Molecular Evolution of Mycoplasma capricolum subsp. capripneumoniae Strains, Based on Polymorphisms in the 16S rRNA Genes

Nucleotide sequence heterogeneity in 16S rRNA genes.Despite the reports of variability within 16S rRNA genes (see Introduction), the extent to which microheterogeneities occur has not been extensively studied. This is partially due to the limitations in the detection techniques commonly used for rDNA sequencing. In the present study, we found that the 16S rRNA genes can differ by >1% between the operons within an individual strain of M. capricolum subsp.capripneumoniae. We have shown for M. capricolumsubsp. capripneumoniae that the interoperon variation in, for instance, strains 4/2LC and GL102 is 1.1% as calculated from data in Table 2. We have also observed a strain-to-strain variation for a specific operon of up to 0.26% between several of the strains (Table2).

The importance of determining the intraspecific variability in 16S rRNA genes for developing reliable phylogenetic hypothesis has recently been pointed out (8, 41). The usefulness of characterizing microheterogeneities is also demonstrated in this work, emphasizing the importance of sequencing the 16S rRNA genes of all operons (if more than one is present) and investigating the strain-to-strain variations in the 16S rRNA genes. It is, of course, also extremely important to deposit sequences in GenBank under the correct species name and strain designation. Consequently, a bacterial DNA or protein sequence should preferably not be accepted in a data bank without a recognized strain designation.

Nucleotide sequences of the 16S rRNA genes of M. capricolum subsp. capripneumoniae.The 16S rRNA gene sequences of both operons from 20 strains of the species M. capricolum subsp. capripneumoniae listed in Table 1were subjected to bidirectional solid-phase rDNA sequencing, which has previously been shown to give very accurate data with the possibility of detecting heterogeneity, e.g., in viral populations and in 16S rRNA genes (see, e.g., references 16, 21, 28, 38, 40-42, and 44). Ten different 16S rRNA polymorphism patterns were found among the 20 M. capricolum subsp.capripneumoniae strains. Furthermore, a sequence length variation of one adenosine in a poly(A) region was found between the 16S rRNA genes of the two rRNA operons in nine of the strains. The poly(A) region is situated between nucleotides 444 and 449 and is not homologous to the poly(A) region with sequence length variations found in M. mycoides subsp. mycoides SC (between positions 1264 and 1270) reported previously (41). Length variations between the 16S rRNA genes of the members of the M. mycoides cluster have so far been observed only within poly(A) regions. These sequence length variations are probably caused by a process known as replication slippage (57).

A primary isolate of a mycoplasma may represent several clonal variants of a particular strain. A sample was therefore also analyzed for clonal variants with respect to 16S rRNA gene sequences among the M. capricolum subsp. capripneumoniae isolates. The 16S rRNA sequences of both operons obtained from a sample from Uganda were compared with those of the cloned strain M79/93 originating from this sample. No sequence differences were observed, which indicates that in a particular sample, only one 16S rRNA clone variant is normally present. The fact that the two nucleotide alternatives in a polymorphism were always present in a 1:1 ratio also supports this observation.

The hypothetical ancestor of the M. capricolum subsp.capripneumoniae strains and two lines of descent.The 24 evolutionary events were compiled and compared as detailed in Table 2. Eleven polymorphic positions were found to be common to all 20 strains, thus defining a hypothetical ancestor from which the M. capricolum subsp. capripneumoniae strains have evolved. The sequences of the two 16S rRNA genes of the hypothetical ancestor can be derived from Table 2. The remaining 12 polymorphisms and the length variation which occurred in only some of the strains most probably arose due to later evolutionary events. The polymorphisms in the 16S rRNA genes of the M. capricolum subsp.capripneumoniae strains (Table 2) were converted into a cladogram (Fig. 2) by comparative analysis. Both the length polymorphism and a nucleotide polymorphism were treated as one event. The horizontal lines are proportional to the number of events. The tree shown in Fig. 2 was rooted by the 11 ancestral evolutionary events and revealed two major lines of descent.

• Open in new tab
• Download powerpoint

Fig. 2.

Phylogenetic tree based on mutational events of the 16S rRNA genes of the M. capricolum subsp.capripneumoniae strains. The tree was constructed by comparative analysis and with the 11 polymorphisms of the putative ancestor used as the root. Two major lines of descent (I and II) can be seen. The positions and types of the polymorphisms are shown on the axis.

