INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Mastitis remains one of the most economically important problems of the dairy cattle industry throughout the world. Milk quality and the prevalence of clinical and subclinical mastitis are major factors in determining farm profitability (16).

Streptococcus agalactiae (Lancefield group B) is a highly contagious obligate parasite of the mammary gland, where it can survive for long periods of time (19). Since this organism is susceptible to treatment with a variety of antimicrobial agents, eradication within a closed herd is possible. With increasing pressures for the reduction of antimicrobial agents in animals as well as in humans, the necessity for improved understanding of the epidemiology of this etiological agent has become apparent (1). Prevalence studies for S. agalactiae in cattle have been conducted in different areas of North America (19, 20, 31). Data about the epidemiology and molecular characteristics of this organism from bovine milk are not available in Canada. In the United States, only a few studies, with a limited number of strains, have been carried out on these subjects (7, 29).

S. agalactiae also causes significant morbidity and mortality in humans, both infants and adults, worldwide (3). In neonates, S. agalactiae is mostly acquired from the mother's vagina in early-onset disease, although nosocomial, community, and breast milk transmissions have been reported (2). In adults, S. agalactiae occurs preferentially in certain individuals, such as diabetics, pregnant and postpartum women, and immunocompromised patients, emphasizing the opportunistic nature of the infection (23). Furthermore, humans act as a significant reservoir of S. agalactiae, since this bacterium may be carried in the vaginas of women without apparent clinical signs (14). Questions have been raised as to whether S. agalactiae is a zoonotic agent or whether host-specific ecovars exist. Controversial reports indicate the absence or the presence of a relationship between human and bovine S. agalactiae isolates (8, 18). Some epidemiological studies on S. agalactiae infections have been based on serotyping techniques, but these traditional procedures are limited in that their discriminatory potential is too low. DNA-based subtyping techniques, such as pulsed-field gel electrophoresis (PFGE) (10), ribotyping (5), restriction enzyme analysis (REA) (9), multilocus enzyme electrophoresis (28), and random amplification of polymorphic DNA (RAPD) (6) have been used efficaciously to subtype S. agalactiae isolates of human origin. Methodologies such as ribotyping and PFGE usually involve time-consuming steps and/or sophisticated equipment. REA has the advantages of simplicity and high discriminatory power but is sometimes difficult to interpret because of the large number of restriction fragments generated. RAPD is an accessible and sensitive method based on the use of arbitrary primers to amplify polymorphic segments of DNA. This technique has been widely used in recent years for detection of diversity among isolates (25, 34, 36).

The objective of this work was to study, by RAPD, the genetic diversity of a collection of S. agalactiae isolates originating from dairy cattle in different parts of Québec, Canada. Data were used to standardize the technique and evaluate the discriminatory power of the primers used. In addition, some isolates from asymptomatic healthy women were also analyzed and compared to bovine isolates.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacteria. Reference strains of S. agalactiae serotypes I_a (SS 615), I_b (SS 618), II (SS 619), III (SS620), IV (3139), and V (SS 1169 [NT1]) were used as positive controls in serotyping and RAPD experiments. All reference strains originated from the Centers for Disease Control and Prevention, Atlanta, Ga., except for S. agalactiae serotype IV (3139), which was kindly sent by J. Henrichsen, Statens Serum Institut, Copenhagen, Denmark.

Serotyping. Human isolates were serotyped on the basis of capsular polysaccharides by the coagglutination method (21). Since most bovine isolates were autoagglutinable, they were serotyped by double diffusion in agarose gels (17). Anti-type I_a, I_b, II, III, IV, and V sera were purchased from Oxoid (Basingstoke, England).

