Streptococcus anginosus of the urogenital tract: evidence of the same strain across anatomical sites of the same females

Abstract

Background

Streptococcus anginosus has been found to colonize multiple anatomical sites of the human body, including the oral cavity, gastrointestinal tract, skin, vagina, blood, and urinary tract. It is frequently isolated from catheterized urine samples obtained from adult females most often with lower urinary tract symptoms, but also on occasion from those without symptoms. Our prior genome analysis of 166 S. anginosus genomes identified two distinct groups, one – which we here call group Urinae – exhibits a tropism for the urinary tract.

Results

Here we sequenced 100 isolates collected from different urogenital sites, including isolates from urine (both catheterized and voided) samples, urethral swabs, perineal swabs, vaginal swabs, and a foreskin swab. These sequenced isolates, as well as 15 isolates previously sequenced by our group, were collected from 50 unique individuals, with isolates from multiple anatomic sites for 26 of these individuals. The majority (89.57%) of the isolates were representatives of the Urinae group and were found in all sample types. Expanding our prior statistical method, we determined that, for 11 females, the same strain of Urinae was isolated from more than one of the urogenital sites.

Conclusions

The identification of S. anginosus group Urinae strains from all urogenital sites sampled signifies that the tropism of this group is not restricted to the urinary tract. Rather, it seems to be a common constituent of the urogenital tract. The identification of the same strain shared across urogenital sample sites from the same individual suggests that the female urogenital sites in fact have interconnected microbiota.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-025-11973-4.

Background

Streptococcus anginosus is a member of the Streptococcus Anginosus Group (SAG), which includes S. anginosus, Streptococcus constellatus, and Streptococcus intermedius. Within S. anginosus species, two named subspecies (anginosus and whileyi) and two named genomosubspecies (AJ1 and vellorensis) exist. This complicated nomenclature demonstrates the challenges faced in classifying SAG and its species and subspecies. Furthermore, S. anginosus is phenotypically inconsistent and has been isolated from a multitude of human anatomical sites, including the oral cavity, gastrointestinal tract, skin, vagina, blood, and urinary tract [1–5]. Our group’s previous investigations of the female bladder microbiota have routinely identified the species S. anginosus in both individuals with urge urinary incontinence at quite high abundance, and sometimes in continent controls, albeit at considerably lower abundance [6–8].

Our previous genomic analysis of 166 S. anginosus genomes found that this species can be divided into two groups based upon the phylogenetic tree of their core genome [5]. The first group included strains isolated from a variety of sites (blood, reparatory tract, oral, fecal, and urine) as well as the species type strain (“Group 1”), while the other was predominated by isolates from the urinary tract, mostly collected by catheter and thus of bladder origin (n = 77; 85.56%) and three strains from vaginal samples (“Group 2”). Most isolates from urine samples (90.59%) were of “Group 2,” suggesting that this group may be specific to the urogenital tract [5]. Subsequently, nine S. anginosus isolates from vaginal samples have been made publicly available [9], although their group membership is unknown. With so few vaginal strains sequenced to date, it is unclear if the species is resident of the vaginal tract or transient.

Prior studies have found the same species of bacteria within the female urinary and vaginal microbiota [10, 11]. Furthermore, similar strains have been isolated from urine and vaginal samples collected from the same individual [10, 12]. Recently, our investigation of lactobacilli isolated from urine samples, perineal swabs, and vaginal swabs from the same female identified instances in which the same strain inhabited two or more of these anatomical sites [13]. These studies suggest that the female urinary and vaginal microbiota are interconnected [10]. This prompted our investigation here which analyzed the genomes of S. anginosus isolates collected from the same individual from multiple urogenital sites. We sequenced 100 S. anginosus isolates, including multiple samples from 26 different females. By narrowing our focus on a single species, we can further investigate the interconnectedness of the microbiota of adjacent urogenital sites.

