Molecular and Cellular Probes (1995) 9, 383-389

Arbitrarily-primed PCR confirms the differentiation of strains of Ureaplasma urealyticum into two biovars

F. Grattard, 1" B. Pozzetto, 1 B. de Barbeyrac, 2 H. Renaudin, 2 M. Clerc, 2 O. G. Gaudin t and C. B4b~ar 2

'Laboratoire de Bact4riologie-Virologie, Universit4 de Saint-Etienne, France, and 2Laboratoire de Bact4rioloEie, Universit4 de Bordeaux, France

(Received 5 June 1995, Accepted 20 June 1995)

Fourteen serotypes are currently recognized in the Ureaplasma urealyticum species. These serotypes have been divided into two genomic clusters or biovars by a large number of typing methods. The parvo-biovar includes strains of serotypes 1,3, 6 and 14 and the T960-biovar, strains belonging to the ten other serotypes. In this study, arbitrarily primed polymerase chain reaction (AP-PCR) has been applied to the analysis of reference strains of the 14 U. urealyticurn serotypes. By using two different sets of 10-mer oligonucleotide primers, the method allowed the clear differentiation between the two known biovars of the species. However, further differentiation within a same biovar was only achieved for a few standard strains of the T960-biovar analysed by using a pairwise combination of primers. The reproducibility of AP-PCR profiles was shown on strains tested after repeated subcultures and with different thermal cyclers. Additional experiments were performed on forty isolates of U. urealyticum recovered from subjects of various origins. They confirmed that AP-PCR was able to identify the strains at the biovar level. With reference to the other typing methods, AP-PCR is easy to perform and can be applied to large numbers of strains for epidemiological purposes. © 1995 Academic Press Limited

KEYWORDS: Ureaplasma urealyticum, arbitrarily-primed PCR, biovar.

INTRODUCTION

Ureaplasma urealyticum is a human genital myco- plasma frequently involved in the colonization of the urogenital tract of women: a recent review reports that 40 to 80% of sexually mature asymptomatic females are colonized.' U. urealyticum is also a patho- gen responsible for nongonococcal urethritis in males. 2 Its role in various diseases related to preg- nancy outcome (infertility, spontaneous abortion, chorioamnionitis, respiratory disease in preterm in- fants, perinatal morbidity and mortality) has often been suspected, despite the absence of virulence markers.'

Subdivision within the U. urealyticum species has

been early suggested on its antigenic heterogeneity? Fourteen serotypes (1 to 14) are currently described on the basis of immunological recognition with type- specific polyclonal antisera. 4s Whether this antigenic heterogeneity reflects differences in pathogenicity is not known; however, serotypes 8 and 4 have been suspected to be more virulent in newborns, 6 and the latter serotype has been isolated more frequently in women with pregnancy complications. 7 Most of the dlfflculhes with serotypmg lie in the fact that the reagents are not currently availlbie and that the typing methods are time-consuming and poorly reproducible on clinical isolates. 8

* Author to whom correspondence should be addressed at: Laboratoire de Bact~riologie-Virologie, Facult~ de/~l~decine J. Lisfranc, 15 rue A. Par~, 42023 Saint-Etienne cedex 2, France.

0890-8508195/060383 +07 $12.00•0 383 © 1995, Ac'ad~mic Press Limited

384 F. Grattard et aL

Genotypic methods have been proposed to in- vestigate the U. urealyticum species, including DNA- DNA hybridization, restriction endonuclease analysis of genomic DNA, rRNA gene probe analysis, 91° pulsed-field gel electrophoresis, u and recently biovar- specific PCR. ~2 All these techniques have been found able to divide this species into two genomic clusters called biovars; the parvo-biovar includes serotypes 1,3, 6 and 14 and the T960-biovar the 10 remaining serotypes. These two biovars differ by their genome size" and their susceptibility to manganese, t3

Arbitrarily primed (AP)-PCR is based on the am- plification of DNA fragments which are recognized by single short oligonucleotides of arbitrarily chosen sequences matching at low stringency to the template DNA? 4~s The method is easy to perform and has been found particularly discriminative in comparison with other markers to differentiate between strains or serovars in several bacterial species including Escherichia coli, ~° Bacillus thuringiensis ~7 or Entero- bacter cloacae. ~8

In this study, we applied this method to the typing of U. urealyticum and examined its ability to distinguish between the serotype standard strains of the species at the biovar and at the serotype levels. We also report its application to the analysis of clinical isolates in various epidemiological conditions.