RAPD fingerprinting. All human isolates and 185 representative bovine isolates were analyzed by RAPD. This analysis was performed as described by Williams et al. (36) with some modifications. The PCR mixture consisted of buffer (10 mM Tris-HCl [pH 8.3], 50 mM KCl; 2.5 mM MgCl₂), 100 µM each of the four deoxynucleoside triphosphates (Pharmacia Biotech Inc., Baie d'Urfé, Québec, Canada), 0.4 µM primer, 50 ng of DNA extracted and purified as previously described (27), and 2.5 U of Taq DNA polymerase (Pharmacia) in a total volume of 25 µl. The primers used are shown in Table 1 and were synthesized by Gibco BRL Custom Primers (Burlington, Ontario, Canada). Each sample was subjected to the first cycle of denaturation (5 min at 94°C) in a DNA Thermal Cycler 480 (Perkin-Elmer Applied Biosystems, Foster City, Calif.). Each of the 35 subsequent cycles consisted of denaturing at 94°C for 30 s, annealing at 35°C for 30 s, and extension at 72°C for 1 min. The last cycle included an extension at 72°C for 5 min. Amplified products were separated by electrophoresis in a 1.4% agarose gel (Sigma) and visualized as white bands on a black background by UV transillumination following ethidium bromide staining. A 1-kb DNA ladder (Gibco) was used in each gel as molecular size standards. A negative control, consisting of the same reaction mixture but with water instead of template DNA, was included in each run. In addition, a positive control, containing the same reaction mixture with a template of DNA from a well-characterized reference strain (S. agalactiae SS 615), was also included. Each isolate was tested under the same conditions at least three times with the selected primers.

Pattern analysis. Photographs of each gel were digitalized with a video camera connected to a microcomputer (Alpha ease; Alpha Innotech Corp., San Leandro, Calif.). After conversion, the data were normalized and analyzed. Degrees of homology were determined by Dice comparisons, and clustering correlation coefficients were calculated by the unweighted pair group method with arithmetic averages. When the calculations were completed, a dendrogram showing the hierarchic representation of linkage level between isolates was drawn. All these processes were carried out with Molecular Analyst Software, Fingerprinting, version 1.12 (Bio-Rad Laboratories, Mississauga, Ontario, Canada).

Discriminatory analysis. The probability that two unrelated isolates sampled from the test population will be placed into different typing groups or clusters was assessed according to the Hunter-Gaston formula (15). This probability is calculated as
where N is the total number of isolates in the sample population, s is the total number of Rapid's patterns described, and n_j is the number of isolates belonging to the jth type.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of informative primers. To identify primers that generate informative arrays of PCR products, eight unrelated S. agalactiae isolates were selected from the entire panel of isolates. They had been isolated from different geographic sites and belonged to different serotypes or were nontypeable (Fig. 1).

View larger version (66K): [in this window] [in a new window]   FIG. 1.   Illustration of the RAPD patterns generated with primers OPS11, OPB17, and OPB18. Lanes 1, reference strain SS615 (serotype Ia); lanes 2 nontypeable bovine isolate from region 3; lanes 3, serotype III bovine isolate from region 2; lanes 4, serotype III bovine isolate from region 7; lanes 5, serotype IV human isolate from region 4; lanes 6, serotype II bovine isolate from region 6; lanes 7, serotype V human isolate from region 4; lanes 8, serotype Ib bovine isolate from region 2; lanes M, 1-kb DNA ladder (DNA molecular size marker).

Serotyping. The double diffusion method was used to serotype bovine isolates because most of them were autoagglutinable. However, nonagglutinable bovine isolates were analyzed by both methods with identical results (data not shown). Results of serotyping for bovine and human isolates can be observed in Table 2. Sixty-two percent of bovine isolates were nontypeable. The remaining bovine isolates were found to belong to four different serotypes. Serotype III was the most prevalent, representing 28% of all bovine isolates. All six serotypes tested were identified among the 38 human isolates (Table 2). Only five human isolates were nontypeable and two autoagglutinated. Serotypes V and III were the most frequently identified serotypes, with prevalences of 26 and 24%, respectively.

Genetic diversity as defined by RAPD fingerprinting. The genetic relationship among all RAPD patterns of S. agalactiae based on the combination of data obtained with the three primers is represented in the dendrogram shown in Fig. 2. Overall, S. agalactiae isolates presented 58% similarity. A total of 94% of the isolates were clustered in four groups (I, II, III, and IV) with 70% similarity among them. Three clusters, A (48 isolates), B (14 isolates), and C (32 isolates), with 79 to 80% similarity, were identified within group IV. The other three groups did not present any clustering. The percent similarity of each group oscillated between 70 and 77%. The heterogeneity of the population was significantly increased by 14 nongrouped isolates.

