Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. Feb 2006; 72(2): 1420–1428.
PMCID: PMC1392973

First Insights into the Evolution of Streptococcus uberis: a Multilocus Sequence Typing Scheme That Enables Investigation of Its Population Biology


Intramammary infection with Streptococcus uberis is a common cause of bovine mastitis throughout the world. Several procedures to differentiate S. uberis isolates have been proposed. However, all are prone to interlaboratory variation, and none is suitable for the description of the population structure. We describe here the development of a multilocus sequence typing (MLST) scheme for S. uberis to help address these issues. The sequences of seven housekeeping gene fragments from each of 160 United Kingdom milk isolates of S. uberis were determined. Between 5 and 17 alleles were obtained per locus, giving the potential to discriminate between 1.3 × 107 sequence types. In this study, 57 sequence types (STs) were identified. Statistical comparisons between the maximum-likelihood trees constructed by using the seven housekeeping gene fragments showed that the congruence was no better than that between each tree and trees of random topology, indicating there had been significant recombination within these loci. The population contained one major lineage (designated the ST-5 complex). This dominated the population, containing 24 STs and representing 112 isolates. The other 33 STs were not assigned to any clonal complex. All of the isolates in the ST-5 lineage carried hasA, a gene that is essential for capsule production. There was no clear association between ST or clonal complex and disease. The S. uberis MLST system offers researchers a valuable tool that allows further investigation of the population biology of this organism and insights into the epidemiology of this disease on a global scale.

Mastitis remains the most economically important infectious disease of dairy cattle throughout the world. The annual loss due to clinical mastitis in the United Kingdom has been estimated at approximately £170 million (21), and between $1.5 to 2.0 billion in the United States (36). These losses can be attributed to a reduction in milk production, the associated costs of treatment, and the culling of persistent and repeatedly infected cows. Microorganisms that cause mastitis can be divided into those that show a contagious route of transmission, such as Staphylococcus aureus and Streptococcus agalactiae, and those that also frequently infect the udder from an environmental reservoir, such as Escherichia coli and Streptococcus uberis. The application of various control measures over the past two decades, based on improved milking practices, postmilking teat disinfection, and routine intramammary antimicrobial treatment after each lactation period, has proved effective against contagious pathogens (23) but had little if any effect on bacteria that infect the mammary gland from an environmental reservoir. S. uberis is currently responsible for ca. 33% of all clinical mastitis in the United Kingdom (14) and occurs at a similar frequency worldwide. The failure to control bovine mastitis caused by S. uberis is largely attributed to insufficient information regarding the epidemiology and pathogenesis of infection (2, 23).

Investigation of S. uberis using various typing schemes has shown evidence of heterogeneity in the population. However, the extent of this is not well defined, and its significance to disease pathogenesis has not been determined. Nonmolecular typing procedures for S. uberis have included phage typing, serotyping, or bacteriocin-like inhibitor typing schemes (13, 22, 33). However, these methods are unsuitable for the analysis of closely related strains. DNA-based methods show more potential, with DNA restriction analysis by pulsed-field gel electrophoresis (PFGE) appearing to be the most reliable and highly discriminatory (1, 4, 16, 28). For example, one study using PFGE divided 343 S. uberis isolates from New Zealand into 330 restriction fragment patterns, with 11 pairs and one group of three strains having similar restriction patterns (4). A PCR-based DNA fingerprinting technique has also been utilized extensively (27). The major problem with these methods is their poor reproducibility and their inability to quantitate the genetic relationships between isolates. Molecular typing methods can be used to assess short and long-term epidemiology, and such levels of discrimination can be achieved by monitoring rapidly evolving variation or variation that accumulates slowly. Techniques that reflect slowly accumulating variation, such as multilocus enzyme electrophoresis (MLEE [29]), are more appropriate for long-term epidemiological understanding. MLEE, as with the other typing methods mentioned, has the major disadvantage of poor portability between laboratories, thereby preventing comparison of results.