MATERIALS AND METHODS

Ureaplasma urealyticum strains

The serotype standard strains used in this study are listed in Table I . They were obtained from D. Taylor- Robinson (St Mary Hospital Medical School, London, UK) and from J. Robertson (University of Alberta, Edmonton, Canada). Some of these strains which were kept lyophilized were also tested after repeated subcultures (serotypes 1,2, 3, 4, 5, 11, 13 and 14).

Forty clinical strains obtained from three different origins were tested by AP-PCR; (i) 10 couples of isolates from the urogenital tract of pregnant women who delivered in the Obstetric Department of our hospital (H6pital Nord, Saint-Etienne) and from the gastric lavage of their neonate; (ii) 10 isolates from sperm specimens of males consulting for infertility; and (iii) 10 isolates from the urogenital tract of HlVl- infected males and females.

Identification of the organisms and culture conditions

The clinical samples were tested for the presence of U. urealyticum by using conventional culture tech- niques, including Shepard broth and A7 culture plates

Table 1. Origin of the serotype standard strains of U. urealyticum

Strain designation Serotype Source*

My10465 1 A DKF1 1 B My9410 2 A T23 2 B My11237 3 A DKF3 3 B My11860 4 A DKF4 4 B My9411 5 A NIH5 5 B My11861 6 A DKF6 6 B My8339 7 A DKF7 7 B T960 8 ATCC 27618 Vancouver 9 B unknown 10 A JSL-K2 11 B JSL-U24 12 B JSL-U38 13 B JSL-U26 14 B

* A: D. Taylor Robinson, MRC Sexually Transmitted Diseases, St Mary Hospital Medical School, London, UK.

B: J. Robertson, Department of Microbiology, University of Alberta, Edmonton, Canada.

(Diagnostics Pasteur, France). Positive samples were identified on metabolic criteria (positivity of the ure- ase test) and on the characteristic morphology of colonies on agar plates. Strains of mycoplasmas were kept at -80°C until use.

Each clinical strain of U. urealyticum was sub- cloned twice. Dilutions of actively growing cultures were spread onto A7 agar plates and incubated at 37°C until growth was detected. Well separated col- onies were picked up and expanded in 50 ml of a modified Shepard broth containing 0.06% urea and 10% horse serum. Reference strains were cultured in 50 ml of the same broth.

DNA preparation for genomic techniques

Bacterial strains were harvested, washed in saline buffer and lysed by SDS. The samples were treated by RNase and proteinase K (Boehringer Mannheim, Meylan, France), and purified by phenol- chloroform-isoamyl alcohol extractions, as described previously? 8

AP-PCR typing

It was adapted from a method reported recently for E. cloacae species. ~8 PCR was carried out in 50 I~1

Typing of U. urealyticum by AP-PCR 385

Table 2. Sequences of oligonucleotides used as primers in AP-PCR (1 to 3) and in 16S rRNA gene PCR (U3, U8 and P6)

Primer Sequence (5'-3')

I AACGCGCAAC

2 ACGTATCTGC

3 CCGCACCCAA

U3 TAGAAAGTCGCTCTTTGTGG

U8 GAAGATGTAGAAAGTCGCGTTTGC

P6 GGTAGGGATACCTTGTTACGACT

Taq polymerase in (10x) buffer (ATGC Bio- technologie). The sequences of the primers, pur- chased from Genset, are listed in Table 2. Each sample was subjected to 25 cycles of amplification. With both pairs of primers, the expected PCR products were detected as a band of 1.3 kb in size on stained agarose gel.