View larger version (2734K): [in this window] [in a new window]   FIG. 2.   Genetic relationship among 223 S. agalactiae isolates (of bovine and human origins) as estimated by clustering analysis of RAPD patterns obtained with three primers. The dendrogram was generated by the unweighted pair group method with arithmetic means. H, carrier woman; B, bovine milk; UT, untypeable; AA, autoagglutinable.

Genetic variation in relation to serotype. The serotype distribution for each RAPD pattern is indicated in Fig. 2. In addition, the relationship between serotype and RAPD group and/or cluster is also observed in Tables 3 and 4 for isolates of bovine and human origin, respectively. Half of serotype III and II isolates of bovine origin were in group II. Nontypeable bovine isolates were proportionally distributed in all groups (Table 3). All serotype I_a isolates and most serotype III isolates recovered from humans were in cluster C. Most human isolates of serotype V were included in group I (Table 4). Clustering was not observed for other serotypes.

Genetic variation of isolates in relation to geographical distribution. In general, clustering was not observed for bovine isolates originating from the same region, except for those from regions 1 and 2 (Fig. 2). In spite of the existing diversity, it was possible to find at least one pair of isolates sharing the same RAPD pattern in most of the regions.

Genetic variation of isolates in relation to host origin. In general, clustering was observed in most S. agalactiae isolates of human origin (Fig. 2). Fifty percent of human isolates were placed in cluster C of group IV, and 37% belonged to group I. The human isolates were then analyzed separately to verify this apparently high homology. Figure 3 shows relatively high variability among human isolates, since only 65% similarity was found. A principal group (group ii), in which approximately 79% of isolates clustered, and one minor group (group i) can be observed (Fig. 3). Two clusters, "a" (9 isolates) and "b" (21 isolates), with 80 to 82% similarity, were identified within group ii. The other group did not present any clear cluster. The percent similarity of each group oscillated between 76 and 77%. None of the human isolates shared identical RAPD patterns with the three primers. In addition, in the dendrogram illustrating the cluster analysis of S. agalactiae isolated from asymptomatic women, most isolates of serotype V appeared in cluster "a" (Fig. 3). This confirmed results obtained with the general dendrogram that included the analysis of bovine and human isolates (Fig. 2).

View larger version (24K): [in this window] [in a new window]   FIG. 3.   Genetic relationship among 38 S. agalactiae isolates from asymptomatic pregnant women as estimated by cluster analysis of RAPD patterns obtained with three primers. The dendrogram was generated by the unweighted pair group method with arithmetic means. UT, untypeable; AA, autoagglutinable.

RAPD typing as an epidemiological tool. The discriminatory capacity of the RAPD typing was determined in order to evaluate the suitability of the primers chosen in this study for the epidemiological analysis of S. agalactiae isolated from milk. It was possible to define 215 RAPD types for the 223 isolates (index of discrimination [D] = 0.9996) by combining data obtained with the three primers, whereas only seven serotypes for the 223 isolates (D = 0.6908) were identified by serotyping.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Limited information was available on the epidemiology of Canadian S. agalactiae isolates recovered from bovine milk. To our knowledge, few studies using DNA-based techniques have been carried out with a large collection of field isolates of bovine origin in North America. Previous studies on S. agalactiae isolates of human origin have suggested that RAPD is superior to serotyping for epidemiological evaluations of this pathogen (6, 24). In the present work, RAPD was used to study a large collection of bovine isolates from Canada. In general, high genetic diversity was found. A possible explanation for this diversity is that different isolates originated from different herds. Similar results were obtained by using other molecular techniques in Australia, Denmark, and the United States, even though these studies were based on lower numbers of herds and field strains (1, 18, 29).

