Molecular Typing and Antimicrobial Susceptibility Profiles of Streptococcus uberis Isolated from Sheep Milk

Intramammary infections are a major problem for dairy sheep farms, and Streptococcus uberis is one of the main etiological agents of ovine mastitis. Surveys on antimicrobial resistance are still limited in sheep and characterization of isolates is important for acquiring information on resistance and for optimizing therapy. In this study, a sampling of 124 S. uberis isolates collected in Sardinia (Italy) from sheep milk was analyzed by multilocus-sequence typing (MLST) and pulsed field gel electrophoresis (PFGE) for genetic relatedness. All isolates were also subjected to antimicrobial susceptibility analysis by the disk diffusion test using a panel of 14 antimicrobials. Resistance genes were detected by PCR assays. MLST analysis revealed that the isolates were grouped into 86 sequence types (STs), of which 73 were new genotypes, indicating a highly diverse population of S. uberis. The most frequently detected lineage was the clonal complex (CC)143, although representing only 13.7% of all characterized isolates. A high level of heterogeneity was also observed among the SmaI PFGE profiles, with 121 unique patterns. Almost all (96.8%) isolates were resistant to at least one antimicrobial, while all exhibited phenotypic susceptibility to oxacillin, amoxicillin-clavulanic acid and ceftiofur. Of the antimicrobials tested, the highest resistance rate was found against streptomycin (93.5%), kanamycin (79.8%) and gentamicin (64.5%), followed by novobiocin (25%) and tetracycline-TE (19.3%). Seventy-four (59.7%) isolates were simultaneously resistant to all aminoglycosides tested. Seventeen isolates (13.7%) exhibited multidrug resistance. All aminoglycosides-resistant isolates were PCR negative for aad-6 and aphA-3′ genes. Among the TE-resistant isolates, the tetM gene was predominant, indicating that the resistance mechanism is mainly mediated by the protection of ribosomes and not through the efflux pump. Three isolates were resistant to erythromycin, and two of them harbored the ermB gene. This is the first study reporting a detailed characterization of the S. uberis strains circulating in Sardinian sheep. Further investigations will be needed to understand the relationships between S. uberis genotypes, mastitis severity, and intra-mammary infection dynamics in the flock, as well as to monitor the evolution of antimicrobial resistance.


Introduction
Mastitis is the most prevalent and costly disease affecting dairy sheep. Sardinia, an island located in the middle of the Mediterranean Sea, has approximately 3.5 million milking Sarda sheep, corresponding to half of the total Italian stock. Shepherding is a relevant part of the regional economy, particularly concerning pecorino cheese production. Therefore, udder health is a critical factor, and control of intra-mammary infections (IMI) is consequently of the greatest importance for dairy farmers. Infectious mastitis outbreaks of small ruminants are usually caused by Gram-positive bacteria, mostly Staphylococcus and Streptococcus species [1][2][3][4]. In a recent study concerning the identification of Streptococcus isolates, we found that Streptococcus (S.) uberis is the pathogen most frequently isolated Pathogens 2021, 10, 1489 2 of 11 from sheep and goat milk samples [5]. S. uberis is considered an environmental pathogen since it is commonly isolated from the environment, milking machines, milkers' hands, and the skin of the teats. Mastitis may therefore arise when the mammary gland is exposed to high loads of this bacterial species.
Genotyping of S. uberis isolated from milk samples is an essential tool in mastitis epidemiology and contributes to our understanding of the pathogen's dissemination. At present, very few studies report the genotypic characterization of S. uberis isolated from the milk of sheep with infectious mastitis [6]. On the other hand, several molecular methods have been applied for typing S. uberis isolated from bovine milk, including multi locus sequence typing (MLST) and pulsed field gel electrophoresis (PFGE) [7][8][9]. PFGE is the most discriminatory method in outbreak analysis [10], whereas MLST is considered the method of choice to investigate populations and population dynamics on a local and global scale [11].
Antimicrobial therapy continues to play an important role in controlling both clinical and sub-clinical streptococcal mastitis. For preventing the emergence and selection of multidrug-resistant pathogens, it is essential to monitor the trends in antimicrobial resistance (AMR) among bacteria. The World Health Organization (WHO) has declared that AMR is one of the top 10 global public health threats facing humanity. The main AMR drivers include the misuse and overuse of antimicrobials; lack of access to clean water, sanitation, and hygiene for both humans and animals; poor infection and disease prevention measures, and control in health-care facilities and farms.
The study aimed to analyze the genetic diversity and the antimicrobial susceptibility profiles of S. uberis field isolates collected during a 7-year period from small ruminants in Sardinia, Italy.

