Bacteriocin Producing Streptococcus agalactiae Strains Isolated from Bovine Mastitis in Brazil

Antibiotic resistance is one of the biggest health challenges of our time. We are now facing a post-antibiotic era in which microbial infections, currently treatable, could become fatal. In this scenario, antimicrobial peptides such as bacteriocins represent an alternative solution to traditional antibiotics because they are produced by many organisms and can inhibit bacteria, fungi, and/or viruses. Herein, we assessed the antimicrobial activity and biotechnological potential of 54 Streptococcus agalactiae strains isolated from bovine mastitis. Deferred plate antagonism assays revealed an inhibition spectrum focused on species of the genus Streptococcus—namely, S. pyogenes, S. agalactiae, S. porcinus, and S. uberis. Three genomes were successfully sequenced, allowing for their taxonomic confirmation via a multilocus sequence analysis (MLSA). Virulence potential and antibiotic resistance assessments showed that strain LGMAI_St_08 is slightly more pathogenic than the others. Moreover, the mreA gene was identified in the three strains. This gene is associated with resistance against erythromycin, azithromycin, and spiramycin. Assessments for secondary metabolites and antimicrobial peptides detected the bacteriocin zoocin A. Finally, comparative genomics evidenced high similarity among the genomes, with more significant similarity between the LGMAI_St_11 and LGMAI_St_14 strains. Thus, the current study shows promising antimicrobial and biotechnological potential for the Streptococcus agalactiae strains.


Introduction
In the past few decades, antibiotics have played an essential role in treating infectious diseases and in antimicrobial prophylaxis for surgical procedures [1]. However, that has not always been the case in history, and microorganisms' infections used to be partially responsible for a lower global average life expectancy. In 1928, Alexander Fleming discovered the first antibiotic, a compound later named as penicillin, which revolutionized treatments of infectious diseases. After its market introduction in the 1940s, mortality rates due to syphilis were drastically reduced. Penicillin's success created an appeal for the discovery of other similar substances. The collective effort to find new antibiotics resulted in the discovery of several other antimicrobial drugs during the antibiotic golden age, such as sulphonamides and streptomycin for treating scarlet fever and tuberculosis, respectively [2].
Unfortunately, most of the original antibiotics and some of their later modifications have been ineffective in treating infectious diseases in recent years. Antibiotic resistance is a natural process in which microorganisms find ways to circumvent the action of drugs designed to kill them [1]. The accelerated selection of resistant strains occurs due to the widely adopted use of antibiotics in humans, animals, and crops [3]. The discovery pace of new classes of antibiotics has slowed ever since the antibiotic golden age to top it off [2]. Therefore, antibiotic resistance is one of the most significant global health challenges of our time. Statistical analyses predict that 10 million people are expected to die every year due to untreated microbial infections by 2050 if antimicrobial resistance continues to rise [4].
Streptococcus agalactiae (also known as group B Streptococcus) is an opportunistic pathogen that colonizes the human urogenital and digestive tracts and is associated with life-threatening complications for invasive infections in pregnant women [5,6]. This bacterial species is also responsible for mastitis infections in cattle and economic loss in dairy farms [7]. Penicillin is the antibiotic used to treat Streptococcus agalactiae infections. However, in cases of penicillin allergy, clindamycin and erythromycin are used as second-line antibiotics. Reports on clinical isolates with resistance to erythromycin and clindamycin are increasingly frequent [5,6], which highlights the need for the monitoring of antibiotic resistance and research on alternative antimicrobial substances for treating these infections.
There are many alternative methods available and under consideration for solving this problem. One approach to solving this crisis involves using antimicrobial peptides [1]. Antimicrobial peptides are small molecules of proteinaceous nature produced by all organisms that help protect them from infections by microorganisms they may encounter [8]. The last antimicrobial peptide approved for use was daptomycin, isolated from soil bacteria Streptomyces roseosporus in 1986. Nonetheless, many antimicrobial peptides are currently under different stages of clinical trials and are expected to be applied in healthcare in the future. Among the 36 antimicrobial peptides under clinical trial in 2019, 5 were of bacterial origin. It is worth mentioning that most of the antimicrobial peptides under clinical trials are of human origin due to their expected low toxicity. However, bacterial antimicrobial peptides are not to be ignored, especially bacteriocins, since they provide a vast diversity of substances with targeted action [9]. Bacteriocins are a class of ribosomally synthesized antimicrobial peptides produced by bacteria and are biologically active against other Gram-positive and Gram-negative bacteria [10].
Lactic acid bacteria represent a group of Gram-positive bacteria, which are important due to their production of lactic acid, organic acids, secondary metabolites, and their potential application as probiotics [11]. This group comprises several genera-namely, Aerococcus, Carnobacterium, Enterococcus, Lactobacillus, Lactococcus, Leuconostoc, Oenococcus, Pediococcus, Streptococcus, Tetragenococcus, and Weisella [12]. Many bacteriocins have been isolated from this group and are used in fermentation processes in the food industry [13].
Considering the antibiotic resistance crisis as a One Health problem, as it involves humans, animals, and the environment, it is of particular ongoing interest to apply an omics bioinformatics approach for the discovery of new antimicrobial drugs and the monitoring of antimicrobial resistance spread [10]. Therefore, we aimed to evaluate the biotechnological potential of Streptococcus agalactiae strains on their production of antimicrobial substances, focusing on the inhibition of pathogenic bacteria. Since the Streptococcus genus has a diversity of bacteriocins described, but there is still a lack of knowledge for the Streptococcus agalactiae species, we also aimed to sequence a few selected Streptococcus agalactiae genomes and apply a genomics approach to assess the presence of gene clusters for antimicrobial substances and secondary metabolites, as well as to monitor antibiotic resistance in these strains.

