Streptococcus vaginalis sp. nov., a novel bacterial species isolated from vaginal swabs of a pregnant woman with diabetes

Sequences targeted at the V3 and V4 16S rRNA hypervariable regions of a streptococcal strain (P1L01T) isolated from vaginal swabs of a pregnant woman with diabetes were 100% similar to those of Streptococcus anginosus subsp. whileyi. However, phylogenetic analysis based on 16S rRNA full-gene sequencing (1562 bp) revealed highest sequence similarity to Streptococcus periodonticum (98.7%), followed by Streptococcus anginosus subsp. whileyi (98.7%), and Streptococcus anginosus subsp. anginosus (98.4%). Phylogenies of housekeeping genes rpoB and groEL were compared to improve classification, and the results showed a clear separation between strain P1L01T and closely related Streptococcus type strains. The complete genome of strain P1L01T consisted of 2,108,769 bp with a G + C content of 38.5 mol%. Average nucleotide identity values, based on genome sequencing, between strain P1L01T and Streptococcus periodonticum KCOM 2412T, Streptococcus anginosus subsp. whileyi CCUG 39159T, and Streptococcus anginosus subsp. anginosus NCTC 10713T were 95.5%, 94.3%, and 95.3%, respectively. The highest in silico DNA–DNA hybridization value with respect to the closest species was 66.2%, i.e., below the species cutoff of 70% hybridization. The main cellular fatty acids of strain P1L01T were 16:0, 18:1ω7c, and 14:0. On the basis of phylogenetic, genotypic and phenotypic data, we propose to classify this isolate as representative of a novel species of the genus Streptococcus, Streptococcus vaginalis sp. nov., in reference to its isolation from vaginal swabs, with strain P1L01T (= NBRC 114754T = BCRC 81289T) as the type strain.


Introduction
The genus Streptococcus, a heterogeneous group of Grampositive bacteria, includes some of the most significant human pathogens -including S. pyogenes, S. pneumoniae (Megged 2020), Streptococcus suis (Yi et al. 2020), S. iniae, S. parauberis , and S. agalactiae (Group B Streptococcus, GBS) (Opinion 2020). Of the 178 different species of the genus Streptococcus (https:// lpsn. dsmz. de/), several are known to cause gynecological and obstetrics infections. Notably, GBS is the predominant infectious organism that causes bacteremia or sepsis in newborns and young infants (Opinion 2020). During labor and delivery, the vertical transmission of GBS from an infected mother to her fetus can also lead to pneumonia or meningitis (Scholl et al. 2016;Yu et al. 2011). For this reason, antibiotics are routinely given for preventing peripartum infections in women who test positive for GBS colonization in the anogenital region (Francois Watkins et al. 2019;Opinion 2020). However, only a limited number of studies have investigated the prevalence of other streptococcal species within the vaginal environment during pregnancy (Al Majid et al. 2020;Kolter and Henneke 2017;Rabe et al. 1988).
The Streptococcus anginosus group is a subgroup of viridans streptococci that comprises three closely related 1 3 species: S. intermedius, S. constellatus, and S. anginosus, of which the last named is divided into two subspecies (S. anginosus subsp. anginosus and S. anginosus subsp. whileyi). In the current study, a streptococcal strain (termed P1L01 T ) was isolated from vaginal swabs of a pregnant woman with type 2 diabetes at 35 weeks of gestation. 16S rRNA gene-based phylogenetic analyses were undertaken, and species that showed > 98.5% similarity were selected for further characterization and comparison (Lim et al. 2019). Phylogenies of housekeeping genes rpoB (Glazunova et al. 2006;Saito et al. 2016) and groEL (Glazunova et al. 2006;Niu et al. 2017;Zbinden et al. 2012) were also analyzed to improve classification. While sequences targeted at the V3 and V4 16S rRNA hypervariable regions were 100% similar to those of Streptococcus anginosus subsp. whileyi, phylogenies of housekeeping genes rpoB and groEL revealed a clear separation between P1L01 T and related type strains of Streptococcus. Herein, we investigated the taxonomic status of strain P1L01 T and propose to classify this isolate as representative of a novel species of the genus Streptococcus, Streptococcus vaginalis sp. nov.

