Genome-Wide Analysis of Synonymous Single Nucleotide Polymorphisms in Mycobacterium tuberculosis Complex Organisms: Resolution of Genetic Relationships Among Closely Related Microbial Strains

Corresponding author: James M. Musser, Rocky Mountain Laboratories, National Institute of Allergy and Infectious Diseases, National Institutes of Health, 903 S. 4th St., Hamilton, MT 59840., jmusser{at}niaid.nih.gov (E-mail)

Communicating editor: Y.-X. FU

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Several human pathogens (e.g., Bacillus anthracis, Yersinia pestis, Bordetella pertussis, Plasmodium falciparum, and Mycobacterium tuberculosis) have very restricted unselected allelic variation in structural genes, which hinders study of the genetic relationships among strains and strain-trait correlations. To address this problem in a representative pathogen, 432 M. tuberculosis complex strains from global sources were genotyped on the basis of 230 synonymous (silent) single nucleotide polymorphisms (sSNPs) identified by comparison of four genome sequences. Eight major clusters of related genotypes were identified in M. tuberculosis sensu stricto, including a single cluster representing organisms responsible for several large outbreaks in the United States and Asia. All M. tuberculosis sensu stricto isolates of previously unknown phylogenetic position could be rapidly and unambiguously assigned to one of the eight major clusters, thus providing a facile strategy for identifying organisms that are clonally related by descent. Common clones of M. tuberculosis sensu stricto and M. bovis are distinct, deeply branching genotypic complexes whose extant members did not emerge directly from one another in the recent past. sSNP genotyping rapidly delineates relationships among closely related strains of pathogenic microbes and allows construction of genetic frameworks for examining the distribution of biomedically relevant traits such as virulence, transmissibility, and host range.

A common theme that has emerged from molecular population genetic analysis of pathogenic bacteria is that biomedically relevant traits, such as host range and virulence, are nonrandomly distributed among phylogenetic lineages (MUSSER 1996 ). Trait-lineage relationships exist, in part, because pathogenic bacterial species often have a level of genetic diversity far in excess of the variation present in higher eukaryotic organisms such as humans. Genetic variation in bacteria is due to differences in gene content and nucleotide variation in or between structural genes. Differences in gene content are caused primarily by gene deletion or by acquisition of mobile elements such as plasmids and bacteriophages (OCHMAN et al. 2000 ). Allelic variation arises from random nucleotide mutation, sometimes followed by selection, and from horizontal gene transfer and intragenic recombination events (REID et al. 2001 ). Two classes of substitutions, referred to as synonymous and nonsynonymous single nucleotide polymorphisms (SNPs), can occur in genes that encode proteins (KIMURA 1983 ; SCHORK et al. 2000 ; GUT 2001 ). Nonsynonymous SNPs (nsSNPs) result in amino acid replacements and hence provide substrate for evolutionary selection. In contrast, synonymous SNPs (sSNPs) do not alter the structure of proteins and are evolutionarily neutral or nearly so (KIMURA 1983 ; SCHORK et al. 2000 ; GUT 2001 ). Because sSNPs are functionally neutral and easy to assay (KIMURA 1983 ; SCHORK et al. 2000 ; GUT 2001 ), they provide useful targets for large-scale molecular population genetic studies examining evolutionary relationships among bacterial strains, especially in strongly clonal species. Moreover, sSNP genotypes can be readily analyzed with computational tools developed over decades of population genetic research. In the postgenomic era, SNPs provide a simple and fast way to compare entire genomes in many bacterial strains.

Mycobacterium tuberculosis is the most successful human pathogen worldwide, responsible for 3 million deaths each year and extensive morbidity and mortality (WORLD HEALTH ORGANIZATION 1998). M. tuberculosis is a member of the M. tuberculosis complex, a group of five closely related "sibling" species [M. tuberculosis sensu stricto (s.s.), M. africanum, M. microti, M. bovis, and M. canettii] that cause tuberculosis in humans and animals (BAESS 1979 ; KAPUR et al. 1994 ; NOLTE and METCHOCK 1995 ; FEIZABADI et al. 1996 ; SREEVATSAN et al. 1997A ; MUSSER et al. 2000 ). Study of 56 genes in several hundred M. tuberculosis complex strains suggested that there is about one synonymous nucleotide substitution per 10,000 nucleotide sites (KAPUR et al. 1994 ; SREEVATSAN et al. 1997A ; MUSSER et al. 2000 ). Restricted allelic variation limits the utility of multilocus sequence analysis (MAIDEN et al. 1998 ) for estimating genetic relationships among M. tuberculosis strains and for studying relationships between strain genotype and patient phenotype.

Although many molecular methods have been used for categorizing M. tuberculosis strains (ROSS et al. 1992 ; GROENEN et al. 1993 ; VAN EMBDEN et al. 1993 ; FROTHINGHAM and MEEKER-O'CONNELL 1998 ; KUREPINA et al. 1998 ; RAMASWAMY and MUSSER 1998 ; KREMER et al. 1999 ; SOINI et al. 2000 ; MAZARS et al. 2001 ; SUPPLY et al. 2001 ), phylogenetic relationships among this group of organisms have not been resolved. nsSNPs located in codon 463 of the katG gene and codon 95 of the gyrA gene permit all M. tuberculosis strains to be assigned to three principal genetic groups (SREEVATSAN et al. 1997A ). The distribution of these nsSNPs provided evidence supporting the hypothesis that principal genetic group 1 is ancestral to group 2 and that group 2 is ancestral to group 3. However, because the original genetic groupings are based on only two nsSNPs, the evolutionary hypothesis of M. tuberculosis outlined by SREEVATSAN et al. 1997A requires additional investigation. Moreover, the worldwide threat of M. tuberculosis to human health emphasizes the need to develop rapid methods to identify genetic relationships among all strains, especially among organisms responsible for large infection outbreaks, drug-resistant strains, and strains that cause severe clinical disease. The availability of genome sequences of two M. tuberculosis strains (COLE et al. 1998 ; http://www.tigr.org) and of partial genome sequences of a third M. tuberculosis strain (http://www.tigrblast.tigr.org) and an M. bovis strain (http://www.sanger.ac.uk/Projects/M_bovis) facilitated resolution of genetic relationships among M. tuberculosis complex organisms by large-scale sSNP analysis. Our results indicate that sSNP genotyping permits all strains of closely related pathogens to be assigned to lineages that are identical or related by descent and removes a critical barrier to population-based studies of the relationships between strain genotypes and patient phenotypes.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacterial strains:
M. tuberculosis complex isolates (n = 432) were recovered worldwide from patients with pulmonary and extrapulmonary tuberculosis and from diseased animals (see Table 1 in the supplementary material at http://www.genetics.org/supplemental). Many of the organisms studied were characterized previously by IS6110 profiling and spoligotyping (BIFANI et al. 1996 ; SREEVATSAN et al. 1997A ; KUREPINA et al. 1998 ; BIFANI et al. 1999 ; YAGANEHDOOST et al. 1999 ; SOINI et al. 2000 , SOINI et al. 2001 ; MATHEMA et al. 2002 ). The strain collection includes 112 isolates representing drug-susceptible and multidrug-resistant strains causing outbreaks in New York City, New Jersey, and Houston, and abundantly occurring strains (n = 104) of the same IS6110 profiles isolated from patients in the United States, Latin America, Africa, and Asia. The 112 isolates represent the known breadth of IS6110 and spoligotype diversity present in the species (see Table 1 in the supplementary material at http://www.genetics.org/supplemental) and will be referred to as the core group of isolates. We also analyzed 90 isolates recovered from patients in Houston enrolled in a population-based study of tuberculosis as well as five pairs of strains cultured from patients with reactivation tuberculosis. These 90 isolates include all organisms with low copy numbers of IS6110 (five or fewer copies) and all strains with unique IS6110 profiles obtained over a 6-month period.

