Mastitis is the disease with the greatest economic effect on the dairy industry in developed countries (Blosser, Reference Blosser1979; Azooz et al., Reference Azooz, El-Wakeel and Yousef2020). Owing to the recent increase in socioeconomic pressures arising from the transition to prudent antibiotic use (Van Werven, Reference Van Werven2018), government policies and public perception are also scrutinising and restricting the dairy sector (Hayman et al., Reference Hayman, Adisasmito, Almuhairi, Behravesh, Bilivogui, Bukachi, Casas, Becerra, Charron, Chaudhary, Janice, Zanella, Cunningham, Dar, Debnath, Dungu, Farag, Gao, Khaitsa, Machalaba, Mackenzie, Wanda Markotter, Mettenleiter, Morand, Smolenskiy, Zhou and Koopmans2023). These pressures and additional economic constraints have compelled farmers to adopt alternative farming practices to reduce costs while meeting the increased demands of modern-day dairy animals (Youngerman et al., Reference Youngerman, Saxton, Oliver and Pighetti2004; Neculai-Valeanu and Ariton, Reference Neculai-Valeanu and Ariton2022). The constant demand for ever higher milk production increases the risk of dairy animals suffering from clinical and subclinical mastitis.

A New Zealand study from 1997, involving 38 dairy herds (n = 798), established that 74.7% of bacteria isolated from clinical mastitis cases were S. uberis (McDougall, Reference McDougall1998), emphasising the potential pathogenicity of this bacterium. Changes in farming and management practices in South Africa can shorten the composting time or impede the composting process for manure-based bedding in total mixed ration systems. Among pasture-based herds inspired by the New Zealand approach, slurry is increasingly used as a fertiliser (Jeppsson et al., Reference Jeppsson, Magnusson, Bergström Nilsson, Ekman, Winblad von Walter, Jansson, Landin, Rosander and Bergsten2024). These higher pathogen loads increase the risk of intramammary infection (IMI; Sherwin et al., Reference Sherwin, Egan, Green and Leigh2021). The growing herd sizes in South Africa, combined with higher stocking densities and heat stress, further increase the burden on animals (Lactodata, 2023). Both of these factors could favour a change in the udder's biome, increasing the risk of pathogens, such as S. uberis (Blignaut et al., Reference Blignaut, Thompson and Petzer2018).

Globally, IMI is more likely to be caused by S. uberis than other pathogens (Petzer et al., Reference Petzer, Karzis, Watermeyer, Van der Schans and Van Reenen2009). In a review article by Lopez-Benavides et al. (Reference Lopez-Benavides, Williamson, Pullinger, Lacy-Hulbert, Cursons and Leigh2007), the biomes vary in dominance of S. uberis. The review indicated that the environment of cattle poses a high risk, with 25% of cows excreting S. uberis in the faeces. This fact could explain the substantial increase in S. uberis IMI when untreated slurry is used on pastures for dairy cows, as described by several authors (Zadoks et al., Reference Zadoks, Allore, Barkema, Sampimon, Gröhn and Schukken2001; McDougall et al., Reference McDougall, Parkinson, Leyland, Anniss and Fenwick2004; Zadoks, Reference Zadoks2007; Tomita et al., Reference Tomita, Meehan, Wongkattiya, Malmo, Pullinger, Leigh and Deighton2008; Sampimon et al., Reference Sampimon, Barkema, Berends, Sol and Lam2009; Reyes et al., Reference Reyes, Rodriguez-Lecompte, Blanchard, McClure and Sánchez2019; Srithanasuwan et al., Reference Srithanasuwan, Pangprasit and Suriyasathaporn2022). Several South African pasture-based dairy farms recently started using slurry on pastures; therefore, it is probable that this practice contributed to the 10% increase in the isolation of S. uberis from IMI from 2016 to 2022 (Karzis & Petzer, submitted).