Multilocus sequence typing (MLST) was designed to overcome the shortfalls of the other molecular typing techniques. The motivation underpinning the original development of MLST for bacteria was to provide a tool to investigate populations and population dynamics on a global scale. This is facilitated by access to the data using the internet. MLST utilizes the same rationale as MLEE, indexing neutral or slowly accumulating genetic variation in housekeeping genes (25). The evolution of housekeeping genes is constrained by their requirement to encode functional products and is not affected by the rapid evolution that may be detected within genes encoding proteins that influence survival in a particular niche. MLST exploits high-throughput nucleotide sequencing technologies to identify this variation at the level of the nucleotide rather than the protein. The method was originally used in the study of Neisseria meningitidis (25) and has subsequently been used to investigate the population structures in numerous pathogenic bacteria including group B Streptococcus (19), Streptococcus suis (20), and Campylobacter jejuni (3). In general, most MLST schemes are based on the sequencing of internal fragments of seven or more housekeeping genes. The loci are chosen on the basis that they are present in every organism, and the observed sequence variation within them is likely to be selectively neutral. Fragments of approximately 400 to 500 bp are used in MLST since such fragments can be rapidly, economically, and accurately sequenced on both strands with a single set of primers. Recently, a MLST scheme for S. uberis has been described that relies on the use of sequences from six loci (38). The list of genes included two open reading frames (pauA and gapC) encoding proteins with a demonstrable extracellular location, both of which play such a role in virulence that they have been proposed as vaccine candidates (9, 23). Another open reading frame (oppF) included in this scheme has also been shown to play a significant role during growth in milk (30) and would consequently be prone to variation according to niche. The observation that pauA, encoding a secreted plasminogen activator, was undergoing positive selection in itself renders this scheme unsuitable for the analysis of population structure and evolutionary relationships (38). Furthermore, it has been noted that in some isolates pauA may be absent or replaced by pauB, which encodes an alternate but unrelated plasminogen activator (35).

We describe here the establishment of a more conventional MLST scheme for the bovine mastitis pathogen, Streptococcus uberis, that enables the investigation of population structure onto which the presence, acquisition, and evolution of biological traits, including virulence, and related genes can be accurately mapped.


Isolation and culture of bacterial isolates.

Strains were obtained from the collections of the Milk Quality Group at the Institute for Animal Health (IAH). These isolates were collected from two herds belonging to the IAH at Compton, as well as from four herds using “organic” farming practice in the United Kingdom between 1999 and 2002, and had previously been identified as S. uberis either by microbiological analysis or use of API 20 Strep test kits (bioMérieux, Marcy l'Etoile, France). In addition, two well-characterized strains were included: 0140J, whose genome has been fully sequenced (http://www.sanger.ac.uk/Projects/S_uberis/), and EF20. These were isolated from cows with clinical mastitis in 1973 and 1970, respectively. 0140J is significantly more pathogenic for lactating animals than EF20 (12). The isolate details are available on the MLST database (pubmlst.org/suberis [identification numbers 1 to 160]). S. uberis isolates were cultured on blood agar plates containing 1% (wt/vol) esculin or in Todd-Hewitt broth (Oxoid) overnight at 37°C. All isolates were confirmed as S. uberis by PCR with species-specific 16S rRNA primers as previously described (11).

Extraction and preparation of chromosomal DNA.

Overnight broth cultures (1.5 ml) were centrifuged at 12,500 × g for 5 min, and the supernatant was removed. Cells were washed with 0.5 ml of TE buffer (10 mM Tris, 5 mM EDTA [pH 7.8]) and repelleted, and the supernatant was removed. The cell walls were disrupted by resuspension in 375 μl of TE buffer containing 30 U of mutanolysin (Sigma; diluted from a 5,000-U/ml stock)/ml and 10 mg of lysozyme (Sigma)/ml and incubated at 37°C for 30 min. Cells were lysed by the addition of 20 μl of sodium dodecyl sulfate (20% [wt/vol] in 50 mM Tris, 20 mM EDTA [pH 7.8]) and 3 μl of proteinase K (Sigma) at 20 mg/ml and agitation. Samples were then incubated at 37°C for 1 h. Saturated NaCl (200 μl, ca. 6.0 M) was added to precipitate protein cell wall material, and tubes were agitated for 15 s. Firm pellets were obtained by centrifugation (12,500 × g) for 10 min. The supernatant (400 to 450 μl) was removed carefully to a fresh, labeled tube. An equal volume of Tris-equilibrated phenol-chloroform-isoamyl alcohol (25:24:1; Sigma) was added, before agitation and centrifugation for 3 min at 8,000 × g to separate the phases. The upper aqueous phase was retained, and DNA was precipitated by adding 2 volumes of ice-cold ethanol and holding the mixture at 4°C for 2 h. Precipitated DNA was pelleted by 5 min of centrifugation at 12,000 × g. Pellets were washed with cold 70% ethanol and then air dried. The DNA was resuspended overnight at 4°C in 50 μl of TE buffer containing 20 μg of RNase A/ml (Sigma, made DNase-free by boiling for 5 min) and then incubated for 15 to 30 min at 37°C before storage at −20°C.

Choice of loci for MLST.

Thirteen candidate loci were initially identified by searching the genome database of strain 0140J (http://www.sanger.ac.uk/Projects/S_uberis/) with sequences of housekeeping genes from other bacteria. Suitable genes were then chosen on the basis of observed sequence diversity in pilot PCR studies with a restricted set of S. uberis isolates. Seven genes were selected based on suitable resolving power and reliable amplification. The following seven loci were selected for the MLST scheme: carbamate kinase (arcC), d-alanine-d-alanine ligase (ddl), glucose kinase (gki), transketolase (recP), thymidine kinase (tdk), triosephosphate isomerase (tpi), and acetyl-coenzyme A acetyltransferase (yqiL).