M. bovis isolates (n = 103) were recovered from 11 host species (cows, pigs, possum, water buffalo, deer, badger, elk, eland, yak, bison, and humans) in 14 countries (United States, Canada, Tanzania, Spain, Australia, New Zealand, Sweden, Wales, Ireland, England, the former Soviet Union, Iran, Japan, and Vietnam). Isolates of M. africanum (n = 8), M. microti (n = 4), and M. canettii (n = 1) also were studied (see Table 2 in the supplementary material at http://www.genetics.org/supplemental). All isolates were assigned to species on the basis of conventional biochemical test results (NOLTE and METCHOCK 1995 ).

Identification and verification of sSNPs:
Several strategies were used to identify sSNPs present in four M. tuberculosis complex genomes (Fig 1). First, the genomes of M. tuberculosis strains H37Rv (COLE et al. 1998 ) and CDC1551 (VALWAY et al. 1998 ; http://www.tigr.org) were aligned with CrossMatch (P. Green, http://www.genome.washington.edu/UWGC/analysistools/Swat.cfm) to identify polymorphic sites, which were differentiated as sSNPs and nsSNPs on the basis of the inferred protein sequences. All sSNPs were verified by sequencing the relevant gene region in strains TN587, CDC1551, and H37Rv, representatives of principal genetic groups 1, 2, and 3, respectively (SREEVATSAN et al. 1997A ). To identify additional sSNPs, the two available M. tuberculosis genome sequences were also aligned with a partially completed M. bovis genome sequence (http://www.sanger.org). sSNPs were differentiated from nsSNPs by comparison of the inferred protein sequences and verified by sequencing the relevant gene region in strains TN587, CDC1551, and H37Rv and in M. bovis strains TN10130 and TN12465. Similar procedures were used to identify and verify sSNPs between H37Rv, CDC1551, and the partially completed genome of M. tuberculosis strain 210 (a principal genetic group 1 organism; BEGGS et al. 2000 ). The PCR primers used to amplify and sequence the gene segments containing putative sSNPs are listed in Table 3 in the supplementary material at http://www.genetics.org/supplemental.

View larger version (27K): In this window In a new window Download PPT slide   Figure 1. Schematic showing distribution of sSNPs analyzed. Complete genome sequences are available for M. tuberculosis s.s. strains H37Rv (COLE et al. 1998 ) and CDC1551 (http://www.tigrblast.tigr.org) and partial genome sequences are available for M. tuberculosis s.s. strain 210 (http://www.tigrblast.tigr.org) and M. bovis AF2122/97 (http://www.sanger.ac.uk/Projects/M_bovis). sSNPs randomly distributed around the chromosome were analyzed. All sSNPs were sequence verified.

sSNP analysis:
The SNaPshot primer extension method (Applied Biosystems, Foster City, CA) was used to analyze sSNPs with the primers listed in Table 3 in the supplementary material at http://www.genetics.org/supplemental. The SNaPshot technique is based on addition of a single fluorescently labeled ddNTP to the 3' end of an unlabeled specific oligonucleotide primer that hybridizes to its target DNA located contiguous to the SNP of interest. Labeling reactions were performed with the SNaPshot kit according to the instructions supplied by the manufacturer, except that the reaction mixture consisted of 2 µl of PCR template, 4 µl of SNaPshot ready reaction premix, 4 µl of dH₂O, and 1 µl 0.2 mM primer. Data were generated with an ABI Prism 3700 automated sequencer (Applied Biosystems).

IS6110 RFLP profiling and spoligotyping:
IS6110 restriction fragment length polymorphism (RFLP) profiling and spoligotyping were conducted by standard methods (VAN EMBDEN et al. 1993 ; VAN SOOLINGEN et al. 1995 ).

Data analysis:
The sSNP data were concatenated, resulting in one character string (nucleotide sequence) for each strain. Phylogenetic and molecular evolution analyses were conducted with MEGA version 2.1 (http://www.megasoftware.net/), using the neighbor-joining method with 1000 bootstrap replicates and distance calculated using the number of different sSNPs.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Confirmation of restricted allelic variation in M. tuberculosis:
Comparison of the 4.4-Mb genome sequences of M. tuberculosis strain H37Rv and CDC1551 identified only 900 SNPs located in open reading frames, confirming the restricted level of structural gene sequence variation reported previously (KAPUR et al. 1994 ; SREEVATSAN et al. 1997A ; MUSSER et al. 2000 ). Approximately 65% of the SNPs were nsSNPs, an unanticipated result given that sSNPs are expected to outnumber nsSNPs except under situations involving positive selection (KIMURA 1983 ). Ninety-one percent of all putative sSNPs studied were confirmed to be true sSNPs (not genome-sequencing artifacts) by sequencing genomic DNA purified from strains H37Rv and CDC1551 (data not shown). A total of 148 sequence-confirmed sSNPs distributed randomly around the chromosome were used for initial analysis.