RESULTS

of a solution containing 50 ng of template DNA, 350 pmol of primers, 200 I~U of each dNTP (Boeh- ringer Mannheim), 1.25 IU of Taq polymerase and (10 x) buffer provided by the manufacturer (ATGC Biotechnologie, Noisy-le-Grand, France). Each sample was subjected to a first cycle of amplification (94°C: 5 min, 35°C: 5 min, 72°C: 5 min) followed by 28 subsequent cycles consisting of denaturation for 1 min at 94°C, low-stringency annealing for 2 min at 35°C and extension for 2 min at 72°C, with a final extension step of 10 min at 72°C. The AP-PCR prod- ucts were separated by electrophoresis on 1.2% agar- ose gels and visualized by ethidium bromide staining. A molecular size marker '100 Base Pair Ladder' (Pharmacia, St Quentin Yvelines, France) was used for reference. A negative control, consisting in the same reaction mixture except that template DNA was replaced by water, was included in each reaction.

Three different thermal cyclers were tested for am- plification: GeneAmp 9600 (Perkin-Elmer Cetus, Emeryville, CA, USA), Thermocis (Cis bio in- ternational, Gif sur Yvette, France) and Prem III (LEP Scientific Ltd., Flobio, France), denoted A, B and C respectively. Fourteen different 10-mer primers, purchased from Genset (Paris, France) were assayed in preliminary experiments. Three primers denoted 1, 2 and 3 were selected for the main part of the study. AP-PCR was performed either with primer 1 used alone or with primers 2 and 3 used in com- bination. The sequences of these primers are listed in Table 2.

16S rRNA gene serovar specific PCR

Choice of primers for AP-PCR

In preliminary experiments, fourteen 10-mer primers were screened with a few representative standard strains of each biovar. The best results were achieved with either primer 1 used alone, or with primers 2 and 3 used in combination. These sets of primers were found able to provide well-resolved patterns ranging from 0.1 to 2 kb in length.

AP-PCR profiles of the serotype standard strains of U. urealyticum

AP-PCR performed with primer 1 was able to dif- ferentiate clearly between the two known biovars of U. urealyticum species, as illustrated in Fig. 1. Strains belonging to serotypes 1, 3, 6 and 14 (parvo-biovar) exhibited similar profiles with two intense bands of 1.8 and 0.5 kb. Another profile with a major band of 1.3 kb was shared by strains belonging to the ten serotypes of the T960-biovar.

With the combination of primers 2 and 3, four different groups of profiles were noted within the 14 reference strains of U. urealyticurn (Fig. 2C). All the strains belonging to the parvo-biovar shared a similar profile with a major band of 0.5 kb. Three distinct profiles could be noted within strains belonging to the T960-biovar: serotypes 4 and 12 shared a similar profile; serotype 13 exhibited a distinct profile and the seven other serotypes shared a third AP-PCR profile characterized by a common band of 1-1 kb (Fig. 2).

The differentiation between the parvo- and the T960- biovar by using 16S rRNA gene PCR was performed according to the method described by Robertson et al., 12 with some modifications. Amplification was carried out in a mixture containing 20 ng of DNA, I pM of each primer (U3 and P6 for the parvo-biovar PCR, or U8 and P6 for the T960-biovar PCR), 200 I~M of each dNTP (Boehringer Mannheim) and 1.25 IU

Reproducibility of the AP-PCR profiles

With each set of primers, the reproducibility of the profiles was studied for the same strain after repeated subcultures and repeated DNA extraction and for different strains belonging to the same serotype. Sim- ilar patterns were observed in both cases (data not shown). Moreover, the same strains were tested with

386

1 2 3 4 5 6

F. Grattard et al.

7 8 9 10 11 12 13 14 Ma

- 1 . 8

- 1 . 3

-0 .8

-0 .5

Fig. 1 Representative AP-PCR profiles of the serotype standard strains of U. urealyticum with primer 1 and thermal cycler A. The numbers correspond to the serotypes of the standard strains. The unmarked lane corresponds to a negative control. Lane Ma, size marker (in kb).

different thermal cyclers, by using identical para- meters of amplification. As shown on Fig. 2 with primers 2 and 3, the three thermal cyclers generated closely related AP-PCR profiles for a same strain.