Streptococcus uberis Genotyping
All 124 S. uberis were typed by both MLST and PFGE analysis. MLST analysis identified 86 different allelic combinations; among these, 13 were STs present in the database while the remaining 73 (84.9%) STs were found for the first time in this study. The ST was considered new if the allelic combination did not overlap with known STs in the S. uberis MLST database. The list of the new IDs assigned to the S. uberis isolates, marked as IZSar, as well as country, year of isolation, MLST profile, ST, and clonal complex (CC) are displayed in https://pubmlst.org/bigsdb?db=pubmlst_suberis_isolates (accessed on 29 April 2021). The most frequently detected STs were ST808, ST1177, ST1192 and ST1247 with 4 isolates each, followed by ST562, ST1174, ST1176, ST1182, ST1188 and ST1242 with 3 isolates (Table S1). The highest number of new alleles was detected for tdk with 10 new alleles, followed by yqil and arcC with 7 and 6 alleles, respectively (Table S1). CC143, with 8 different STs grouping 17 isolates, was the most abundant clonal complex among the ovine S. uberis isolates, followed by CC86 (four STs and 5 isolates) and CC5 (one ST and 1 isolate) (Table S2). The remaining STs did not belong to any CC. The goeBURST full MST analysis identified a total of 20 possible ST founders and 4 within CC143 (ST294, ST808, ST1192 and ST1195). The founder ST294 contained 4 other STs: 2 with a single locus variant (SLV) and 2 with double locus variants (DLVs). The founder ST808 had 6 STs: 2 with a SLV, 3 with DLVs and 1 with triple locus variants (TLVs) (Figure 1). Figure 2 illustrates the relationship between the 86 STs of this study and those (n = 51) extracted from the S. uberis MLST database and related to other Italian regions. These S. uberis were mainly isolated from bovine mastitis. Unlike what has been found in ovine mastitis, the CC5 was predominant in S. uberis isolates from bovine mastitis ( Figure 2). The goeBURST analysis showed that the majority of Italian cattle strains are grouped in the ST 1199 which is a SLV of the predicted founder ST808 ( Figure 2).
All 124 S. uberis isolates were also typed by PFGE. Figure 3 shows the percentage similarity among patterns determined by cluster analysis in a dendrogram. A high level of polymorphism was observed among the SmaI profiles; in fact, a total of 121 pulsotypes (PTs) were unique within this dataset. By using a cut off similarity value of 75% in the UPGMA dendrogram, 68 S. uberis grouped into 29 PFGE clusters (from 1 to 29) whereas 56 isolates had a lower level of similarity in comparison to other isolates and were not assigned into clusters.   All 124 S. uberis isolates were also typed by PFGE. Figure 3 shows the percentage similarity among patterns determined by cluster analysis in a dendrogram. A high level of polymorphism was observed among the SmaI profiles; in fact, a total of 121 pulsotypes (PTs) were unique within this dataset. By using a cut off similarity value of 75% in the UPGMA dendrogram, 68 S. uberis grouped into 29 PFGE clusters (from 1 to 29) whereas 56 isolates had a lower level of similarity in comparison to other isolates and were not assigned into clusters.   (6), erythromycin (7), trimethoprim-sulphamethoxazole (8), amoxicillin-clavulanic acid (9), cephalothin (10), oxacillin (11), novobiocin (12), ceftiofur (13) and pirlimycin (14).