Amplification parameters and nucleotide sequence determination.

PCR products were amplified with oligonucleotide primers designed from the 0140J genome sequence (http://www.sanger.ac.uk/Projects/S_uberis/). The primers and annealing temperatures are listed in Table Table1.1. Each 30-μl reaction contained 5 μl of DNA (ca. 150 ng), 15 μl of 2× Taq PCR Mastermix (QIAGEN, United Kingdom), 4 μl of sterile distilled water, and 3 μl of forward and reverse primers (10-pmol/μl stock solutions). The cycling parameters were initial denaturation for 5 min at 94°C, followed by 40 cycles of denaturation at 94°C for 30 s, annealing for 30 s at the appropriate temperature (Table (Table1),1), and extension at 72°C for 45 s. A final extension was performed at 72°C for 7 min. A GeneAmp 2700 Thermal Cycler (Applied Biosystems, United Kingdom) was used for all PCRs.

Oligonucleotide primers for amplification and sequencing

Amplification products were purified by using the MinElute-96 kit (QIAGEN) as described in the manufacturer's protocols. The purified DNA was resuspended in 25 μl of sterile distilled water and stored in microtiter plates at −70°C until sequence reactions could be performed.

Sequencing reactions used 1 μl of purified PCR product (ca. 10 ng), 1.7 μl of BigDye v3.0 Ready Reaction Mix (Applied Biosystems, United Kingdom), 0.8 μl of sterile distilled water, and 3.2 pmol of primer in accordance with the manufacturer's instructions. Both strands of each gene fragment were sequenced. The cycling conditions and subsequent removal of unincorporated dye terminators were as described in the manufacturer's protocols. Sequencing reactions were resolved on an ABI Prism 3700 DNA Analyzer (Applied Biosystems).

Data analysis.

Sequences were assembled from the resultant chromatograms by using the Sequencher v4.1 software package (Gene Codes Corp.). For each of the seven loci, every different sequence obtained from the 160 isolates was assigned as a distinct allele. Each isolate is defined by an allelic profile consisting of seven integers, which corresponds to the allele numbers at the seven loci in the order arcC, ddl, gki, recP, tdk, tpi, and yqiL. Each unique allelic profile is assigned as a sequence type (ST). An MLST database containing the sequence of all alleles, the allelic profiles (STs), and information about the S. uberis isolates is maintained at Oxford University and can be found on the S. uberis pages of the MLST Web site (pubmlst.org/suberis).

In-frame concatenated sequences of different combinations of loci were exported from the MLST database using the built-in concatenation function of the Web site. The maximum-likelihood (ML) phylogenetic trees were reconstructed by using the HKY85 model of DNA substitution with the value of the shape parameter (α) of a discrete approximation to a gamma distribution of rate heterogeneity among sites estimated from the empirical data during tree reconstruction. This model was selected by using Modeltest. One thousand replicate neighbor-joining bootstrap trees, using the ML substitution model, were reconstructed to determine whether any incongruencies observed between trees using different combinations of loci were due to low resolution or were strongly supported. These analyses were performed using PAUP* version 4 (beta 10) (32).

Previously described statistical methods were used to further examine the extent of congruence among the gene trees (6, 15). First, the Shimodaira-Hasegawa test was used to determine whether there were significant differences among the tree topologies inferred for each gene. This analysis was undertaken by estimating the ML tree for each of the seven genes and then comparing, in turn, the difference in log likelihood (Δ −ln L) between each of the seven topologies on each of the seven genes. If each gene tree has the same phylogeny (i.e., phylogenetic congruence), as expected under entirely clonal evolution, then they should not differ significantly in likelihood. To further assess the extent of congruence among the seven ML gene trees, we used the randomization test (6, 15). In this case the Δ −ln L values for each of the seven ML trees fitted to each of the seven genes were compared to the equivalent values computed for 200 random trees created from each gene. If the Δ −ln L values for the ML trees fall within the range obtained from the random trees, then there is no more congruence among the trees than expected by chance alone. All of these analyses were carried out in PAUP*.

The relatedness of isolates was analyzed using the Sequence Type Analysis and Recombinational Test (START) software, a collection of tools for analysis of MLST data (available from pubmlst.org [17, 31]). Related STs were clustered into clonal complexes or lineages using the START program BURST. In addition, the recently described enhanced version of BURST, eBURST, was used for comparison (eburst.mlst.net [7]). The START software was also used to determine the ratios of nonsynonymous to synonymous polymorphisms (dN/dS ratios) for each locus (26).

Statistical comparisons of the disease status of different groups of isolates were undertaken by chi-square analysis.

Detection of hasA gene by PCR and Southern hybridization.