Strains, Media, and Growth Conditions
In this study, 54 Streptococcus agalactiae strains isolated from bovine mastitis were tested for their ability to produce antimicrobial substances. These strains were isolated from dairy farms in the states of São Paulo and Minas Gerais in Brazil (Table 1) and phenotypically identified via biochemical tests in a previous study [26]. Seventy-five strains were used as indicator bacteria in the deferred growth inhibition assays, accounting for Listeria innocua (n = 1), Micrococcus sp. (n = 1), Lactococcus lactis (n = 1), Cellulomonas fimi (n = 2), Klebsiella pneumoniae (n = 2), Staphylococcus aureus (n = 5), Streptococcus uberis (n = 12), Streptococcus porcinus (n = 12), Streptococcus agalactiae (n = 17), and Streptococcus pyogenes (n = 20). Klebsiella pneumoniae KPC strain shows resistance to carbapenems, and Staphylococcus aureus ATCC 33591 shows resistance to methicillin. Micrococcus sp. and Cellulomonas fimi strains were used due to their known susceptibility to diverse antimicrobials. Information on antibiotic resistance to the other indicator bacteria was not determined.
Streptococcus spp., Listeria innocua, and Lactococcus lactis were grown on Brain and Heart Infusion media. Staphylococcus aureus and Klebsiella pneumoniae were grown in TSB media. Micrococcus sp. and Cellulomonas fimi were grown on LB media. The antagonism tests were conducted on BHI media, and all bacteria were incubated at 37 • C 18 h before experiments.

Deferred Growth Inhibition Assays
Deferred growth inhibition assays were performed to test the 54 Streptococcus agalactiae strains for their ability to produce antimicrobial substances. The inhibition assays were performed following the spot-on-lawn method, as previously described in [27]. For this, Streptococcus agalactiae strains were inoculated in 4 mL of BHI broth and incubated at 37 • C overnight. From those cultures, 10 µL was collected and inoculated on sterile BHI plates in the form of circular spots. Then, the plates were incubated at 37 • C overnight. After that, the plates were exposed to cotton balls soaked in 1 mL of chloroform for 30 min to kill the bacteria. After treatment with chloroform vapor, indicator strains were inoculated in 4 mL of BHI semisolid medium and poured on top of the plates, which were incubated at 37 • C overnight. The absence of growth of the indicator strains on top of the Streptococcus agalactiae spots was considered a positive result of inhibition. In the presence of a clear zone of inhibition surrounding the Streptococcus agalactiae spots, the size of the inhibition zones was measured in millimeters. The sterile BHI medium was used as the negative control.

Strains
LGMAI_St_08, LGMAI_St_11, and LGMAI_St_14 were selected for wholegenome sequencing due to their action spectra.
Genomic DNA was extracted using the phenol-chloroform method [28], with some modifications, and quantified by Qubit (Invitrogen, Waltham, MA, USA) fluorimetry. An amount of 5 µg/µL of gDNA was considered for the construction of paired-end sequencing libraries (2 × 150 bp) of 450 bp insert size, following the manufacturer's protocol for the NEBNext ® Fast DNA Fragmentation and Library Preparation Kit (New England Biolabs Inc., Ipswich, MA, USA). Final library-quality analysis was performed via 2100 bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) with read length gDNA size control using agarose gel electrophoresis. All strains were sequenced on the Illumina Hi-Seq 2500 platform, as recommended by the manufacturer.

Genome Assembly
The first step of the genome assembly process was to check the reads' quality through FastQC [29] and remove the adapters with the Adapter Removal [30] software. The estimated best k-mers were selected by KmerStream [31], followed by the assembly using Edena [32] and SPAdes [33]. Then, the results were combined and the CD-HIT [34], package PSI-CD-HIT, was used to remove the redundant contigs, producing a final contigs file. Assembled genomes were then annotated using Prokka software [35].