AP-PCR profiles of clinical isolates of U. urealyticum

Clinical isolates from various origins were investigated both by AP-PCR (performed with primer 1 and thermal cycler A) and by 16S rRNA gene specific PCR. Con- cordant results were obtained in all cases. All the strains isolated from pregnant women and their neo- nates (group I) were shown to belong to the parvo- biovar. Seven of ten strains isolated from subjects consulting for infertility (group 11), and six of ten strains isolated from HIV-1 infected subjects (group III) exhibited a parvo-biovar pattern. The AP-PCR profiles of some representative isolates from each group of subjects are shown in Fig. 3. In all ex- periments, water controls remained negative.

DISCUSSION AND CONCLUSIONS

AP-PCR has been reported to be useful for detecting polymorphism in bacteria without having any data about the genome sequence. This method is easy to perform and can be more discriminative than other genotypic methods in epidemiological investigations provided correct primers have been selected. 14-17 The present study reports the evaluation of AP-PCR in the characterization of U. urealyticum strains. With reference to other typing methods based on hy- bridization and/or electrophoresis, s-u AP-PCR can be

performed in a few hours and needs only small amounts of bacterial DNA. Thus this technique is particularly useful in the case of organisms which grow slowly and has already been applied with suc- cess to the typing of other fastidious bacteria such as Borrelia burgdorferi, ~ Mycobacterium tuberculosis 2° or Mycoplasma gallisepticum. 2~

Our data demonstrates that AP-PCR is able to separate the serotype standard strains of U. urea- lyticum into two groups, corresponding to the two genomic clusters or biovars of the species. In- terestingly, unlike the other methods of biovar de- termination using PCR with primers derived from 1 6S rRNA gene ~2 or urease gene, 22 the present method requires a single amplification step to recognize the two biovars. Indeed AP-PCR was found able to type clinical isolates from different epidemiological origins and produced results in complete agreement with the 16S rRNA gene PCR.

It has previously been shown that variations in the AP-PCR fingerprints could be related to some parameters of amplification such as length or con- centration of primers, commercial origin of Taq polymerase or technical specifications of thermal cyclers. 2324 Our results demonstrate that AP-PCR per- formed according to the conditions presented in the Materials and Methods section was able to generate reproducible profiles, even with the use of different thermal cyclers. Despite the extra time required for the extraction of DNA, this step is crucial for the re- producibil i ty of AP-PCR patterns, as already observed by others. 24 Using relatively large amounts of primers may also contribute to increase both the complexity and the reproducibility of the profiles. 24

The use of a combination of primers to characterize the U. urealyticum species was initially justified by

Typing of U. urealyticum by AP-PCR

(A) 1 3 2 13 4 12 5 7 8 9 10 11 Ma

387

1.1

0.8

0.5

(B) 1 3 2 13 4 12 5 7 8 9 10 11 Ma

1.1

0.8

0.5

(C) 1 3 6 14 2 13 4 12 5 7 8 9 10 11 Ma

1.1 0.8

0.5

Fig. 2 Representative AP-PCR profiles of the serotype standard strains of U. urealyticum with the combination of primers 2 and 3. The profiles were obtained by using thermal cyclers A, B or C (see materials and methods). Symbols as in Fig. 1.

the small number of bands obtained with single short primers. For standard strains belonging to the T960- biovar, which includes strains with more hetero- geneous genome sizes, u we show that the pair-wise combination of primers gave more discriminative profiles. Teng etal. have recently proposed to separate

the T960-biovar into two subgroups by using Dral digestion of a PCR product of the multiple-banded surface antigen gene. 2s Interestingly, with the ex- ception of serotype 10, our AP-PCR results are in accordance with those of Teng et al. 2s concerning the inclusion of serotypes 2, 5, 7, 8, 9 and 11 into a same

388

(A)

a b c d e f

I - w g m m m

u D ~ m ~

F. Grattard et al.

(B)

g h i j k 1 m M a

(C)