PCR amplification of the hasA gene (which is required for capsule formation) was carried out using S. uberis genomic DNA as the template as described for the MLST genes, except that an annealing temperature of 50°C was used. The oligonucleotide primers used were as follows: forward primer; 5′-GAAAGGTCTGATGCTGATG, and reverse primer; 5′-TCATCCCCTATGCTTACAG. The hasA genotype of strains that gave negative PCR results was confirmed by Southern hybridization. Briefly, genomic DNA was digested with HindIII, separated by agarose gel electrophoresis, transferred to a nylon membrane (Immobilon-Ny+; Millipore Corp.), and hybridized by standard procedures. A digoxigenin-labeled hasA probe was used. This was generated by PCR from S. uberis 0140J with the primers shown above using a PCR DIG Probe Synthesis Kit (Roche) according to the manufacturer's instructions.


Development of a MLST scheme for S. uberis.

Chromosomal DNA was obtained from 160 isolates confirmed to be S. uberis. The sequences of the seven loci were determined for each isolate, and allelic profiles were assigned. The alleles defined for the MLST scheme were based on gene regions with sequence lengths of between 357 bp (ddl) and 500 bp (tdk). Between 5 (tpi) and 17 (arcC) alleles were observed for each locus (Table (Table2).2). The average number of alleles at each locus was 10.4, providing the potential to distinguish approximately 1.3 × 107 different STs. The proportion of variable nucleotide sites in the selected loci ranged from 1.1% (tpi) to 4.2% (tdk). All seven loci appeared to be under stabilizing selective pressure, since most of the substitutions were synonymous, as indicated by the dN/dS ratios being substantially less than 1 (Table (Table22).

Characteristics of S. uberis MLST loci

The 160 isolates were resolved into 57 STs (Table (Table3),3), 39 of which only occurred once. The genome sequence strain, 0140J (http://www.sanger.ac.uk/Projects/S_uberis/), was assigned as ST-1 and was found to be unique in this data set. Sixty-five isolates (40.6%) were represented by one of four STs: ST-5, ST-6, ST-26, and ST-24. The most prevalent ST (ST-5) was identified 21 times, followed by ST-6 (18 isolates), ST-26 (15 isolates), and ST-24 (11 isolates). A database was set up (pubmlst.org/suberis [18]) to enable global access to the MLST data. This database contains details of the allele sequences obtained, the allelic profiles and ST assignments, and details of the sequenced isolates (the isolates described in the present study were assigned identification numbers of 1 to 160).

Characteristics of S. uberis isolates analyzed by MLST in this study

Population structure.

ML trees were constructed from the concatenated sequences of three random loci for the 57 STs (four combinations of three loci were analyzed; Fig. Fig.1).1). The topologies of the trees showed clear differences, with instances of STs that have been paired in some trees but not in others. For example, ST-39 and ST-41 were paired in the recP-tpi-arcC tree (with a bootstrap value of 74.3%) but were completely separated in the other trees. Similarly, ST-45 and ST-52 were paired in the gki-tpi-ddl tree (bootstrap value of 91.1%) but separated in the ddl-tdk-arcC and recP-tpi-arcC trees. Although the bootstrap values indicate low resolution in the trees, these findings suggest that the population structure is not strictly clonal and that its evolution has involved recombination.

FIG. 1.
ML trees of concatenated sequences from combinations of three loci. ST designations are given in Table Table3.3. All horizontal branch lengths are drawn to scale.

To analyze this further, statistical tests of congruence were performed as described in Materials and Methods. The ML trees inferred for each of the seven gene fragments were significantly different in log likelihood (Table (Table4).4). Furthermore, for each of the seven genes the ML trees were no more similar in likelihood than the 200 random trees for each data set, an indication that there has been substantial recombination in the evolutionary history of S. uberis.

Tests for congruence among S. uberis ML gene trees

Lineage assignment.

The lineage of the STs was investigated by using the clustering programs, BURST and eBURST. BURST analysis was performed by using the default value for group definition (where STs are grouped if they share at least five alleles with one other member of the group). This predicted that the majority of the isolates were members of the same lineage. The relationships between the STs in this group are shown in Fig. Fig.2.2. ST-5 was the predicted founder of this lineage, since it had the most single locus variants and was frequently isolated. Similarly, e-BURST was performed using the default value for group definition (this method is more conservative and only groups STs that share six alleles). This also predicted that ST-5 was the ancestor of a lineage and suggested the likely evolutionary descent of this lineage (Fig. (Fig.3).3). We have therefore assigned STs that were closely related to ST-5 to a clonal complex (designated the ST-5 clonal complex). Members of the complex were defined as STs that shared at least four alleles with ST-5. The definition of clonal complexes is somewhat arbitrary, but this definition allows clonal complex assignment to be carried out automatically on any new STs entered into the database.

FIG. 2.
BURST results: the relationships between the STs in the largest BURST group (group definition of five identical loci). Central STs represent predicted group or subgroup founders. The inner rings contain single locus variants, the outer rings represent ...
FIG. 3.
eBURST analysis: the predicted evolutionary descent of STs within the largest e-BURST group (group definition of six identical loci). The predicted founder (ST-5) is in boldface. The area of the circles representing STs correlates with the number of isolates ...