Sample isolation
As recommended by the National Health Bureau Care Program (Taiwan), vaginal swabs of a Taiwanese pregnant woman with type 2 diabetes were obtained at 35 weeks of gestation as part of screening for GBS colonization. Swab specimens were submitted for bacterial species identification. Under institutional review board approval (IRB number: 201701371A3), the patient gave written informed consent. Strain P1L01 T was isolated from cultures grown at 37 °C in brain-heart infusion (BHI) medium under anaerobic conditions. As the patient was at high risk for preterm labor, steroids and antibiotics were administered at 37 weeks of gestation to accelerate lung maturation and prevent peripartum infections.

Genomic features and gene phylogeny
Genomic DNA was extracted from strain P1L01 T using the QIAGEN DNeasy Blood & Tissue kit (QIAGEN, Valencia, CA, USA). Phylogenies of housekeeping genes rpoB and groEL were compared to improve classification within the genus Streptococcus (Glazunova et al. 2006;Niu et al. 2017;Saito et al. 2016;Zbinden et al. 2012). Genomic sequences of the 16S rRNA, rpoB, and groEL genes were obtained from the DNA Data Bank of Japan (DDBJ; http:// www. ddbj. nig. ac. jp/) using ARSA (accession number: NZ_CP034543, NZ_CP012805, and NZ_LR134283). Neighbor-joining phylogenetic analyses, based on 16S rRNA gene sequence comparison, were performed using the EzBioCloud database (http:// www. ezbio cloud. net/ ident ify). Sequence homology analyses for the rpoB and groEL genes were undertaken using the Genetyx-Win software (version 5.1; Genetyx Corporation, Tokyo, Japan).
The draft genome sequence of strain P1L01 T was obtained with an Illumina MiniSeq system (Illumina, San Diego, CA, USA) using the 600-cycle MiSeq Reagent Kit v3. De novo assembly of reads was performed using the SPAdes (version 3.10.1) (Bankevich et al. 2012). Gene annotation of the draft sequence was carried out using the NCBI Prokaryotic Genomes Automatic Annotation Pipeline (http:// www. ncbi. nlm. nih. gov/ genom es/ static/ Pipel ine. html) (Klimke et al. 2009). Average nucleotide identity (ANI) values was calculated with Orthologous Average Nucleotide Identity tool (Lee et al. 2016). Digital DNA-DNA hybridization (dDDH) values were determined using Formula 2 of the Genome-to-Genome Distance (GGD) Calculator (Meier-Kolthoff et al. 2013).
Sequence alignment was carried out using the CLUSTAL W software (Thompson et al. 1997) and the MEGA X package (Kumar et al. 2018) was used for all analyses. Phylogenetic trees were constructed with different methods-including maximum likelihood (ML), minimum evolution (ME) (Rzhetsky and Nei 1992) and neighbor joining (NJ) with the Kimura's two-parameter model (Kimura 1980;Saitou and Nei 1987)-using 1000 bootstrap replications.

Phenotyping
Sheep blood (final concentration, 5% in BHI medium) was used for hemolytic activity assays. Fermentation capacity was examined using a commercial kit (API 50 CHL; bio-Mérieux, Lyon, France) according to the manufacturer's protocol. Halotolerance (2.5, 3.5, 4.5, 5.5 and 6.5% NaCl) and ranges of permissive growth temperatures (10, 20, 30, 37, 45, 50 °C) and pH (5.0 to 10.0 in increments of 0.5 pH units as adjusted with HCl and NaOH) were determined as described by Kozaki et al. (1992). The biochemical profiles of P1L01 T and closely related strains (S. periodonticum, S. anginosus subsp. whileyi, and S. anginosus subsp. anginosus) were assessed using the API ZYM and API Rapid ID 32 Strep systems (bioMérieux). Whole-cell fatty acid methyl ester (FAME) profiling was performed using the Sherlock Microbial Identification system (MIDI). After incubation of bacterial cells on solid BHI medium for 48 h at 30 °C under anaerobic conditions, FAME extraction and preparation were carried out as described by Sasser (1990).