M a n o p q r s t u v w

1.8

1.3

0 .8

0 .5

Fig. 3 Representative AP-PCR profiles obtained with primer 1 and thermal cycler A for strains of U. urealyticum isolated from mothers and their neonates (A), from subjects consulting for infertility (B) and from HIV-1 infected subjects (C). All the strains from group I (lanes a to f), 4 strains from group 11 (lanes g to j) and 6 strains from group III (lanes n to p and r to t) exhibit a parvo-biovar pattern. Lanes Ma, size marker (in kb).

subgroup. However, the differences observed by AP- PCR within serotypes of the T960-biovar for standard strains would need to be furtherly confirmed on clinical isolates of the same biovar previously char- acterized by serotyping.

The antigenic heterogeneity of the strains may de- rive from the higher degree of diversity found in the U. urealyticum species. The relationship between invasiveness of strains and particular serotypes has been suggested, despite controversial results, s One of the major problems with serotyping is the difficulty to meet objective diagnostic criteria. 8 Moreover, sero- typing remains poorly reproducible and cannot be applied in most laboratories because antisera are not commercially available. By contracts, AP-PCR may constitute a very useful tool to evaluate the import- ance of these two distinct clusters in the pathogenicity of U. urealyticurn within a large number of labora- tories. Harasawa et al. have already stressed that

the relationship between U. urealyticum and human diseases should be understood in terms of genomic clusters rather than of serovars2

Finally, the application of AP-PCR to the typing oF U. urealyticum gives further evidence for the sep- aration of strains into two genomic clusters. These data raise once more the question of dividing the species into two distinct taxonomic groups.

REFERENCES

1. Cassell, G. H., Waites, K. B., Watson, H. L., Crouse, D. T. & Harasawa, R. (1993). Ureaplasma urealyticum intrauterine infection: role in prematurity and disease in newborns. Clinical Microbiology Reviews 6, 69-87.

2. Brown, M. B., Cassell, G. H., Taylor-Robinson, D. & Shepard, M. C. (1983). Measurement of antibody to Ureaplasma urealyticum by an enzyme-linked assay and detection of antibody responses in patients with

Typing of U. urealytlcum by AP-PCR 389

nongonococcal urethritis. Journal of Clinical Micro- biology 17, 288-95.

3. Shepard, M. C., Lunceford, C. D., Ford, D. K. et al. (1974). Ureaplasma urealyticum gen. nov., sp. nov.: proposed nomenclature for the human T (T-strain) mycoplasmas. International Journal of Systemic Bac- teriology 24, 160-71.

4. Horowitz, S. A., Duffy, L., Garrett, B., Stephens, J., Davies, J. K. & Cassell, G. H. (1987). Can group- and serovar-specific proteins be detected in Ureaplasma urealyticum? Pediatric Infectious Diseases 5, $325-31.

5. Robertson, J. A. & Stemke, G. W. (1982). Expanded serotyping scheme for Ureaplasma urealyticum strains isolated from humans. Journal of Clinical Microbiology 15, 873-8.

6. Quinn, P. A., Li, C. S., Th'ng, C., Dunn, M & Butany, J. (1993). Serological response to Ureaplasma urea- lyticum in the neonate. Clinical Infectious Diseases 17, $136-43.

7. Naessens, A., Foulon, W., Breynaert, J. & Lauwers, S. (1988). Serotypes of Ureaplasma urealyticum isolated from normal pregnant women and patients with preg- nancy complications. Journal of Clinical Microbiology 26, 319-22.

8. Stemke, G. W. & Robertson, J. A. (1985). Problems associated with serotyping strains of Ureaplasma urea- lyticum. Diagnostic Microbiology and Infectious Dis- ease 3, 311-20.

9. Harasawa, R., Dybvig, K., Watson, H. L. & Cassell, G. H. (1991 ). Two genomic clusters among 14 serovars of Ureaplasma urealyticum. Systematic Applied Micro- biology 14, 393-6.

10. Razin, S. & Yogev, D. (1986). Genetic relatedness among Ureaplasma urealyticum serotypes (serovars). Pediatric Infectious Diseases 5, $300-4.