The ST-5 complex dominated the population and contained 112 isolates (70%) and 24 different STs. It includes the genome sequence strain, 0140J. The remaining 33 STs (48 isolates) were not assigned to a clonal complex at this time, since the BURST analysis of this data set did not clearly predict any other founders or reveal other groups containing more than three STs.

These results show that the MLST scheme described here is considerably better suited to the identification of clusters of related genotypes and for the analysis of the population structure of S. uberis than that recently published (38).

Relationship between ST and disease status of the cow at sampling.

Of the 160 isolates investigated, 51 (31.9%) were from clinical cases of mastitis and 82 (51.2%) were from subclinical infections (those with no obvious signs of disease) (Table (Table3).3). In addition, 24 isolates (15.0%) were from animals that were still infected (subclinically) despite having been given antibiotic treatment for clinical mastitis, and three were from animals with unknown disease status. Analysis of the MLST results did not reveal any clear association between disease status and ST. This analysis showed that the 11 most common STs had all been isolated from both clinical and subclinical intramammary infections. Similarly, there was no significant difference in disease status between ST-5 clonal complex isolates and other isolates.

Relationship between ST and herd.

Of the 160 total isolates, 131 (81.9%) were obtained from cows within the two IAH herds and 29 were obtained from organic herds in South West England (see isolate database at pubmlst.org/suberis). Examination of the most common STs (those with at least four isolates) revealed that 7 of 11 STs had been isolated from both IAH and non-IAH herds (Table (Table5).5). These included the most common STs, ST-5 and ST-6. The other four (ST-8, ST-10, ST-22, and ST-23) had only been isolated from IAH herds. The predominant ST from the IAH herds was ST-5 (20 of 131 isolates or 95.2% of ST-5 isolates were from IAH), whereas the most common STs from non-IAH herds were ST-4 and ST-6 (both single locus variants of ST-5).

Association between ST and farm

Relationship between lineage and presence of capsule gene, hasA.

Capsulation in S. uberis is dependent on the expression of two genes present as single copies at different loci, hasA (hyaluronate synthase) and hasC (UDP-glucose pyrophosphorylase) (34). Naturally occurring acapsular isolates appear to lack hasA, whereas hasC was shown to be ubiquitous (8). To determine whether there was a correlation between the presence or absence of hasA and sequence type or clonal complex, the 160 S. uberis isolates were analyzed by PCR. A product was amplified from most of the isolates, but 21 were negative. The absence of hasA in these 21 strains was confirmed by Southern hybridization (data not shown).

Carriage of hasA correlated with ST and lineage (Table (Table3).3). With the exception of ST-29, all members of an ST were congruent (either positive or negative) for hasA. Furthermore, all of the 112 isolates in the ST-5 clonal complex possessed hasA. In contrast, the isolates that were not assigned to a clonal complex gave variable results. Interestingly, the majority of hasA negative strains (16 of 21 or 76.2%) were isolated from subclinical infections.


The primary aims of this study were to develop an unambiguous and discriminatory typing scheme for Streptococcus uberis and to increase the understanding of the population structure of this important bovine pathogen. This was achieved by using MLST, a molecular typing method that has been used successfully in the characterization of many bacterial species and fungi (2, 5, 19, 20, 23). The data generated from the S. uberis isolates in the present study has been used to construct a database on a central MLST Web site that permits the direct comparison of data generated in other laboratories (pubmlst.org/suberis). This database is fully accessible and searchable and can be added to by other researchers around the globe. This provides a framework upon which the distribution of genes involved in pathogenicity can be superimposed and that can be expanded in future studies to improve understanding of the population structure and global epidemiology of this pathogen.

For the development of a MLST scheme for S. uberis, seven housekeeping gene loci, which could be amplified and sequenced from a random collection of S. uberis isolates, were chosen. At the beginning of the present study, the genome sequence of S. uberis was incomplete, so the distribution of the loci throughout the genome was not known. The completion of the genome sequence showed that the minimum distance between loci was 25 kb. The seven loci chosen did not encode surface proteins or proteins involved in pathogenicity and were not horizontally acquired genes. Their dN/dS ratios were substantially less than 1, which shows they are not subject to positive selection. For the 160 isolates included in the present study, an average of 10.4 alleles were obtained for each locus, which is consistent with previous typing studies suggesting that S. uberis is a diverse species (1, 4, 28). The percentage of variable nucleotide sites (1.1 to 4.2%) is slightly higher than that seen in group B streptococcus (19) (1.1 to 2.5%) but considerably lower than that of S. suis (20) (11.0 to 29.0%), although it should be noted that our data were based on a relatively small number of farms. In addition, analysis of the sequences using ML trees provided evidence of significant recombination within S. uberis. This is likely to have occurred by transduction since S. uberis is not naturally competent but does carry phages (13).