Taxonomic Identification
We performed a taxonomic identification to properly classify the genomes before their submission to the GenBank database. This classification was carried out using the multilocus sequence analysis (MLSA) method. In addition to the 3 Streptococcus genomes sequenced in our study, another 36 Streptococcus species genomes were used for this analysis. Lactococcus lactis subsp. lactis and Lactococcus lactis subsp. cremoris were used as outgroups to root the phylogenetic tree ( Table 2). The 16S rRNA, shikimate dehydrogenase (aroE), glucokinase (gki), phenylalanine-tRNA ligase alpha subunit (pheS), and recA genes were extracted from these genomes and used to construct the concatenated sequences for the phylogenetic analysis. The sequence alignment via Clustal W and phylogenetic tree inference were made in MEGA X software [36]. Phylogenetic tree inference was based on a maximum likelihood analysis with 1000 bootstraps, using the general time reversible model, with a gamma distribution and considering some nucleotide sites to be invariant. Streptococcus agalactiae possesses a capsular polysaccharide, which is a virulence factor that grants the bacteria with immune system evasion and host cell invasion. There are, in total, ten different capsular types, and they are used in the species serotyping. The main serotypes identified in bovine mastitis infections are Ia, II, III, and V, whereas those associated with infections in humans are mainly Ia, Ib, II, III, and V [7]. Thus, we also performed a whole-genome sequence data-based identification of GBS capsular serotypes, as previously described by [37], to determine the isolates serotypes. For this analysis, LGMAI's Streptococcus agalactiae subject genomes were blasted against query references for the ten different GBS capsular serotypes, as described in [37].

Assessments for Virulence Factors and Antibiotic Resistance
Bacteriocins are ribosomally synthesized antimicrobial peptides produced by bacteria.
For instance, such compounds may grant bacteriocin-producing bacteria advantages in replication and settlement to different sites [38,39]. Thus, these substances have similar characteristics to bacterial virulence factors. Therefore, the ABRicate software [40] was used to analyze the presence of virulence factors in the Streptococcus agalactiae genomes that could be linked to the inhibition phenotypes observed. Furthermore, taking into consideration the rising levels of antimicrobial resistance identified in this species, the software was also used to look for the presence of antibiotic resistance genes in the strains' genomes. ARG-ANNOT (Antibiotic Resistance Gene-ANNOTation) [41], CARD (Comprehensive Antibiotic Resistance Database) [42], NCBI AMRFinderPlus tool [43], and Resfinder [44] were used for the assessment of antibiotic resistance genes. The VFDB (Virulence Factors Database) [45] tool was used to identify virulence factors, and the GIPSy (Genomic Island Prediction Software) [46] software was used to predict pathogenicity and resistance islands. Streptococcus thermophilus JIM 8232 (assembly accession number GCA_000253395.1) was used as the non-pathogenic reference genome required for the GIPSy analysis.

Genome Mining for Antimicrobial Substances
The antibiotics and secondary metabolites analysis shell (antiSMASH) platform [47] was used to identify secondary metabolite biosynthesis gene clusters. The BAGEL4 [48] webserver was used to identify putative biosynthesis gene clusters of ribosomally synthesized and post-translationally modified peptides (RiPPs) and unmodified bacteriocins. Similarly, the BACTIBASE [49] web platform was used to identify bacteriocin genes and provide information about such peptides' physicochemical and structural properties. For all these analyses, the Streptococcus agalactiae genome fasta files were used as input for the matching search with the databases. Undetected accessory genes were manually searched in the genomes' annotation files.
Multiple sequence alignment for the bacteriocin zoocin A was performed using twelve zoocin A sequences obtained from the Uniprot database (Table 3).

Genome Comparison and Functional Annotation
The eggNOG-mapper tool [50] was used to analyze the Streptococcus agalactiae genomes and identify different clusters of orthologous groups (COGs). Genome comparisons were performed with the OrthoVenn2 web platform [51]. This platform annotates orthologous gene clusters and provides a comparison of multiple protein datasets via Venn diagrams. The BRIG software [52] was used to analyze the nucleotide identity among the LGMAI Streptococcus agalactiae genomes and 15 other S. agalactiae genomes obtained from NCBI (Table 4). For the nucleotide identity comparison, the LGMAI_St_08 genome was used as the reference genome due to its larger genome size. Furthermore, predicted genomic islands were plotted onto the BRIG similarity plots to observe their conservation among the different Streptococcus agalactiae strains analyzed.

