INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Evidence has been presented that the species currently known as Ureaplasma urealyticum should be separated into two new species, namely, U. parvum (previously U. urealyticum biovar 1) and U. urealyticum (previously U. urealyticum biovar 2) (9). Ureaplasmas are commensals in the genital tract, recognized causes of disease (15), and suspected contributors to a number of other pathological conditions (1). Because they are commonly found in healthy people, their pathogenic role can be difficult to prove (10, 18). The majority of human ureaplasma isolates belong to the proposed new species U. parvum (1, 3, 5, 7, 8), which includes serovars 1, 3, 6, and 14. U. urealyticum (biovar 2) is isolated less often but is not uncommon. Some ureaplasma serovars have been associated with disease syndromes more commonly than with normal flora (19), but data are limited because of difficulties with conventional serotyping methods. Rapid molecular methods for identification of species and subtypes would be of great value in studies of the epidemiology and pathogenesis of infections with U. parvum and U. urealyticum (8).

Recently, PCR-based methods have been used successfully to distinguish the two Ureaplasma species (biovars), but there is a need to improve their specificity and sensitivity (2, 8). Target sequences of the 16S rRNA gene and 16S rRNA-23S rRNA intergenic spacer regions (6, 12), the urease gene subunits (2, 11), and the 5' ends of the multiple-banded antigen (MBA) genes (8, 16, 17) have all been used in PCR-based assays to differentiate U. parvum from U. urealyticum. Previously, we have sequenced portions of these genes from all 14 Ureaplasma serovars (9) and described PCR-based assays for the identification of U. parvum and U. urealyticum and the subtyping of U. parvum (8). In the present study, we evaluated the specificity of a large range of primers and used a small subset to develop an algorithm for the detection, species identification, and subtyping of Ureaplasma species.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. Two sets of reference strains of all 14 serovars of U. urealyticum (including biovars 1 and 2, referred to hereafter as U. parvum and U. urealyticum, respectively) were used as previously described (8, 9). These included one set obtained directly from the American Type Culture Collection (ATCC reference strains) and another kindly provided by H. L. Watson, Department of Microbiology, University of Alabama at Birmingham (UAB reference strains). Additional reference strains from the ATCC were used to test the specificity of primers: Mycoplasma pneumoniae strains M129 (ATCC 29342) and FH (ATCC 15531), Mycoplasma genitalium (ATCC 33530), Mycoplasma fermentans (ATCC 19989), Mycoplasma hyorhinis (ATCC 17981), and Acholeplasma laidlawii (ATCC 23206). Mycoplasma hominis isolates were grown from clinical specimens on A7 agar in our laboratory and identified by colonial morphology and partial sequencing of the 16S rRNA gene and the 16S rRNA-23S rRNA gene spacer regions.

Clinical isolates and specimens. A total of 78 Ureaplasma isolates obtained from vaginal swabs of pregnant women and recently cultured in our laboratory and 185 vaginal swabs obtained from pregnant women and women attending a sexually transmitted disease clinic and in which Ureaplasma species had been previously detected (8) were used in this study.

Oligonucleotide primers. The 29 individual primers used in this study to amplify portions of three genes of all 14 serovars are shown in Table 1. They include 4 primers that have been previously described (12, 16, 17) and 25 new primers designed by us. The nomenclature of our primers is based on specificity (e.g., UU and UP for U. urealyticum and U. parvum, respectively), gene target (e.g., UM for the MBA genes), the direction of the sequence (S, sense; A, antisense), and the numbered base position at which the primer sequence starts (9).

DNA preparation and PCRs. DNA preparation and PCR were performed as previously described (8, 9). In each reaction, positive and negative controls were processed in parallel with the tested samples to detect false-negative results or contamination. Melting temperature values are shown in Table 1. The annealing temperature was 58°C for all primer pairs except UPS2-UPA2, UUS2-UUA2, UMS-170-UMA263, and UMS-61-UMA263, for which the annealing temperature was 55°C.