The analysis performed in the present study revealed that the 160 isolates could be resolved into 57 STs, of which ST-5 was the most prevalent. The four most common STs represented 40.6% of the strain collection. Clustering analysis revealed that there was one major lineage within this collection of isolates. This lineage (the ST-5 clonal complex) contained the six most common STs, which suggests it may be well fitted to infection of the bovine mammary gland. In contrast, most of the unassigned STs contained only one or two isolates. Interestingly, the well-characterized virulent strain 0140J is a member of the ST-5 complex, whereas the less-virulent EF20 is an unrelated isolate.

Comparison of the isolates obtained from IAH and non-IAH herds revealed clear differences in the ST profile between the two groups. Although several STs were common to both groups of isolates, the predominant ST from the IAH herds (ST-5) was only isolated once in the non-IAH herds. ST-4 and ST-6 (single locus variants of ST-5) were more prevalent in these herds. A similar outcome was observed in a previous study of S. uberis isolates from two Dutch dairy herds using RAPD (random amplification of polymorphic DNA) fingerprinting (37). In that study, different predominant isolates were identified in each of the herds, although 2 of 17 RAPD types were found in both herds. In contrast, although typing studies based on restriction profiles have identified predominant isolates within herds, these have often failed to find evidence of the same type in different herds (1, 16). Thus, typing of S. uberis by MLST is clearly more suitable than PFGE for long-term epidemiology.

Analysis of the data with respect to the disease status of the cow from which the isolates were obtained showed that none of the more common STs was exclusively associated with clinical mastitis or with subclinical infection but did reveal some interesting trends. In particular, ST-23 was more commonly associated with subclinical infection than the other STs, suggesting it could be less pathogenic. In contrast ST-10 was usually isolated from clinical and posttreatment cases, suggesting it could have increased virulence. The significance of these findings will require examination of a larger number of isolates. However, no significant differences in disease status were observed between the ST-5 lineage compared to other isolates. The IAH herd isolates showed a higher proportion of clinical and posttreatment isolates (50.8%) than the non-IAH herd isolates (34.5%). This finding may be a result of the management practices used by the different farms or increased vigilance allowing detection of clinical signs of disease at the IAH. Previous studies have shown differences in the disease treatment levels between farms with different management regimes, with organic herds treating less often than conventional herds (10). However, it was unclear if this accurately reflected the level of disease in each type of herd.

Although multiple isolates from individual cows at different stages of infection were not analyzed in this work, previous studies where infections by S. uberis were followed from one lactation to the next showed that in the 13 quarters infected in both lactations, different strains were detected in 12 quarters (28). MLST will be a useful tool for typing strains from persistent and recurring cases of mastitis, and will allow a definitive determination of whether the infecting isolates are the same strain or if a new infection occurs.

Information about the population structure of S. uberis obtained by MLST will aid the understanding of its pathogenicity, since the carriage of potential virulence genes by particular lineages can now be investigated. This is exemplified by the analysis of hasA carriage, a gene required for capsule biosynthesis (34). The role of capsule in pathogenicity is currently unclear. Early studies showed that the production of capsule by S. uberis correlated with the ability to resist phagocytosis by neutrophils, with the removal of the capsule allowing strains to be readily phagocytosed (24). In contrast, later studies showed that acapsular mutants were equally able to persist in cows after experimental challenge, demonstrating that the capsule is not required for the development of infection and clinical mastitis in this model (8). In the present study, a population negative for hasA, and therefore devoid of capsule, was identified. Highly significant differences were seen in the ST profile obtained from isolates with or without the hasA gene, with all 112 ST-5 complex strains having hasA. Examination of the status of cows from which the hasA-negative isolates were obtained showed the majority (16 of 21 [76.2%]) to be from subclinical infections, with only two isolates from clinical cases and three isolates from posttreatment cases. Interestingly, presence of hasA was also shown to correlate with isolation from clinical disease in a selection of isolates from Denmark (8), indicating that our observations are not a phenomenon local to the United Kingdom. Since capsule appears unlikely to contribute to disease directly it may be of interest to determine which genes other than hasA are also linked to lineage. It is unclear whether hasA has spread horizontally through the population or whether it has been lost from some STs.

In conclusion, we have developed the first classical MLST scheme for a major bovine pathogen of economic importance, S. uberis, and demonstrated its potential for future studies of global epidemiology, population biology, and pathogenesis.


This study was supported by BBSRC grant 201/S51848.

The assistance of Elizabeth Berry and Kirsty Kliem in the collection of samples is greatly appreciated.