Antimicrobial Activity Spectrum
The deferred growth inhibition assays demonstrated that 14 out of the 54 Streptococcus agalactiae strains inhibited at least 1 of the indicator strains. Streptococcus agalactiae I14 and Streptococcus pyogenes were the only indicator strains inhibited (Table 5).
All four Streptococcus agalactiae producing strains selected for the additional tests inhibited at least 5 Streptococcus porcinus, 6 Streptococcus pyogenes, 15 Streptococcus agalactiae, 4 Streptococcus uberis, Cellulomonas fimi, and the Micrococcus sp. Neither Klebsiella pneumoniae nor Staphylococcus aureus was inhibited by the Streptococcus agalactiae strains ( Table 6). The four Streptococcus agalactiae strains did not show differences in terms of inhibition zones size, with all of them presenting mean values of 2 mm. Table 6. Additional deferred growth inhibition assays results. Additional tests were performed with strains LGMAI_St_08 (3772), LGMAI_St_11 (3800), and LGMAI_St_14 (3799). (+) indicates clear inhibition zones of at least 2 mm in size; (−) indicates absence of inhibition.

Indicator Strains
Streptococcus agalactiae Strains

Genome Sequencing
The genomes belonging to the LGMAI_St_08, LGMAI_St_11, and LGMAI_St_14 strains were sequenced. The LGMAI_St_11 strains had the smallest genome size with 2,292,224 bp, while the LGMAI_St_08 strains had the largest genome size with 2,397,674 bp. Furthermore, GC content ranged from 35.6% to 35.7%. All three genomes presented a good number of predicted genes based on reference genomes. Complete detailed statistics on the resulting genomes are shown in Table 7. Table 7. Genome statistics obtained from assemblers and side analyses.

Parameter Strains
LGMAI_St_08 LGMAI_St_11 LGMAI_St_14 Quality assessment for assembly completeness performed with the Busco software showed similar results for all three genomes, identifying all 124 universal single-copy orthologs for bacteria, among which 122 are complete, and 2 are fragmented.

Taxonomic Identification
We inferred a phylogenetic tree to better understand the Streptococcus species' relationships and confirm the Streptococcus agalactiae strains' biochemical identification. The phylogenetic inference was performed via a multilocus sequence analysis using the maximum likelihood estimation method and the general time-reversible substitution model. This multilocus sequence analysis confirmed that the bovine strains belong to the Streptococcus agalactiae species, with strong bootstrap statistics supporting the branch. Moreover, the separation of the Streptococcus species in putative monophyletic groups, such as anginosus, bovis, mitis, mutans, pyogenes, salivarius, and sanguinis clusters was also detected. Streptococcus parasanguinis ATCC 15912, however, did not fall into the salivarius group ( Figure 1). All three LGMAI's Streptococcus agalactiae subject genomes matched the serotype III query sequence with 99% identity, superseding the analysis's requirement. Only one mutation was observed at position 76 ( Figure 2). Furthermore, none of the genomes additionally matched the serotype II query sequence, as sometimes observed, allowing for a clear identification as serotype III. All three LGMAI's Streptococcus agalactiae subject genomes matched the serotype III query sequence with 99% identity, superseding the analysis's requirement. Only one mutation was observed at position 76 ( Figure 2). Furthermore, none of the genomes additionally matched the serotype II query sequence, as sometimes observed, allowing for a clear identification as serotype III.

Antibiotic Resistance Genes and Virulence Factors
The assessments of antibiotic resistance genes performed with the CARD and Resfinder databases identified the genes mprF and mreA, with 99.49% and 99.79% identity, respectively, in all three genomes analyzed. The mprF gene encodes a phosphatidylglycerol lysyltransferase enzyme. The mreA gene encodes the protein responsible for riboflavin biosynthesis. No other antibiotic resistance genes were found either in the NCBI AMRFind-erPlus tool or in the ARG-ANNOT databases.
The Virulence Factors Database identified 24 genes in the LGMAI_St_11 and LGMAI_St_14 genomes and 25 genes in the LGMAI_St_08 genome, representing important virulence factors such as the cps locus, the operon cyl, the hyaluronate lyase enzyme, and the CAMP factor. The cfa/cfb gene was detected exclusively in the LGMAI_St_08 genome (Table 8).