Methods

Isolation of strains

Our study includes 100 isolates from prior IRB-approved studies (IRB nos., Loyola University Chicago – 204133, 204195, 206129, 206469, 207102, 207152, 209983, and 210444; University of California, San Francisco – 15-16363; University of Iowa – 201608793) [8, 14–21]. Briefly, for voided and catheterized urine specimens, samples were processed using the Expanded Quantitative Urine Culture (EQUC) protocol [16, 22]. For strains from vaginal and perineal swabs, samples were collected using a BD ESwab^® collection kit (catalog no. 220245). Each swab was swirled in 1 ml of bacterial preservative (liquid Amies) and an aliquot was provided for EQUC processing. All isolates were identified via MALDI-TOF, as previously described [22], and stored at −80 °C in the Loyola Urinary Education and Research Collaborative (LUEREC) collection at Maywood, IL USA.

DNA extraction and sequencing

Each strain was streaked on a Columbia nalidixic acid (CNA) agar plate and incubated for 24 h at 35 °C with 5% CO₂ at the Loyola University Chicago Health Sciences Campus (Maywood, IL USA) and then supplied to the Putonti lab at Loyola University Chicago’s Lake Shore Campus (Chicago, IL USA). There, a single colony was selected and incubated in Brain Heart Infusion media (BHI) (BD Biosciences) and incubated for 24 h at 35 °C with 5% CO₂. DNA was extracted from the liquid culture using the Qiagen DNeasy Blood and Tissue Microbial Kit, following the manufacturer’s protocol for Gram-positive bacteria, and quantified using a Qubit fluorometer. DNA was sent to SeqCoast Genomics, LLC (Portsmouth, NH USA). There, libraries were constructed using an Illumina DNA Prep tagmentation kit and unique dual indexes. Sequencing was performed on the Illumina NextSeq2000 platform (San Diego, CA USA) using a 300-cycle flow cell kit, which produced 2 × 150 bp paired reads. Read demultiplexing, read trimming, and run analytics were performed using DRAGEN v3.10.12 (Illumina) and then delivered from SeqCoast. Each sample was sequenced with the intent to produce a minimum of 200 Mbp (100x coverage); as Additional file 1 shows, this was exceeded for all but one strain.

Genome assembly and annotation

Raw reads were trimmed using bbduk (part of the BBMap suite, v39.01; sourceforge.net/projects/bbmap/) with the following parameters: minlength = 30, qtrim = rl, maq = 20, maxns = 0, trimq = 20. Trimmed reads were then assembled using SPAdes v3.15.5 with the --only-assembler option [23]. Contigs less than 1000 bp in length were trimmed from the assemblies. Contamination and completeness were computed using CheckM [24] through BV-BRC [25]. Species designations were confirmed through BV-BRC and by computing the average nucleotide identity (ANI) via FastANI v1.32 [26]. Annotation was performed during submission to the NCBI Assembly database by the Prokaryotic Genome Annotation Pipeline (PGAP) [27]. (Note, some assemblies were annotated by different versions of PGAP depending upon when they were submitted.) Genome assembly statistics are listed in Additional file 1.

Phylogenetic analysis

A pangenome was derived using Anvi’o v7.2 [28]. Complete genomes assembled as part of this study and publicly available S. anginosus genomes were included in this analysis. Publicly available genomes included in our analysis are listed in Additional file 2. These genomes include isolates sequenced by others as well as our group. Publicly available genomes from the participants included in this study are indicated in Additional file 1. The core genome was computed to include the set of single-copy genes conserved among all genomes, and the aligned amino acid sequences of the core were retrieved from the pangenome. A phylogenetic tree was derived using the FastTree v.2.1.11 [29] plug-in through Geneious Prime v.2022.2.1. The tree was visualized with iTOL v. 6.9 [30]. Within the resulting tree, paired samples were color-coded using Adobe Illustrator v28.5.

Statistical analysis

To simulate genome assemblies from the same isolate, the fastq files for the 115 genomes examined in this study (Additional file 1) were randomly subsampled using seqtk (https://github.com/lh3/seqtk) and its sample command. We randomly subsampled 150,000 read pairs from each R1 and R2 fastq file 50 times. The seqtk sample command was seeded using 50 unique seed values generated via the Python random library. In total, 50 pairs of subsampled reads were produced for each genome. Next, each subsampled fastq file pair was trimmed and assembled, as described above for genome assembly and annotation. Pairwise comparisons were done between these simulated genomes using two metrics. ANI was computed using FastANI v1.32 [26] with default parameters (k = 16). Mash distance was computed using mash v2.2 [31] using the dist command and the parameter “-s 1000000” for the sketch size.