Gene phylogeny
The 16S rRNA gene sequences (1562 bp) of strain P1L01 T were compared with those stored in the EzBioCloud database. Maximum-likelihood phylogenetic tree analysis based on the 16S rRNA genes of the isolated strains and their closest related species is depicted in Fig. 1. Three species-Streptococcus periodonticum JCM 33300 T (Japan Collection of Microorganisms, Ibaraki, Japan), Streptococcus anginosus subsp. whileyi DSM 25818 T (DSMZ GmbH, Braunschweig, Germany), and Streptococcus anginosus subsp. anginosus BCRC 14730 T (BCRC, Hsinchu, Taiwan)-were most related to strain P1L01 T . The genotypic characteristics and similarity of the 16S rDNA, rpoB and groEL genes of strain P1L01 T and closely related species and subspecies are shown in Table 1. The highest sequence similarity was observed with Streptococcus periodonticum (98.7%), followed by Streptococcus anginosus subsp. whileyi (98.7%), and Streptococcus anginosus subsp. anginosus (98.4%; Table 1). Similar findings were obtained using the NJ and ME algorithms (Supplementary Figures S1 and S2). The concatenated sequences (5190 bp) of two housekeeping genes (rpoB, 3567 bp; groEL, 1,623 bp) in strain P1L01 T had 93.6-95.1% identity with those of S. periodonticum, S. anginosus subsp. whileyi and S. anginosus subsp. anginosus (Table 1).
Phylogenetic trees based on concatenated nucleotide alignments of housekeeping genes were constructed using the ML, NJ, and ME algorithms. The results showed a clear separation between P1L01 T and Streptococcus reference strains ( Fig. 2 and Supplementary Figures S3 and S4).

Genome features
The complete genome of the P1L01 T strain-which consisted of 2,108,769 bp with a G + C content of 38.5 mol%contained 1808 protein-coding genes and 53 predicted RNA genes. ANI values, based on genome sequencing, between strain P1L01 T and Streptococcus periodonticum, Streptococcus anginosus subsp. whileyi, and Streptococcus anginosus subsp. anginosus were 95.5%, 94.3%, and 95.3%, respectively. The dDDH value with respect to the closest species was 66.2% (Table 1).

Biochemical features
The biochemical characteristics of strain P1L01 T were determined in parallel to those of Streptococcus periodonticum, Streptococcus anginosus subsp. whileyi, and Streptococcus anginosus subsp. anginosus strains. Significant differences for strain P1L01 T were observed compared with other strains with respect to fermentation capacity and enzyme activities ( Table 2). All negative traits from API 50 CHL kit are listed in Supplementary Table 1. The major cellular fatty acids of strain P1L01 T were C 16:0 (40.7%), C 18:1 ω7c (18.0%), and   Table 2).

Phenotypic features
The cells were γ-hemolytic on BHI agar containing 5% sheep blood under aerobic or anaerobic conditions. After incubation in BHI broth anaerobically at different temperatures, strain P1L01 T grew well at 30 and 37 °C for 48 h and grew weakly at 20 and 45 °C. Growth could be observed at pH 4.5-8.5. Growth at various NaCl concentrations on BHI broth was assessed and the results showed that growth occurred in the presence of up to 5.5% NaCl.