Strains
LGMAI_St_08 LGMAI_St_11 LGMAI_St_14 We used the GIPSy software to predict resistance and pathogenicity islands in the LGMAI's Streptococcus agalactiae genomes. The analysis identified a total of a hundred and twenty putative genomic islands. After filtering the result for scores strong and normal, we registered eight genomic islands for the LGMAI_St_08 strain, seventeen for the LGMAI_St_11 strain, and fourteen for the LGMAI_St_14 strain (Table 9). These predicted genomic islands were then plotted onto a circular genome map using the BRIG software to evaluate their conservation amongst different Streptococcus agalactiae strains.
The BAGEL4 web server identified one area of interest in each genome. This area represents the gene cluster that regulates the production of the bacteriocin zoocin A. The zoocin A core peptide gene showed an identity of 58.76% with an e-value of 3 × 10 −117 . Similarly, the BACTIBASE database detected the antimicrobial peptide zoocin A in all three genomes analyzed. The zoocin A amino acid sequence showed an identity of 60% with an e-value of 3.44045 × 10 −101 .
Protein sequence analysis on the InterPro database recognized a leader peptide sequence and two domains in the LGMAI's putative zoocin A: the peptidase M23 catalytic domain and the lytic exoenzyme target recognition domain. Multiple sequence alignment between the LGMAI's putative zoocin A and twelve other zoocin A sequences obtained from the Uniprot database showed amino acid conservation of the two domains (Figures 3  and 4). Multiple sequence alignment revealed that the LGMAI's putative zoocin A amino acid sequences were completely identical.  The zoocin A immunity factor (Zif) is a FemABX-like aminoacyl-transferase. The search for FemABX-like proteins in the genomes' annotation files revealed the presence of FemB aminoacyl-transferases. LGMAI's putative Zif sequences analysis on the InterPro database identified protein and superfamily regions, t-RNA binding arm, and coil (Table 10). These three sequences are identical and composed of 410 amino acid residues with 50% identity and 65% similarity to the Zif protein (UniProt entry P74894) from Streptococcus equi subsp. zooepidemicus.

Genome Comparison and Functional Annotation
The eggNOG-mapper tool registered 2169 orthologous groups for the LGMAI_St_08 strain, 2130 for the LGMAI_St_11 strain, and 2127 for the LGMAI_St_14 strain. Metabolism was the most abundant category among these groups (Table 11). Table 11. Distribution of the orthologous groups identified by the eggNOG-mapper tool among the 4 major COG categories.

Strains
LGMAI_St_08 LGMAI_St_11 LGMAI_St_14 Overall, all three genomes showed similar results for the annotation of clusters of orthologous groups. Function unknown (S), replication, recombination, and repair (L), and transcription (K) were the most plentiful functional categories observed. Furthermore, we observed no clusters designated to the general function prediction only (R), chromatin structure and dynamics (B), cytoskeleton (Z), and nuclear structure (Y) functional categories ( Figure 5). The genome comparison analysis showed that the three strains are very similar amongst themselves. A total of 2275 clusters are shared by all three strains, accounting for 6853 proteins (LGMAI_St_08: 2291 (33.43%), LGMAI_St_11: 2281 (33.28%), and LGMAI_St_14: 2281 (33.28%)). Pairwise comparisons of orthologous groups indicate that LGMAI_St_11 and LGMAI_St_14 are closer than the LGMAI_St_08 strain. They share 52 groups, accounting for 106 proteins (LGMAI_St_11: 52 (49.06%) and LGMAI_St_14: 54 (50.94%)).
LGMAI_St_08 registered 16 exclusive orthologous groups composed of 43 proteins ( Figure 6). Molecular function clusters seem to be conserved, strictly represented in the overlap of the three strains. The overlap between strains LGMAI_St_08 and LGMAI_St_14 is the smallest, accounting for proteins related to extracellular region. Moreover, proteins responsible for cellular components, especially the categories for cell parts and cell membrane were also conserved. Most of the 16 orthologous groups found exclusively in LGMAI_St_08 belong to the biological processes category, with only one group for molecular function related to ion binding.
The circular genome map generated by the BRIG software shows that the three LGMAI strains share high similarities amongst themselves. In addition, they possess genomic regions that are not present in some of the other 15 genomes used for comparison in the analysis. Interestingly, some of the predicted pathogenicity and resistance islands overlap with parts of those regions (Figure 7). . Circular genome map of the three LGMAI strains and 15 other well-represented Streptococcus agalactiae genomes obtained from NCBI. From the inside out, the genomes used in this analysis are in the legend order. The blue, red, and green inner circles represent the LGMAI strains. Outside each of the LGMAI's genomes, predicted pathogenicity and resistance islands are plotted in the same colors of their corresponding strains. Letters indicate the genomic island types: A-resistance islands, B-pathogenicity islands, and C-miscellaneous islands. Letters' colors match their corresponding genomes. This circular genome map shows nucleotide identity between the reference genome and the other genomes used in the analysis. Stronger colors register greater identities.

