Modification of the CpsA Protein Reveals a Role in Alteration of the Streptococcus agalactiae Cell Envelope

Abstract

The bacterial cell envelope is a crucial first line of defense for a systemic pathogen, with production of capsular polysaccharides and maintenance of the peptidoglycan cell wall serving essential roles in survival in the host environment. The LytR-CpsA-Psr proteins are important for cell envelope maintenance in many Gram-positive species. In this study, we examined the role of the extracellular domain of the CpsA protein of the zoonotic pathogen group B Streptococcus in capsule production and cell wall integrity. CpsA has multiple functional domains, including a DNA-binding/transcriptional activation domain and a large extracellular domain. We demonstrated that episomal expression of extracellularly truncated CpsA causes a dominant-negative effect on capsule production when expressed in the wild-type strain. Regions of the extracellular domain essential to this phenotype were identified. The dominant-negative effect could be recapitulated by addition of purified CpsA protein or a short CpsA peptide to cultures of wild-type bacteria. Changes in cell wall morphology were also observed when the dominant-negative peptide was added to wild-type cultures. Fluorescently labeled CpsA peptide could be visualized bound at the mid-cell region near the division septae, suggesting a novel role for CpsA in cell division. Finally, expression of truncated CpsA also led to attenuation of virulence in zebrafish models of infection, to levels below that of a cpsA deletion strain, demonstrating the key role of the extracellular domain in virulence of GBS.

INTRODUCTION

Many bacterial pathogens can exist in the host as part of the normal microflora and only cause disease when they gain access to a normally sterile site, such as the bloodstream or cerebral spinal fluid, or when the host becomes immunocompromised. Group B Streptococcus (GBS), Streptococcus agalactiae, can colonize the human vaginal and rectal mucosa as a commensal organism (1–3) but can also cause a range of invasive infections in immunocompromised and older adults (4–6). GBS is also a broad-host-range pathogen capable of causing severe sepsis and meningoencephalitis in fish (7–12), bovine mastitis (13–15), and human neonatal sepsis and meningitis (16–21) and has been recently implicated in chorioamniosis (22, 23), causing premature birth (24, 25) and stillbirth (25, 26). The current strategy to prevent GBS infection of neonates is intrapartum antibiotic prophylaxis (IAP) (27), i.e., intravenous antibiotics given to a woman during labor and delivery. IAP is administered either to women after a positive GBS culture during prenatal screening (27, 28) or if certain risk-based criteria are met (29, 30). Fortunately, this has led to a reduction in the incidence of GBS disease in the first week of life (early onset disease), but it has not changed the incidence of GBS disease in the first 3 months of life (late onset disease) (18, 21, 27).

One mechanism allowing bacteria to successfully function as both a commensal and a pathogen is regulation of a polysaccharide capsule expression. This is an essential event for GBS immunoevasion and survival in the bloodstream (31) but is not essential for colonization, since a GBS isolate with a complete deletion of the cps locus is capable of colonizing the vaginal mucosa (32). GBS has nine serotypes defined by the genes of the cps locus, leading to different polysaccharide compositions (33). The first regulatory gene of this locus, cpsA, is conserved across all serotypes and plays a role in transcriptional activation of this locus (34).

CpsA is a member of the LCP (LytR-CpsA-Psr) protein family. These are found in a broad range of Gram-positive bacterial species (35) and are important for linkage of cell wall carbohydrates to the cell wall, cell wall maintenance, and cell division (36). Cell wall stability and proper control of cell division is important for bacterial survival and proliferation in the stressful environment of the host. LCP family proteins have one or more transmembrane domains and a large extracellular domain containing the conserved LCP domain (35), which when characterized appears to have a catalytic function of ligation of cell wall carbohydrates to the peptidoglycan cell wall (36). The CpsA protein, however, is unique among the family of LCP proteins containing in addition to the extracellular LCP domain, an additional extracellular domain, called the accessory domain.

Analysis of the GBS CpsA protein (37) demonstrated a bifunctional role for CpsA in capsule expression and cell wall stability. Deletion of the cpsA gene led to a reduction in capsule production and to the bacteria forming unusually long chains of cocci. A role for CpsA in transcriptional regulation of the capsule operon was demonstrated by showing the ability of the full-length CpsA to bind specifically to the two promoters of the cps locus, whereas truncated CpsA missing the extracellular domain is still capable of DNA binding but has lost specificity. This suggested that the extracellular domain is serving a role in transcriptional regulation of this locus, possibly through promotion of proper protein conformation of CpsA for specific DNA binding (37).