Discussion
Diabetes in pregnancy including gestation diabetes, type 1 and type 2 diabetes is the most common metabolic complications of pregnancy and is increasing in prevalence (American Diabetes Association Diabetes Care 2021; Dunlop et al. 2015). Diabetes in pregnancy is known to increase the risk of adverse maternal and fetal outcomes. While maternal microbiome has been shown to play an important role in affecting pregnancy outcomes (Tremaroli 2012, Tilg 2014, Dunlop et al. 2015, we have previously demonstrated that Prevotella bivia dominated in pregnant women with severe hypertension (Lin et al. 2020). Diabetic mothers are, however, particularly prone to develop cervicovaginal infections such as GBS, mycoplasma hominis, and ureaplasma urealyticum (Dunlop et al. 2015;Lukic et al 2017). The isolation of novel Streptococcus vaginalis sp. nov. in this study corroborated the concept of dysbiosis in impaired glucose metabolism of pregnancy. Next-generation sequencing (NGS) is increasingly being applied for the analysis of complex microbial communities (Boers et al. 2019;Graspeuntner et al. 2018). Most published studies in the field have relied on sequences targeted at the V3 and V4 16S rRNA hypervariable regions (319F-806 R). During the analysis of vaginal swabs obtained from a Taiwanese pregnant woman with type 2 diabetes, we isolated a strain (P1L01 T ) that could not be identified conclusively to the species level based on 16S V3-V4 rRNA gene sequence analysis. The highest sequence homology at these regions was obtained with Streptococcus anginosus subsp. whileyi and Streptococcus anginosus subsp. anginosus (100% and 99.8%, respectively). On analyzing the sequences targeted at the V6 region (926F-1100R), strain P1L01 T was distinctly different from S. anginosus subsp. whileyi (similarity: 90.2%), S. anginosus subsp. anginosus (similarity: 90.2%), and S. periodonticum (similarity: 94.3%). The highest ANI and dDDH values among strain P1L01 T and the S. periodonticum type strain were 95.5% and 66.2%, respectively. While dDDH values were lower than the commonly accepted cutoff (70%), the ANI value between strain P1L01 T and S. periodonticum KCOM 2412 T was of uncertain significance (Goris et al. 2007;Rossello-Mora and Amann 2015). Recent studies have suggested that bacterial isolates with ANI values < 95-96% or dDDH values < 70% with respect to type strains of their closest related species and genera should be classified in novel species Chun et al. 2018;Lim et al. 2019). Thus, a hypothesis was set forth that strain P1L01 T might represent a novel species of the genus Streptococcus.
Additional phenotyping of strain P1L01 T further supported this possibility. First, strain P1L01 T was γ-hemolytic, whereas the other three strains investigated in this study showed α-hemolytic activity. Second, biochemical analysis of strain P1L01 T demonstrated distinct characteristics compared to other strains in terms of the carbohydrate metabolism-with clear differences in acid production from melibiose, N-acetyl glucosamine, amygdaline, d-raffinose and β-gentiobiose. Additionally, strain P1L01 T can be distinguished from other strains through α-galactosidase and alanyl-phenylalanyl-proline arylamidase activities.
Based on phylogenetic, genotypic and phenotypic data, we therefore propose to classify this isolate as representative of a novel species of the genus Streptococcus, Streptococcus vaginalis sp. nov., in reference to its isolation from vaginal swabs. However, the question as to whether this novel species should be considered a pathogen or a cause In conclusion, we have confidently recognized a novel species of the genus Streptococcus, termed Streptococcus vaginalis sp. nov., using phylogenetic, genotypic and phenotypic identification methods. The novel species was isolated from vaginal swabs of a pregnant woman with diabetes and was most closely related to S. anginosus. The role of Streptococcus vaginalis sp. nov. in the pathogenesis of gynecological and/or obstetrics infections is currently questionable. More studies are necessary to clarify whether the novel species should be considered a pathogen or a cause of human infection.
The type strain, P1L01 T (= NBRC 114754 T = BCRC 81289 T ), was isolated from vaginal swabs of a pregnant woman with diabetes in Taiwan.