Discussion
The deferred growth inhibition assays demonstrated that the LGMAI Streptococcus agalactiae strains produce bacteriocin-like substances. Additionally, the antagonism tests indicated an activity spectrum focused mainly on Streptococcal species, specifically Streptococcus pyogenes, Streptococcus agalactiae, Streptococcus porcinus, and Streptococcus uberis.
As frequently observed in studies of bacteriocins, these bacterial antimicrobial peptides are usually more prone to inhibiting closely related species [53]. In agreement with such expectations, we observed that the Streptococcus agalactiae bacteriocin-like substance showed an inhibition spectrum targeting mainly streptococcal species. From a clinical perspective, these results are interesting and promising because Streptococcus is a genus that comprises many relevant pathogenic bacteria [54].
Streptococcus porcinus accommodates similar beta-hemolytic streptococci that belong to Lancefield groups E, P, U, and V [55]. In humans, Streptococcus porcinus is responsible for urogenital tract infections in women. In pigs, it is associated with abortion events, lymph node and throat abscesses, pneumonia, and endocarditis. It is also related to cases of endometritis in swine after artificial insemination in farms. Antibiotic resistance is prevalent in this species, with some strains resistant to erythromycin, streptomycin, tetracycline, aminoglycosides, quinolones, macrolides, and others [56].
For instance, Streptococcus uberis and Streptococcus agalactiae are responsible for mastitis infections in animals [57,58], leading to both economic losses for farmers and health complications for the infected cattle. Antibiotic therapy is still one of the main ways to manage mastitis infections in cows. However, the issue of antimicrobial resistance asks for reductions and better usage of antibiotics, and that is also the case for antibiotics applied as preventive measures against infectious diseases in farms [59]. As a result, some alternative therapies to traditional antibiotics are considered, including phage therapy and phage enzymes, probiotics, and bacteriocins [60].
Since topical administration is more feasible for antimicrobial peptides [9], one could speculate topical application of the LGMAI strains' substances to protect cattle udders and prevent infections that eventually result in mastitis and its subsequent implications. Nisin, for example, is a licensed food preservative used in over 50 countries [13]. Tests with nisin-containing sanitizer solutions and wipes have been carried out and showed satisfactory results on their efficacy in cleaning udder surfaces and preventing mastitis infections in cattle [60].
Not only do these species infect animals, but also humans. Streptococcus agalactiae (GBS) is a naturally occurring bacteria of the gastrointestinal and vaginal tracts that commonly produce no harm to their hosts. However, GBS infections can progress and result in bacteremia in adults and severe invasive GBS disease in newborns, this latter frequently fatal [58]. In particular, Streptococcus pyogenes is a pathogenic species infecting humans exclusively. Mainly found on the nasopharynx, these bacteria are responsible for causing mild to severe diseases, such as pharyngitis, scarlet fever, type II necrotizing fasciitis, and toxic shock syndrome [61]. Altogether, there are several streptococcal infections to be addressed in humans too. Additionally, CDC's 2019 report on antimicrobial resistance classifies Streptococcus pyogenes and Streptococcus agalactiae as concerning threats. Their increasing resistance to clindamycin and erythromycin limits treatment options, requiring monitoring and preventive actions.
It is worth mentioning that the Streptococcus agalactiae strains were isolated from bovine mastitis. Therefore, their ability to inhibit other mastitis-causing streptococcal species, such as Streptococcus uberis and other Streptococcus agalactiae strains, possibly represents a way of competition for the cow's udder site. This characteristic could be explored to address a search for alternative methods for dealing with mastitis infections in animals. Furthermore, future additional tests against other mastitis-causing bacteria should be considered.
The genomes presented GC content equivalent to the expected GC content for the Streptococcus agalactiae species and generated a good number of predicted genes/proteins. In addition, all three genomes passed the quality assessment for genome completeness with the Busco software, registering all the 124 universal single-copy orthologs, with only 2 of them fragmented.
The housekeeping genes aroE, gki, pheS, and recA, as well as the 16S rDNA, have already been used for taxonomic studies regarding the Streptococcus genus [62]. Here, we used these genes to generate concatenated sequences for the phylogenetic analysis of the LGMAI strains. For this, we performed a multilocus sequence analysis using the maximum likelihood method. The inferred phylogenetic tree allowed for the precise identification of the strains as Streptococcus agalactiae, with good bootstrap statistics support. Streptococcus species into the anginosus, bovis, mitis, mutans, pyogenes, salivarius, and sanguinis groups, as previously described [63,64], were also clearly retrieved. However, Streptococcus parasanguinis ATTC 15912 did not fall into the sanguinis group, requesting a more robust MLSA analysis with more orthologous proteins to better elucidate its group assignment.
Taxonomic studies for streptococcal bacteria using whole-genome sequencing (WGS) data are necessary because they enlighten much of the genus' relationships, especially the relationships within the viridans streptococci, which are sometimes confusing [63]. Antibiotic resistance analysis revealed two genes, mreA, and mprF, in all three LGMAI strains' genomes. The mreA gene encodes a riboflavin biosynthesis protein with flavokinase activity. This gene acts in a multidrug efflux pump manner and is associated with broad-spectrum resistance to many antimicrobial compounds, such as erythromycin, azithromycin, spiramycin, clindamycin, some cephalosporins, rifampicin, and others [65,66]. The mprF gene encodes a phosphatidylglycerol lysyltransferase that synthesizes lysylphosphatidylglycerol (LPG). LPG synthesis promotes resistance to various cationic antimicrobial peptides, such as defensins, cathelicidins, and bacteriocins, and thus also contributing to bacterial virulence [67,68]. The Centers for Disease Control and Prevention classify Streptococcus agalactiae as a concerning threat due to increasing resistance levels to clindamycin and erythromycin [69]. The mreA gene shows major significance to this extent since it is responsible for conferring resistance to erythromycin and clindamycin.
The virulence assessment analysis generated similar results in the three LGMAI genomes, accounting for important GBS virulence factors, such as the cps locus, cyl operon, and hylB gene. In addition, the cfa/cfb gene was found exclusively in the LGMAI_St_08 genome. The analysis identified 12 genes belonging to the cps locus-i.e., cpsA, cpsB, cpsC, cpsD, cpsE, cpsF, cpsK, cpsL, neuA, neuB, neuC, and neuD. All 12 genes mentioned are commonly shared by the 10 GBS serotypes [70]. This region encodes the gene products responsible for assembling of the GBS capsular polysaccharide (CPS), which is rich in sialic acid and encases Streptococcus agalactiae, facilitating bacteria immune evasion [71].