1. Baseggio, N., P. D. Mansell, J. W. Browning, and G. F. Browning. 1997. Strain differentiation of isolates of streptococci from bovine mastitis by pulsed-field gel electrophoresis. Mol. Cell Probes 11:349-354. [PubMed]
2. Bramley, A. J. 1984. Streptococcus uberis udder infection-a major barrier to reducing mastitis incidence. Br. Vet. J. 140:328-335. [PubMed]
3. Dingle, K. E., F. M. Colles, D. R. A. Wareing, R. Ure, A. J. Fox, F. E. Bolton, H. J. Bootsma, R. J. L. Willems, R. Urwin, and M. C. J. Maiden. 2001. Multilocus sequence typing system for Campylobacter jejuni. J. Clin. Microbiol. 39:14-23. [PMC free article] [PubMed]
4. Douglas, V. L., S. G. Fenwick, D. U. Pfeiffer, N. B. Williamson, and C. W. Holmes. 2000. Genomic typing of Streptococcus uberis isolates from cases of mastitis, in New Zealand dairy cows, using pulsed-field gel electrophoresis. Vet. Microbiol. 75:27-41. [PubMed]
5. Enright, M. C., and B. G. Spratt. 1998. A multilocus sequence typing scheme for Streptococcus pneumoniae: identification of clones associated with serious invasive disease. Microbiology 144:3049-3060. [PubMed]
6. Feil, E. J., E. C. Holmes, D. E. Bessen, M.-S. Chan, N. P. J. Day, M. C. Enright, R. Goldstein, D. W. Hood, A. Kalia, C. E. Moore, J. Zhou, and B. J. Spratt. 2001. Recombination within natural populations of pathogenic bacteria: short-term empirical estimates and long-term phylogenetic consequences. Proc. Natl. Acad. Sci. USA 98:182-187. [PMC free article] [PubMed]
7. Feil, E. J., B. C. Li, D. M. Aanensen, W. P. Hanage, and B. G. Spratt. 2004. eBURST: inferring patterns of evolutionary descent among clusters of related bacterial genotypes from multilocus sequence typing data. J. Bacteriol. 186:1518-1530. [PMC free article] [PubMed]
8. Field, T. R., P. N. Ward, L. H. Pedersen, and J. A. Leigh. 2003. The hyaluronic acid capsule of Streptococcus uberis is not required for the development of infection and clinical mastitis. Infect. Immun. 71:132-139. [PMC free article] [PubMed]
9. Fontaine, M. C., J. Perez-Casal, X.-M. Song, J. Shelford, P. J. Willson, and A. A. Potter. 2002. Immunization of dairy cattle with recombinant Streptococcus uberis GapC or a chimeric CAMP antigen confers protection against heterologous bacterial challenge. Vaccine 20:2278-2286. [PubMed]
10. Hardeng, F., and V. L. Edge. 2001. Mastitis, ketosis, and milk fever in 31 organic and 93 conventional Norwegian dairy herds. J. Dairy Sci. 84:2673-2679. [PubMed]
11. Hassan, A. A., I. U. Khan, A. Abdulmawjood, and C. Lämmler. 2001. Evaluation of PCR methods for rapid identification and differentiation of Streptococcus uberis and Streptococcus parauberis. J. Clin. Microbiol. 39:1618-1621. [PMC free article] [PubMed]
12. Hill, A. W. 1988. Pathogenicity of two strains of Streptococcus uberis infused into lactating and non-lactating bovine mammary glands. Res. Vet. Sci. 45:400-404. [PubMed]
13. Hill, A. W., and C. A. Brady. 1989. A note on the isolation and propagation of lytic phages from Streptococcus uberis and their potential for strain typing. J. Appl. Bacteriol. 67:425-431. [PubMed]
14. Hillerton, J. E., M. F. S. Shearn, R. M. Teverson, S. Langridge, and J. M. Booth. 1993. Effect of pre-milking teat dipping on clinical mastitis on dairy farms in England. J. Dairy Res. 60:31-41. [PubMed]
15. Holmes, E. C., R. Urwin, and M. C. J. Maiden. 1999. The influence of recombination on the population structure and evolution of the human pathogen Neisseria meningitidis. Mol. Biol. Evol. 16:741-749. [PubMed]
16. Jayarao, B. M., E. E. Schilling, and S. P. Oliver. 1993. Genomic deoxyribonucleic acid restriction fragment length polymorphism of Streptococcus uberis: evidence of clonal diversity. J. Dairy Sci. 76:468-474. [PubMed]
17. Jolley, K. A., E. J. Feil, M.-S. Chan, and M. C. J. Maiden. 2001. Sequence type analysis and recombinational tests (START). Bioinformatics 17:1230-1231. [PubMed]
18. Jolley, K. A., M.-S. Chan, and M. C. J. Maiden. 2004. mlstdbNET- distributed multi-locus sequence typing (MLST) databases. BMC Bioinformatics 5:86-93. [PMC free article] [PubMed]