Antimicrobial Susceptibility and Resistance Genes
Almost all S. uberis isolates (120/124, 96.8%) were resistant to at least one antimicrobial. Resistance was exhibited to all the antimicrobials tested with the exception of AMC, OX and EFT. Of the antimicrobials tested, the highest resistance rate was recorded against S (116/124, 93.5%), K (99/124, 79.8%) and CN (80/124, 64.5%), followed by NV (31/124, 25%), and TE (24/124, 19.3%). Seventy-four (59.7%) isolates were simultaneously resistant to all aminoglycosides. Seventeen isolates (13.7%) were MDR: ten were resistant to three different classes, six to four different classes and one to five different classes (P, TE, S, SXT and NV) (Figure 4). No specific correlation was found between ST/CC/ PFGE and resistance to aminoglycosides or other antimicrobials. On the basis of inhibition zone diameters, established by guidelines of CLSI, many isolates presented intermediate values with respect to E and TE: 17 (13.7%) and 39 (31.5%), respectively. All of them were analyzed by PCR for the corresponding resistance genes. Of the three E-resistant isolates, two harboured ermB gene while one carried ermC gene. The ermA, ermTR and mef A genes were not detected among the analyzed isolates. All E-intermediate isolates were PCR negative. Among the twenty-four S. uberis isolates resistant to TE, ten were positive for tetM, six for tetO, two for tetK and one for tetO. One isolate carried simultaneously tetK and tetM genes, one for both tetO and tetS, and another one for both tetk and tetO genes. Seven isolates were non-typeable using the primers for these tet genes. Among the 39 TE-intermediate isolates, 3 were PCR positive for tetK and 2 for tetM. One isolate possessed both tetK and tetM genes. Of the six P-resistant isolates, five harbored the blaZ gene, encoding β-lactamase, which is responsible for enzymatic hydrolysis of the ring of β-lactam antibiotics. All aminoglycosides-resistant isolates were PCR negative for aad-6 and aphA-3 genes. Among the twenty-four S. uberis isolates resistant to TE, ten were positive for tetM, six for tetO, two for tetK and one for tetO. One isolate carried simultaneously tetK and tetM genes, one for both tetO and tetS, and another one for both tetk and tetO genes. Seven isolates were non-typeable using the primers for these tet genes. Among the 39 TE-intermediate isolates, 3 were PCR positive for tetK and 2 for tetM. One isolate possessed both tetK and tetM genes. Of the six P-resistant isolates, five harbored the blaZ gene, encoding β-lactamase, which is responsible for enzymatic hydrolysis of the ring of β-lactam antibiotics. All aminoglycosides-resistant isolates were PCR negative for aad-6 and aphA-3′ genes.

Discussion
The present study contributes to understand the population structure of S. uberis involved in ovine mastitis in Sardinia, Italy. Several studies are available on S. uberis from dairy cows [8,9,[12][13][14][15][16], while only limited reports are available on the genotyping and antimicrobial resistance of S. uberis from small ruminants [6,17]. In this work, which represents the first study dealing with molecular typing of ovine S. uberis in Italy, the genetic relatedness of isolates was established by using MLST and PFGE.