The analysis also identified 11 cyl operon genes-that is, cylX, cylD, cylG, acpC, cylZ, cylA, cylB, cylF, cylI, cylJ, and cylK. These genes are responsible for the synthesizing of the ornithine rhamnolipid pigment, also referred to as a hemolytic pigment due to its hemolytic activity [72]. This pigment contains a 12 double-bond polyene chain with potent cytotoxic activity and antioxidant properties, assisting infection and immune evasion processes [71]. The hemolytic pigment is associated with penetration of the human placenta and induction of loss of barrier function in human amniotic epithelial cells, promoting colonization of the feminine genital tract [72].
Cooperatively, fbsB, hylB, cfa/cfb genes, and pilus island PI-2b are vital during GBS infection in bovine mastitis [7]. The hylB gene encodes a hyaluronidase that acts as an endoglycosidase and cleaves glycosaminoglycan chains. Over-expression of the hyaluronidase has been demonstrated to confer hypervirulence for strains, even in those cases that lack the hemolytic pigment [71]. The degradation of the host hyaluronic acid releases its components, which bind to TLR2/TLR4 receptors and block their signaling pathways. Thus, this protein is crucial for immunosuppression during GBS infection [72].
The cfa/cfb gene was found exclusively in the LGMAI_St_08 genome. This gene produces the CAMP factor of Streptococcus agalactiae, which acts as a pore-forming toxin, ultimately lysing the host red blood cells [73]. The CAMP factor binds to the Fc site of IgM and IgG similarly to the Staphylococcus aureus protein A, which is responsible for immune evasion [74]. Both hylB and cfa/cfb genes cause tissue damage and are essential to GBS infection in cattle [7]. The identified virulence genes comprise the main virulence factors of Streptococcus agalactiae, but none of them show any correlation with the inhibition phenotype observed in the deferred growth assays.
The GIPSy software predicted eight genomic islands for the LGMAI_St_08 strain, seventeen for the LGMAI_St_11 strain, and fourteen for the LGMAI_St_14 strain. Comparative analysis, via mean nucleotide identity, revealed that most of the identified genomic islands represent conserved areas among Streptococcus agalactiae. Some of the LGMAI strains' putative islands, however, are not present in the genomes used for comparison.
Despite the vital role of bacteriophage sequences in Streptococcus agalactiae strains' variability and evolution [75], we could not recover any intact phage sequences from any of the three LGMAI strains' genomes. Nonetheless, two questionable phage sequences were identified in the LGMAI_St_08 genome. Otherwise, only incomplete bacteriophage sequences were registered for LGMAI_St_11 and LGMAI_St_14 strains.
The antiSMASH analysis of antibiotics and secondary metabolites identified putative sites for producing type III polyketide synthases, aryl polyenes, and class V lanthipeptide. Polyketides are natural products that possess biologically active properties, such as antineoplastic, cholesterol-reducing, and antimicrobial [76]. On the other hand, aryl polyenes represent bacterial pigments similar to the carotenoids that protect bacteria against reactive oxygen species [77]. Lastly, lanthipeptides are ribosomally synthesized peptides that contain the lanthionine amino acid on their structure. These peptides may have several biologically active properties, including antimicrobial. Lanthipeptides with antimicrobial activity are called lantibiotics [78]. However, no core peptide was identified for these sites.
The antiSMASH webserver also recognized one biosynthetic cluster for the nonribosomal peptide-synthetase (NRPS) equibactin. The equibactin system was first observed in Streptococcus equi subsp. equi and is associated with iron uptake [79]. It is worth mentioning that iron acquisition is a vital process for pathogenic bacteria because available iron is usually scarce in mammalian cells. Equibactin is analogous to the yersiniabactin system from Yersinia sp. Therefore, equibactin may play a role in Streptococcus agalactiae virulence [80].
The BAGEL4 database recognized the zoocin A core peptide, showing 58.76% identity, in all three LGMAI genomes. The bacteriocin zoocin A is a lytic exoenzyme first identified in Streptococcus equi subsp. zooepidemicus 4881. This bacteriocin is an endopeptidase responsible for hydrolyzing the peptidoglycan structure of susceptible organisms [81]. Moreover, the zoocin A activity spectrum has been shown to strongly target Streptococcus spp., as also observed in the antagonism assays performed in this study [82]. In the same fashion, the BACTIBASE database analysis also showed correspondence to the zoocin A core peptide from Streptococcus equi subsp. zooepidemicus, with 60% identity and 66% similarity. This antimicrobial peptide contains two domains: an N-terminal catalytic domain (peptidase M23) and a C-terminal target recognition domain (TRD) [83]. Such domains were successfully identified via InterPro analysis of the zoocin A protein sequences.
It is noteworthy that the BACTIBASE database does not account for zoocin A inhibitory activity against Micrococcus luteus and Streptococcus uberis. The LGMAI bacteriocin-like substances studied herein did inhibit Micrococcus sp. and Streptococcus uberis indicator strains, suggesting sequence variation and evolution for the putative zoocin A antimicrobial substance identified in the LGMAI strains.
Apart from the zoocin A core peptide, the immunity factor, also referred to as Zif, protects the producing bacteria and is a vital part of this bacteriocin activity scheme. The Zif factor is a FemABX-like protein that blocks binding to the catalytic and recognition domains via adding an extra l-alanine during the peptidoglycan biosynthesis [84]. The Prokka software annotated aminoacyl transferases FemB in all three LGMAI genomes. These sequences were submitted to families and domains analysis on the InterPro database, which resulted in identifying the FemABX peptidyl transferase family of proteins matching positions.
The analysis with the eggNOG-mapper tool showed a similar distribution of orthologous groups among the three LGMAI strains. Nevertheless, the LGMAI_St_11 and LGMAI_St_14 genomes share more orthologous groups between each other than they do with the LGMAI_St_08 genome.
Comparative genomics with the OrthoVenn2 webserver corroborated the previous observations. The three genomes share many orthologous groups, with a slightly more significant similarity between the LGMAI_St_11 and LGMAI_St_14 genomes. In addition, proteins for molecular function and cellular components are more conserved, while proteins for metabolic processes are more variable than those mentioned above.
Furthermore, it is interesting to mention that the LGMAI_St_08 genome registered 16 exclusive orthologous functional groups. Most of these groups are related to biological processes, which may be associated with adaptations to the LGMAI_St_08 environment.