19. Jones, N., J. F. Bohnsack, S. Takahashi, K. A. Oliver, M.-S. Chan, F. Kunst, P. Glaser, C. Rusniok, D. W. M. Crook, R. M. Harding, N. Bisharat, and B. G. Spratt. 2003. Multilocus sequence typing system for group B streptococcus. J. Clin. Microbiol. 41:2530-2536. [PMC free article] [PubMed]
20. King, S. J., J. A. Leigh, P. J. Heath, I. Luque, C. Tarradas, C. G. Dowson, and A. M. Whatmore. 2002. Development of a multilocus sequence typing scheme for the pig pathogen Streptococcus suis: identification of virulent clones and potential capsular serotype exchange. J. Clin. Microbiol. 40:3671-3680. [PMC free article] [PubMed]
21. Kossaibati, M. A., and R. J. Esslemont. 1997. The costs of production diseases in dairy herds in England. Vet. J. 154:41-51. [PubMed]
22. Lämmler, C. 1991. Biochemical and serological properties of Streptococcus uberis. Zentbl. Veterinarmed. B. 38:737-742. [PubMed]
23. Leigh, J. A. 1999. Streptococcus uberis: a permanent barrier to the control of bovine mastitis? Vet. J. 157:225-238. [PubMed]
24. Leigh, J. A., and T. R. Field. 1994. Streptococcus uberis resists the bactericidal action of neutrophils despite the presence of bound immunoglobulin. Infect. Immun. 62:1854-1859. [PMC free article] [PubMed]
25. Maiden, M. C., J. A. Bygraves, E. Feil, G. Morelli, J. E. Russell, R. Urwin, Q. Zhang, J. Zhou, K. Zurth, D. A. Caugant, I. M. Feavers, M. Achtman, and B. G. Spratt. 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. [PMC free article] [PubMed]
26. Nei, M., and T. Gojobori. 1986. Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol. Biol. Evol. 3:418-426. [PubMed]
27. Oliver, S. P., B. E. Gillespie, and B. M. Jayarao. 1998. Detection of new and persistent Streptococcus uberis and Streptococcus dysgalactiae intramammary infections by polymerase chain reaction-based DNA fingerprinting. FEMS Microbiol. Lett. 160:69-73. [PubMed]
28. Phuektes, P., P. D. Mansell, R. S. Dyson, N. D. Hooper, J. S. Dick, and G. F. Browning. 2001. Molecular epidemiology of Streptococcus uberis isolates from dairy cows with mastitis. J. Clin. Microbiol. 39:1460-1466. [PMC free article] [PubMed]
29. Selander, R. K., D. A. Caugant, H. Ochman, J. M. Musser, M. N. Gilmour, and T. S. Whittam. 1986. Methods of Multilocus Enzyme Electrophoresis for bacterial population genetics and systematics. Appl. Environ. Microbiol. 51:873-884. [PMC free article] [PubMed]
30. Smith, A. J., A. J. Kitt, P. N. Ward, and J. A. Leigh. 2002. Isolation and characterization of a mutant strain of Streptococcus uberis, which fails to utilize a plasmin derived β-casein peptide for the acquisition of methionine. J. Appl. Microbiol. 93:631-639. [PubMed]
31. Sneath, P. H. A., and R. R. Sokal. 1973. Numerical taxonomy, p. 230-234. W. H. Freeman and Co., San Francisco, Calif.
32. Swofford, D. L. 2003. PAUP*: phylogenetic analysis using parsimony (*and other methods), version 4. Sinauer Associates, Sunderland, Mass.
33. Tagg, J. R., and L. G. Vugler. 1986. An inhibitor typing scheme for Streptococcus uberis. J. Dairy Res. 53:451-456. [PubMed]
34. Ward, P. N., T. R. Field, W. G. F. Ditcham, E. Maguin, and J. A. Leigh. 2001. Identification and disruption of two discrete loci encoding hyaluronic acid capsule biosynthesis genes hasA, hasB, and hasC in Streptococcus uberis. Infect. Immun. 69:392-399. [PMC free article] [PubMed]
35. Ward, P. N., and J. A. Leigh. 2002. Characterization of PauB, a novel broad-spectrum plasminogen activator from Streptococcus uberis. J. Bacteriol. 184:119-125. [PMC free article] [PubMed]
36. Wells, S. J., S. L. Ott, and A. H. Seitzinger. 1998. Key health issues for dairy cattle-new and old. J. Dairy Sci. 81:3029-3035. [PubMed]
37. Zadoks, R. N., B. E. Gillespie, H. W. Barkema, O. C. Sampimon, S. P. Oliver, and Y. H. Schukken. 2003. Clinical, epidemiological, and molecular characteristics of Streptococcus uberis infections in dairy herds. Epidemiol. Infect. 130:335-349. [PMC free article] [PubMed]
38. Zadoks, R. N., Y. H. Schukken, and M. Wiedmann. 2005. Multilocus sequence typing of Streptococcus uberis provides sensitive and epidemiologically relevant subtype information and reveals positive selection in the virulence gene pauA. J. Clin. Microbiol. 43:2407-2417. [PMC free article] [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...