Analysis was conducted independently for the genomes of S. anginosus Group 1 (herein called Anginosus) and the S. anginosus Group 2 (called Urinae). For each group, ANI values were pooled together and bootstrapped with replacement 10,000 times. This also was done for the mash distance values. Analysis of Variance (ANOVA) was performed to test for strain level variation. We used kernel density estimation to construct an empirical distribution of ANI/mash distance values for each group. The pairwise ANI/mash distance values of genome assemblies were projected onto the empirical distribution of ANI/mash distance values for that group for easy visualization. Additionally, the null hypothesis tested is that the strains are statistically the same. This was tested against the one-sided alternative that the observed ANI value is an outlier on the left tail of this distribution or that the observed mash distance value is an outlier on the right tail of this distribution, i.e., the ANI/mash distance value corresponds to 2 dissimilar strains. To test this hypothesis, an empirical p value was calculated by estimating the empirical probability using the pemp function (EnvStats) in R. All graphs were generated using R.

All code for subsampling, written in Python, and statistical analysis, written in R, is available through GitHub (https://github.com/putonti/LactoCompare).

Genome comparisons of technical replicates

S. anginosus UMB4687B a member of the Urinae group, was selected to test the ANI and mash values of the same strain extracted and sequenced independently ten times. The strain was grown, the DNA was extracted and sequenced, and the reads were trimmed and assembled as described above. ANI and mash values were computed and compared to the distribution for the simulations for sequences identified in the Urinae group. Raw reads have been deposited in BioProject PRJNA1150124.

Results

S. anginosus genomes from the urogenital tract

Here, we sequenced 100 isolates, identified via MALDI-TOF as S. anginosus. In addition to 15 isolates previously sequenced by our group, these isolates were collected from different anatomical sites from 50 unique individuals. As noted in Additional file 1, the 115 genomes examined here were isolated from urine (catheterized and voided) samples, vaginal swabs, perineal swabs, urethral swabs, and one sample from a foreskin swab. These samples were collected from asymptomatic (“healthy”) individuals as well as those with overactive bladder symptoms (OAB), stress urinary incontinence (SUI), urge urinary incontinence (UUI), or interstitial cystitis/painful bladder syndrome (IC/PBS). For 8 individuals, S. anginosus was isolated and successfully sequenced at more than one collection timepoint. These include participants with OAB in a study of the effects of vaginal estrogen cream on symptoms [15] (participant ID starts with “EST”) and participants with IC/PBS in a study of the effects of acupuncture on symptoms ([19]; treatment study unpublished). For 26 individuals, at 32 unique collection timepoints, we sequenced genomes of S. anginosus from more than one collection sample types.

Genome analysis via ANI revealed that these isolates are members either of Group 1 proposed by Prasad et al. [5], which includes the previously characterized two subspecies and two genomosubspecies, or members of Group 2 proposed by Prasad et al. [5]. Because Group 2 is predominately comprised of urinary isolates, we will here refer to the group as “Urinae” and to Group 1 as “Anginosus,” as it includes the characterized subspecies and genomosubspecies for S. anginosus. The majority (89.57%) of the isolates examined here are representatives of the Urinae group (Table 1).

Table 1.

Number of genomes sequenced for each group from different urogenital sample types

Sample Type	Group		Total
	Urinae	Anginosus
Catheterized Urine	32	2	34
Voided Urine	24	3	27
Urethral Swab	2	1	3
Perineal Swab	14	2	16
Vaginal Swab	30	4	34
Foreskin	1	0	1
Total	103	12	115

The pangenome for these 115 genomes, as well as 153 publicly available genomes, was computed. The single-copy core genome was identified and used to derive a phylogenetic tree. Shown in Fig. 1, most of the genomes isolated from the same individual and timepoint have a core genome most similar to other genomes from that individual and timepoint. Furthermore, paired samples from the same individual but collected at different timepoints most often clade together. However, exceptions to these observations exist. The two genomes from participant EST33 at Baseline (indicated in Fig. 1 by the blue arrows) do not clade together. The genomes from participant EST49 at Baseline (indicated in Fig. 1 by the brown arrows) clade separately; the sample from the voided urine does not in the clade with the other two samples (catheterized urine and vaginal swab) from this participant’s collection. While the genomes from EST15 at Baseline (Fig. 1, teal triangles; n = 2) clade together and the genomes from EST15 at 12-week (Fig. 1, teal closed circles; n = 4) clade together, samples from these two timepoints from this participant are in two different areas of the Urinae subtree. However, the majority of genomes from the same individual, regardless of isolation site or collection timepoint, clade together, indicating that these isolates share a common ancestor.

