A Vaginal Tract Signal Detected by the Group B Streptococcus SaeRS System Elicits Transcriptomic Changes and Enhances Murine Colonization

ABSTRACT

Streptococcus agalactiae (group B streptococcus [GBS]) can colonize the human vaginal tract, leading to both superficial and serious infections in adults and neonates. To study bacterial colonization of the reproductive tract in a mammalian system, we employed a murine vaginal carriage model. Using transcriptome sequencing (RNA-Seq), the transcriptome of GBS growing in vivo during vaginal carriage was determined. Over one-quarter of the genes in GBS were found to be differentially regulated during in vivo colonization compared to laboratory cultures. A two-component system (TCS) homologous to the staphylococcal virulence regulator SaeRS was identified as being upregulated in vivo. One of the SaeRS targets, pbsP, a proposed GBS vaccine candidate, is shown to be important for colonization of the vaginal tract. A component of vaginal lavage fluid acts as a signal to turn on pbsP expression via SaeRS. These data demonstrate the ability to quantify RNA expression directly from the murine vaginal tract and identify novel genes involved in vaginal colonization by GBS. They also provide more information about the regulation of an important virulence and colonization factor of GBS, pbsP, by the TCS SaeRS.

INTRODUCTION

Streptococcus agalactiae (group B streptococcus [GBS]) is an important human pathogen most known for its ability to cause deadly neonatal infections. The primary risk factor to newborns is maternal colonization with GBS in the genitourinary tract (1). In 2010, the CDC revised guidelines to prevent these infections, calling for universal screening of all pregnant women between 35 and 37 weeks of gestation (2). Treatment of GBS in pregnant females involves intrapartum intravenous antibiotics, which has decreased the incidence of neonatal GBS sepsis but has not affected rates of late-onset disease in newborns over 1 week old (3). Maternal intrapartum prophylactic antibiotics have also been shown to have deleterious effects on the intestinal flora of newborns, including a decrease in the frequency of beneficial bifidobacterial species (4). A GBS vaccine has been proposed as a more advantageous strategy to prevent maternal colonization rather than treating infection once it is detected. Development of a GBS vaccine or other anticolonization strategies requires a more thorough understanding of the genetic profiles of GBS during vaginal carriage.

The vaginal environment is made up of a complex and dynamic microbial community. Environmental stressors on GBS colonizing the vaginal tract include changes during the menstrual cycle in pH, the normal colonizing flora, and host innate immune factors, such as interleukin-17 (IL-17) (5–7). The molecular components necessary for vaginal colonization by GBS have been the subject of study in recent years, with individual adhesins such as serine-rich-repeat proteins and regulators being associated with increased vaginal carriage in murine models (5, 8–10). Here we determine, for the first time, the full transcriptional profile of GBS strain A909 during murine vaginal colonization compared to laboratory culture conditions. Transcriptome sequencing (RNA-Seq) studies described here show that numerous global changes occur in bacterial transcription during vaginal colonization, including widespread metabolic shifts, differential expression of numerous transcriptional regulators, and the upregulation of many putative adherence factors. These data will be invaluable for future studies examining GBS colonization factors as well as indicating potential vaccine targets and therapeutics aimed at preventing GBS vaginal colonization in women.

Our findings show that a two-component system (TCS) homologous to the SaeRS virulence-associated TCS in Staphylococcus aureus was highly upregulated in GBS during vaginal colonization. As a canonical TCS in S. aureus, SaeS serves as a sensor histidine kinase, modulating the phosphorylation state of its cognate response regulator, SaeR, in response to environmental signals. Phosphorylated SaeR binds DNA and acts as a direct transcriptional regulator of a wide variety of virulence factors in S. aureus, including alpha-hemolysin and coagulase, among other important exoproteins, and the role of the SaeRS TCS in S. aureus virulence has been demonstrated by numerous studies (11–15). Because genes of the S. aureus SaeRS regulon have no apparent homologs in GBS, the role that SaeRS plays in gene regulation and the signal that it senses in GBS is unknown. Here, we identify genes controlled by the SaeRS TCS system during growth in a murine model of vaginal colonization and demonstrate that at least one of these genes is an important factor in GBS colonization or survival in the vaginal tract. Finally, we show that a signal present in vaginal lavage (VL) fluid from mice is sufficient to induce SaeRS-dependent gene expression.

RESULTS

Transcriptomic analysis of GBS during vaginal colonization signals a shift in genetic programming.

The pathogen GBS is most recognized as a vaginal colonizer, so we used RNA-Seq to measure genome-wide mRNA levels during growth in a murine vaginal colonization model and compared them to those occurring during growth under laboratory culturing conditions. GBS cultures grown statically at 37°C in a chemically defined medium (CDM) were compared to bacteria collected from the vaginal tract 48 h following initial inoculation. Approximately one-third of the entire genome of strain A909, 731 genes, were identified as being differentially expressed, with a false-discovery rate (q value) of <0.05, in the vaginal tract compared to genes in log-phase growth in CDM. Of these, 630 were differentially regulated >2-fold (Fig. 1A; see also Table S1 in the supplemental material). Pathway enrichment against KEGG metabolic pathways was performed using Kyoto Encyclopedia of Genes and Genomes (KEGG) orthologous (KO) group assignments; metabolic pathways that are differentially expressed during vaginal growth are shown in Fig. 1B.

FIG 1.

Transcriptional changes during in vivo growth. (A) Volcano plot of transcriptomic changes in liquid versus mouse. Log₂ changes in gene expression were plotted versus the −log₁₀ false discovery rate (FDR). Genes shown in red have an FDR of <0.05 and a fold change of >2. Three genes discussed in the text are labeled with gene numbers. (B) Pathways statistically overrepresented during growth in vivo.

Phosphotransferase systems (PTSs) are almost universally highly upregulated during in vivo colonization (Table 1 and Fig. 1B), with some systems being upregulated >500-fold. PTSs are one important method that bacteria use to sense carbon source availability and transport nutrients into the cell.

TABLE 1.

