Alterations in Genes rib, scpB and Pilus Island Decrease the Prevalence of Predominant Serotype V, Not III and VI, of Streptococcus agalactiae from 2008 to 2012

Streptococcus agalactiae (GBS) can infect newborns, pregnant women and immunocompromised or elderly people. This study aimed to investigate differences in three pilus genes and virulence genes pavA, cfb, rib and scpB and changes in predominant serotypes III, V and VI from 2008 to 2012. The susceptibilities to penicillin, ceftriaxone, azithromycin, erythromycin, clindamycin, levofloxacin and moxifloxacin of 145 GBS strains of serotype III, V and VI strains from 2008 and 2012 were determined using disc diffusion method. PCR identification of ST-17, the pilus genes and virulence genes; multilocus sequence typing (MLST); and conserved domain and phylogenetic analysis of scpB-1 and scpB-2 proteins were performed. A dramatic number reduction was observed in serotype V, not III and V, from 2008 to 2012. The rate of resistance to azithromycin, clindamycin and erythromycin was the highest in serotype V. ST-17 was only found in serotype III with pilus genes PI-1+PI-2b. The major pilus genotype was PI-1+PI-2a. Serotype V without the rib gene was reduced in number between two studied years. Compared to scpB-1, scpB-2 had a 128-bp deletion in a PA C5a-like peptidase domain and putative integrin-binding motif RGD. In conclusion, reduction in serotype V may be due to presence of scpB-2 or lack of genes scpB and rib.


Introduction
Streptococcus agalactiae (group B streptococcus; GBS) is a major cause of early-onset or late-onset neonatal sepsis or meningitis and infections in pregnant women and nonpregnant adults, especially those with underlying diseases such as diabetes and in immunocompromised or elderly people [1,2]. Penicillin is recommended to treat perinatal GBS infection [3]; nevertheless, erythromycin and clindamycin are alternative drugs to treat patients that are allergic to penicillin. Therefore, erythromycin-or clindamycin-resistant GBS has increased recently [4,5].
Several methods have been developed to detect GBS infection in humans and animals and to differentiate GBS strains. Ten GBS serotypes can be differentiated by multiplex PCR targeting capsular polysaccharide genes [6]. Furthermore, multiple locus sequence typing (MLST) can be used to determine the sequence patterns (ST) and clonal complexes (CCs), that can be used to trace the disease outbreak and determine the phylogenetic tree [7]. GBS ST-17 strains are highly invasive and virulent for neonatal infections and can be found mainly in serotype III and rarely in serotypes Ia and VI [8]. Therefore, rapid PCR identification of ST-17 has been developed using the gbs2018 gene as a target [9].

Antimicrobial Susceptibility Analysis
The number of strains differed among serotypes with little increase for serotypes III and VI and a dramatic decrease in serotype V ( Table 1). All strains were sensitive to penicillin, ceftriaxone, levofloxacin and moxifloxacin, and 53.1% of the strains were resistant at least to one of three antimicrobials azithromycin, clindamycin and erythromycin. The resistance rate was 51% for azithromycin, 49% for clindamycin and 49.7% for erythromycin overall and differed among three serotypes, being the highest in serotype V and the lowest in serotype VI. Most resistant strains were resistant to all three antimicrobials but differed in prevalence between the two periods with an increase from 29.4% to 38.9% in serotype III, a decrease from 60.3% to 53.3% in serotype V and a decrease from 41.2% to 25% in serotype VI.

Virulence Gene Analysis
Four virulence genes, namely pavA, cfb, scpB and rib, in S. agalactiae were detected with different patterns: one PCR product for pavA and cfb; three PCR products for scpB, namely 567-bp scpB-1, 439-bp scpB-2 and no PCR product; and two PCR products for rib,