M. tuberculosis s.s. isolates comprise eight major clusters of genotypes:
A core group of 112 isolates representing the breadth of diversity in the species, as assessed by IS6110 copy number and profile, spoligotype, and principal genetic group, was genotyped for the 148 sSNPs. Of the 148 sSNPs, 35 were polymorphic only in strain H37Rv or CDC1551. The topology of the phylogenetic tree generated from the data set that included all 148 sSNPs was virtually identical to the tree based on the 113 sSNPs. The effect of including the 35 sSNPs present only in strains H37Rv or CDC1551 was to accentuate the branch distance between these 2 isolates and related genotypes (data not shown). The phylogenetic tree generated on the basis of the remaining 113 sSNPs in the core group of 112 strains contained eight major clusters of related genotypes arbitrarily designated I–VIII and marked by high bootstrap values (Fig 2). Clusters I and II contain all isolates of principal genetic group 1; clusters III–VI are composed of isolates of principal genetic group 2; and clusters VII and VIII contain all isolates of principal genetic group 3. Hence, principal genetic groups and major sSNP genotypic clusters correlated strongly. The sSNP analysis also demonstrated that principal genetic group 3 organisms (clusters VII and VIII) arose from a principal group 2 precursor cell and identified the ancestor as a cluster VI-related organism.

View larger version (22K): In this window In a new window Download PPT slide   Figure 2. Phylogenetic tree showing estimates of genetic relationships among 112 M. tuberculosis s.s. isolates, based on analysis of 113 sSNPs. The tree was generated using the neighbor-joining method with 1000 bootstrap replicates and distance calculation using the number of different sSNP loci (http://www.megasoftware.net/). The eight major clusters were arbitrarily designated I–VIII. The group designations refer to the three principal genetic groups of M. tuberculosis described previously on the basis of SNPs located at codon 463 in the katG gene and codon 95 in the gyrA gene (SREEVATSAN et al. 1997A ). Each strain is listed with a principal genetic group number (G1, G2, or G3); the strain number; and the number of IS6110 hybridizing bands (one- or two-digit number).

Analysis of sSNPs identified by comparing the genome sequence of strains H37Rv and CDC1551 with a principal genetic group 1 M. tuberculosis strain and an M. bovis strain:
Previous studies (SREEVATSAN et al. 1997A ; MUSSER et al. 2000 ; SOINI et al. 2000 , SOINI et al. 2001 ) of genetic variation among M. tuberculosis strains indicated that isolates of principal genetic group 1 were more polymorphic than suggested by analysis of the 113 sSNPs. For example, principal genetic group 1 organisms had the greatest breadth of IS6110 copy number (SREEVATSAN et al. 1997A ) and substantial variation in spoligotype pattern (SOINI et al. 2000 , SOINI et al. 2001 ). One hypothesis to explain the relative lack of sSNPs identified among principal genetic group 1 organisms is that the strategy used to identify sSNPs minimized the likelihood of detecting nucleotide sites that are polymorphic among group 1 organisms. This would be the case if principal genetic group 1 organisms accumulated many sSNPs after diverging from the last common ancestor giving rise to strain CDC1551 (principal genetic group 2) and strain H37Rv (principal genetic group 3). In essence, the relative lack of sSNP diversity among principal group 1 organisms could be an artifact caused by identifying SNPs solely on the basis of comparing genomes from principal genetic group 2 (CDC1551) and group 3 (H37Rv) organisms.

Principal genetic group 1 organisms have been responsible for many large tuberculosis outbreaks in the United States and for a very large proportion of cases in the former Soviet Union and many Asian countries (BIFANI et al. 1996 , BIFANI et al. 1999 ; SREEVATSAN et al. 1997A ; ANH et al. 2000 ; BEGGS et al. 2000 ; PARK et al. 2000 ; SOINI et al. 2000 ; CHAN et al. 2001 ). Resolution of clonal relationships among principal genetic group 1 organisms is important because of their medical importance worldwide. To test the hypothesis that principal genetic group 1 organisms carry additional sSNP diversity, the two completed M. tuberculosis genome sequences (H37Rv and CDC1551) were aligned and compared, in turn, with partially completed genome sequences of strain 210 (principal genetic group 1; http://www.tigrblast.tigr.org) and of M. bovis strain AF2122/97 (http://www.sanger.ac.uk/Projects/M_bovis). Analysis of a random sample of 104 sequence-verified sSNPs among 91 strains (all group 1 organisms among the core group of 112 strains and representative isolates of each major group 2 and group 3 lineage and a selection of M. bovis isolates) confirmed the hypothesis. In addition, several distinct subpopulations of principal genetic group 1 organisms were resolved (Fig 3).

View larger version (37K): In this window In a new window Download PPT slide   Figure 3. Comparison of two phylogenetic trees showing the relationships among principal genetic group 1 organisms. The phylogenetic trees were constructed using (A) 113 sSNPs obtained from the comparison of genomes of principal genetic group 2 (strain CDC1551) and group 3 (strain H37Rv) organisms and (B) these 113 sSNPs plus 104 new sSNPs identified by comparing the above-described genomes (CDC1551 and H37Rv) with the partial genome sequences of M. tuberculosis group 1 strain 210 and M. bovis strain AF2122/97. Both trees were generated using the neighbor-joining method with 1000 bootstrap replicates and distance was calculated using the number of sSNPs (http://www.megasoftware.net/). The IS6110 RFLP profile and copy number are indicated next to each strain. The figure shows that the lack of sSNP diversity among principal group 1 organisms (A) is caused by identifying SNPs on the basis of comparing genomes of principal genetic group 2 and 3 organisms. Inclusion of 104 additional sSNPs found by comparing M. tuberculosis groups 2 and 3 with M. tuberculosis group 1 and M. bovis genomes reveals genetic subpopulations of principal genetic group 1 with similar or identical IS6110 RFLP profiles (W, 003, 033; 002, 022, N2; 007, 015, KY, AM).

Phylogenetically informative sSNPs for high-throughput strain genotyping and population-based studies of tuberculosis biology:
Population-based studies of M. tuberculosis epidemiology have been hindered because at least 50% of strains have unique or low-copy-number IS6110 RFLP profiles (SMALL et al. 1993 ; BISHAI et al. 1998 ; LOCKMAN et al. 2001 ) and hence their genetic relationships to other strains are not known. This problem has greatly limited the ability to study correlations between strain genotypes and patient phenotypes, because data for at least 50% of patients are uninformative.

