruli ulcer. Genome Res. (2007) 17(2):192–200.
Syvanen M. Horizontal gene transfer: evidence and possible consequences. Annu Rev Genet. (1994) 28:237–261.[Web of Science][Medline]
Thomas CM, Nielsen K. Mechanisms of, and barriers to, horizontal gene transfer between bacteria. Nature Rev Microbiol. (2005) 3:711–721.[CrossRef]
Wolf YI, Aravind L, Grishin NV, Koonin EV. Evolution of aminoacyl-tRNA synthetases - Analysis of unique domain architectures and phylogenetic trees reveals a complex history of horizontal gene transfer events. Genome Res. (1999) 9:689–710.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. Castellanos, A. Aranaz, K. A. Gould, R. Linedale, K. Stevenson, J. Alvarez, L. Dominguez, L. de Juan, J. Hinds, and T. J. Bull Discovery of Stable and Variable Differences in the Mycobacterium avium subsp. paratuberculosis Type I, II, and III Genomes by Pan-Genome Microarray Analysis Appl. Envir. Microbiol., February 1, 2009; 75(3): 676 - 686. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. S. Snyder, S. McGowan, M. Rogers, E. Duro, E. O'Farrell, and N. J. Saunders The Repertoire of Minimal Mobile Elements in the Neisseria Species and Evidence That These Are Involved in Horizontal Gene Transfer in Other Bacteria Mol. Biol. Evol., December 1, 2007; 24(12): 2802 - 2815. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||