Evolutionary line I contained the 9 M. capricolum subsp.capripneumoniae strains which shared the synapomorphy (444–449)_A/− (Table 2; Fig. 2), situated in a poly(A) segment of the molecule in a region which is characterized by bilaterally bulged residues (Fig. 3). It remains to be shown whether the presence or absence of a sixth adenosine in this region of the 16S rRNA molecule will affect the function of the mature small ribosomal subunit from the strains of line I. The strains of line I showed a very low interstrain variability, and only two additional polymorphic positions, 509T/C and 875G/A were found, thus bifurcating into sublines IA and IB (Table 2; Fig. 2). Strain 95043 from Niger differed from the hypothetical ancestor only by the truncation of one adenosine in the 16S rRNA gene of therrnB operon.

• Open in new tab
• Download powerpoint

Fig. 3.

Secondary-structure model of the 16S rRNA molecule from the rrnB operon from M. capricolum subsp.capripneumoniae. All identified polymorphisms are indicated with arrows. They are numbered, and their composition (Table 2) is given according to the letter code of the International Union of Biochemistry. The nucleotide residues of the rrnA andrrnB operons are written at the arrows as N/N, respectively. The segment containing the sequence length variation betweenrrnA and rrnB is denoted by an arc. This model has been adapted from the secondary-structure model of the 16S rRNA molecule of M. capricolum subsp. capricolumdescribed by Gutell et al. (13, 14).

The 11 M. capricolum subsp. capripneumoniaestrains of evolutionary line II shared the two polymorphisms 404C/A and 1080C/A (Table 2; Fig. 2). The earliest member of line II was M. capricolum subsp. capripneumoniae 9081 from Oman, and line II branched into the two sublines IIA and IIB with a fork of three additional branches at the node of IIB (Fig. 2). This line is by far the most heterogeneous, and the evolutionary drift seems to be more pronounced for the 16S rRNA genes in the strains of line II than in those of line I. The two lines are separated by only three mutational events but can nevertheless be recognized as true phylogenetic entities, since none of the polymorphisms present in only some of the strains were shared among M. capricolum subsp.capripneumoniae strains of the two lines (Table 2). This observation can therefore be regarded as a consistency measure for the two lines, and it also indicates that sequence homogenization due to gene conversion has not occurred as was observed for the tufgenes of Salmonella typhimurium (1).

The 16S rRNA gene sequences of the five M. capricolum subsp.capripneumoniae strains, which were used in a recent study on the DNA relatedness of strains from the M. capricolumspecies group (5), have also been determined in this work. Although the hybridization values were very similar for all these strains (87 to 89%), the 16S rRNA sequence data clearly showed that two of the strains (F38^T and 7/1a) belonged to line II and three of the strains (G1943/80, G280/80, and G94/83) belonged to line I (Fig. 2; Table 2).

Polymorphisms as epidemiological markers and diagnostic targets.All members of the M. mycoides cluster are closely related as judged by biochemistry (10, 41), serological reactions (25, 35), DNA-DNA reassociation studies (2), and 16S rRNA sequence analysis (41). It is therefore important to identify polymorphisms in the 16S rRNA genes, because they can be used in diagnostic applications to distinguish between closely related species (45). The polymorphisms may also be useful as epidemiological markers. Two major lines of descent have been defined for M. capricolum subsp.capripneumoniae strains in this work. Representatives of line I were found only in Africa, whereas representatives of line II were found in both Africa and Asia (Fig. 1). A representative of line I, strain 95043, was isolated in Niger in 1995, and a strain of subline IB (Fig. 2; Table 2) was isolated in the neighboring country of Chad. Several strains of subline IA were isolated in Kenya and Uganda. The occurrence of the same type in neighboring countries, such as IA in Kenya and Uganda and IIB in Kenya and Sudan, might well be explained by trade and other movement of animals across these borders. An area of eastern Uganda and western Kenya is, for instance, populated by the same tribe, and the border is not well respected.