To test the hypothesis that this barrier to population-based studies can be removed by analysis of sSNPs, we identified 27 highly informative sSNPs and characterized 90 isolates from Houston. These 27 sSNPs were chosen such that they proportionally represent each of the eight genetic clusters and all major subclusters. Moreover, analysis of the 112 core strains for these 27 sSNPs produced a phylogenetic tree with the same overall topology as the comprehensive sSNP data set (data not shown). The 90 isolates represent a 6-month strain sample that included all organisms with low copy numbers of IS6110 (five or fewer copies) and all strains with unique IS6110 RFLP profiles. All 90 isolates could be assigned readily to one of the eight major phylogenetic clusters (data not shown). Hence, sSNP genotyping successfully resolved the phylogenetic position of these organisms, thereby facilitating population-based studies of tuberculosis.

Phylogenetic clusters and IS6110 RFLP profiles:
Because of the considerable controversy regarding the amount of phylogenetic signal present in IS6110 profiles (FOMUKONG et al. 1997 ; SREEVATSAN et al. 1997A ; MAZARS et al. 2001 ), we next examined the distribution of IS6110 RFLP profiles among the strains assigned to each of the eight major phylogenetic clusters of M. tuberculosis. This insertion sequence is present in virtually all M. tuberculosis isolates and is highly polymorphic in copy number and position in the chromosome (VAN EMBDEN et al. 1993 ; FOMUKONG et al. 1997 ; SREEVATSAN et al. 1997A ; MAZARS et al. 2001 ). IS6110 profiling has been widely used in epidemiologic studies to infer genetic relationships among strains (SMALL et al. 1993 ; VAN EMBDEN et al. 1993 ; BIFANI et al. 1996 , BIFANI et al. 1999 ; BISHAI et al. 1998 ; YAGANEHDOOST et al. 1999 ; LOCKMAN et al. 2001 ). Several investigators have assigned M. tuberculosis s.s. to two primary phylogenetic lineages on the basis of low and high copy numbers of IS6110 (FOMUKONG et al. 1997 ; MAZARS et al. 2001 ), but the distribution of nsSNPs in katG and gyrA is at variance with this idea (SREEVATSAN et al. 1997A ).

Two important findings were revealed by our sSNP analysis. First, we found that there is no simple relationship between IS6110 copy number and phylogenetic lineage. For example, isolates with relatively few copies of IS6110 were present in very distinct phylogenetic lineages (Fig 4). Importantly, all isolates of principal genetic group 1 were genetically allied in two clusters by the sSNP analysis; however, these organisms contained from 0 to >20 copies of IS6110. Hence, IS6110 copy number alone is not phylogenetically informative.

View larger version (18K): In this window In a new window Download PPT slide   Figure 4. Relationships among IS6110 copy number and eight major genetic clusters of M. tuberculosis s.s. identified by analysis of 113 sSNPs. The tree is modified from the one shown in Fig 2. A sample of 112 core isolates of M. tuberculosis s.s. from global sources representing the span of species diversity was studied for 148 sSNPs randomly distributed around the chromosome. Thirty-five of the sSNPs were polymorphic in only strain H37Rv or CDC1551 ("monomorphic sSNPs") and were not considered further. The phylogenetic tree is based on the remaining 113 sSNPs and was generated using the neighbor-joining method with 1000 bootstrap replicates and distance calculation using the number of different sSNP loci (http://www.megasoftware.net/). The eight major clusters were arbitrarily designated I–VIII. The group designations refer to the three principal genetic groups of M. tuberculosis described previously on the basis of SNPs located at codon 463 in the katG gene and codon 95 in the gyrA gene (SREEVATSAN et al. 1997A ). The number of IS6110 hybridizing copies identified in strains assigned to each of the eight major genetic clusters of M. tuberculosis is shown.

Second, sSNP analysis permitted unambiguous identification of phylogenetic relationships among strains with similar or identical IS6110 profiles from diverse geographic areas. This could not be accomplished previously because of the possibility that similarity of the IS6110 RFLP profile among strains with no known direct epidemiologic link was caused by convergence to the same RFLP profile, rather than by identity by descent. We identified one subcluster containing strains with the W IS6110 profile (organisms responsible for many hundreds of cases of tuberculosis in New York City; BIFANI et al. 1996 ) and strains of the closely related IS6110 profiles 003 and 033, which together caused 167 cases of tuberculosis in Harris County, Texas, between 1995 and 2001 (Fig 3). Similarly, another subcluster contained strains with related IS6110 RFLP profiles designated 002, 022, and N2 that have caused abundant disease in Harris County and New York City (n = 262 cases). In another example, strains with IS6110 profiles designated 007 and 015 caused 169 cases of tuberculosis in Harris County between 1995 and 2001. sSNP genotyping demonstrated that these organisms were clonally related to strains from Siberia and New York City, with IS6110 patterns designated KY and AM, respectively (Fig 3).

Utility of sSNP genotyping for epidemiologic study purposes:
We next assessed the utility of sSNP genotyping for epidemiologic purposes. Strains grouped together on the basis of IS6110 RFLP profile and spoligotype, and known to be related by conventional epidemiologic investigation strategies such as contact tracing (YAGANEHDOOST et al. 1999 ), were studied for sSNP genotype. Strains assigned to 11 documented case clusters (11 distinct IS6110 RFLP profiles) were analyzed. Each IS6110 print group contained strains collected over at least 3 years. All isolates with identical IS6110 pattern and spoligotype had the same sSNP genotype (data not shown). To further investigate the epidemiologic utility of sSNP analysis, we examined the sSNP profile of five pairs of strains cultured from patients with reactivation tuberculosis. Paired isolates cultured from the same patient had sSNP genotypes identical to one another, as did six isolates of M. bovis Bacillus Calmette-Guérin with very different laboratory passage histories. Together, these results confirm that sSNP profiles are stable genetic markers and demonstrate that sSNP genotyping is useful for epidemiologic strain tracing.

sSNP genotyping of M. tuberculosis complex organisms:
The five sibling species of the M. tuberculosis complex are closely related (BAESS 1979 ; NOLTE and METCHOCK 1995 ), but overall estimates of phylogenetic relationships among them have not been clarified because appropriate molecular methods have not been available. To address this problem, we determined the sSNP genotypes of an additional 128 M. tuberculosis complex strains recovered from many host species and geographic localities, including organisms classified by conventional diagnostic methods (NOLTE and METCHOCK 1995 ) such as M. africanum (n = 8 isolates), M. microti (n = 4), M. canettii (n = 1), M. bovis (n = 103), and M. tuberculosis s.s. (n = 12; see Table 2 in the supplementary material at http://www.genetics.org/supplemental). sSNP analysis demonstrated that isolates of three of these non-M. tuberculosis s.s. sibling species are genetically allied with isolates of principal genetic group 1 M. tuberculosis (Fig 5). Many isolates of M. africanum are very closely related to isolates of M. tuberculosis s.s. with few copies of IS6110. Isolates of M. microti are allied with isolates of M. tuberculosis recovered from infected seals from several continents (COUSINS et al. 1993 ; LIEBANA et al. 1996 ; ZUMARRAGA et al. 1999 ) and with several isolates of M. africanum from humans (Fig 5). The results are consistent with limited data suggesting that tuberculosis in seals is caused by genetically distinct isolates of M. tuberculosis (COUSINS et al. 1993 ; LIEBANA et al. 1996 ; ZUMARRAGA et al. 1999 ). The M. microti-M. africanum-seal bacillus cluster of isolates comprises a very distinct sSNP group (bootstrap value = 99) that forms a phylogenetic bridge between M. tuberculosis s.s. (human specialist) and M. bovis (infecting predominantly nonhuman host species).