Gene categories differentially expressed during vaginal colonization

Gene category and no. (strain A909)	Gene name	Descriptiona	Fold change in vaginal colonization
Phosphotransferase systems
sak_0257		Trehalose-specific IIBCA component	−2.700
sak_0261		Sugar-specific IIC component	5.300
sak_0354		IIBC component	−20.520
sak_0398		IIA component, lactose/cellobiose family	31.250
sak_0399		IIB component, lactose/cellobiose family	18.700
sak_0400		IIC component, lactose/cellobiose family	26.070
sak_0524		Galactitol-specific IIA component	31.890
sak_0525		Galactitol-specific IIB component	53.170
sak_0526		Glucose-specific IIABC component	100.730
sak_0528		Galactitol-specific IIA component	52.970
sak_0529		Galactitol-specific IIC component	68.580
sak_0530		Galactitol-specific IIB component	140.590
sak_1377		Fructose-specific IIABC component	7.730
sak_1702		Sucrose-specific IIABC component	8.400
sak_1759		Fructose-specific, IIC component	5.800
sak_1825		Galactitol-specific IIC component	23.700
sak_1833		IIA component	684.810
sak_1834		IIB component, lactose/cellobiose family	948.230
sak_1835		Sugar-specific IIC component	88.170
sak_1893		IIC component	84.130
sak_1894		IIB component	76.440
sak_1895		IIA component	65.180
sak_1908		IID component, mannose/fructose/sorbose family	209.060
sak_1909		IIC component, mannose/fructose/sorbose family	123.500
sak_1910		IIB component, mannose/fructose/sorbose family	88.780
sak_1911		IIA component, mannose/fructose/sorbose family	26.670
sak_1920		Glucose-specific IIABC component	16.460
ABC transporters
sak_0166	rbsB	Ribose ABC transporter, ribose-binding protein	102.631
sak_0167	rbsC	Ribose ABC transporter, permease protein	69.165
sak_0168	rbsA	Ribose ABC transporter, ATP-binding protein	61.861
sak_0169	rbsD	Ribose ABC transporter protein RbsD	31.775
sak_0207	oppD	Oligopeptide ABC transporter, permease protein	−2.191
sak_0208		Oligopeptide ABC transporter, permease protein OppC	−2.085
sak_0209	oppD	Oligopeptide ABC transporter, ATP-binding protein	−2.245
sak_0210	oppF	Oligopeptide ABC transporter, ATP-binding protein	−2.498
sak_0302		QAT family ABC transporter, permease	−2.163
sak_0303		QAT family ABC transporter, ATP-binding	−2.255
sak_0472		BioY family protein	4.521
sak_0561		ABC transporter, ATP-binding/permease protein	−4.218
sak_0562		ABC transporter, ATP-binding/permease protein	−4.435
sak_0795	cylA	ABC transporter, ATP-binding protein CylA	2.903
sak_0796	cylB	ABC transporter, permease protein CylB	3.487
sak_0899		Amino acid ABC transporter, ATP-binding protein, putative	−5.711
sak_0900		Amino acid ABC transporter, permease protein, putative	−2.611
sak_0901		ABC transporter, substrate-binding protein	−2.416
sak_1074		ABC transporter, substrate-binding protein	−8.399
sak_1103		Iron chelate uptake ABC transporter, ATP-binding protein	3.031
sak_1104		Iron chelate uptake ABC transporter, permease protein	3.352
sak_1105		Iron chelate uptake ABC transporter, permease protein	2.848
sak_1426	fhuD	Ferrichrome ABC transporter, ferrichrome-binding protein	3.423
sak_1427	fhuB	Ferrichrome ABC transporter, permease protein	4.479
sak_1476		Cyclodextrin ABC transporter, permease protein	4.363
sak_1477		Cyclodextrin ABC transporter, permease protein	4.452
sak_1538	nikE	Nickel ABC transporter, ATP-binding protein	38.719
sak_1539	nikD	Nickel ABC transporter, ATP-binding protein	44.320
sak_1540	nikC	Nickel ABC transporter, permease protein	68.911
sak_1541	nikB	Nickel ABC transporter, permease protein	78.738
sak_1542	nikA	Nickel ABC transporter, nickel-binding protein	53.693
sak_1554	mtsC	Metal ABC transporter, permease protein	47.051
sak_1555	mtsB	Metal ABC transporter, ATP-binding protein	52.891
sak_1556	mtsA	Metal ABC transporter, metal-binding lipoprotein	36.614
sak_1594	livF	HAAT family ABC transporter, ATP-binding protein	−8.380
sak_1595	livG	HAAT family ABC transporter, ATP-binding protein	−7.300
sak_1596	livM	HAAT family ABC transporter, permease	−10.973
sak_1597	livH	HAAT family ABC transporter, permease	−10.450
sak_1646		HAAT family ABC transporter, permease	−2.663
sak_1647		ABC transporter, ATP-binding protein	−2.131
sak_1747	cydC	ABC transporter, ATP-binding protein CydC	−2.267
sak_1884	msmK	Sugar ABC transporter, ATP-binding protein	4.983
sak_1925	pstA	Phosphate ABC transporter, permease protein PstA	18.213
sak_1927		Phosphate ABC transporter, phosphate-binding protein	10.791
sak_1966		ABC transporter, ATP-binding protein	−3.675
sak_2068		Glycine betaine/carnitine/choline ABC transporter, ATP-binding protein	−2.434
sak_2069		Glycine betaine/carnitine/choline ABC transporter, permease	−2.196
Starch and sucrose metabolism
sak_1472	glgP	Glycogen/starch/alpha-glucan phosphorylase	4.686
sak_0976	glgB	1,4-Alpha-glucan branching enzyme	3.913
sak_0979	glgA	Glycogen synthase	5.437
sak_1473	malQ	4-Alpha-glucanotransferase	4.785
sak_0977	glgC	Glucose-1-phosphate adenylyltransferase	5.077
sak_0978	glgD	Glucose-1-phosphate adenylyltransferase, GlgD subunit	5.519
sak_1703	scrB	Sucrose-6-phosphate hydrolase	3.684
sak_1188		Glycosyl hydrolase, family 1	−2.833
sak_0258		Alpha amylase family protein	−2.605
sak_1155		Phosphoglucomutase/phosphomannomutase family protein	3.877
sak_0398		PTS system, IIA component, lactose/cellobiose family	31.254
sak_0399		PTS system, IIB component, lactose/cellobiose family	18.697
sak_0400		PTS system, IIC component, lactose/cellobiose family	26.073
sak_1920		PTS system, glucose-specific IIABC component, putative	16.457
sak_1702		PTS system, sucrose-specific IIABC component	8.405
sak_0354		PTS system, IIBC component	−20.528
sak_0257		PTS system, trehalose-specific IIBCA component	−2.697
Fatty acid biosynthesis and metabolism
sak_0416	fabM	Enoyl-CoA hydratase/isomerase family protein	−20.503
sak_0417	fabT	MarR transcriptional regulator	−11.350
sak_0418	fabH	3-Oxoacyl-(acyl-carrier-protein) synthase III	−11.029
sak_0419	acpP	Acyl carrier protein	−4.304
sak_0420	fabK	Enoyl-(acyl-carrier-protein) reductase II	−19.083
sak_0421	fabD	Malonyl-CoA-acyl carrier protein transacylase	−15.727
sak_0422	fabG	3-Oxoacyl-(acyl-carrier-protein) reductase	−21.685
sak_0423	fabF	3-Oxoacyl-(acyl-carrier-protein) synthase II	−10.549
sak_0424	accB	Acetyl-CoA carboxylase, biotin carboxyl carrier protein	−10.054
sak_0425	fabZ	Beta-hydroxyacyl-(acyl-carrier-protein) dehydratase fabz	−12.538
sak_0426	accC	Acetyl-CoA carboxylase, biotin carboxylase	−8.781
sak_0427	accD	Acetyl-CoA carboxylase, carboxyl transferase, beta subunit	−10.492
sak_0428	accA	Acetyl-CoA carboxylase, carboxyl transferase, alpha subunit	−8.249
Fructose and mannose metabolism
sak_1036	pfk	6-Phosphofructokinase	−2.172
sak_1888	lacC	Tagatose-6-phosphate kinase	102.565
sak_0537	galT	Galactose-1-phosphate uridylyltransferase	30.080
sak_1703	scrB	Sucrose-6-phosphate hydrolase	3.680
sak_1887	lacD	Tagatose 1,6-diphosphate aldolase	104.990
sak_0538	galE	UDP-glucose 4-epimerase	13.700
sak_1889	lacB	Galactose-6-phosphate isomerase, LacB subunit	55.168
sak_1890	lacA	Galactose-6-phosphate isomerase, LacA subunit	123.335
sak_1155		Phosphoglucomutase/phosphomannomutase family protein	3.880
sak_0524		PTS system, galactitol-specific IIA component, putative	31.886
sak_0528		PTS system, galactitol-specific IIA component, putative	52.973
sak_0525		PTS system, galactitol-specific IIB component, putative	53.173
sak_0530		PTS system, galactitol-specific IIB component, putative	140.593
sak_0526		PTS system, galactitol-specific IIC component, putative	100.735
sak_0529		PTS system, galactitol-specific IIC component	68.575
sak_1825		PTS system, IIC component, putative	23.697
Oxidative phosphorylation
sak_1750	cydA	Cytochrome d ubiquinol oxidase, subunit II	−3.030
sak_1749	cydB	Cytochrome d ubiquinol oxidase, subunit II	−2.081
sak_1036	pfk	6-Phosphofructokinase	−2.170
sak_1378		1-Phosphofructokinase, putative	6.540
sak_0527	rhaD	Rhamnulose-1-phosphate aldolase	83.546
sak_0594		Phosphoglucomutase/phosphomannomutase family protein	21.790
sak_0981	atpB	ATP synthase F0, A subunit	−3.005
sak_0982	atpF	ATP synthase F0, B subunit	−2.699
sak_0980	atpE	ATP synthase F0, C subunit	−2.417
sak_0983	atpH	ATP synthase F1, delta subunit	−2.672
sak_0985	atpG	ATP synthase F1, gamma subunit	−2.254
sak_1377		PTS system, fructose-specific IIABC component	7.730
sak_1759		PTS system, fructose-specific IIC component	5.800
sak_1911		PTS system, IIA component, mannose/fructose/sorbose family	26.668
sak_1910		PTS system, IIB component, mannose/fructose/sorbose family	88.776
sak_1909		PTS system, IIC component, mannose/fructose/sorbose family	123.500
sak_1908		PTS system, IID component, mannose/fructose/sorbose family	209.064
sak_1751		Pyridine nucleotide-disulfide oxidoreductase family protein	−3.130
sak_0666	fbp	Fructose-1,6-bisphosphatase	3.520
Two-component systems
sak_0467	saeR	DNA-binding response regulator	8.469
sak_0468	saeS	Sensor histidine kinase	6.033
sak_1880		Sensor histidine kinase, putative	3.625
sak_1881		Response regulator	4.978
sak_1917	rgfC	Histidine kinase	3.879
sak_1921		Sensor histidine kinase	2.575
sak_2061		DNA-binding response regulator	−3.831
sak_2062		Sensor histidine kinase	−2.506
sak_2066		Sensor histidine kinase, putative	53.384