Subtyping of U. urealyticum by direct sequencing. We previously showed (9) that the 10 serovars of the new species U. urealyticum could be divided into three subtypes based on sequence differences in amplified fragments of the 5' ends of the MBA genes at six positions between 112 and 251. With primers UMS-170 and UMA263, amplified fragments of reference serovars, clinical isolates, and specimens were sequenced by use of an ABI 373A sequencing machine with Applied Biosystems Taq DyeDeoxy Terminator Cycle-Sequencing Ready Reaction Kits according to the manufacturer's instructions. All of the U. urealyticum isolates and specimens were sequenced with primer UMS-170; some were also sequenced with primer UMA263 to confirm the results.

Algorithm for identification and subtyping of U. parvum and U. urealyticum. In order to develop a practical identification and subtyping scheme, we designed an algorithm (Fig. 1) for clinical use and then applied it to all known Ureaplasma serovars in the ATCC and UAB reference sets and to clinical isolates and clinical specimens.

View larger version (24K): [in this window] [in a new window]   FIG. 1.   Algorithm for identification and subtyping of U. parvum and U. urealyticum. For identification of U. parvum and U. urealyticum (A), the result (band size) was 326 bp for U. parvum serovars 1 and 3/14 and 327 bp for U. parvum serovar 6.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

PCRs. No amplification occurred when DNA extracted from Mycoplasma or Acholeplasma species was tested with any of 17 primer pairs. Five primer pairs (UPS1-UPA, UPS1-UPA1, UPS-UPSA, UPS2-UPA2, and UMS-57-UMA222) were specific for and amplified all 4 serovars of U. parvum, and four (U8-UUA, UUS-UUSA, UUS2-UUA2, and UMS-61-UMA263) were specific for and amplified all 10 serovars of U. urealyticum.

Specificity of U. parvum and U. urealyticum identification and subtyping primers. All the ATCC and UAB reference strains of U. parvum and U. urealyticum were correctly identified with the species- and subtype-specific primers as shown in the algorithm (Fig. 1).

U. parvum and U. urealyticum identification and subtyping results for clinical isolates and clinical specimens. Of the 78 clinical isolates, 62 (79.5%) were identified as U. parvum, 15 (19.2%) were identified as U. urealyticum, and 1 (1.3%) was mixed. Of 185 vaginal swabs that had been shown previously to contain Ureaplasma species, 151 (81.6%) contained U. parvum only, 20 (10.8%) contained U. urealyticum only, and 14 (7.6%) contained both. Results of subtyping of clinical isolates and Ureaplasma species in vaginal swabs are shown in Table 2.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In the past, U. urealyticum was divided into two biovars by various phenotypic and genetic methods (9). Recently, we reported new data supporting the proposal that these two biovars be designated separate species, namely, U. parvum (previously U. urealyticum biovar 1) and U. urealyticum (previously U. urealyticum biovar 2) (9). Several primer sets have been described for the identification of biovars of the old species U. urealyticum (or the new species U. parvum and U. urealyticum). However, they were not based on the sequences of all 14 serovars (2, 12, 13, 16, 17), and some lacked specificity or the ability to detect all serovars (2, 12). Better PCR-based methods for both identification and subtyping are needed to facilitate studies of the relationship between Ureaplasma species or subtype and disease (4).

Many of the primer targets used in this study were based on our previous observation that the heterogeneity of the intergenic spacer regions is greater than that within the genes (9). We believed that primers based on these regions would be more discriminatory for the identification and subtyping of Ureaplasma species (9). These included primers UPSA-UUSA (16S rRNA-23S rRNA gene spacer regions); UPS2-UPA2 and UUS2-UUA2 (ureA-ureB and ureB-ureC gene spacer regions, respectively); and UMS-57, UMS-61, UMS-83, UMS-54, UMS-112, and UMS-112' (upstream of the MBA genes) (Table 1).