View larger version (16K): In this window In a new window Download PPT slide   Figure 5. Phylogenetic tree showing estimates of genetic relationships among 330 M. tuberculosis complex isolates, based on study of 90 sSNPs. The phylogenetic tree, generated using the neighbor-joining method with 1000 bootstrap replicates and distance calculated using the number of different sSNP loci (http://www.megasoftware.net/), was rooted with M. canettii, the most divergent member of the M. tuberculosis complex (VAN SOOLINGEN et al. 1997 ; KREMER et al. 1999 ; VAN EMBDEN et al. 2000 ; BROSCH et al. 2002 ). To more clearly show the evolutionary relationships occurring among members of the M. tuberculosis complex, certain branches of the phylogenetic tree were compressed: M. tuberculosis isolates assigned to cluster II–VIII (Fig 2) and M. bovis isolates showing little or no sSNPs divergence. M. tuberculosis and M. bovis are well differentiated from one another, and strains of M. africanum, M. microti, and M. tuberculosis isolated from seals are clearly located between these two members of the complex.

One genetic group of M. bovis warrants special comment. Four isolates recovered from cattle in Malawi are closely allied to, but distinct from, the cluster containing virtually all M. bovis isolates from several continents and many host species (Fig 5). The Malawi isolates have a distinct spoligotype and IS6110 profile compared to other M. bovis isolates (KREMER et al. 1999 ; data not shown). Importantly, these isolates have the allele of the pyrazinamidase gene (pncA) commonly present in M. tuberculosis s.s. (SREEVATSAN et al. 1997B ) and the 1973-bp variant of Rv3135 found in M. bovis isolates and many principal genetic group 1 organisms (MUSSER et al. 2000 ). The data are consistent with the hypothesis that these Malawi isolates represent an immediate relative of the great majority of extant M. bovis isolates causing disease in developed countries.

Inasmuch as sSNP genotyping provided new information about overall evolutionary relationships among four of the five M. tuberculosis complex organisms, we extended the investigation to M. canettii, the most divergent member of the complex (VAN SOOLINGEN et al. 1997 ; KREMER et al. 1999 ; VAN EMBDEN et al. 2000 ). The rooted tree revealed a deep branching of M. canettii (or a related last common ancestor) into two distinct lineages, one leading to M. tuberculosis s.s. and the second resulting in the M. microti-M. africanum-seal bacillus cluster (Fig 5). This latter cluster then gave rise to the large majority of organisms responsible for extant tuberculosis in animals in developed countries. In summary, sSNP genotyping resolved phylogenetic relationships among all five members of the M. tuberculosis complex.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Advantages of sSNP genotyping:
sSNPs afford many advantages for analysis of phylogenetic relationships among microbial strains, especially closely related clonal organisms such as the M. tuberculosis complex. Most or all sSNPs are selectively neutral and hence minimally subject to convergence, a process that can obscure or distort evolutionary relationships (KIMURA 1983 ). Binary data are obtained, which means that the information is readily amenable to storage, retrieval, analysis by personal computers equipped with simple software, and comparison between different laboratories and studies. Importantly, the considerable biomedical interest in human SNPs (SCHORK et al. 2000 ; DALY et al. 2001 ; GUT 2001 ; JOHNSON et al. 2001 ) means that microbial pathogen research will benefit extensively from the ongoing development and implementation of methods to index very large numbers of SNPs efficiently, inexpensively, and automatically (KWOK 2001 ). Because of the many advantages of using sSNP analysis for estimating genetic relationships among strains, and because of the public health and medical research need to classify strains into identical or closely similar genetic groups, it is likely that in the future M. tuberculosis complex strains will be routinely categorized on the basis of sSNP genotypes. This process will simplify population genetic analysis and epidemiologic investigations. Indeed, the cost of SNP genotyping is decreasing rapidly, suggesting that routine profiling of large samples of SNPs in many species of closely related pathogens will become feasible.

sSNP-based population genetic framework:
One goal of bacterial population genetic research is to understand the relationships between genetic diversity, clonal lineages, and biomedically relevant phenotypes such as virulence, transmissibility, host specialization, and evolutionary success (MUSSER 1996 ). The estimates of overall chromosomal relationships for all strains provided by sSNP genotyping makes possible the mapping of traits onto the M. tuberculosis phylogenetic tree, and this now can be done for all isolates, heretofore impossible.

IS6110 RFLP profiling (IS6110 fingerprinting) is used extensively to categorize strains for epidemiologic studies and to infer relationships among strains (SMALL et al. 1993 ; VAN EMBDEN et al. 1993 ; BIFANI et al. 1996 , BIFANI et al. 1999 ; BISHAI et al. 1998 ; YAGANEHDOOST et al. 1999 ; LOCKMAN et al. 2001 ). Strains with more than five copies of IS6110 and identical RFLP profiles are considered to be clonally related by descent. This view is supported by our sSNP analysis, but the converse is not always true. That is, strains with very different IS6110 copy numbers and RFLP patterns also may be clonally related by descent, as exemplified by strains typed as principal genetic group 1 (Fig 5). Importantly, sSNP genotyping clearly ruled out the idea that strains of M. tuberculosis s.s. with many IS6110 copies ("high-copy strains") and few IS6110 copies ("low-copy strains") are genetically distinct populations, thereby resolving a long-standing controversy (FOMUKONG et al. 1997 ; SREEVATSAN et al. 1997A ; MAZARS et al. 2001 ).