Strains of line II were found in countries bordering maritime routes (Mediterranean sea: Turkey, Tunisia; Indian Ocean: Kenya, Oman; Red Sea: Ethiopia, Sudan). Strain 9081, with a mutation pattern of line II, was isolated in Oman in 1990 and differed from the hypothetical ancestor by two polymorphisms in positions 404 and 1080. The polymorphism type pattern of subline IIA is the closest to that of line II and differs by one extra polymorphism in position 232. Strains of this type were isolated in Dubai in 1991. The polymorphism type pattern of the subline IIB differed from that of line II by the two polymorphisms in positions 94 and 1403. Strains of this type have been isolated in Kenya in 1976 (strain F38) and Sudan in 1989. Strains of types IIB1 and IIB2 differed from the strains of type IIB by one polymorphism in positions 538 and 684, respectively. A strain of type IIB1 was isolated in Ethiopia in 1992, and representatives of type IIB2 were isolated in Tunisia 1980 and Oman 1988. The Oman strain 7/1a originated from a goat which was imported from Turkey. Strains with a polymorphism pattern of type IIB3 differed from type IIB by two polymorphisms in positions 1255 and 1297. Representatives of type IIB3 were isolated in Oman.

Three types of M. capricolum subsp.capripneumoniae were found in Oman, namely, II, IIB2, and IIB3. A plausible explanation for this would be the extensive importation of goats from several countries in the Arabic peninsula for the religious feasts. These animals are normally slaughtered, but some might be kept alive and can thus infect indigenous herds. It is also noteworthy that two countries which were most likely to have CCPP-infected animals in the 19th century, Algeria (51) and South Africa (20), might have had strains of type IIB2. Algeria has a border with Tunisia, where type IIB2 was found in 1980, and South Africa imported Angora goats, which most probably brought CCPP from Turkey in 1881. The Turkish goat imported to Oman in 1988 was also found to be of type IIB2. In general, no correlation could be seen between the polymorphism pattern and the time of isolation of the strains.

Secondary structure analysis of the polymorphic positions and their implications.The positions of the 24 mutational events in the 16S rRNA molecule of M. capricolum subsp.capripneumoniae are shown in the secondary-structure model in Fig. 3. The nucleotide substitution rates according to a recently published variability map of the 16S rRNA molecule (54) are included in Table 2, showing the variability in each of the polymorphic positions. For comparison, consensus 16S rRNA sequences from an alignment of the mollicutes (32, 39) were computed by using different cutoff values for which a residue in a certain position was present. These percentages are included in Table 2, indicating that positions of high, intermediate, and low variability found in (eu)bacteria (54) also seem to hold for the variability in the 16S rRNA genes of mycoplasmas. Of the 11 polymorphisms of the ancestor, 6 (positions 180, 452, 672, 844, 1151, and 1446) were located in the high-variability groups 1 and 2 as defined by Van de Peer et al. (54). In contrast, 6 of the 12 microheterogeneities found in only some of the M. capricolum subsp.capripneumoniae strains (i.e., positions 94, 404, 509, 538, 684, and 897) belonged to group 5 or 6 of low variability. Moreover, frequently observed polymorphisms situated in stems were found predominantly in positions of relatively high variability (positions 452, 672, 1080, 1151, 1403, and 1446), while rarely found microheterogeneities often were situated in highly conserved locales in the 16S rRNA molecule (positions 538, 871, 875, 897, and 1255). It is noteworthy that all but 875 of the rarely found polymorphisms in sites of low variability occur in the starting or ending base pair of a stem (positions 538, 871, 897, and 1255). The reason for this is not known, but the base pairing at these positions might be rather flexible, at least in these mycoplasmas. This observation indicates that polymorphisms shared among all strains occur in rather variable positions while rarely found polymorphisms are present in more highly conserved positions of the 16S rRNA molecule of M. capricolum subsp. capripneumoniae.

A total of 13 polymorphisms are situated in stems (Fig. 3). A compensatory mutation in the opposite position to stabilize the actual stem in the other molecule was never observed. However, the general alternatives to noncanonical base pairing (13, 54) were followed in most cases. Therefore, a guanosine, a uridine, and an adenosine residue in one strand can have C or U, A or G, and U or C, respectively, as their counterparts. Moreover, the polymorphisms 509T/C, 875G/A, and 1255A/G indicated that cytidines can tolerate A or G as pairing nucleotides and that a C · G base pair can be substituted with a C · A pair, in certain positions. Therefore, a C · A pair might not necessarily affect the stability of a stem (54), and plausible explanations are a protonated C or a tautomeric configuration of A or C in these positions, which have been suggested in previous reports (19, 54). Interestingly, over 40 polymorphisms have been localized to stem regions of the 16S rRNA molecule of mycoplasmas (39, 41, 44, also see above) and without a compensatory mutation in the corresponding nucleotide position. This finding is contradictory to the polymorphisms in the 16S rRNA genes of the recently described B. sporothermodurans (42), where most of the microheterogeneities found in stems were associated with compensatory mutations in the complementary nucleotide position.