Fig. 1.

Phylogenetic tree based upon the single-copy core genome rooted between the Anginosus group and Urinae group. Genomes sequenced as part of this work are indicated with a colored line. Those that are from individuals and timepoints that are singletons (i.e., there are no genomes from other sample types) are indicated as “Unpaired” and with an open circle. Genomes from isolates of the same individual are indicated with the same color; those genomes from an individual with multiple collection timepoints are differentiated by different symbols (closed circle, triangle, cross). Participant id and timepoint are listed next to the symbol in the legend. Timepoints include “Baseline” (during enrollment in the study), “12-week” (12 weeks after enrollment), “EOT” (end of treatment), and “1” (first collection for the PEMI study). Nodes with arrows indicate genomes from the same individual and collection timepoint that do not clade together in the core tree

Is the same strain found in different anatomical sites?

To determine if two genomes from two different anatomical sites were representatives of the same strain, we first generated a distribution of ANI and mash distance values from simulated genomes representative of the two groups. Discussed in detail in the Methods, we randomly subsampled the raw reads from each of the 115 genomes listed in Additional file 1 to produce 50 simulated genomes. The simulated genomes from the same source genome were then compared using both the ANI and mash distance metric, capturing similarity between shared sequences and distance over the entire genome sequence, respectively. These values were used to create distributions (Fig. 2). Using these distributions, we then calculated the empirical p value to qualify our distinction between unrelated, related, and identical genomes from different anatomical sites.

Fig. 2.

Distribution of ANI and mash distance values from pairwise ANI and mash distance calculations of simulated genomes for Anginosus (A and C) and Urinae (B and D)

Empirical p values were computed using these distributions for the genomes from isolates of the same individual. This was done for both the ANI and mash distance metrics. All three Anginosus participant collections had p values > 0.05, and thus we failed to reject the null hypothesis that the strains are the same, based upon the ANI metric. Therefore, we posit that that these isolates are related if not representatives of the same strain. However, when considering the mash distance metric, two of the Anginosus participant collections, EST04 12-week and PEMIO31 timepoint 1, had empirical p values of 0, below the 0.05 threshold, suggesting that they are not the same strain. For the Urinae participant collections, four had samples with empirical p values for the ANI metric < 0.05 (Table 2). This includes the Urinae isolates from the catheterized urine samples from EST18 at Baseline (UMB3301 and UMB3306), which differ from the other Urinae isolates from this participant and collection timepoint, and two that were identified as different based upon their core genome (Fig. 1). These four Urinae participant collections also were found to have an empirical p value < 0.05 for the mash distance metric. Additionally, mash analysis of the PBSA06 at 12-week pair of samples had an empirical p value < 0.05 (Table 2). All ANI, mash distance, and their respective empirical p values can be found in Additional file 3.

Table 2.

Participant collections in which different Urinae strains were isolated from the urogenital tract

Participant	Collection Timepoint	Sample(s)	Empirical p (ANI)	Empirical p (mash distance)
EST18	Baseline	Catheterized Urine (x2)	0.00020	0
EST33	Baseline	Catheterized Urine, Vaginal Swab	0.00020	0
EST39	Baseline	Voided Urine, Perineal Swab	0.01896	0.03421-0.03500
EST49	Baseline	Catheterized Urine, Voided Urine, Vaginal Swab	0.00020–0.02760	0.00000-0.03490
PBSA06	12-week	Voided Urine, Vaginal Swab	n/a	0.03480

While the empirical p value enables us to identify unrelated strains, values greater than the 0.05 threshold include both related and identical strains. To tease apart these two groups, we independently extracted and sequenced 10 times a single culture of strain UMB4687B, a representative of the Urinae group. Pairwise comparisons of the complete genome assemblies were computed, again using both ANI and mash, and compared to the distributions previously computed (Fig. 2) to compute the empirical p. As Fig. 3 shows, the ANI values for the 10 replicates are in the densest area of the distribution for both metrics.