Several treatment practices may encourage the selection of virulence and resistance in S. uberis pathogens (Hoque et al., Reference Hoque, Istiaq, Rahman, Islam, Anwar, Siddiki, Sultana, Crandall and Hossain2020; Zhang et al., Reference Zhang, Feng, Xin-pu, Jin-yin, Ling, Yu-long, Yong, Xu-rong and Hong-sheng2020). Drug resistance (Zhang et al., Reference Zhang, Feng, Xin-pu, Jin-yin, Ling, Yu-long, Yong, Xu-rong and Hong-sheng2020) in S. uberis species has been observed globally, with cases reported in South America (Gianneechini et al., Reference Gianneechini, Concha and Franklin2002), New Zealand (Petrovski et al., Reference Petrovski, Williamson, Lopez-Villalobos, Parkinson and Tucker2011; McDougall et al., Reference McDougall, Clausen, Ha, Gibson, Bryan, Hadjirin, Lay, Raisen, Ba, Restif, Parkhill and Holmes2020), Canada (Cameron et al., Reference Cameron, Saab, Heider, McClure, Rodriguez-Lecompte and Sanchez2016) and North Africa (Saed and Ibrahim, Reference Saed and Ibrahim2020). The successful treatment of S. uberis, a bacterium that has the potential to reside in the udder parenchyma as a part of its defence and resistance mechanism, is a complex process. This is owing to virulence properties, such as the Streptococcus uberis adhesion molecule (Almeida et al., Reference Almeida, Luther, Park and Oliver2006), biofilm expression (Krömker and Matthies, Reference Krömker and Matthies2014; Magagula et al., Reference Magagula, Petzer, Famuyide and Karzis2023) and antibiotic resistance genes (Reyes et al., Reference Reyes, Rodriguez-Lecompte, Blanchard, McClure and Sánchez2019; Srithanasuwan et al., Reference Srithanasuwan, Pangprasit and Suriyasathaporn2022).

In South Africa, molecular epidemiology data on S. uberis in dairy cows are currently absent, unlike in other countries, where numerous studies have been conducted (Jolley et al., Reference Jolley, Bray and Maiden2018; Zhang et al., Reference Zhang, Feng, Xin-pu, Jin-yin, Ling, Yu-long, Yong, Xu-rong and Hong-sheng2020; Srithanasuwan et al., Reference Srithanasuwan, Pangprasit and Suriyasathaporn2022). Identifying strains through molecular characterisation can enhance the general understanding of the dynamic evolution of S. uberis biomes (Lopez-Benavides et al., Reference Lopez-Benavides, Williamson, Pullinger, Lacy-Hulbert, Cursons and Leigh2007). The latter may assist to distinguish between persistent (host adapted) and transient (recurrent or environmental) IMI of S. uberis isolated from the same cow. Persistent S. uberis IMI being when the same S. uberis strain is isolated on consecutive samplings from the same cow. However, transient IMI in a cow will be an infection from a different S. uberis strain identified on consecutive samplings. Consecutive sampling should be within acceptable time periods (McDougall et al., Reference McDougall, Parkinson, Leyland, Anniss and Fenwick2004).

Given the persistent and transient S. uberis strains (Jeppsson et al., Reference Jeppsson, Magnusson, Bergström Nilsson, Ekman, Winblad von Walter, Jansson, Landin, Rosander and Bergsten2024), it is increasingly important to identify the specific strain and its associated virulence profile. Phenotypic diagnosis fails to distinguish between S. uberis strains, which can lead to the elevated and incorrect diagnosis of persistent IMI cases in cows. Genomic characterisation, such as multilocus sequence typing (MLST) using a seven-gene approach for housekeeping of genetic material and the evaluation of what is required and expressed by all single cells and cells of an organism under normal and pathophysiological conditions, can be used for identification to strain type level (Butte et al., Reference Butte, Dzau and Glueck2001; Coffey et al., Reference Coffey, Pullinger, Urwin, Jolley, Wilson, Maiden and Leigh2006). This method can serve as a foundation for further whole-genome studies and identifying virulence factors in S. uberis. Such information is currently lacking in South Africa.