sSNP analysis provided new insight into genetic relationships among M. tuberculosis complex organisms. We found that isolates of M. tuberculosis s.s. and M. bovis are well differentiated from one another and located in distinct branches of the sSNP-based phylogenetic tree. Moreover, rooting of the tree with M. canettii revealed that strains of M. microti and M. tuberculosis isolates recovered from seals with tuberculosis and some M. africanum isolates were placed phylogenetically between M. tuberculosis s.s. and most M. bovis isolates. Thus, the present genome-wide sSNPs study adds some important insights to evolutionary relationships recently described by other authors (BROSCH et al. 2002 ; MOSTOWY et al. 2002 ). Taken together, these results reveal that repeatedly isolated clones of M. bovis causing disease in cattle and other animals in many developed countries and M. tuberculosis isolated worldwide are distinct, deeply branching genotypic complexes whose extant members did not emerge directly from one another in the recent past, as is generally thought to be the case (DIAMOND 1997 ; STEAD 1997 ).

sSNP genotyping for rapid characterization of strains of pathogenic microbes:
Many circumstances require rapid and unambiguous characterization of the genetic profile (chromosomal fingerprint) of microbial strains, including highly virulent pathogens. High-throughput sSNP genotyping is an attractive method for conducting analyses of this type because many hundreds to thousands of SNPs can be indexed in less than a few days. It is reasonable to anticipate that in a putative outbreak situation this magnitude of sSNP genotyping would be required.

Our study was greatly assisted by the availability of genome sequence data from four strains that represent distinct phylogenetic lineages of the M. tuberculosis complex. Were this not the case, the insights obtained would be far more circumscribed and the conclusions less robust. For example, analysis of sSNPs identified solely by comparison of the genomes of strains H37Rv and CDC1551 failed to reveal the complexity of relationships among and between principal genetic group 1 organisms and isolates of the other M. tuberculosis complex members. sSNP-based studies of genetic relationships in other pathogens should take into consideration the benefits afforded by selection of sSNPs on the basis of comparison of multiple divergent chromosomes. In this regard, data obtained from the increasing number of comparative genome sequencing projects in many pathogens will benefit subsequent sSNP studies, and more so if strains from distinct phylogenetic lineages are sequenced.

In summary, sSNP genotyping provided many new insights into phylogenetic relationships among members of the M. tuberculosis complex, which are closely related and heretofore could not be unambiguously assigned to distinct lineages. sSNP genotyping of strains of pathogenic microbes is a powerful new strategy that is generally applicable and especially useful for indexing genetic relationships among closely related organisms such as Yersinia pestis (ACHTMAN et al. 1999 ), Bacillus anthracis (KEIM et al. 1997 , KEIM et al. 2000 ), Bordetella species (MUSSER et al. 1986 ), M. leprae (CLARK-CURTISS and WALSH 1989 ), and Plasmodium falciparum (RICH et al. 1998 ; VOLKMAN et al. 2001 ), agents that have considerable detrimental impact on human health.

	FOOTNOTES

¹ These authors contributed equally to this work.
² Present address: Civil and Environmental Engineering, University of Washington, Seattle, WA 98195.

	ACKNOWLEDGMENTS

We thank T. Bowland and A. Lee for technical assistance; N. Williams-Bouyer for assistance with large strain collections; J. Driscoll for providing spoligotype data for some of the isolates; S. D. Reid for assistance with phylogenetic analysis; and V. Deretic, N. P. Hoe, H. Ochman, S. D. Reid, and K. Virtaneva for critical review of the manuscript. This research was supported in part by National Institutes Health contract N01-AO-02738.

Manuscript received May 7, 2002; Accepted for publication September 24, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ACHTMAN, M., K. ZURTH, G. MORELLI, G. TORREA, and A. GUIYOULE et al., 1999  Yersinia pestis, the cause of plague, is a recently emerged clone of Yersinia pseudotuberculosis.. Proc. Natl. Acad. Sci. USA 96:14043-14048.[Abstract/Free Full Text]

ANH, D. D., M. W. BORGDORFF, L. N. VAN, N. T. LAN, and T. VAN GORKOM et al., 2000  Mycobacterium tuberculosis Beijing genotype emerging in Vietnam. Emerg. Infect. Dis. 6:302-305.[Medline]

BAESS, I., 1979  Deoxyribonucleic acid relatedness among species of slowly-growing mycobacteria. Acta Pathol. Microbiol. Scand. Sect. B 87:221-226.[Medline]

BEGGS, M. L., K. D. EISENACH, and M. D. CAVE, 2000  Mapping of IS6110 insertion sites in two epidemic strains of Mycobacterium tuberculosis.. J. Clin. Microbiol. 38:2923-2928.[Abstract/Free Full Text]

BIFANI, P. J., B. B. PLIKAYTIS, V. KAPUR, K. STOCKBAUER, and X. PAN et al., 1996  Origin and interstate spread of a New York City multidrug-resistant Mycobacterium tuberculosis clone family. JAMA 275:452-457.[Abstract]

BIFANI, P. J., B. MATHEMA, Z. LIU, S. L. MOGHAZEH, and B. SHOPSIN et al., 1999  Identification of a W variant outbreak of Mycobacterium tuberculosis via population-based molecular epidemiology. JAMA 282:2321-2327.[Abstract/Free Full Text]

BISHAI, W. R., N. M. GRAHAM, S. HARRINGTON, D. S. POPE, and N. HOOPER et al., 1998  Molecular and geographic patterns of tuberculosis transmission after 15 years of directly observed therapy. JAMA 280:1679-1684.[Abstract/Free Full Text]

BROSCH, R., S. V. GORDON, M. MARMIESSE, P. BRODIN, and C. BUCHRIESER et al., 2002  A new evolutionary scenario for the Mycobacterium tuberculosis complex. Proc. Natl. Acad. Sci. USA 99:3684-3689.[Abstract/Free Full Text]

CHAN, M. Y., M. BORGDORFF, C. W. YIP, P. E. DE HAAS, and W. S. WONG et al., 2001  Seventy percent of the Mycobacterium tuberculosis isolates in Hong Kong represent the Beijing genotype. Epidemiol. Infect. 127:169-171.[Medline]

CLARK-CURTISS, J. E. and G. P. WALSH, 1989  Conservation of genomic sequences among isolates of Mycobacterium leprae.. J. Bacteriol. 171:4844-4851.[Abstract/Free Full Text]

COLE, S. T., R. BROSCH, J. PARKHILL, T. GARNIER, and C. CHURCHER et al., 1998  Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393:537-544.[Medline]

COUSINS, D. V., S. N. WILLIAMS, R. REUTER, D. FORSHAW, and B. CHADWICK et al., 1993  Tuberculosis in wild seals and characterisation of the seal bacillus. Aust. Vet. J. 70:92-97.[Medline]

DALY, M. J., J. D. RIOUX, S. F. SCHAFFNER, T. J. HUDSON, and E. S. LANDER, 2001  High-resolution haplotype structure in the human genome. Nat. Genet. 29:229-232.[Medline]

DIAMOND, J., 1997 Guns, Germs and Steel: The Fates of Human Societies. W. W. Norton, New York.