Fig. 3.

Distributions of simulated genomes from the10 sequences of S. anginosus UMB4687B. A ANI comparisons. B mash distance comparisons. Red lines in both panels indicate the pairwise ANI and mash values of the ten complete genome assembles, respectively

We used the results obtained from sequencing the same culture multiple times to inform our determination of a threshold to distinguish between related and identical strains found in multiple sampled sites containing Urinae isolates. Here, we chose to focus on Urinae as most of our isolates were representatives of this group (Table 1). The empirical p values for the ANI comparisons (Fig. 3A, red lines) and mash distance comparisons (Fig. 3B, red lines) ranged from 0.7614 to 1.0000 and 0.6937 to 0.9993, respectively (Additional file 4). To be the most conservative, we decided to require a pairing to have an empirical p value greater than or equal to the ranges of both the minimum ANI and the minimum mash distance – ANI empirical p ≥ 0.7614 and mash distance empirical p ≥ 0.6937. Of the 29 participant collections in which Urinae was isolated from more than one sampled site, 11 contained the same strain across sites (Table 3). It should be noted that for several of these participant collections, additional anatomical sites were sampled, and isolates were sequenced; however, these additional sites failed to meet our chosen threshold. For instance, EST12 at Baseline included samples from catheterized urine, perineal swabs, and vaginal swabs. While all three sites met the ANI empirical p value threshold, the pairwise comparisons of the catheterized urine genome and the genomes of the two other sites did not meet the mash distance empirical p value threshold. Thus, we hypothesize that the perineal swab and vaginal swab genomes represent an identical strain whereas the genome from the catheterized urine sample represents a related strain.

Table 3.

Participant collections with the same Urinae strain identified in more than one of the sampled sites

Participant	Collection Timepoint	Samples	Average ANI	Average Mash distance
EST12	Baseline	Perineal Swab, Vaginal Swab	99.9982	5.54 × 10−5
EST14	12-week	Voided Urine, Perineal Swab, Vaginal Swab	99.9995	6.80 × 10−6
EST14	Baseline	Voided Urine, Vaginal Swab	99.9991	1.38 × 10−5
EST15	12-week	Catheterized Urine, Voided Urine, Perineal Swab, Vaginal Swab	99.9984	1.51 × 10−5
EST18	Baseline	Voided Urine, Perineal Swab	99.9992	7.50 × 10−6
EST19	Baseline	Catheterized Urine, Voided Urine, Perineal Swab, Vaginal Swab	99.9974	1.64 × 10−5
PBSA07	EOT	Voided Urine, Vaginal Swab	99.9988	1.94 × 10−5
PBSA16	EOT	Voided Urine, Vaginal Swab	99.9990	2.54 × 10−5
PEMI13	Baseline	Catheterized Urine, Perineal Swab	99.9984	1.71 × 10−5
PEMISF3	1	Perineal Swab, Vaginal Swab	99.9991	4.41 × 10−6
PEMISF9	Baseline	Catheterized Urine, Perineal Swab	99.9980	4.41 × 10−6

Discussion

Our previous analysis suggested that the Urinae group of S. anginosus had a tropism for the urinary tract, as the majority of strains within this group were from urine isolates [5]. However, very few strains at that time were available from other urogenital sites (vaginal = 3). In addition to sequencing 61 S. anginosus isolates from urine samples, our current study includes isolates from urethral, perineal, vaginal, and foreskin swabs. The overwhelming majority (89.97%) of these isolates are members of the Urinae group (Table 1). There is an inherent bias in our samples though as many individuals were sampled both multiple times and at multiple locations; however, we did not identify any individuals who carried both Urinae and Anginosus.