This study was the first of its kind to be conducted in a South African dairy herd. The aim was to assess the prevalence of S. uberis STs in a herd and to compare these results with those from other countries in the PubMLST/GenBank database. This longitudinal study also enabled the investigation of the epidemiological behaviour of these S. uberis STs to determine whether the STs were mainly persistent or transient. Bacteria that became persistent adapted to the host environment, whereas the transient bacteria remained in the udder for a limited period and may be more indicative of bacteria preferring to live in the environment.

Accordingly, veterinarians and farmers may be better equipped to understand the risk-reward balance associated with various farming practices, such as green manure bedding and untreated slurry pasture fertilisation, both in South Africa and in future contexts.

Materials and methods

Experimental design

A retrospective longitudinal study was conducted on S. uberis isolated from composite cow milk samples from cows considered being healthy or experiencing subclinical and clinical mastitis from a large commercial dairy herd in South Africa.

Test herd

This pasture-based herd was a closed herd comprising approximately 1005 Jersey cows in milk, with year-round calving. It is located in the Eastern Cape, one of South Africa's three major dairy regions. For the past 15 years, microbiology and cytology have been performed on milk samples collected quarterly from all lactating cows as part of an ongoing proactive udder health management strategy. At the time of the study (2021), the prevalence of S. uberis IMI isolated during the routine quarterly herd sampling was 7.44% (Table 1).

Table 1.

Streptococcus uberis isolated during 2021 from routine quarterly herd examinations and monthly clinical mastitis cases

Table 1 Long description

Quarterly routine herd sampling and monthly clinical mastitis cultures are summarized for 2021, reporting how often Streptococcus uberis was isolated and how many cows had repeat isolations. Routine sampling tested 4,020 cows across the year, with S. uberis isolated 301 times overall, or 7.48 percent. By quarter, routine isolation rates were similar: 7.35 percent (Jan to Mar), 8.09 percent (Apr to Jun), 7.18 percent (Jul to Sept), and 7.12 percent (Oct to Dec). Clinical mastitis cultures yielded 64 S. uberis isolates, representing 21.26 percent of clinical cases overall, with a decline from 34.61 percent in Jan to Mar to 12.67 percent in Jul to Sept, then 15.27 percent in Oct to Dec. Repeat-isolation patterns differed: in routine sampling, 142 cows had one isolation, 35 had two consecutive, 13 had three consecutive, and 7 had four consecutive; additional nonconsecutive repeats were 8 cows with two and 2 cows with three. In clinical cases, 48 cows had one S. uberis case, 14 had two consecutive, and 1 cow each had three consecutive and four consecutive, with no nonconsecutive repeats reported. Percentages for repeat categories are reported relative to S. uberis isolates or clinical S. uberis cases, so they are not directly comparable to the overall testing denominators.

Sampling for molecular characterisation

Streptococcus uberis, isolated from two routine herd examinations conducted during July and October 2021 and from clinical mastitis cases that occurred during July 2021, was available for characterisation. Sixty-three S. uberis isolates were selected at random from these isolates for MLST determination. The authors specifically selected seven S. uberis isolates in 2024 (February to June) from the same herd, which were obtained from the same cows in consecutive samplings, to increase the sample size for the longitudinal investigation.

Sampling method

Milk samples (foremilk) were collected aseptically, kept on ice and transported to the Milk Laboratory, Faculty of Veterinary Science, Onderstepoort (SANAS-accredited) within 48 hours, for microbiological and cytological analysis. Foremilk samples from clinical mastitis cases were collected upon detection (altered milk quality, i.e., floccules or udder changes), frozen and sent to the laboratory for analysis in batches at monthly intervals.

Microbiology and cytology

Bacterial identification of isolates from routine herd samples was performed phenotypically, using classical microbiological culture methods. These included assessments of colony morphology, haemolysis, pigmentation, catalase, staphylase/coagulase, maltose and potassium hydroxide tests. Identification followed the guidelines outlined by the National Mastitis Council (2017). Verification of the S. uberis selected for MLST was conducted using matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (Bruker Daltonics, Bremen, Germany).