FEIZABADI, M. M., I. D. ROBERTSON, D. V. COUSINS, and D. J. HAMPSON, 1996  Genomic analysis of Mycobacterium bovis and other members of the Mycobacterium tuberculosis complex by isoenzyme analysis and pulsed-field gel electrophoresis. J. Clin. Microbiol. 34:1136-1142.[Abstract]

FOMUKONG, N., M. BEGGS, H. EL HAJJ, G. TEMPLETON, and K. EISENACH et al., 1997  Differences in the prevalence of IS6110 insertion sites in Mycobacterium tuberculosis strains: low and high copy number of IS6110.. Tuber. Lung Dis. 78:109-116.[Medline]

FROTHINGHAM, R. and W. A. MEEKER-O'CONNELL, 1998  Genetic diversity in the Mycobacterium tuberculosis complex based on variable numbers of tandem DNA repeats. Microbiology 144:1189-1196.[Medline]

GROENEN, P. M., A. E. BUNSCHOTEN, D. VAN SOOLINGEN, and J. D. VAN EMBDEN, 1993  Nature of DNA polymorphism in the direct repeat cluster of Mycobacterium tuberculosis: application for strain differentiation by a novel typing method. Mol. Microbiol. 10:1057-1065.[Medline]

GUT, I. G., 2001  Automation in genotyping of single nucleotide polymorphisms. Hum. Mutat. 17:475-492.[Medline]

JOHNSON, G. C., L. ESPOSITO, B. J. BARRATT, A. N. SMITH, and J. HEWARD et al., 2001  Haplotype tagging for the identification of common disease genes. Nat. Genet. 29:233-237.[Medline]

KAPUR, V., T. S. WHITTAM, and J. M. MUSSER, 1994  Is Mycobacterium tuberculosis 15,000 years old? J. Infect. Dis. 170:1348-1349.[Medline]

KEIM, P., A. KALIF, J. SCHUPP, K. HILL, and S. E. TRAVIS et al., 1997  Molecular evolution and diversity in Bacillus anthracis as detected by amplified fragment length polymorphism markers. J. Bacteriol. 179:818-824.[Abstract/Free Full Text]

KEIM, P., L. B. PRICE, A. M. KLEVYTSKA, K. L. SMITH, and J. M. SCHUPP et al., 2000  Multiple-locus variable-number tandem repeat analysis reveals genetic relationships within Bacillus anthracis.. J. Bacteriol. 182:2928-2936.[Abstract/Free Full Text]

KIMURA, M., 1983 The Neutral Theory of Molecular Evolution. Cambridge University Press, Cambridge, UK.

KREMER, K., D. VAN SOOLINGEN, R. FROTHINGHAM, W. H. HAAS, and P. W. HERMANS et al., 1999  Comparison of methods based on different molecular epidemiological markers for typing of Mycobacterium tuberculosis complex strains: interlaboratory study of discriminatory power and reproducibility. J. Clin. Microbiol. 37:2607-2618.[Abstract/Free Full Text]

KUREPINA, N. E., S. SREEVATSAN, B. B. PLIKAYTIS, P. J. BIFANI, and N. D. CONNELL et al., 1998  Characterization of the phylogenetic distribution and chromosomal insertion sites of five IS6110 elements in Mycobacterium tuberculosis: non-random integration in the dnaA-dnaN region. Tuber. Lung Dis. 79:31-42.[Medline]

KWOK, P. Y., 2001  Methods for genotyping single nucleotide polymorphisms. Annu. Rev. Genomics Hum. Genet. 2:235-258.[Medline]

LIEBANA, E., A. ARANAZ, B. FRANCIS, and D. COUSINS, 1996  Assessment of genetic markers for species differentiation within the Mycobacterium tuberculosis complex. J. Clin. Microbiol. 34:933-938.[Abstract]

LOCKMAN, S., J. D. SHEPPARD, C. R. BRADEN, M. J. MWASEKAGA, and C. L. WOODLEY et al., 2001  Molecular and conventional epidemiology of Mycobacterium tuberculosis in Botswana: a population-based prospective study of 301 pulmonary tuberculosis patients. J. Clin. Microbiol. 39:1042-1047.[Abstract/Free Full Text]

MAIDEN, M. C., J. A. BYGRAVES, E. FEIL, G. MORELLI, and J. E. RUSSELL et al., 1998  Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc. Natl. Acad. Sci. USA 95:3140-3145.[Abstract/Free Full Text]

MATHEMA, B., P. J. BIFANI, J. DRISCOLL, L. STEINLEIN, and N. KUREPINA et al., 2002  Identification and evolution of an IS6110 low-copy-number Mycobacterium tuberculosis cluster. J. Infect. Dis. 185:641-649.[Medline]

MAZARS, E., S. LESJEAN, A. L. BANULS, M. GILBERT, and V. VINCENT et al., 2001  High-resolution minisatellite-based typing as a portable approach to global analysis of Mycobacterium tuberculosis molecular epidemiology. Proc. Natl. Acad. Sci. USA 98:1901-1906.[Abstract/Free Full Text]

MOSTOWY, S., D. COUSINS, J. BRINKMAN, A. ARANAZ, and M. A. BEHR, 2002  Genomic deletions suggest a phylogeny for the Mycobacterium tuberculosis complex. J. Infect. Dis. 186:74-80.[Medline]

MUSSER, J. M., 1996  Molecular population genetic analysis of emerged bacterial pathogens: selected insights. Emerg. Infect. Dis. 2:1-17.[Medline]

MUSSER, J. M., E. L. HEWLETT, M. S. PEPPLER, and R. K. SELANDER, 1986  Genetic diversity and relationships in populations of Bordetella spp. J. Bacteriol. 166:230-237.[Abstract/Free Full Text]

MUSSER, J. M., A. AMIN, and S. RAMASWAMY, 2000  Negligible genetic diversity of Mycobacterium tuberculosis host immune system protein targets: evidence of limited selective pressure. Genetics 155:7-16.[Abstract/Free Full Text]

NOLTE, F. S., and B. METCHOCK, 1995 Mycobacterium, pp. 400–437 in Manual of Clinical Microbiology, edited by P. R. MURRAY, E. J. BARON, M. A. PFALLER, F. C. TENOVER and L. H. YOLKEN. American Society for Microbiology Press, Washington, DC.