11. Robertson, J. A., Pyle, L. E., Stemke, G. W. & Finch, L. R. (1990). Human ureaplasmas show diverse genome sizes by pulsed-field gel electrophoresis. Nucleic Acids Re- search 18, 1451-5.

12. Robertson, J. A., Vekris, A., B~b~ar, C. & Stemke, G. W. (1993). Polymerase chain reaction using 16S rRNA gene sequences distinguishes the two biovars of Ureaplasma urealyticum. Journal of Clinical Micro- biology 31,824-30.

13. Robertson, J. A. & Chen, M. H. (1984). Effects of manganese on the growth and morphology of Urea- plasma urealyticum. Journal of Clinical Microbiology 19, 857-62.

14. Welsh, J. & McClelland, M. (1990). Fingerprinting gen- omes using PCR with arbitrary primars. Nucleic Acids Research 18, 7213-24.

15. Williams, J. G. K., Kubelick, A. R., Livak, K. J., Rafalski, J. A. & Tingey, S. V. (1990). DNA polymorphisms

amplified by arbitrary primers are useful genetic markers. Nucleic Acids Research 18, 6531-5.

16. Wang, C., Whittam, T. S., Berg, C. M. & Berg, D. E. (1993). RAPD (arbitrary primer) PCR is more sensitive than multilocus enzyme electrophoresis for dis- tinguishing related bacterial strains. Nucleic Acids Re- search 21, 5930-3.

17. Brousseau, R., Saint-Onge, A., Pr~fontaine, G., Masson, L. & Cabana, J. (1993). Arbitrary primer polymerase chain reaction, a powerful method to identify Bacillus thuringiensis serovars and strains. Applied and En- vironmental Microbiology 59, 114-19.

18. Grattard, F., Pozzetto, B., Berthelot, Ph., et al. (1994). Arbitrarily primed PCR, ribotyping, and plasmid pattern analysis applied to investigation of a nosocomial out- break due to Enterobacter cloacae in a neonatal Intens- ive Care Unit. Journal of Clinical Microbiology 32, 596-602.

19. Welsh, J., Pretzman, C., Postic, D., Saint Girons, I., Baranton, G. & McClelland, M. (1992). Genomic fin- gerprinting by arbitrarily primed polymerase chain re- action resolves Borrelia burgdorferi in three distinct phyletic groups. International Journal of Systemic Bac- teriology 42, 370-7.

20. Linton, C. J., Jalal, H., Leeming, J. P. & Millar, M. R. (1994). Rapid discrimination of Mycobacterium tuberculosis strains by random amplified polymorphic DNA analysis. Journal of Clinical Microbiology 32, 2169-74.

21. Gear,/, S. J., Forsyth, M. H., Saoud, S. A., Wang, G., Berg, D. E. & Berg, C. M. (1994). Mycoplasma gal- lisepticurn strain differentiation by arbitrary primer PCR (RAPD) fingerprinting. Molecular and Cellular Probes 8, 311-16.

22. Blanchard, A., Hentschel, J., Duffy, L. Baldus, K. & Cassell, G. H. (1993). Detection of Ureaplasma urea- lyticum by polymerase chain reaction in the urogenital tract of adults, in amniotic fluid, and in the respiratory tract of newborns. Clinical Infectious Disease 17, $148- 53.

23. MacPherson, J. M., Eckstein, P. E., Scoles, G. J. & Gajadhar, A. A. (1993). Variability of random amplified polymorphic DNA assay among thermal cyclers, and effects of primer and DNA concentration. Molecular and Cellular Probes 7, 293-9.

24. McClelland, M. & Welsh, J. (1994). DNA fingerprinting by arbitrarily primed PCR. PCR Methods and Ap- plications 4, $59-65.

25. Teng, L.-J., Zheng, X., Glass, J. I., Watson, H. L., Tsai, J. & Cassell, G. H. (1994). Ureaplasma urealyticum biovar specificity and diversity are encoded in multiple- banded antigen gene. Journal of Clinical Microbiology 32, 1465-9.