The isolation of Urinae from multiple urogenital sites suggests that Urinae is a urogenital subgroup of S. anginosus. While this suggests that the Urinae group has adapted specifically to this environment, future studies of the gut and oral microbiome, especially of the individuals who have Urinae in their urinary tract or vaginal microbiome, is needed. Our prior comparative analysis of S. anginosus genomes did include ten strains isolated from the gut/fecal samples within the Urinae group [5]; however, gut/fecal samples were more commonly associated with the Anginosus group. Thus, while Urinae appears to have a tropism for the urogenital tract, it is possible that members of the Urinae group can colonize the other anatomical sites from which the Anginosus group strains have been isolated.

Being able to proliferate in the vaginal and urinary environment requires a level of flexibility characteristic of species in the Streptococcus genus, which consists of many species that have high competency and mutability like S. pneumoniae or S. pyogenes [32–34]. The ability of this genus to persist in the human microbiota displays a long history of commensality, which has recently been challenged by the increased association of S. anginosus with disease, most notably its association with cystic fibrosis [35–37]. Specific to the urogenital tract, it is prevalent and abundant among females with anaerobic vaginitis (AV) [38] and UUI symptoms [6, 7], but it also is one of the most frequently detected species in urine samples of asymptomatic presumably healthy individuals, albeit at considerably reduced abundance [8]. The identification of the Urinae group requires further investigation into whether unsequenced disease/symptom-associated urogenital isolates are members of the Anginosus group, or whether they belong to Urinae. AV is associated with a less acidic environment [39] and S. anginosus has been shown to be aciduric [40] as, along with other streptococci, it is a producer of l(+)-lactic acid [41]. Thus, it would be interesting to determine if AV-associated S. anginosus are members of the Anginosus or Urinae group and if their abundance relates to the vaginal environment of AV.

The connectivity of the urinary tract and vaginal microbiota is also explored in this study. Prior studies have identified some species that can inhabit both communities [10–12]. Most recently, we found evidence of identical strains of Lactobacillus species from urinary and vaginal samples collected from the same individual [13]. Here, we can expand this “connection” to also include the community of the perineum (Table 3). Prior work has suggested exchanges between the urinary, vaginal, and rectal microbial communities; after all, the prevailing hypothesis about UTI is that it is caused by a recent acquisition of bacteria from the gut microbiome [42, 43]. As listed in Table 3, we see several instances of the same Urinae strain found at multiple sampled sites for participants in the vaginal estrogen treatment study (participant ID starts with “EST”), in which females with OAB symptoms were treated with estrogen cream, applied twice weekly for 12 weeks [15]. Because the same strain at multiple sites also was identified within other study cohorts, including asymptomatic individuals (participant ID starts with “PEMI” in Table 3), we cannot attribute the presence of the same strain at multiple anatomical sites as the result of the treatment regimen. The data proves that a highly similar bacterium can reside in both the vaginal and urinary tracts. Thus, the relationship of the biomes is certain: they are interconnected, at least for some species.

Here, we expanded our prior statistical method for distinguishing between unrelated, related, and identical strains, which relied entirely on ANI measurements [13], as we also quantified the difference between genomes via the mash distance. Since mash compares the entire genome and not just shared regions, it would be more sensitive to recently acquired DNA via horizontal gene transfer, which may have a phenotypic effect. Furthermore, our empirical tests using UMB4687B, a representative of the Urinae group, enabled us to differentiate between closely related and identical strains (Fig. 3), resulting in the identification of 11 instances in which the same strain inhabits more than one anatomical site of the individual. This suggests that the strains can travel from one anatomical site to another. Although streptococci have no appendage dependent motility, there has been one study that observed a gliding-type motility for four S. anginosus isolates from urogenital samples [44]. Interestingly, the only isolate that did not exhibit this gliding-type motility (later determined to be “sliding” rather than gliding [45]) was the one non-urogenital isolate. While we can speculate that this motility is present in the Urinae group, thus enabling it to colonize multiple urogenital sites, further testing is necessary.

What is the same strain is a question that is as philosophical as it is scientific. ISJME defines a strain as follows: “A strain is made up of the descendants of a single isolation in pure culture. A strain is usually made up of a succession of cultures and is often derived from a single colony.” [46]. In fact, this definition comes from the first edition of Bergey’s Manual of Systematic Bacteriology published in 1984, a decade before we had the first genome of a cellular organism. Experimental evolution has, however, enabled us to see the evolution of different strains from a single colony source (e.g [47, 48]). Here we have theoretically tested this boundary of strain, proposing that values that fall within the distribution of our simulated subsampling data are the same strain or likely closely related with empirical tests enabling us to distinguish between the two.