^aHAAT, hydrophobic amino acid uptake; CoA, coenzyme A; QAT, quaternary amine uptake.

The fab operon, which regulates fatty acid biosynthesis, was highly downregulated during vaginal colonization (Table 1 and Fig. 1B). Lowered expression of genes in this operon is thought to increase the amount of long-chain unsaturated fatty acids, which can be protective during growth at a low pH (16, 17). As the human vaginal tract is normally at pH ranges between 3.8 and 4.5, an overall low expression of the fab operon may provide GBS a competitive advantage in the acidic environment.

Differential regulation of two-component systems during growth in vivo.

A recent study defined 21 two-component systems highly conserved in GBS (18). A909 contains 20 of these TCSs, and in our analysis, three complete TCSs were differentially expressed between conditions of lab culturing and host colonization (Table 1), with an additional three histidine kinases being differentially expressed without differential expression of the cognate response regulator. All but one of these differentially expressed TCSs were upregulated during in vivo growth.

One of the most highly upregulated TCSs is SAK_0467/0468. This TCS is homologous to the well-characterized SaeR (48% identical) and SaeS (34% identical) TCSs found in Staphylococcus aureus. In S. aureus, SaeRS regulates numerous virulence factors, such as alpha-hemolysin, lipase protein A, and adhesins, among others (13, 15, 19). The organization of the TCS locus in S. aureus is different from a canonical TCS in that it contains two additional genes (saeP and saeQ) immediately upstream and overlapping with the saeR gene (Fig. 2A). In S. aureus, saeP and saeQ are required to activate SaeS phosphatase activity, allowing the removal of the activating phosphate on SaeR, effectively recycling it and turning off SaeR activity (20). Alignment of the SaeR gene sequences in GBS (Streptococcus agalactiae) and S. aureus is shown in Fig. 2B, with important residues highlighted. In the GBS homologous system, saeP and saeQ are not present, and in their place is a large gene, sak_0466, encoding a putative cell wall surface anchor family protein recently named plasminogen binding surface protein (PbsP). PbsP was shown to be important for both plasminogen binding and dissemination of GBS during invasive disease in a different strain of GBS, NEM316 (21).

FIG 2.

saeRS loci and SaeR protein sequences in Streptococcus agalactiae and Staphylococcus aureus. (A) saeRS loci of S. aureus and S. agalactiae, including the surrounding region. (B) SaeR protein sequence alignment between S. agalactiae and S. aureus. Residues marked in yellow were shown to be important for DNA binding, and the residue marked in blue is the predicted phosphorylated aspartic acid residue for the SaeR protein in S. aureus (44).

The genetic pathways controlled by SaeRS are different in S. aureus and in GBS, as the S. aureus regulon, which is well documented, comprises genes that have no homologs in GBS. Because SaeRS is upregulated during vaginal growth and is likely to regulate a distinct set of genes, we examined transcriptomic changes between A909 and an isogenic ΔsaeR strain during log-phase growth in two different laboratory media (chemically defined medium [CDM] and Todd-Hewitt broth supplemented with yeast extract [THY]) and during colonization of the vaginal tract.