OCHMAN, H., J. G. LAWRENCE, and E. A. GROISMAN, 2000  Lateral gene transfer and the nature of bacterial innovation. Nature 405:299-304.[Medline]

PARK, Y. K., G. H. BAI, and S. J. KIM, 2000  Restriction fragment length polymorphism analysis of Mycobacterium tuberculosis isolated from countries in the western Pacific region. J. Clin. Microbiol. 38:191-197.[Abstract/Free Full Text]

RAMASWAMY, S. and J. M. MUSSER, 1998  Molecular genetic basis of antimicrobial agent resistance in Mycobacterium tuberculosis: 1998 update. Tuber. Lung Dis. 79:3-29.[Medline]

REID, S. D., N. P. HOE, L. M. SMOOT, and J. M. MUSSER, 2001  Group A streptococcus: allelic variation, population genetics, and host-pathogen interactions. J. Clin. Invest. 107:393-399.[Medline]

RICH, S. M., M. C. LICHT, R. R. HUDSON, and F. J. AYALA, 1998  Malaria's Eve: evidence of a recent population bottleneck throughout the world populations of Plasmodium falciparum.. Proc. Natl. Acad. Sci. USA 95:4425-4430.[Abstract/Free Full Text]

ROSS, B. C., K. RAIOS, K. JACKSON, and B. DWYER, 1992  Molecular cloning of a highly repeated DNA element from Mycobacterium tuberculosis and its use as an epidemiological tool. J. Clin. Microbiol. 30:942-946.[Abstract/Free Full Text]

SCHORK, N. J., D. FALLIN, and J. S. LANCHBURY, 2000  Single nucleotide polymorphisms and the future of genetic epidemiology. Clin. Genet. 58:250-264.[Medline]

SMALL, P. M., R. W. SHAFER, P. C. HOPEWELL, S. P. SINGH, and M. J. MURPHY et al., 1993  Exogenous reinfection with multidrug-resistant Mycobacterium tuberculosis in patients with advanced HIV infection. N. Engl. J. Med. 328:1137-1144.[Abstract/Free Full Text]

SOINI, H., X. PAN, A. AMIN, E. A. GRAVISS, and A. SIDDIQUI et al., 2000  Characterization of Mycobacterium tuberculosis isolates from patients in Houston, Texas, by spoligotyping. J. Clin. Microbiol. 38:669-676.[Abstract/Free Full Text]

SOINI, H., X. PAN, L. TEETER, J. M. MUSSER, and E. A. GRAVISS, 2001  Transmission dynamics and molecular characterization of Mycobacterium tuberculosis isolates with low copy numbers of IS6110.. J. Clin. Microbiol. 39:217-221.[Abstract/Free Full Text]

SREEVATSAN, S., X. PAN, K. E. STOCKBAUER, N. D. CONNELL, and B. N. KREISWIRTH et al., 1997a  Restricted structural gene polymorphism in the Mycobacterium tuberculosis complex indicates evolutionarily recent global dissemination. Proc. Natl. Acad. Sci. USA 94:9869-9874.[Abstract/Free Full Text]

SREEVATSAN, S., X. PAN, Y. ZHANG, B. N. KREISWIRTH, and J. M. MUSSER, 1997b  Mutations associated with pyrazinamide resistance in pncA of Mycobacterium tuberculosis complex organisms. Antimicrob. Agents Chemother. 41:636-640.[Abstract]

STEAD, W. W., 1997  The origin and erratic global spread of tuberculosis. How the past explains the present and is the key to the future. Clin. Chest Med. 18:65-77.[Medline]

SUPPLY, P., S. LESJEAN, E. SAVINE, K. KREMER, and D. VAN SOOLINGEN et al., 2001  Automated high-throughput genotyping for study of global epidemiology of Mycobacterium tuberculosis based on mycobacterial interspersed repetitive units. J. Clin. Microbiol. 39:3563-3571.[Abstract/Free Full Text]

VALWAY, S. E., M. P. SANCHEZ, T. F. SHINNICK, I. ORME, and T. AGERTON et al., 1998  An outbreak involving extensive transmission of a virulent strain of Mycobacterium tuberculosis.. N. Engl. J. Med. 338:633-639.[Abstract/Free Full Text]

VAN EMBDEN, J. D., M. D. CAVE, J. T. CRAWFORD, J. W. DALE, and K. D. EISENACH et al., 1993  Strain identification of Mycobacterium tuberculosis by DNA fingerprinting: recommendations for a standardized methodology. J. Clin. Microbiol. 31:406-409.[Abstract/Free Full Text]

VAN EMBDEN, J. D., T. VAN GORKOM, K. KREMER, R. JANSEN, and B. A. VAN DER ZEIJST et al., 2000  Genetic variation and evolutionary origin of the direct repeat locus of Mycobacterium tuberculosis complex bacteria. J. Bacteriol. 182:2393-2401.[Abstract/Free Full Text]

VAN SOOLINGEN, D., L. QIAN, P. E. DE HAAS, J. T. DOUGLAS, and H. TRAORE et al., 1995  Predominance of a single genotype of Mycobacterium tuberculosis in countries of east Asia. J. Clin. Microbiol. 33:3234-3238.[Abstract]

VAN SOOLINGEN, D., T. HOOGENBOEZEM, P. E. DE HAAS, P. W. HERMANS, and M. A. KOEDAM et al., 1997  A novel pathogenic taxon of the Mycobacterium tuberculosis complex, Canetti: characterization of an exceptional isolate from Africa. Int. J. Syst. Bacteriol. 47:1236-1245.[Abstract/Free Full Text]

VOLKMAN, S. K., A. E. BARRY, E. J. LYONS, K. M. NIELSEN, and S. M. THOMAS et al., 2001  Recent origin of Plasmodium falciparum from a single progenitor. Science 293:482-484.[Abstract/Free Full Text]

WORLD HEALTH ORGANIZATION, 1998 World Health Report. World Health Organization, Geneva.

YAGANEHDOOST, A., E. A. GRAVISS, M. W. ROSS, G. J. ADAMS, and S. RAMASWAMY et al., 1999  Complex transmission dynamics of clonally related virulent Mycobacterium tuberculosis associated with barhopping by predominantly human immunodeficiency virus-positive gay men. J. Infect. Dis. 180:1245-1251.