Previous analysis of GBS CpsA protein function revealed that episomal expression of the CpsA-Full protein in an isogenic cpsA deletion strain could complement the cpsA deletion for wild-type capsule expression (37). However, episomal expression of a truncated form of CpsA missing the most C-terminal (LCP) extracellular domain and having only the extracellular accessory domain (CpsA-245, Fig. 1) could not complement for capsule production in the cpsA deletion strain (37). Episomal expression of these same protein constructs in the wild-type GBS strain had quite different results. Although the CpsA-Full protein had no effect on wild-type capsule expression, episomal expression of the CpsA-245 protein caused a dominant-negative effect on capsule production, resulting in decreased capsule production and increased chain length (37). Therefore, the truncated CpsA-245 protein effectively inhibited the ability of the wild-type strain to produce capsule even in the presence of the wild-type CpsA protein. This indicated that the extracellular domains of the CpsA protein were important for both the capsule expression and cell wall maintenance functions (37).

FIG 1.

GBS CpsA protein. (A) Membrane topology of CpsA. (B) Domains of CpsA and location of truncations.

In this study, we identified the regions of the extracellular accessory domain of the GBS CpsA protein that are required for the dominant-negative effect. Furthermore, we determined that the dominant-negative phenotype could be induced by extracellular addition of the truncated CpsA protein or a short peptide of CpsA. Visualization of peptide binding revealed that this region of CpsA is in contact with the midcell septal region of the bacterial cell wall. Lastly, we provide evidence that deletion of the cpsA gene or expression of a truncated CpsA protein leads to attenuation of virulence.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Escherichia coli strains TOP10 and Electromax DH5α (Invitrogen) were used for the construction and maintenance of plasmids. These strains were grown under aerobic conditions with shaking in liquid Luria-Bertani (LB) medium (Acumedia) or aerobically on LB agarose plates: LB medium supplemented with 1.5% technical agar (Acumedia) and supplemented with antibiotics at the following concentrations: ampicillin, 100 μg/ml; chloramphenicol, 20 μg/ml; or erythromycin, 750 μg/ml. S. agalactiae strain 515 and the isogenic cpsA mutant (34) were grown statically in closed tubes in Todd-Hewitt media (Alpha Biosciences) plus 0.2% yeast extract (THY; Acumedia) with the addition of 1.4% bacteriological agar (Acumedia) for solid media and then grown anaerobically or in 5% CO₂ and supplemented with antibiotics at the following concentrations: chloramphenicol, 3 μg/ml; or erythromycin, 2 μg/ml. Cultures grown in the presence of purified protein or synthetic peptide were grown in medium supplemented with HALT protease inhibitor (Thermo Fisher).

CpsA strain construction.

The CpsA insert strain was constructed using primers 5′ GBS cpsA-insert PstI and 3′ GBS cpsA-insert XmaI (Table 1) by amplifying a 454-bp fragment using GBS 515 genomic DNA as the template. The purified fragment was cloned into the PstI and XmaI sites of vector pUC19-Erm (38). The plasmid was transformed into GBS strain 515, and single-crossover recombinants were selected by plating on THY-Erm (2 μg/ml) plus agarose plates. CpsA truncations were made using a template encoding an MBP-cpsA (plasmid pGBS cpsA-full [37]) using the primers 5′ MBP RBS BamHI (see Table 1) and 3′ primer, depending on the truncation (see Table 1), and cloned into vector pLZ12-rofA for constitutive expression. Plasmids were transformed into GBS strain 515 and its isogenic ΔcpsA derivative as previously described (37).

TABLE 1.

Primers used in this study

Primer	Sequence (5′-3′)a
5′ GBS cpsA-insert PstI	AAAACTGCAGCGTTGCTACTACTTTATATGG
3′ GBS cpsA-insert XmaI	TCCCCCCGGGGTATCAATACCGCTAAATAG
5′ MBP RBS BamHI	CGCGGATCCGCGGATAACAATTTCACACAGG
3′ GBS cpsA full PstI	AAAACTGCAGTTATTCCTCCATTGTGTTC
3′ GBS cpsA 245 PstI	AAAACTGCAGTTATGTTGATATAGAGCCAAAAG
3′ GBS cpsA 210 PstI	AAAACTGCAGTTAAATCGTTTTTATCTGCG
3′ GBS cpsA 187 PstI	AAAACTGCAGTCATAAAACCATAGCCTGACTATC
3′ GBS cpsA 153 PstI #2	AAAACTGCAGCATTCTATCAAATCAGTAATATTTC AG
3′ GBS cpsA 132 PstI	AAAACTGCAGTCAAGCTTCTATATTGG
3′ GBS cpsA 117 PstI	AAAACTGCAGTTACTCAATTTCAGAGTATGAAGC

^aRestriction sites are underlined.

Percoll buoyant density assays.