The SaeR-regulated transcriptome differs greatly between in vitro and in vivo conditions.

Transcriptomic data indicated that SaeRS regulates, either directly or indirectly, a large portion of the genome of GBS under certain conditions. A q value of 0.05 and a cutoff of ≥2-fold change were the criteria that we used to define differentially expressed genes (DEGs). Compared to the wild-type (WT) strain, during log-phase growth, 301 genes were differentially expressed in the ΔsaeR strain in CDM and 466 genes were differentially expressed in THY. Of those that were different in CDM, 136 genes (∼45%) were also changed in THY, though a small number were regulated in opposite directions. The results of these experiments are summarized in Fig. 3A and B, and detailed transcriptomic information can be found in Table S4 in the supplemental material. Although many DEGs were observed in liquid culture, the magnitudes of differences were generally low, with only 6/301 genes in CDM and 29/466 genes in THY showing differences of >10-fold. The clear majority of DEGs were seen to differ only 2- to 5-fold, which may indicate an indirect mechanism of regulation. Despite the large number of genes differentially regulated by SaeRS in liquid culture, only three genes were differentially expressed between the wild-type and the ΔsaeR strains during growth in the vaginal tract (Fig. 3C). In the ΔsaeR mutant, levels of sak_1753 mRNA were more than 800 times lower than for the wild type during vaginal colonization and levels of pbsP transcript were more than 200-fold lower. The third gene differentially regulated between wild-type and ΔsaeR strains was a predicted small open reading frame (ORF), sak_RS00910, which is located between sak_0183 and sak_0184. This ORF encodes a putative peptide of 38 amino acids with no predicted function or homolog. The sak_RS00910 transcript was downregulated a more modest 7.7-fold in the ΔsaeR mutant. Two of these three DEGs, sak_0466 (pbsP) and sak_1753, were among the top 10 most highly upregulated genes seen in the wild type during growth in the vaginal tract compared to lab culture (marked in Fig. 1A). pbsP was upregulated over 250-fold during vaginal colonization compared to growth in CDM, and sak_1753 was the most highly differentially expressed gene, upregulated over 3,000-fold in vivo (see Table S3 in the supplemental material). This evidence indicates that SaeR is responsive to conditions in the vaginal tract and that targets of SaeR regulation could play an important role during growth in vivo.

FIG 3.

The SaeR regulon. (A) Scatter plot shows differentially expressed genes (FDR < 0.05) between the WT and the isogenic ΔsaeR mutant as the log₂-fold change in chemically defined medium (CDM) versus rich medium (THY). (B) Venn diagram showing differentially expressed gene overlap in CDM versus THY. (C) Scatter plot showing gene expression in WT A909 versus the isogenic ΔsaeR mutant. The four genes with an FDR P value of <0.05 are marked in red.

Role of PbsP in vaginal colonization.

PbsP, a documented surface protein containing a sortase-dependent LPXTG motif, was shown to be important in respiratory tract colonization and invasive disease models of infection for GBS and Streptococcus pneumoniae (21–23). Considering that we found pbsP transcript levels to be substantially increased during growth of GBS in the vaginal tract, we asked whether removal of the gene would affect its ability to colonize the vaginal tract. We employed a murine model of vaginal carriage to look at colonization levels. In this model, mice are injected with estrogen to synchronize the estrous cycle 1 day prior to vaginal inoculation with bacteria. Colonization is subsequently measured over time as CFU in vaginal washes. We generated an in-frame deletion mutant, ΔpbsP, and compared its ability with that of the wild type to colonize and remain within the murine vaginal tract over the course of several days (Fig. 4). Initial colonization levels 24 h after inoculation were equivalent in WT A909 and the ΔpbsP mutant, but the numbers of viable ΔpbsP bacteria recovered on day 2 postinoculation were significantly decreased compared to the WT, indicating a colonization defect in the ΔpbsP mutant (Fig. 4). By later time points, many mice were no longer colonized and the trending decrease in ΔpbsP mutant colonies was present but no longer significant. Addition of a plasmid containing a copy of PbsP under a constitutive PrecA promoter complemented the colonization defect.

FIG 4.

Role of pbsP in vaginal colonization. Colonization levels in CFU/vaginal wash of WT A909 versus the isogenic ΔpbsP mutant on days 1, 2, 3, and 5. Statistical significance was assessed using a nonparametric Mann-Whitney test on raw values. *, P < 0.05.

Regulation of pbsP and sak_1753 by SaeRS via a signal present in vaginal lavage fluid.

The signal recognized by SaeS in GBS is not known, but we hypothesized that a component of the vaginal tract, which may comprise a soluble or cell-associated factor not found in laboratory media, may be sensed by SaeS and relayed to SaeR to regulate expression of pbsP and sak_1753. To determine whether such a signal was present in material of the vaginal tract, VL fluids from several mice were collected and pooled. For these experiments, mice were not synchronized for estrus, so lavage fluids from 10 mice were pooled to decrease estrous cycle variability. Wild-type or ΔsaeR cells were then suspended in either VL or phosphate-buffered saline (PBS). Following a 1-h incubation, transcript levels of pbsP and sak_1753 were assessed by quantitative real-time (RT)-PCR. In WT cells, incubation with VL caused increased transcription of both pbsP and sak_1753, approximately 20- to 30-fold over that of levels seen in PBS, corroborating the initial RNA-Seq results described above (Fig. 5A). In ΔsaeR cells, an increase in transcript amounts for either of these two genes following incubation with VL was not observed (Fig. 5A). These data indicate that the SaeRS system senses a component of VL fluid and this signal promotes upregulation of pbsP and sak_1753 via the response regulator SaeR.

FIG 5.

Vaginal lavage fluid induces increased gene expression of pbsP and sak_1753 via SaeR. (A) qPCR transcript levels of the indicated genes following incubation of WT A909 or ΔsaeR mutant cells with vaginal lavage fluid at 37°C for 60 min. (B) Vaginal lavage fluid was treated as indicated with heat, pronase, or filtration prior to incubation with A909. qPCR transcript levels are shown for the indicated genes. All transcript levels were compared to those of a housekeeping gene, gyrA, and are shown as a ratio (to PBS-treated control levels). Student's t test was used to analyzed statistical significance *, P < 0.05; **, P < 0.005. (C) EMSA gel showing increased shift of the pbsP promoter DNA probed with increased concentrations of SaeR D53E protein. This shift is no longer visible after addition of 10-fold excess unlabeled target DNA but not with 10-fold excess unlabeled random DNA.

To help determine the nature of the signal sensed by SaeRS, vaginal lavage fluid samples were treated by heating at 100°C for 30 min, were incubated at 37°C for 30 min with 1 mg/ml pronase enzyme, or were passed through a 3-kDa filter. Following treatments, VL samples were tested for the ability to upregulate expression of pbsP and sak_1753. Lavage fluid treated with heat or pronase no longer retained signaling ability, but filtrates passing through a 3-kDa molecular mass cutoff remained able to induce increased gene expression, although to a lower level than untreated VL, approximately 4- to 8-fold above levels seen in PBS-treated cells (Fig. 5B). These findings indicated that the signal is likely to comprise a small, heat-labile peptide of <30 residues.