All polymorphisms of the M. capricolum subsp.capripneumoniae strains were checked against the corresponding positions of the 16S rRNA mutation database which provides a list of mutated positions in Escherichia coli(52). Only position 897 in M. capricolum subsp.capripneumoniae (912 in E. coli)T/C was listed in the database. A change of a C to a U in this position in E. coli has been shown to confer resistance to streptomycin (34). The actual polymorphism 897(912)T/C is present in strain GL102, a laboratory strain originating from strain Gabés, which has been passaged 102 times in the laboratory. It has not been possible to find whether streptomycin has been used during any of the passages. Nevertheless, this transition indicated that mutations may also occur in positions of low variability (Table 2) with a significant rate in the 16S rRNA genes.

We have so far characterized over 80 microheterogeneities in the 16S rRNA genes of different Mycoplasma species isolated from different hosts (16, 39, 41, 44). A comparison of these polymorphisms with those determined in this study confirmed that microheterogeneities in the 16S rRNA genes of mycoplasmas are distributed throughout the molecule but are rarely seen in universal regions (11) of the molecule. Polymorphisms with identical nucleotide compositions and present in different species have been found in only 3 of the more than 80 positions.

Evolution of the 16S rRNA genes of M. capricolum subsp.capripneumoniae: rrnA operon versusrrnB operon.The 16S rRNA sequences of therrnA operon were 99.73 to 99.93% similar for the differentM. capricolum subsp. capripneumoniae strains. Identical values were found for the rrnB operon. The percent similarity ranged from 98.77 to 99.20% between the two operons. This means that the similarity between 16S rRNA genes within a strain is lower (11 to 17 nucleotide differences) than that between 16S rRNA genes of homologous operons in different strains (0 to 4 nucleotide differences). A phylogenetic tree was constructed from 16S rRNA sequences of the individual operons (Fig.4). Mycoplasma putrefaciensserved as the outgroup, and the branches are denoted according to the corresponding polymorphism pattern (Table 1). The phylogenetic analysis revealed two distinct entities, one of which consisted of sequences from the rrnA operon and the other consisted of sequences from the rrnB operon. A slightly higher evolutionary rate was observed for the sequences of the rrnB operon as judged from the branch lengths (Fig. 4). The tree implies that therrnB operon has evolved by one to three further nucleotide changes. Strikingly, the topologies for the respective clade are very similar, and the line I strains form deep branches of both clades. Descendants belonging to line II form more recent branches, whereM. capricolum subsp. capripneumoniae 9081 of subline II is placed intermediately in the respective operonal cluster. This shows that the 16S rRNA genes from the two operons coevolved and that both are phylogenetically informative. The tree illustrates the importance of using 16S rRNA sequences from homologous operons to infer the phylogeny when polymorphisms are present and the species to be analyzed are closely related.

• Open in new tab
• Download powerpoint

Fig. 4.

Phylogenetic tree based on 16S rRNA sequences of the individual operons obtained by the neighbor-joining method (46). The tree revealed two distinct clades representing sequences of the rrnA and rrnB operons, indicating that the individual operons are monophyletic and that gene conversion has not taken place.

Conclusion.The evolution within mycoplasmas has been reported to be unusually rapid (56). In a recent study, we suggested that members of the M. mycoides cluster could be used as a model system for molecular evolution by mapping the polymorphisms in the two 16S rRNA genes (41). In the present work, we have characterized microheterogeneities in strains of M. capricolum subsp. capripneumoniae, a subspecies of theM. capricolum species group (41) of the M. mycoides cluster (55), and have constructed two kinds of trees. One tree was based on mutational events from consensus sequences, and this tree formed two lines of descent without underlying synapomorphies (close parallelism as a result of common inherited genetic factors causing incomplete synapomorphy). In a tree based on individual operons, both 16S rRNA genes were found to evolve and to reflect similar phylogeny. Therefore, strains of M. capricolum subsp. capripneumoniae constitute a very useful model for studies of molecular evolution. M. capricolum subsp. capripneumoniae may also be a suitable subspecies for comparing the evolution of other genes with the evolution observed for the 16S rRNA genes.