Differences between the two proposed new species in the 16S rRNA genes have been described previously and used to design biovar-specific (or species-specific) primers (12). The primer pair U8-P6, designed to amplify U. urealyticum (biovar 2), also amplified DNA from M. pneumoniae (12), which is found not infrequently in the genital tract (14). We designed a new primer, UUA, which was paired with U8. This pair was specific for U. urealyticum (biovar 2) and did not amplify DNA from two M. pneumoniae ATCC strains or other Mycoplasma species tested. Two new U. parvum (biovar 1)-specific primer pairs, UPS1-UPA and UPS-UPA, also based on the 16S rRNA gene, did not amplify DNA from either U. urealyticum (biovar 2) or the Mycoplasma species tested.

The 16S rRNA and 23S rRNA gene spacer regions are normally continuous with but more heterogeneous than the 16S rRNA genes (6, 9). Primers spanning these regionsUPS-UPSA for U. parvum (biovar 1) and UUS-UUSA for U. urealyticum (biovar 2)were specific for all serovars within the corresponding species (biovar) for both ATCC and UAB reference strains.

Primers based on urease gene sequences and used to differentiate U. parvum and U. urealyticum have been described previously (2, 13). We designed two additional species-specific primer sets, UPS2-UPA2 (U. parvum specific) and UUS2-UUA2 (U. urealyticum specific), targeting the ureA-ureB and ureB-ureC gene spacer regions, respectively. They were specific for all serovars within the corresponding species (biovar) for both ATCC and UAB reference strains.

The MBA genes contain both species- and serovar-defining regions (19, 20). Several primer sets based on MBA sequences have been described previously for differentiating U. parvum and U. urealyticum (9, 16, 17). To improve sensitivity and specificity and to provide more choice, we designed additional species-specific primers based on our sequencing results. The primer pair UMS-57-UMA222 was specific for U. parvum, and UMS-61-UMA263 was specific for U. urealyticum.

Although the 5' ends of the MBA genes are highly conserved among the 10 serovars of U. urealyticum, we were able to distinguish three subtypes based on sequence data (9). There is more sequence variation in the 5' ends of the MBA genes of the four serovars of U. parvum, which allow them to be separated into three groups. Primer pairs with specificity for individual serovars or subtypes of U. parvum and U. urealyticum were used in combination and supplemented, when necessary, by sequencing of UMS-170-UMA 263 amplicons to characterize human ureaplasmas, as shown in Fig. 1.

Having confirmed the sensitivity and specificity of the new primer pairs, we designed an algorithm (Fig. 1) for the identification and subtyping of U. parvum and U. urealyticum, using a selection of the most suitable primers. Its utility was evaluated with ATCC and UAB reference strains, stored clinical isolates, and Ureaplasma-positive clinical specimens. The results suggested that the algorithm would be particularly suitable for use in epidemiological studies. Identification and subtyping of clinical isolates and specimens confirmed the previous finding that U. parvum is found much more commonly (87% of isolates or specimens overall) than U. urealyticum (19%) among vaginal flora (3). Both Ureaplasma species were detected in 1 clinical isolate (1%) and 14 vaginal swabs (8%) (6% overall). Among the U. parvum strains detected, serovars 3/14 (48%) and 1 (43%) were found more commonly than serovar 6 (23%); 8% of isolates and 16% of vaginal swabs (14% overall) contained two or more subtypes. U. urealyticum subtypes 1 (34%) and 2 (62%) were found more commonly than subtype 3 (6%). Only one clinical isolate contained mixed U. urealyticum subtypes.

In summary, we have designed a series of new primer pairs based on previously reported sequences of three important ureaplasma genes and adjoining regions and modified some previously published primers to improve their sensitivity and specificity. Our initial evaluation of the algorithm for the identification and subtyping of U. parvum and U. urealyticum, using a selected set of primers with ATCC and UAB reference strains, clinical isolates, and clinical specimens, confirmed their specificity and sensitivity. Further evaluation of their sensitivity for direct examination of clinical specimens is required. However, we believe that, in the future, they will assist in studies of the epidemiology, pathogenicity, and clinical significance of Ureaplasma species in humans and will provide significant advantages over conventional serotyping methods.