To determine whether SaeR regulation of pbsP was direct, an electrophoretic mobility shift assay (EMSA) was performed. Because previous studies have demonstrated that SaeR must be phosphorylated in order to bind DNA (24), we attempted to phosphorylate SaeR in vitro using acetyl phosphate. Although we did not observe a shift when using this treatment, we were unable to confirm phosphorylation of SaeR. As it has been shown for several response regulators that replacement of the phosphorylated aspartic acid residue with a glutamate, which provides an extended side chain length, can often result in a constitutively active protein (25), we expressed and purified the SaeR Asp-to-Glu (D53E) variant. Direct binding of the pbsP promoter region was observed via an EMSA (Fig. 5C).

DISCUSSION

To persist on mucosal surfaces such as the vaginal tract, bacteria must have mechanisms to adhere to tissues, evade antimicrobial agents, and adapt to changing environments and nutrients. Identifying these mechanisms would greatly aid the ability to develop treatments that target colonizing pathogens. In the case of GBS, the vaginal environment has been the primary subject of study, as maternal vaginal colonization can result in severe neonatal infections. Although some colonization factors have been described for GBS, an overall picture of the gene profile during in vivo growth is an important missing piece of the puzzle for a full understanding of GBS carriage and infection. Here we present in vivo RNA-Seq data quantifying the transcriptome of GBS during murine vaginal carriage.

The genetic programs governing in vivo growth are shown to be substantially different from what is seen during growth under laboratory conditions, with over one-quarter of the genome showing differential regulation between these two environments. Membrane components such as sensor kinases and nutrient transport systems were generally highly upregulated in the vaginal environment (Fig. 1B). In both rich and chemically defined media, where nutrients are abundant, it may be unnecessary for bacteria to induce expression of PTS systems. In the host, on the other hand, the upregulation of phosphotransferase systems may be essential for survival.

Metabolic systems were also found to be highly differentially expressed. Of particular importance, the fab fatty acid synthesis operon was substantially downregulated during growth in the vaginal tract compared to what was seen under lab culturing conditions. In S. pneumoniae, fatty acid regulation is controlled by the MarR family transcriptional regulator FabT. Deletion of FabT in S. pneumoniae results in upregulation of all fab genes in the fatty acid synthesis operon besides fabM, indicating that FabT acts as a repressor in S. pneumoniae. The authors proposed that in a FabT-deficient mutant strain, upregulation of the fab genes leads to decreased amounts of membrane unsaturated fatty acids (UFA), which results in sensitivity to acid (17). An increase in long-chain fatty acids also results in growth benefits in Streptococcus mutans in acidified environments (16). In the vaginal tract, where sensitivity to acid would be highly detrimental, we found that all genes in the fatty acid synthesis operon, including fabM and fabT, were highly downregulated compared to growth in either CDM or THY, indicating a switch to longer-chain UFA in the vaginal tract than in the laboratory media. Downregulation of the fab operon could lead to increased UFA in the membrane and protect the bacteria from the acidic vaginal environment. The decrease in fabT levels that we observed is surprising if FabT acts as a transcriptional repressor in GBS; however, if FabT is a relatively stable protein, perhaps increased fabT transcription occurs in the cells during initial stages of colonization, prior to the 48-hour sampling time point, leading to an overall transcriptional repression of the entire operon, including fabT.

It is likely that many of the transcriptional differences that we observed in laboratory log-phase growth and in vivo growth are due to changes in the growth state of the organisms. It is possible that the in vivo environment would more closely mimic a stationary-phase growth in liquid culture. The switch from log-phase to stationary-phase growth in bacteria is often characterized by a decrease in expression of ribosomal genes, which in this case have been depleted from our input libraries. A repression in cellular processes associated with cell replication and growth as well as transcription and translation is often observed in the log- to stationary-growth phase shift (26, 27). A previous study by Sitkiewicz and Musser examined gene expression changes in GBS during different growth phases in liquid culture and described a characteristic gene expression profile during stationary phase, including upregulation of particular metabolic genes and stress response genes as well as differential expression of numerous virulence factors (28). We observe similar patterns of gene expression for some of these genes, including high upregulation of the arginine/ornithine carbamoyltransferases (sak_2063-2065 and sak_2123-2125) and the glp operon (sak_0345-0347) during in vivo growth. Conversely, some of the genes found to be differentially expressed in stationary versus log phase in that study are not regulated similarly under our conditions; fba (sak_0178), gap (sak_1790), pgk (sak_1788), and eno (sak_0713), all found to be differentially expressed in stationary-phase liquid culture, are not differentially expressed in vivo compared to what is seen in log-phase growth. Some genes are even regulated in the opposite manner; gbs1539 was found to be upregulated during stationary-phase growth in liquid culture but downregulated in vivo. These data suggest that although the bacterial cells grown in vivo have certain characteristics of stationary-phase growth, it would be an oversimplification to assume that the hallmarks of stationary-phase growth are recapitulated in vivo.

Several TCSs were also found to be differentially regulated during in vivo growth. One such system, highly upregulated in the vaginal tract, was identified as a homolog of the S. aureus virulence regulatory TCS, SaeRS. Despite a high level of sequence homology between the SaeRS proteins of S. aureus and GBS, the S. aureus SaeRS regulon is without homologs in GBS and the known SaeR DNA-binding sequence, GTTAAN₆GTTAA (24), was not found at promoter regions of genes observed to be differentially expressed by SaeR in our RNA-Seq data. Alignment of the SaeR sequences of S. aureus and GBS shows that the residues associated with DNA binding in S. aureus (H198, R199, R201, and W218) are conserved, with the exception of residue R199 of S. aureus, thought to be less important for binding, which is an alanine in GBS (Fig. 2A). However, regions of SaeR flanking the DNA binding helices are not similar and may impact the identity of DNA recognition sequences between the orthologs. We have demonstrated that SaeR D53E is able to bind upstream of pbsP. Further work is under way to identify the exact DNA binding site recognized by SaeR in GBS to more easily determine which genes this response regulator directly controls. Although we hypothesize that SaeS acts as the sensor kinase to phosphorylate SaeR, as is the case in S. aureus, we cannot rule out the possibility that SaeR acts promiscuously with other sensor kinases to regulate gene expression.