Buoyant density was determined using linear Percoll (GE Healthcare) gradients as previously described (37). Briefly, 2 ml of Percoll, supplemented with 0.15 M NaCl, diluted to low (1.085 g/cm³) density was carefully overlaid onto 2 ml of Percoll diluted to high (1.120 g/cm³) density in a 5-ml tube and placed at a 15° angle overnight to allow formation of a continuous linear gradient. Gradients were placed vertically and allowed to settle at least 30 min prior to use. Overnight bacterial cultures were normalized to an optical density at 600 nm (OD₆₀₀) of 0.6 and concentrated to a volume of 50 μl in phosphate-buffered saline (PBS). The culture was carefully added to top of gradient, and gradients were centrifuged for 30 min at room temperature and at 5,000 rpm in a swinging bucket (Eppendorf centrifuge 5403 with rotor 16A4-44). The distance traveled in the gradient was measured from the bottom of the meniscus to the center of the cell band, and the density was determined by calculation of a linear curve from distance traveled by beads of known density (GE Healthcare). Enzyme-linked immunosorbent assay measurements with antibody to the GBS serotype 1a capsule (kindly provided by Dennis Kasper) was used to confirm that the amount of capsule reported in the buoyant density measurements accurately depict the amount of capsule on the strains, including significance measurements (see Table S1 in the supplemental material).

Protein purification.

Proteins were purified by using the pMal-c2x system (NEB) as previously described (37). Elution fractions containing protein were detected by the BCA assay (Pierce) and examined by SDS-PAGE to determine protein purity and integrity. Fractions were concentrated as necessary with a 50-kDa centrifugal concentrator (Millipore) and stored at −80°C in 50% glycerol.

CpsA peptide assays.

A peptide using the sequence from amino acids 210 to 245 of the CpsA accessory domain (INRKNTNHKEGVFNIYISGITDTF) was synthesized (GenScript) and resuspended in sterile double-distilled water at a concentration of 2 mg/ml. Peptide was labeled with fluorescein isothiocyanate (FITC) as follows. To 192 μl of peptide (2 mg/ml) in 568 μl of carbonate-bicarbonate buffer (0.1 M, pH 9.0), 40 μl of freshly prepared fluorescein-5-isothiocyanate stock solution (16 mM in dimethyl formamide [DMF]) was added, and the reaction mixture was incubated in a 37°C rotary shaker for 3 h. The reaction mixture was loaded onto a 5-ml Sephadex G-25 column, and the column was washed with deionized water while fractions were collected. Each fraction was monitored at 280 nm and 495 nm, and fractions containing both absorbance values were combined and lyophilized. The resulting orange powder was redissolved in water, and the concentration of labeled peptide (C) was calculated using the following equation: C = (A₂₈₀ – A₄₉₅ × 0.35)/128, where C is the molar peptide concentration, and A₂₈₀ and A₄₉₅ are the absorbances at 280 and 495 nm, respectively.

Phase-contrast microscopy was performed on overnight cultures of wild-type bacteria grown in the presence or absence of 200 pmol of unlabeled CpsA peptide/ml. Next, 5 μl of overnight culture was mounted with a coverslip and observed using a Zeiss Axioskop 40 and imaged with AxioVision 4.7. For visualization of peptide binding to bacteria, cultures were grown in the presence of FITC-labeled peptide overnight at the concentration indicated and normalized to an OD₆₀₀ of 1.0, followed by three washes in PBS. Then, 5-μl portions of washed bacteria were observed as a wet mount using a Zeiss Axioskop 40 and imaged with AxioVision 4.7. For competitive binding analysis, cultures were grown overnight in the presence of 2.5 μg of FITC-labeled peptide/ml with or without a 25-fold molar excess of unlabeled peptide. The cultures were normalized to an OD₆₀₀ of 1.0, followed by three washes in PBS. Fluorescence of a 100-μl sample in triplicate was read in a Tecan Spectrafluor Plus fluorometer with Magellan6 software using the fluorescein default settings.

Zebrafish infections.

As described previously (39), adult zebrafish were injected intramuscularly with midlog bacterial cultures at an infectious dose of 10⁶ CFU per fish. For dissemination analysis, fish were sacrificed at 24 h, and viable CFU in the brain, heart, and spleen were determined by plating on CNA selective agar (Acumedia).

Statistical analysis.

All numerical results were analyzed using a one-way analysis of variance, and post hoc analyses were performed using the Bonferroni and Holms multicomparison test.