Somatic cell count (SCC) analysis was done using fluoro-optoelectronic technology with the Fossomatic (FC) (Rhine Ruhr, Wendywood, South Africa). Streptococcus uberis isolates were transferred to cryotubes containing tryptic soy broth with 10% glucose and stored in an approved biobank at −80°C (Milk Laboratory, Department of Production Animal Studies, University of Pretoria).

Multilocus sequence typing

Molecular characterisation was performed at Inqaba Biotec™ (Pretoria, South Africa) on 70 isolates: 12 from clinical mastitis, 40 from subclinical cases (≥300,000 cells/ml) and 18 from milk with low SCC (<300 cells/ml) using MLST from well-established methods (Jolley et al., Reference Jolley, Chan and Maiden2004, Reference Jolley, Bray and Maiden2018; Coffey et al., Reference Coffey, Pullinger, Urwin, Jolley, Wilson, Maiden and Leigh2006).

Genomic DNA was extracted from the 70 isolates using the Quick-DNA™ Fungal/Bacterial Miniprep Kit (Zymo Research, Catalogue No. D6005) following the manufacturer’s instructions (https://files.zymoresearch.com/protocols/_d6005_quick-dna_fungal-bacterial_miniprep_kit.pdf) and according to the study by Coffey et al. (Reference Coffey, Pullinger, Urwin, Jolley, Wilson, Maiden and Leigh2006).

In this study, the MLST target regions comprised seven housekeeping genes: glucose kinase (gki), transketolase (recP), D-ala-D-ala ligase (ddl), thymidine kinase (tdk), carbamate kinase (arcC), triosephosphate isomerase (tpi) and acetyl-CoA acetyltransferase (yqiL). These regions amounted to 3338 base pairs per isolate. The DNA was amplified and purified for sequencing.

Isolates’ DNA was prepared using the Exonuclease I and Shrimp Alkaline Phosphatase (ExoSAP) protocol (Zymo Research, ZR-96 DNA Sequencing Clean-up Kit™, Catalogue No. D4050). The integrity of genomic DNA was assessed using a 1% agarose gel (CSL-AG500, Cleaver Scientific Ltd) stained with EZ-Vision® Bluelight DNA dye. Sequencing was performed in a forward direction using the Nimagen BrilliantDye™ Terminator Cycle Sequencing Kit V3.1 (BRD3-100/1000) and the ABI 3730xl Genetic Analyser (Applied Biosystems, Thermo Fisher Scientific).

The Fast Alignment of Sequences Tool (FASTA) sequence alignment data for all novel alleles in the MLST dataset were submitted to the PubMLST database. These data were curated, assigned ST numbers and grouped into available clonal complex (CC) where possible. The data are available on PubMLST at the link provided under ‘country = South Africa’ https://pubmlst.org/bigsdb?db=pubmlst_suberis_isolates&l=1&page=query. A dendrogram was created to display the relative similarity among the tested isolates (Figure S1).

Statistical analysis

The software CLC Main Workbench 24.0 was used to construct the dendrogram, employing the neighbour-joining method to calculate the relevant similarities among the tested isolates. This was generated without an existing tree, using the Hasegawa–Kishino–Yano nucleotide substitution model and the WAG protein substitution model. Other parameters included a transition/transversion ratio of 2.0; a gamma distribution parameter of 1.0; 100 replicates; four substitution rate categories; rate of variation; no gamma distribution and bootstrap analysis.

Ethics consideration

Ethical approval was obtained from the research and animal ethics committees of the Faculty of Veterinary Science, University of Pretoria (REC 040-23).

Results

Results from conventional microbiology for routine quarterly testing and clinical mastitis cases

During the four quarterly routine herd tests in 2021, 301 S. uberis were isolated from 207 (20.6%) of the 1005 cows sampled quarterly. This represented an average S. uberis IMI of 7.44% in the study herd during 2021. Although the S. uberis IMI was relatively low in the herd, this bacterium was responsible for 21.26% of clinical mastitis cases identified in the same year (Table 1).