RNA-Seq analysis was undertaken to identify transcripts whose levels were influenced by SaeRS under different growth conditions, including during growth in the vaginal tract. SaeRS appears to be involved in the regulation of many GBS genes during growth in liquid culture but accounted for a very small and specific genetic program during growth in the vaginal tract. Under this condition, SaeR regulated sak_1753, a hypothetical gene with no known homologs, and a putative small peptide-encoding gene, sak_RS00910, which is located between sak_0183 and sak_0184. Importantly, SaeR also regulated sak_0466 (pbsP), which has been shown to be important for dissemination of GBS during invasive disease, and which we show here to be an important colonization factor for GBS. In liquid culture, the majority of differentially expressed genes were changed by a small degree, generally between 2- and 5-fold, whereas the three DEGs observed during vaginal growth were downregulated 7-fold, >200-fold, and >800-fold. It is possible that the changes seen in liquid culture were the result of indirect rather than direct regulation in the vaginal tract. It is also possible that SaeRS is responsible for a vastly different genetic program depending on environmental signals.

Two of the genes regulated by SaeR during vaginal colonization, sak_1753 and sak_0466 (pbsP), were among the most highly differentially expressed genes in the vaginal tract of WT strains compared to those growing in liquid. The role of sak_1753 in vaginal carriage is currently being examined, although there are very few known homologs and no predicted function for the protein product. PbsP, on the other hand, has recently been identified as a possible GBS vaccine target and was shown to be important in hematogenous dissemination during an intravenous infection. It has also been found to be upregulated during growth in human blood (29) and at 40°C (30). PbsP is highly conserved in GBS strains and present in all sequenced GBS clinical isolates from humans. The role of PbsP in invasive disease was shown in GBS strain NEM316, a clonal complex 23 isolate of capsular serotype III (21). A909, a clonal complex 19 isolate with capsular serotype Ia, was used in these studies, indicating the importance of this gene in multiple strains of GBS in vivo.

In GBS, pbsP is located immediately upstream of saeRS and contains a sortase cell wall-anchoring LPXTG motif, and levels of PbsP protein are decreased in an srtA mutant strain (31). A recent paper examined the role of PbsP in an intravenous model of GBS infection. In this model, PbsP was important for binding of plasminogen, dissemination to organs such as the kidney, and transendothelial migration into the brain (21). In S. pneumoniae, a homologous gene named pneumococcal adherence and virulence factor B, pavB (also termed pfbB) was found to be important for adherence to respiratory epithelial cell lines via fibronectin and plasminogen and for colonization of the nasopharynx (22, 23).

Here we demonstrate that PbsP is also important in maintenance of colonization in the vaginal tract and that its expression levels, controlled by SaeRS, are highly upregulated during incubation with murine vaginal lavage fluid. On day 1 following vaginal inoculation, numbers of A909 and A909ΔpbsP isolates were similar. We hypothesize that during this 24-hour period, bacteria were not necessarily being actively removed from the vaginal tract, and thus the A909ΔpbsP mutants were able to persist in the vaginal environment. Following the day 1 vaginal wash, the bacteria would be required to actively adhere to the tissue to remain in the vaginal tract, emphasizing the defect in colonization by the mutant. Thus, its role does not appear to be limited to invasive disease. In their 2016 article, Buscetta et al. suggest that PbsP plays a role in favoring systemic spread of GBS and invasive infection (21). Because their model begins with intravenous (i.v.) infection rather than mucosal carriage, it is possible that they are observing an unintended side effect of PbsP during systemic infection rather than the primary role as an adhesin on mucosal surfaces. Previous studies have linked the expression of pbsP to the master virulence regulator CovRS, as levels of pbsP transcript and protein are higher in a ΔcovRS strain (21, 32, 33). Direct binding of CovR to the promoter of pbsP was not observed, and it was postulated that the regulation may be indirect (21). It is possible that SaeRS is being regulated by CovRS, although changes in transcript levels of saeRS were not observed in a ΔcovRS mutant (32). Changes in transcript levels of psbP were altered to a much greater extent in a ΔsaeR mutant than in the ΔcovRS mutant, and we postulate that the SaeRS system is more likely to be a direct regulator of psbP than CovRS. Our EMSA data also indicate that phosphomimetic SaeR D53E is a direct regulator of pbsP via promoter binding. Based on these data, we propose that SaeR is phosphorylated by SaeS following interaction with a signal present in the vaginal tract and that phosphor-SaeR directly binds upstream of pbsP to increase gene expression. Alignment of the ∼200 bp upstream of the ATG start site of pbsP and sak_1753 shows a surprisingly high degree of sequence similarity (see Fig. S1 in the supplemental material). Further work will be needed to characterize the exact DNA binding sequence.

Our studies indicate that the signal is likely a small heat-labile peptide of less than 3 kDa. It was shown in S. aureus that subinhibitory concentrations of the human neutrophil peptide (HNP) α-defensins could induce expression of SaeRS-controlled promoters (34). This finding led the authors to hypothesize that host-produced antimicrobial peptides could serve as a signal to activate the SaeRS system. It is possible that α-defensins or a similar small antimicrobial peptide also serves as a signal for SaeRS in S. agalactiae. The extracellular linker region of SaeS (WFNGHMTLT) was found to be important for responding to HNP in S. aureus (35). In S. agalactiae, the linker region has some similarity (GGLNHMLIET), indicating that the signal sensed by these proteins may be similar but not identical. Current studies are focusing on identifying the exact nature of the signal sensed by SaeS in VL and determining whether the signal is related to estrous cycle changes. A model of these data is shown in Fig. 6. Signals encountered in the host environment cause numerous transcriptional changes within the GBS cell. One two-component system, SaeRS, was found to be upregulated during growth in the host. This system senses a host signal during colonization of the vaginal tract and induces expression of sak_1753, a gene encoding an unknown hypothetical protein, and pbsP, encoding a putative adhesin and colonization factor, likely via direct binding. We demonstrate the PbsP is important for vaginal colonization. These data provide important insight into the molecular programming of GBS during growth and colonization in the vaginal tract.

FIG 6.

Model of S. agalactiae murine vaginal colonization, illustrating vaginal colonization of GBS, signaling via SaeRS, and host colonization via PbsP.

MATERIALS AND METHODS

Bacterial strains, media, plasmids, and primers.

All strains and plasmids are shown in Table S1, and the primers used in this study are described in Table S2 in the supplemental material. Escherichia coli strain BH10C (36) was cultivated in Luria-Bertani (LB) medium or on LB agar. When necessary, antibiotics were included at the following concentrations for E. coli propagation: chloramphenicol (Cm), 10 μg ml⁻¹; kanamycin (Kan), 150 μg ml⁻¹; erythromycin (Erm), 100 μg ml⁻¹. All GBS strains used in this study were derived from the clinical isolate A909 (37). GBS were routinely grown in Todd-Hewitt medium (BD Biosciences) supplemented with 2% (wt/vol) yeast extract (Amresco) (THY) or a chemically defined medium (CDM) modified from that described by van de Rijn and Kessler RE (38, 39). The exact composition of CDM and the protocol for preparation have been previously published (39). Plating was done on THY agar plates or CHROMagar StrepB agar plates (CHROMagar). When necessary, antibiotics were included at the following concentrations for GBS propagation: chloramphenicol (Cm), 3 μg ml⁻¹; kanamycin (Kan), 150 μg ml⁻¹; erythromycin (Erm), 10 μg ml⁻¹; spectinomycin (Spec), 100 μg ml⁻¹; ampicillin (Amp), 50 μg ml⁻¹.