S. agalactiae can be differentiated on the basis of distinct polysaccharide surface antigens. In this study, as in others (4, 11), most bovine isolates were nontypeable by polysaccharide antigens. Only 38% of isolates of bovine origin were typeable; serotype III was the most important. Previous studies showed a relative heterogeneity in the distribution of different serotypes of bovine isolates (4, 18, 26, 35). The importance of invasive serotype III strains is well known among human isolates (28), but the significance for bovine isolates is unknown yet.

In general, no evident correlation could be established between serotyping and RAPD patterns. Before this study, data which combine genomic diversity and antigenic typing were not available for bovine isolates. Results showed genetic heterogeneity not only among different serotypes but also among isolates belonging to same serotype. This suggests that the RAPD technique may be more accurate than capsular serotyping in differentiating S. agalactiae isolates from a bovine population. RAPD of S. agalactiae of bovine origin therefore appears to be of great value for epidemiological studies.

Clustering was not observed for bovine isolates originating from the same region, except for those from regions 1 and 2. This result is consistent with a previous report of Rivas et al. (29), who analyzed S. agalactiae of bovine origin by automated ribotyping. They could not find one ribotype in all three regions delineated in New York State. In the present study, at least two isolates with an identical RAPD pattern were found in each region. This fact might suggest that, in some instances, there may be a common source of S. agalactiae in different herds from the same region.

The serotype distribution of S. agalactiae of human origin appears to have changed over time. Until recently, the predominant serotypes that were detected among clinical isolates by the Centers for Disease Control and Prevention and other laboratories were I_a and III (3, 13, 14). A striking change, however, occurred in the 1990s, when the percentage of serotype V climbed from 2.6% in 1992 to 14% in 1993 and then to 20% in 1994 (12). The reasons for this increase are still unclear. Interestingly, serotypes V and III were identified in the present study as the most frequent serotypes among isolates from carrier women, with prevalences of 26 and 24%, respectively.

Reports on the genetic diversity of S. agalactiae isolated from healthy women are controversial. Huet et al. concluded that the genetic polymorphism of isolates from carrier women, as evaluated by ribotyping, is relatively limited (14). However, this technique appears to have low discriminatory power when it is used alone for epidemiological studies of S. agalactiae (14). On the other hand, Helmig et al. observed considerable heterogeneity in a population of S. agalactiae isolates from asymptomatic women (13). In agreement with other studies (7, 33), data presented here indicate that isolates from asymptomatic women have a slightly closer relationship than isolates of bovine origin. In spite of some clustering of human isolates, relatively high diversity was seen among them.

In this study, only one pair of human (serotype V) and bovine (nontypeable) isolates showing the same RAPD pattern was found. This suggests the possibility of a common origin for both isolates. This is in agreement with the results of Jensen and Aarestrup, who detected identical ribotypes for isolates from milk and dairy workers (18). Despite the fact that a common source of human and bovine isolates is possible (18), results obtained in this work do not allow confirmation of this hypothesis. Isolates belonging to different serotypes but indistinguishable by genetic analysis have already been described (2, 18). One possible explanation is the ability of S. agalactiae to regulate capsule expression in a phase shift-like manner (32). The ability to phase shift may be of particular interest in S. agalactiae mastitis, since bacterial adherence is an important factor in the pathogenesis of bovine mastitis, and the adhesion of S. agalactiae to epithelial cells seems to be inversely proportional to the degree of encapsulation (30).

The selection of primers is critical in maximizing the discriminatory power of RAPD typing. An index of discrimination (D) greater than 0.90 is necessary for interpreting typing results with confidence (15). Two previous studies have reported genetic analysis of S. agalactiae isolates of human origin by RAPD (6, 24). In one of those studies, a partially degenerated oligonucleotide with a D of 0.98 was used (24), whereas in the other, a combination of four primers with a D of 0.90 was obtained (6). Our data suggest that the RAPD typing generated by the combination of OPS11, OPB17, and OPB18 primers (D = 0.9996) has increased the ability of the methodology to detect variability between isolates. Potential applications include identification of isolates that appear to have broad geographic distribution, suggesting interfarm transfer, and discrimination among recurrent versus new intramammary infections. Such information may allow the establishment of control and eradication programs at the herd level. Furthermore, RAPD typing may be used to study the relationship between human and bovine infection.