MLST was performed on 70 S. uberis isolates: 12 from clinical cases, 40 from subclinical cases and 18 from milk with low SCC. A dendrogram was constructed using the isolates’ STs and is presented in Figure S1. A total of 42 S. uberis sequence types (STs) were identified from this one herd. Of these, 41 were novel strains (Table 2, Table S1 and Figure S1), while only 1 strain (ST 1613) was present in the PubMLST database and had previously been isolated from bovine milk in Australia (McDougall et al., Reference McDougall, Parkinson, Leyland, Anniss and Fenwick2004). Six ST 1613 were isolated from four cows, with two cows yielding the strain twice, all from subclinical mastitis cases.

Table 2.

Multilocus sequence typing with sequence type (ST) and clonal complex (CC) results of the 70 Streptococcus uberis isolates

Table 2 Long description

Multilocus sequence typing results are listed for 70 Streptococcus uberis isolates, linking each isolate ID to cow ID, sampling date, recurrence status, udder health category, sequence type number, and clonal complex or nearest match. Samples span July 2021 and October 2021, with additional follow-up isolates in early to mid 2024 for a few cows. Many cows have a single isolate and a single ST, indicating high diversity across the herd. Recurring isolates appear in multiple cows and are flagged as either similar or different across time; some cows keep the same ST across two dates (for example cow 1234 with ST 2106, cow 11,232 with ST 2132, and cow 11,251 with ST 1613a), while others switch STs between episodes (for example cow 1327 and cow 15,310). Cow 1788 has multiple clinical mastitis isolates across 2021 to 2024 with different STs, suggesting repeated infections rather than persistence of one strain. Clonal complex assignments are often absent, but when present they frequently reference ST-5 or ST-86, with occasional ST-143; several entries note only a nearest match rather than a definitive complex. Udder health status varies across subclinical, clinical, and low somatic cell categories, and the same ST can occur in different health states, so the table supports tracking strain recurrence but does not by itself establish causation between ST and disease severity.

^a Pre-existing ST; SIM, similar; DIF, different; CM, clinical mastitis; SC, subclinical mastitis; LSC, low somatic cell.

Three common CCs, ST-5 CC (12/70 isolates), ST-86 CC (9/70 isolates) and ST-143 CC (2/70 isolates) (pubmlst.org), were identified among the ST strains (Table 3).

Table 3.

Streptococcus uberis strains isolated from consecutive intramammary infections over time, indicating persistent (same strains) and transient (different strains) intramammary infections

Table 3 Long description

Consecutive milk samplings are summarized for cows where more than one Streptococcus uberis isolate was obtained, listing cow ID, total phenotypic isolations during the study, sampling dates used for MLST, sequence types, clonal complexes, and whether cases were subclinical or clinical mastitis. Seven cows had the same sequence type on consecutive samplings, suggesting persistence: 1234 (ST2106), 1318 (ST2127), 11232 (ST2132), 11251 (ST1613), 13204 (ST2105), 17280 (ST2160), and 17352 (ST1613). In this persistent group, most outcomes were two subclinical cases; cow 17280 included one subclinical and one clinical case. Seven cows had different sequence types across consecutive samplings, suggesting transient infection: 1327, 1788, 1981, 12264, 15310, 17151, and 17175, with some cows showing multiple sequence types across three or four sampling points. Clinical mastitis appears mainly in the transient group, notably cow 1327 with two clinical cases and cow 1788 with three clinical cases. Clonal complexes vary, with several entries marked as unknown and others assigned to complexes such as ST-5, ST-83, ST-86, and ST-143. Interpretation is limited to strain typing across the listed sampling dates and does not by itself confirm transmission routes or the exact duration of infection between samplings.

^a PubMLST data (Streptococcus uberis isolates (pubmlst.org)).

^b Unknown clonal complexes.

MLST data for cows from which S. uberis was isolated in consecutive samplings

Among the 14 cows with multiple S. uberis isolations in consecutive samplings, 7 had similar and 7 yielded dissimilar strains (Tables 2 and 3), revealing that only 7/53 (13.2%) cows tested exhibited persistent IMI, while most had transient IMI.