Mouse model of vaginal colonization.

Female outbred CD1 mice aged 7 to 12 weeks were used for all experiments. Experiments were performed as previously described (5, 8, 9). Briefly, 1 day prior to inoculation (day −1), mice were given an intraperitoneal injection of 0.5 mg β-estradiol valerate (Acros Organics) suspended in 100 μl filter-sterilized sesame oil (Sigma) to synchronize estrus. On day 0, mice were vaginally inoculated with 10⁷ CFU of bacteria in 10 μl PBS. On days 1, 2, 3, and 5, the vaginal lumen was washed with 50 μl sterile PBS, using a pipette to gently circulate the fluid approximately six times. To enumerate the bacteria released from the vaginal lumen, the PBS-based lavage fluid was collected and placed on ice for no more than 30 min before serial dilutions in PBS were plated onto CHROMagar StrepB plates to obtain CFU counts. For enumeration of A909ΔpbsP cells with the complementation vector, cells were also plated on THY Spec plates to ensure retention of the complementing plasmid.

RNA collection.

For RNA collection from the murine host for RNA-Seq, vaginal lavage fluid from mice colonized with bacteria for 48 h was taken directly from the mice and placed into a tube containing 1 ml TRIzol LS reagent (Ambion) and placed on ice. Lavage fluid from a total of 10 mice was placed into the same tube of TRIzol LS reagent, giving a final volume of 1 ml lavage fluid and 1 ml TRIzol LS reagent. These samples were kept on ice for no more than 30 min. Samples were centrifuged for 1 min at 14,000 × g, the TRIzol LS Reagent was removed, and RNA preparation was completed using the Ambion RiboPure Bacteria kit (AM1925; Thermo Fisher). Briefly, 250 μl of RNAWiz was used to resuspend the bacterial pellets. RNA was then purified according to the manufacturer's protocol using 10 min of bead beating using a MiniBeadbeater (BioSpec) set to homogenize to lyse the cells. Following collection, RNA was treated with DNase I for 30 min at 37°C. Vaginal lavage fluid from 10 mice generally gave an average of ∼0.5 to 5 μg of total RNA. For collection of RNA from planktonic cultures, bacteria were grown statically overnight at 37°C in THY broth supplemented with the appropriate antibiotics. Following overnight growth, cultures were back diluted 1:50 in freshly prepared CDM broth. Once the cells reached an optical density at 600 nm (OD₆₀₀) value of 0.3 to 0.6, cultures were spun down at the same OD, and RNA was prepared from bacterial pellets as described above using the RiboPure Bacteria kit and treated with DNase I. For quantitative PCR (qPCR), planktonic cultures were grown as above, back diluted in CDM, and grown to an OD₆₀₀ value of 0.3 to 0.6. One milliliter of cells at an OD₆₀₀ of 0.4 was then pelleted and resuspended in either 1 ml of PBS or 1 ml of vaginal lavage fluid. Cells were incubated at 37°C for 1 h and then pelleted, and RNA was collected and treated with DNase I as described above.

Preparation of eukaryotic and ribosomally depleted cDNA libraries for RNA-Seq.

Prior to the library creation, RNA samples collected from the murine vaginal tract were depleted of eukaryotic RNA using the MICROBEnrich kit (AM1901; Thermo Fisher). Both vaginal and cultured samples were then taken through the rRNA removal MICROBExpress kit (AM1905; Thermo Fisher) according to the manufacturer's instructions. RNA integrity and eukaryotic and rRNA depletion were then assessed by measurement on both a TapeStation 2200 (Agilent) and a Qubit RNA high-sensitivity fluorometer (MBL) by the DNA Services Facility (DNAS) at the University of Illinois at Chicago Center for Genomic Research. If a high concentration of rRNA was still detected, samples were taken through the MICROBExpress kit an additional time.

RNA-Seq libraries were generated from a small amount of RNA (10 to 400 ng) using the KAPA Stranded RNA-Seq Library Preparation kit for Illumina Platforms (KR0934; KAPABiosystems). Briefly, 10 to 400 ng of rRNA-depleted RNA was fragmented by heating fragmentation buffer at 94°C for 6 min. First-strand cDNA was synthesized using random primers followed by 2nd-strand synthesis and marking. A poly(A) tail was then added to double-stranded cDNA (dscDNA) fragments, and Illumina adapters were ligated onto the library fragments. The library was then subjected to two rounds of cleanup using magnetic beads for DNA purification (P920-30; 101Bio) followed by library amplification for 10 cycles and an additional magnetic-bead cleanup step. All libraries were then sent to the DNAS facility for quality control (QC) and quantification on the Tapestation 2200 instrument (Agilent). At least 20 μl of 50 nM libraries was sent to the University of Chicago Genomics Facility for sequencing and analysis.

RNA sequencing and analysis.

Raw sequencing data were analyzed by the University of Illinois at Chicago Research Resources Center. Data were aligned to the GBS A909 genome using BWA MEM (40), and gene expression was quantified using featureCounts (41). Gene expression was normalized to counts per million sequences, and differential expression analysis was performed on raw counts using edgeR (42), and the false-discovery rate (FDR) correction was used to adjust P values for multiple testing. Significantly differentially expressed genes (DEGs) with FDR of <0.05 were selected for pathway analysis. The KEGG gene and pathway annotations for A909 (T number T00278) were downloaded from the KEGG website (http://www.genome.jp/dbget-bin/www_bget?gn:T00278), and all genes were assigned to their KEGG Orthology (KO) group. All metabolic pathways associated with any KO in the A909 genome were downloaded, and a list of all KOs in each pathway was created to serve as a pathway database. Enrichment statistics for each pathway were computed with respect to our differentially expressed genes by comparing the fraction of KOs in a pathway that were DEGs to the overall fraction of DEGs. We computed P values using Fisher's exact test and corrected for multiple testing with the FDR correction over all pathways.

Construction of saeR and pbsP mutants and pLC010 complementation plasmid.