Virulence Gene Analysis
Four virulence genes, namely pavA, cfb, scpB and rib, in S. agalactiae were detected with different patterns: one PCR product for pavA and cfb; three PCR products for scpB, namely 567-bp scpB-1, 439-bp scpB-2 and no PCR product; and two PCR products for rib, namely 369-bp PCR product and no PCR product ( Figure 2). The prevalence of these genes was 100% for both pavA and cfb, followed by 97.2% for scpB and only 34.5% for rib (Table 3). In scpB, more strains carried scpB-1 (60%) than scpB-2 (37.2%), which was only observed in serotype V with number reduction from 2008 to 2012, and no scpB was found in serotype V in 2008, not 2012. All serotype III and VI strains carried the scpB-1. An increase in the rib gene was observed from 21.7% to 56.6% between the two periods, due to a reduction in no-rib strains in serotype V from 2008 to 2012. Furthermore, the prevalence of rib differed among the three serotypes, being more than 88% in serotype III in both periods and exhibiting a decrease and an increase between the two periods in serotypes V and VI, respectively.  namely 369-bp PCR product and no PCR product ( Figure 2). The prevalence of these gen was 100% for both pavA and cfb, followed by 97.2% for scpB and only 34.5% for rib (Tab 3). In scpB, more strains carried scpB-1 (60%) than scpB-2 (37.2%), which was only observe in serotype V with number reduction from 2008 to 2012, and no scpB was found in ser type V in 2008, not 2012. All serotype III and VI strains carried the scpB-1. An increase the rib gene was observed from 21.7% to 56.6% between the two periods, due to a redu tion in no-rib strains in serotype V from 2008 to 2012. Furthermore, the prevalence of r differed among the three serotypes, being more than 88% in serotype III in both perio and exhibiting a decrease and an increase between the two periods in serotypes V and V respectively.

Amino Acid Comparison and Conserved Domain (CDD) Analysis of scpB Protein
The genes scpB-1 and scpB-2 of three serotypes were sequenced and compared wi sequences of accession numbers AF189002.2 (scpB-1) and FJ752116 (scpB-2). The 567-b scpB-1 was amplified at nucleotide positions from 969 to 1535, and 439-bp scpB-2 showe a 128-bp deletion at nucleotide positions from 1058 to 1185 and at amino acid sequenc from 353 to 395 (Figure 3). CDD analysis of scpB protein found a protease-associated d main containing proteins like Streptococcus pyogenes C5a peptidase domain (PA C5a-li peptidase domain) at amino acid positions 346-475 and putative integrin binding mo RGD at amino acid positions 389-391. The scpB-2 gene lacked this motif and partial pep dase domain. Using AF189002.2 as a standard, mutations were determined at L352I an Q447P in serotype III and at D342A, H344Q, N400K and R437K in serotype V (scpB-2 gen

Amino Acid Comparison and Conserved Domain (CDD) Analysis of scpB Protein
The genes scpB-1 and scpB-2 of three serotypes were sequenced and compared with sequences of accession numbers AF189002.2 (scpB-1) and FJ752116 (scpB-2). The 567-bp scpB-1 was amplified at nucleotide positions from 969 to 1535, and 439-bp scpB-2 showed a 128-bp deletion at nucleotide positions from 1058 to 1185 and at amino acid sequences from 353 to 395 ( Figure 3). CDD analysis of scpB protein found a protease-associated domain containing proteins like Streptococcus pyogenes C5a peptidase domain (PA C5alike peptidase domain) at amino acid positions 346-475 and putative integrin binding motif RGD at amino acid positions 389-391. The scpB-2 gene lacked this motif and partial peptidase domain. Using AF189002.2 as a standard, mutations were determined at L352I and Q447P in serotype III and at D342A, H344Q, N400K and R437K in serotype V (scpB-2 gene) and in serotype VI G776 strain (scpB-1 gene), and no mutations were determined in strains of serotypes V and VI (scpB-1 gene).

Phylogenetic Analysis
The phylogenetic analysis of amino acid sequences separated all strains into two groups, A (scpB-1) and B (scpB-2, except strain G776) (Figure 4). Group A was categorized into two subgroups, A1 (serotype III) and A2 (serotypes V and VI). Furthermore, group B was separated into subgroup B1 (serotype V with five mutations) and subgroup B2 (serotypes V and VI with four mutations). Interestingly, the scpB-1 gene in the serotype VI G776 strain had mutations identical to those in the scpB-2 gene, and another serotype VI G260 had no mutations in scpB-1. and in serotype VI G776 strain (scpB-1 gene), and no mutations were determined in strains of serotypes V and VI (scpB-1 gene).