Conclusions
The deferred growth inhibition assays showed that the LGMAI Streptococcus agalactiae strains produce antimicrobial substances with an activity spectrum targeting streptococcal species-specifically, Streptococcus pyogenes, Streptococcus agalactiae, Streptococcus uberis, and Streptococcus porcinus. The LGMAI strains also inhibited Micrococcus sp. and Cellulomonas fimi. The multilocus sequence analysis and the CDC's WGS GBS serotyping determined that the three strains are Streptococcus agalactiae serotype III.
Additionally, based on functional annotations and comparative genomics, the LGMAI_St_11 and LGMAI_St_14 strains share more similarities between them. These results are in agreement with their closer sampling geographic locality. Moreover, the three strains show high conservation of genes related to molecular function and cellular components.
Furthermore, assessments for antibiotic resistance genes identified the mreA gene in the three LGMAI genomes. Such a gene encodes a flavokinase and is related to resistance to erythromycin. Because of elevated resistance to erythromycin, Streptococcus agalactiae is already considered a concerning threat. Therefore, this finding contributes to the monitoring of antimicrobial resistance in GBS. Moreover, assessments for virulence factors recovered almost all of the main GBS virulence factors. The FbsA, FbsB, Lmb adherence proteins, pilus islands, and the cba gene remained unidentified.
Finally, putative sites were found for antibiotics and secondary metabolites with diverse biological activities, specifically antimicrobial, antioxidant, and iron uptake. Furthermore, the gene for the bacteriocin zoocin A, which targets streptococcal species, was also detected. Altogether, the antibiosis results and the zooA gene endorse the production of an antimicrobial substance by the LGMAI Streptococcus agalactiae strains. This substance targets pathogenic streptococci-notably, Streptococcus pyogenes, Streptococcus agalactiae, Streptococcus uberis, and Streptococcus porcinus. Data Availability Statement: The genomes' sequencing project has been deposited in the NCBI under BioProject accession number PRJNA637496. The Streptococcus agalactiae LGMAI_St_08, LGMAI_St_11, and LGMAI_St_14 genomes have been deposited in the NCBI under the accession numbers JAI-WPA000000000, JAIWPB000000000, and JAIWPC000000000, respectively.