For generation of a plasmid to replace saeR with aphA3 (kanamycin resistance gene), upstream and downstream DNA fragments flanking the saeR gene were amplified by PCR using primer pairs LC112/LC115 and LC113/LC116, respectively. These products were purified and then fused in a second PCR using the outside primers LC112/LC116. This fusion product was cloned into a temperature-sensitive pJC159 (Cm^r) vector using internal enzymes HindIII and ClaI following by ligation. Using inverse PCR, this plasmid was replicated linearly using primers LC129/LC130 with MluI cut sites between the upstream and downstream saeR sequences. Separately, the kanamycin resistance gene aphA3 was amplified from pOSKAR using primers JC292/JC304 with MluI cut sites. The linearized plasmid and aphA3 gene were digested with MluI and ligated. This plasmid was electroporated into competent BH10C E. coli cells and plated onto LB plates supplemented with Cm and Kan for propagation. The plasmid was electroporated into electrocompetent A909 cells, and a two-step temperature-dependent selection process was used to isolate mutants of interest (43). Briefly, cells containing each deletion construct were grown at the permissive temperature (30°C) and then shifted to 37°C and plated on the THY plates containing Cm and Kan to select for bacteria in which the plasmid had integrated at one of the flanking regions. Cells were then grown at the permissive temperature to allow the plasmid to recombine out of the chromosome, and loss of Cm resistance but maintenance of Kan resistance was used to identify a successful second crossover event and loss of the mutation vector. Genotypes were confirmed by PCR and sequencing. The creation of A909Δ pbsP::cat was done using Gibson assembly of four fragments. The temperature-sensitive pJC162 (Erm^r) vector was digested with EcoRI and PstI. PCR was done on the 984 bp immediately upstream of pbsP using primers LC169/LC170, on the chloramphenicol resistance gene cat using primers LC171/LC172, and on the 1,070 bp immediately downstream of pbsP using primers LC173/LC174. The three PCR products and the digested vector were combined with 2× HiFi DNA Assembly master mix (E2621; New England BioLabs [NEB]) and incubated at 50°C for 60 min. This mixture was then electroporated into E. coli BH10C cells and plated onto LB plates supplemented with Erm and Cm, and the knockout was created as described above, selecting for Cm colonies that lost the Erm resistance marker.

To create the complementation plasmid pLC010, the pbsP gene was amplified from A909 genomic DNA using primers LC253/LC254. Plasmid pJC303, a pLZ-12-Spec-based plasmid with a constitutive recA promoter, was digested with NotI and BamHI. The digested plasmid and pbsP PCR product were assembled using a Gibson reaction with 2× NEBuilder Hifi DNA Assembly master mix (E2621; New England BioLabs). The resulting plasmid, pLC010, was electroporated into E. coli BH10C cells. Following propagation in E. coli, the plasmid was sequenced and electroporated into electrocompetent A909ΔpbsP cells and recovered on THY Cm, Spec plates.

Preparation of cDNA and qPCR.

All primers used in qPCR are listed in Table S2. cDNA for qPCR was made using DNase I-treated RNA from cells incubated with either PBS or pooled vaginal lavage fluids. cDNA preparation was done using the Superscript III first-strand synthesis system (18080051; Thermo Fisher Scientific) according to the manufacturer's instructions, including treatment with RNase H. Amplification was done using antisense gene-specific primers LC060 (gyrA), LC133 (pbsP), and LC199 (sak_1753). cDNA was diluted between 1:2 and 1:5, depending on the desired concentration, and used for qPCR. qRT-PCR was done using the Fast SYBR green master mix (4385612; Applied Biosystems) and a CFX Connect Real Time PCR detection system (1855200; Bio-Rad). The gyrase A gene (gyrA), a housekeeping gene not seen to have differential expression during growth in the vaginal tract, was used as a reference gene. All samples were run in triplicate technical replicates on a single plate, and triplicate biological replicates were used to determine final statistics.

SaeR D53E protein expression and purification.

Cloning of saeR D53E into the pET21a expression vector containing a 6His tag was done using the NEBuilder HiFi DNA assembly master mix (New England BioLabs). A 692-bp gblock fragment of saeR (sak_0467) containing a D53E mutation was ordered from Integrated DNA Technologies. The gblock fragment was amplified by PCR using primers LC190/LC191. This product and pET21a vector were digested with XhoI and XbaI. The plasmid and DNA were ligated using 2× Hifi master mix (NEB). The resulting plasmid was electroporated into E. coli BL21(DE3) cells and selected on LB agar plates containing ampicillin. BL21 cells containing the expression plasmid were grown overnight in LB containing Amp and back diluted to 1:20 into 1 liter fresh LB medium and grown at 37°C. When cells reached an OD₆₀₀ of approximately 0.5, 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) was added. Cells were incubated at 28°C for 6 h and then harvested and frozen overnight at −20°C. Cells were lysed in 20 mM Tris (pH 8.0)–100 μg/ml lysozyme using sonication. Lysed cells were spun down, and supernatant was passed through a 0.4-μm filter. Protein was purified using a 3-ml HisPur Ni-nitrilotriacetic acid (Ni-NTA) column (88226; Thermo Fisher). Glycerol (20%) was added to purified protein preparations prior to storage at −80°C.

Electrophoretic mobility shift assay (EMSA).

Primers used for EMSA probe DNA are listed in Table S2. The pbsP promoter DNA probe (195 bp) was amplified using primers LC203/LC204. LC203 includes a 5′ fluorescent 6-carboxyfluorescein (6-FAM) tag (Sigma). Unlabeled pbsP promoter DNA probe (191 bp) was amplified using unlabeled primers LC204/LC142. Unlabeled nonbinding DNA (100 bp) was amplified using unlabeled primers LC231/232. EMSA reaction volumes included 5 nM labeled probe, 50 mM Tris (pH 8.0), 50 mM KCl, 10 mM MgCl₂, 0.5 mM EDTA, 0.2 mM dithiothreitol (DTT), 0.05% Triton X-100, 12% glycerol, 10 μg bovine serum albumin (BSA), and 50 ng salmon sperm DNA. Protein was added in the concentrations shown in Fig. 5C. When needed, unlabeled pbsP promoter DNA or unlabeled nonbinding DNA was added as shown in Fig. 5C. Reaction mixtures were incubated at room temperature for 30 min. Xylene cyanol (5%, 1 μl) was added to each reaction mixture, and 10 μl was run on a 6% nondenaturing Tris-borate polyacrylamide gel (120-V prerun for 30 min, 120-V run for 75 min). The gel was sandwiched between clear polypropylene sheet protectors and imaged using a Typhoon Trio Variable Mode Imager (GE Healthcare) using fluorescent excitation/emission wavelengths of 488/520 nm.

Ethics statement.

All mouse experimentation was approved by the University of Illinois at Chicago Animal Care and Use Committee (ACC) and IACUC under protocol number 16-068. All animal work was carried out using accepted veterinary standards in accordance with the Animal Care Policies of the University of Illinois at Chicago Office of Animal Care and Institutional Biosafety Committee and IACUC. This institution has Animal Welfare Assurance Number A3460.01 on file with the Office of Laboratory Animal Welfare, NIH.

Accession number(s).

Sequence data have been deposited in the NCBI GEO database under accession no. GSE109680.