Discussion
The present study contributes to understand the population structure of S. uberis involved in ovine mastitis in Sardinia, Italy. Several studies are available on S. uberis from dairy cows [8,9,[12][13][14][15][16], while only limited reports are available on the genotyping and antimicrobial resistance of S. uberis from small ruminants [6,17]. In this work, which represents the first study dealing with molecular typing of ovine S. uberis in Italy, the genetic relatedness of isolates was established by using MLST and PFGE. The data generated by the MLST analysis was used to implement the Streptococcus uberis database (https://pubmlst.org/organisms/streptococcus-uberis) (accessed on 29 April 2021) and to compare all isolates collected in Italy, even if of different animal origin. The MLST approach is well suited for such comparative studies; in addition to distinguishing continuous and new infections, it also allows to investigate the genetic lineage of the isolates. Our results showed that the 124 isolates were grouped into 86 STs, of which 73 were new STs. Among these, only ST294 (Table S1) is common to the 14 ovine isolates from the Lazio region of Italy analyzed in a previous study [6]. ST294 is considered a possible ST founder by the goeBURST full MST algorithm (Figure 1). ST294, ST384, ST562, ST808, ST1112, ST1189, ST1192 and ST1195 were grouped into CC143, which represents the most widespread lineage in Sardinia, although including only 13.7% of all isolates considered in this study (Figure 2). CC86 and CC5 were detected at very low rates of 4% and 0.8%, respectively. Although there are no other data concerning S. uberis of small ruminants, lineages CC143 and CC86 were highly associated with bovine IMI in Portugal [18], India [19], Italy [6] and Australia [16], whereas CC5 is the most prevalent clonal complex among bovine mastitis isolates collected in the UK [20] and Swiss Midlands [7].
According to Tomita et al. [12], CC5 and CC143 are correlated with clinical and subclinical bovine mastitis and may represent lineages of virulent isolates, while isolates belonging to CC86 may be associated with low somatic cell count cows. This hypothesis is based on the detection of specific virulence factors in 95% of the isolates grouped in CC5 and CC143 and only in 25% in those belonging to CC86. Work is in progress for detecting virulence marker genes in our isolates.
In agreement with other authors [14,18,21] on molecular typing by PFGE, macrorestriction analysis revealed a high degree of genetic diversity even among S. uberis from small ruminant origin. Using a similarity value of ≥75%, the isolates were distributed in 29 small clusters. Two isolates, belonging to cluster 3, exhibited the same PFGE profile, and were collected from different animals of the same flock. Two other isolates, belonging to cluster 13, presented the same PFGE pattern. In this latter case, isolates were recovered from distinct sheep farms distant; about 23 km, as the crow flies. The last two isolates with the same PFGE pattern, belonging to cluster 19, were collected from two separate samplings, made 10 months apart, on the same sheep farm. The occurrence of isolates with identical PFGE profiles could be due to their transmission from sheep to sheep through the milking process (1st and 3rd cases) or to the commercial movement of infected animals between farms (2nd case). According to the criteria described by Tenover et al. [10], isolates showing indistinguishable macrorestriction patterns should be considered the same strain. However, in the comparison between PFGE and MLST, we found that, in the 1st case, the isolates presented different STs (ST1176 and ST1179), even if it concerned only the yqiL allele; in the 2nd case, STs also were different (ST1182 and ST1179) with 5 allelic variants of ddl, gki, tdk, tpi and yqiL genes. In the 3rd case, STs were identical. In light of these results, it would be correct to consider the same strain only the isolates belonging to the 3rd case. Whole genome sequencing (WGS) characterization will be required to shed further light on this hypothesis.
In the present study, we evaluated the AMR of the 124 S. uberis isolates to 14 antimicrobial agents. A recognized limitation of AMR tests is the lack of small ruminants-specific interpretive criteria to categorize isolates as susceptible, resistant, or intermediate. In this study, the cut-off values used are based on Streptococcus/Staphylococcus from human and bovine origin [22]. Using these breakpoints, a high percentage (59.7%) of isolates were found to be simultaneously resistant to the three aminoglycosides tested, with peaks of 93.5%, 79.8% and 64.5% for S, K and CN, respectively. High percentages of resistance to these antimicrobials using the same method were observed by other researchers [17,23]. An acquired high level of resistance to streptomycin can be explained by the extensive use of this antimicrobial in combination with penicillin in the treatment of ovine mastitis. However, we did not find the presence of genes associated with resistance to aminoglycosides (i.e., aad-6 and aphA-3') among the isolates analyzed in our research. Taber et al. [24] Pathogens 2021, 10, 1489 7 of 11 hypothesized a "natural" resistance of streptococci to aminoglycosides resulting in a low level of intrinsic resistance to these drugs. Further studies are needed to determine the presence of other resistance determinants and the mechanism of gene transfer. In this work, 19.3% of S. uberis isolates were phenotypically resistant to tetracycline; this resistance mainly depends on ribosomal protection proteins, mediated by tetM gene (41.6%) or tetO (25%) determinants. This finding is consistent with some previous surveys on Streptococcus spp. collected from dairy cattle in Poland [25], in the United States [26] and China [27]. Therefore, it can be stated that the tetM gene is involved in tetracycline resistance not only in human streptococci [28] but also in those isolated from bovine and ovine mastitis. For macrolides, we investigated the presence of ermA, ermB, ermC, ermTR and mef A genes that codify methylase which reduce the binding of antibiotics to the target site. Only 3 S. uberis were E-resistant isolates: two harbored ermB while one carried ermC. These findings were consistent with previous studies, which found a predominance of ermB among bovine isolates [23,25,27]. A noteworthy result in our study was the high resistance (25%) of S. uberis isolates to novobiocin, which is not reported in the literature.