Prior comparative analyses of bacterial, and more recently viral, genomes and metagenomes have found a natural gap in ANI values between genomes of the same species [49, 50]. In bacteria, this gap occurs between ANI values of 99.2% and 99.8%, with a value of > 99.99% proposed as the barrier between strains [50]. Our method identifies a slightly higher threshold for the Urinae genomes (Table 3). Even the authors who proposed the 99.99% threshold note that this value may vary for a species of interest [50]. In our analyses none of the comparisons identified two strains as being identical, either an ANI of 100% or a mash distance of 0. It is important to remember that these species are evolving in their environments influenced by biotic and abiotic factors. In our own prior work that considered urogenital samples of lactobacilli from the same female, we identified the same strain in two samples collected 12 weeks apart; SNPs were observed over the course of the collection period [13]. While gene acquisition or loss will not affect the ANI metric, as it only quantifies similarity between shared regions, it will impact the mash distance. Thus, in considering both of these metrics, we can consider both strain evolution via mutation as well as horizontal gene transfer.

As our prior analysis and our core genome analysis here shows, the S. anginosus species includes multiple clades, including clades associated with named subspecies (anginosus and whileyi) and genomosubspecies (AJ1 and vellorensis). The Urinae group also includes several clades (Fig. 1), although the branches are not as deep as those that distinguish the subspecies and genomosubspecies of the S. anginosus group Anginosus branch. The presence of S. anginosus in diverse biomes must be supported with biodiversity within the species. Based upon the isolation of Urinae to date, members of this group suggest a tropism for the urogenital environment. Further research should investigate the presence of Urinae outside of the urogenital tract.

Conclusions

While we previously found members of the S. anginosus group Urinae exhibited a tropism for the urinary tract, here we find that this group can be found throughout the female urogenital tract. Furthermore, when the same group (Urinae or Anginosus) are found in both the vaginal and urinary tracts, they are often very similar to one another. In many instances they are so similar that they may be assumed to be members of the same strain. The multiple occurrences of the same strain found at multiple anatomical sites within the same individual strongly suggest that the urinary tract and vaginal microbiomes are interconnected. The improved statistical method presented here provides a robust means for strain identification. This paper builds upon the current hypothesis of niche connectivity and encourages further research to consider them one urogenital microbiome.

Authors’ contributions

H.A. performed experiments and analysis, contributed to writing the first draft, and prepared Fig. 1. J.G.-I. performed experiments and analysis. L.B. and S.B. performed statistical analyses and prepared Figs. 2 and 3. A.J.W. and C.P. conceived of the project. A.J.W. contributed samples and provided feedback throughout the experiments. C.P. contributed to writing the first draft and prepared Fig. 1. All authors reviewed the final manuscript.

Funding

This work was supported by the NIDDK-funded R25 program at Loyola University Chicago (NIH R25DK122954; PI Putonti). HA was supported by a Mulcahy fellowship from Loyola University Chicago.

Data availability

All sequence data generated as part of this study is publicly available. This includes both raw reads and assembled genomes. Additional File 1 lists the SRA and Assembly accession numbers for the sequenced genome isolates. All code used for genome comparisons is available through GitHub (https://github.com/putonti/LactoCompare). Raw reads for the ten extractions and sequencing runs for S. anginosus UMB4687B have been deposited in BioProject PRJNA1150124.

Declarations

Ethics approval and consent to participate

All isolates sequenced as part of this project were from prior IRB-approved studies (IRB nos., Loyola University Chicago – 204133, 204195, 206129, 206469, 207102, 207152, 209983, and 210444; University of California, San Francisco – 15-16363; University of Iowa – 201608793). Samples were obtained with informed consent from participants.

Consent for publication

Not applicable.

Competing interests

A.J.W. discloses membership on the scientific advisory boards of Cerillo, Pathnostics, and Urobiome Therapeutics, as well as funding from Pathnostics. The other authors have no disclosures.