Discussion
Group B Streptococcus (GBS) serotypes evolve differently. Earlier, we reported the change of predominant serotypes from Ib and V to III and VI due to a difference in genomic variation with a dramatic decrease in serotype V, not in serotypes III and VI (Table  1) [1]. Other factors may be involved in such a change in predominant serotypes. Antibiotic susceptibility analysis demonstrated that these three serotypes differed in their rate of resistance to clindamycin, erythromycin and azithromycin; further, the rate of re- Figure 4. Phylogenetic tree of the scpB protein of 8 strains in this study and 19 strains from Genbank. The phylogenetic tree was constructed using the unweighted pair group method with arithmetic mean (UPGMA) with bootstrap tests (1000 replicates) shown above the branches. Evolutionary distances were calculated using the p-distance method, in units of amino acid differences per site. All ambiguous positions for each sequence pair were deleted (pairwise deletion option).

Discussion
Group B Streptococcus (GBS) serotypes evolve differently. Earlier, we reported the change of predominant serotypes from Ib and V to III and VI due to a difference in genomic variation with a dramatic decrease in serotype V, not in serotypes III and VI (Table 1) [1]. Other factors may be involved in such a change in predominant serotypes. Antibiotic susceptibility analysis demonstrated that these three serotypes differed in their rate of resistance to clindamycin, erythromycin and azithromycin; further, the rate of resistance to clindamycin and erythromycin was higher than that of isolates from studies conducted in Switzerland and Korea (except serotype V) (Table 4) [16,17]. The overall rate of resistance to clindamycin and erythromycin found in this study was lower than that found in studies in China [8,18] and the strains in the USA [5] and higher than that in Brazil and Iran [19,20]. As mentioned above, the highest antimicrobial resistance rate was not associated with number reduction in serotype V, suggesting other factors may be involved. Highly invasive GBS serotype III/ST-17 is the main cause of neonatal sepsis or meningitis in China, France and Serbia [21][22][23]. In this study, we confirmed that PCR identification of ST-17 is a useful method and that ST-17 was only found in serotype III, not in serotypes Ia and VI (Table 2); these results differed from those in a previous report of the existence of ST-17 in serotypes Ia and VI in China [8]. Pili play an important role in bacterial attachment to epithelial cells. GBS contains three pilus genes, namely PI-1, PI-2a and PI-2b [10]. PI-1+PI-2a is the major pilus type in China, Iran and Taiwan (Table 2) [18,20] and is responsible for colonization and biofilm formation [24]. However, serotype III/ST-17 showed two pilus types, PI-1+PI-2b and PI-2b, with PI-2b being the predominant (Table 2). These results confirm PI-2b in highly invasive ST-17 strains, which is important for GBS and host interaction during GBS infection [11]. Furthermore, PI-2b can be found in serotypes Ia and II [20], serotype V ( Table 2) and serotype VI [25]. In this study, serotype V with PI-2b was observed in 2008 and disappeared in 2012 ( Table 2), suggesting that PI-2b is not important for adherence or invasion of serotype V and that other virulence genes are involved in the reduction in serotype V.
All serotype III/ST-17 strains had four virulence genes, pavA, cfb, rib and scpB. Despite the presence of virulence genes pavA and cfb in all strains, the prevalence of rib and scpB had no change in serotype III and increased in serotypes V and VI between the two studied years ( Table 3). As a member of the Alp protein family with highly repetitive sequences stimulating protective immunity, the rib gene was identified in 69.8% of GBS isolates and was associated with serotypes II, III and V, especially in the serotype III/ST-17 strains [8,12,26]. In Romania, the rib gene was found in 73.7% of serotype III strains but was absent in serotype V strains [27], while in China, the rib gene was found in 70.6% of serotype III strains and only in 10.5% of serotype V strains [8]. These results demonstrate the serotype association of the rib gene. In this study, the rib gene was found in 88.5% of serotype III strains, 21.6% of serotype VI strains and 15% of serotype V strains, and the prevalence of the rib gene between the two periods did not change in serotype III and increased in serotype V and VI (Table 3). However, a significant reduction in number from 55 to 7 was found in serotype V from 2008 to 2012. These results imply that the rib gene is an important virulence factor. CDD analysis determined that the scpB gene encodes a surface-localized serine protease: C5a-like peptidase, which inactivates human C5a to inhibit complement activation, with a putative integrin-binding motif RGD (Figure 3) [26]. Commonly, GBS strains carry the scpB gene [28][29][30]. In the present study, we identified two scpB genes: full-length scpB-1 and deletion scpB-2 ( Figure 3). scpB-2 was reported in the USA in 2009 [31] and lacked the partial C5a-like peptidase and putative integrin-binding motif RGD (Figure 3). Further, scpB-2 gene was only found in serotype V; a change in number from 42 to 5 was found for strains with the scpB-2 gene, and an increase from 2 to 5 for those strains with the scpB-2 and rib genes in serotype V (Table 3). These data suggest that genes scpB-1 and rib are important virulence factors. The amino acid comparison and phylogenetic analysis of scpB genes demonstrated serotype-associated mutations and the evolution of scpB-2 in serotype V possibly from scpB-1 in serotype VI (Figure 3 and 4). This phenomenon may be due to phase variations.