Bacterial Isolates and DNA Extraction
In this study, we analyzed 124 S. uberis isolated between January 2011-December 2017 from sheep milk collected in different provinces of Sardinia (Italy). The isolates belonged to a collection of Streptococcus species used for the preparation of inactivated autogenous vaccines, due to the presence of clinical or subclinical mastitis cases in the flock. Information about the type of mastitis was not provided by the farm veterinarian. With the exception of four isolates, all the others were from different dairy farms. Bacterial isolation and identification at the species level by PCR-RFLP and MALDI-TOF MS analysis, have been described in a previous study [5].
Bacterial DNA was extracted from all isolates and from S. uberis reference strain ATCC 700407 according to Onni et al. [29]. Briefly, isolates were grown in 5 mL Brain Heart Infusion broth (BHI, Oxoid Ltd., Basingstoke, UK) at 37 • C for 18 h with shaking. A 100 µL aliquot was pelleted by centrifugation at 7000 rpm for 2 min. The pellet was resuspended in 49.5 µL deionized sterile water supplemented with 0.5 µL lysostaphin (1 mg/mL) dissolved in 20 mM sodium acetate, and incubated at 37 • C for 10 min. After addition of 1 µL proteinase K (5 mg/mL) (Roche Diagnostics, Monza, Italy) and 150 µL of 0.1 M Tris-HCl (pH 7.5), samples were further incubated for 10 min. Finally, samples were boiled for 5 min and then stored at −20 • C.

MLST
Amplification of the seven housekeeping genes [carbamate kinase (arcC), D-ala-Dala ligase (ddl), glucose kinase (gki), transketolase (recP), thymidine kinase (tdk), triose phosphate isomerase (tpi) and acetyl CoA acetyl-transferase (yqiL)] was performed using the primer sequences as described by Coffey et al. [11]. Gene fragments were amplified by conventional PCR (GeneAmp PCR System 9700, Applied Biosystems, Foster City, CA, USA), then purified and sent to BMRGenomics (https://www.bmr-genomics.it/) (accessed on 28 September 2020) for DNA sequencing. The allelic profile of each isolate was identified and, where possible, strains were assigned to sequence type (ST) using the S. uberis MLST database (https://pubmlst.org/organisms/ streptococcus-uberis) (accessed on 28 September 2021). Alleles not identified as well as unknown allelic profile were submitted to the database and new STs were generated. Phyloviz version 2.0 (www. phyloviz.net) (accessed on 29 April 2021) and goeBURST algorithm [30] were used to visualize the data. The same algorithm was also used to compare our strains with other S. uberis of ruminant origin isolated in Italy, and present in the MLST database (accessed on 29 April 2021).

PFGE
All 124 S. uberis isolates were genotyped by PFGE. DNA was extracted with Bio-Rad CHEF genomic DNA plug kit (BioRad, Segrate, Italy) according to the manufacturer's instructions. Each plug was digested with 20 U of SmaI (Roche) for 3 h at 25 • C and washed with 5 mL of TE buffer for 10 min at room temperature before being loaded on a 1% certified Megabase agarose (BioRad) gel. The lambda PFG ladder (New England Biolabs, Ipswich, MA, USA) was used as a molecular size standard. Electrophoresis was carried out in a contour-clamped homogeneous electric field (CHEF)-Mapper system (BioRad) at 14 • C in 0.5× TBE buffer. DNA fragments were separated after 18 h migration with 6 V/cm, 120 • at pulse times of 10-45 sec. Gels were stained with ethidium bromide and photographed with Alliance LD2 gel documentation system (UVITEC, Cambridge, UK) in TIFF format. The S. uberis digitalized PFGE patterns were analyzed with the Gel Compar II software (Applied Maths, Sint-Martens-Latem, Belgium). Dendrograms were generated by the unweighted pair group method with arithmetic averages (UPMGA) using the Dice correlation coefficient with a position tolerance of 1.5%. Isolates with ≥90% similar profiles were considered to represent the same clone [10].

Conclusions
To sum up, genotyping by MLST and PFGE analysis indicates the epidemiological complexity of S. uberis isolated from ovine milk. The lack of a predominant strain and the demonstration that a wide variety of isolates are capable of infecting the mammary gland suggests that most of the S. uberis IMI cases are due to contamination of the mammary gland from the environment. Nevertheless, further research is needed to shed light on the possible transmission of some strains within or among flocks. The high resistance of the isolates toward aminoglycosides, novobiocin and tetracycline confirms that greater care should be taken when these antimicrobials are used in the veterinary practice. Continuous monitoring should be carried out to maintain an updated understanding of the level of antimicrobial resistance among dairy sheep in Sardinia.