Bacterial Isolates
In total, 145 GBS isolates of serotypes III, V and VI were collected from Chiayi Chang Gung Memorial Hospital in 2008 and 2012 [1]. The GBS genomic DNA was purified using a DNA extraction kit (Quality Systems, Taipei, Taiwan). Purified DNA was separated in 0.8% agarose gel in 0.5× TBE buffer at 100 V for 35 min. After staining with ethidium bromide, images were recorded, analyzed and quantified using ImageJ free software (https://imagej.nih.gov/ij/; Accessed on 03 June 2022).

Antimicrobial Susceptibility
Susceptibility to penicillin, ceftriaxone, azithromycin, erythromycin, clindamycin, levofloxacin and moxifloxacin was determined by disc diffusion method according to guidelines of the Clinical and Laboratory Standards Institute (CLSI) [32]. Streptococcus pneumoniae ATCC49619 was used as standard.

Pilus Island Genes, Virulence Genes and Sequencing
The pilus island genes PI-1, PI-2a and PI-2b were identified by multiplex PCR with the primer sets listed in Table 5. PI-1-negative isolates were confirmed further by additional PCR reaction using primers PI1_all [33]. Virulence genes rib, pavA, scpB and cfb were identified by previous PCR methods with the primers listed in Table 5 [30]. PCR products were separated in 1.5% agarose gel in 0.5× TBE buffer at 100 V for 35 min. After staining with ethidium bromide, images were recorded and analyzed. We randomly selected different PCR products of scpB-1 and scpB-2 amplified from three serotypes for sequencing. The forward and reverse sequences were assembled with the SeqMan software of Lasergene DNAStar 7.1. Assembled sequences were annotated using the NCBI BLAST program (https://blast.ncbi.nlm.nih.gov/Blast.cgi; Accessed on 10 March 2022).

Phylogenetic Analysis of scpB Gene
The amino acid sequences of 8 strains were aligned with those sequences of 19 strains from 10 countries including the United States, Canada, Japan, China and Singapore from the NCBI Genbank database. Phylogenetic analysis of scpB protein was performed using the software Molecular Evolutionary Genetic Analysis (MEGA 11, https://megasoftware.net, 12 May 2022). The phylogenetic tree was constructed using (UPGMA) with bootstrap tests (1000 replicates). The branch lengths of this tree are the same units used to infer Pathogens 2022, 11, 1145 9 of 10 evolutionary distances in phylogenetic trees. Evolutionary distances were calculated using the p-distance method with 7 threads of system resource usage. All ambiguous positions for each sequence pair were deleted (pairwise deletion option).

Conclusions
In conclusion, the predominant change in three serotypes between two periods may be due to the presence of ST-17 in serotype III with PI-1+PI-2b, scpB-1 and rib genes in serotypes III and VI. The scpB-1 protein contains a C5a-like peptidase and lacks this domain in scpB-2. Therefore, either the lack of the scpB gene and rib gene or the presence of scpB-2 may have some influence on the reduction in the number of serotype V strains.