RESULTS

To further investigate the function of the extracellular accessory domain of the GBS CpsA protein, a nonpolar cpsA insertion mutation on the chromosome was constructed and tested for the ability to produce capsule. This mutant strain expresses a truncated form of the CpsA protein, missing the C-terminal extracellular LCP domain (similar to the CpsA-245 truncation, Fig. 1). Buoyant density analysis revealed that the CpsA-insert strain produces capsule levels even further reduced than those of the cpsA deletion strain (Fig. 2). These results, coupled with our previous data using the episomally expressed CpsA-245 protein in the wild-type GBS strain as described above, suggest that the CpsA extracellular accessory domain, when in the absence of the LCP domain, does not support capsule production.

FIG 2.

Capsule production. The capsule production of GBS wild-type, the ΔcpsA, and cpsA insert (i.e., strain 515 with pUC19-Erm plasmid inserted into and disrupting cpsA) strains was measured by Percoll buoyant density centrifugation. Error bars represent the standard deviations. **, P < 0.01.

To narrow down the region of the accessory domain responsible for the dominant-negative effect, serial truncations of the extracellular domain of the CpsA-245 protein were made (see Fig. 1). These constructs were expressed episomally in the wild-type GBS strain and the ΔcpsA strain (Fig. 3). Only the CpsA-117 and CpsA-full were able to complement the cpsA deletion strain. The CpsA-210, CpsA-187, and CpsA-132 proteins when expressed episomally in the wild type had wild-type capsule levels. However, the CpsA-245 protein episomally expressed in the wild-type strain had significantly reduced (P < 0.05) capsule levels, commensurate with the cpsA deletion strain (Fig. 3). The CpsA-153 protein episomally expressed in the wild-type strain also showed a strong degree of inhibition but was not statistically significant. These results indicate that the regions of CpsA important for the dominant-negative effect are found in amino acids 210 to 245.

FIG 3.

Capsule production by strains with episomally expressed CpsA truncations. Capsule production was measured by Percoll buoyant density centrifugation. Notations on the left indicate CpsA version expressed from a plasmid in the indicated strain. Black columns represent the wild-type strains expressing the indicated construct. Gray columns represent the ΔcpsA strains expressing the indicated construct. Densities were normalized to the wild type with empty vector. Error bars represent the standard deviations. **, significantly (P < 0.01) different from wild-type capsule levels; *, significantly (P < 0.05) different from wild-type levels; ##, significantly (P < 0.01) different from ΔcpsA capsule levels.

We next questioned whether the dominant-negative phenotype caused by expression of the CpsA-245 protein in the presence of the wild-type CpsA protein (wild-type strain) could exert its inhibitory effect if added exogenously or if it needed to be expressed by the bacteria. Therefore, purified proteins of various CpsA truncations were added to a culture of wild-type bacteria and assessed for capsule production by buoyant density centrifugation (Fig. 4). The addition of purified CpsA-117 protein and the CpsA-full protein, which had no effect on capsule levels when expressed episomally in the wild-type strain (Fig. 3), also had no effect on capsule levels when supplied as purified proteins to the cultures (Fig. 4). However, the addition of purified CpsA-245 protein, which causes a dominant-negative effect on capsule levels when expressed episomally in the wild-type strain, caused a similar reduction in capsule levels when added exogenously to a culture of wild-type bacteria. As a negative control, cultures were grown with the purified SalR protein, a Streptococcus pyogenes protein constructed and purified by the same process and had no effect on capsule expression. Therefore, the dominant-negative phenotype caused by expression of the CpsA-245 protein is capable of causing the same result when added exogenously from the extracellular environment.

FIG 4.

Capsule production in the presence of purified protein. Capsule production was measured by Percoll buoyant density centrifugation. Cultures were grown with the indicated protein at a concentration of 20 pmol/ml, in the presence of protease inhibitors. “Vector” strains have empty plasmid. Error bars represent standard deviation. **, P < 0.01.

The serial truncations of the extracellular domains determined that the most C-terminal end of the accessory domain was important for the dominant-negative effect. Therefore, the CpsA peptide, a synthetic peptide corresponding to amino acids 218 to 240, was constructed. When wild-type cultures were grown in the presence of the CpsA peptide, capsule production was significantly reduced (Fig. 5, P < 0.01). However, the addition of a scrambled peptide with the same amino acids in different order had no effect on capsule production. These results further confirm the ability of the dominant-negative factor to function from the extracellular environment. Moreover, these results demonstrate that the N-terminal DNA binding and transmembrane domains of the CpsA protein are not required for the dominant-negative effect to occur.

FIG 5.

Capsule production in the presence of CpsA peptide. Capsule production was measured by using Percoll buoyant density centrifugation. Cultures were grown with the indicated peptide at a concentration of 500 pmol/ml in the presence of protease inhibitors. “Vector” strains have empty plasmid. Error bars represent the standard deviations. *, P < 0.05; **, P < 0.01.

Previous analysis of cpsA mutations or episomal expression of truncated forms of the CpsA protein revealed that, in addition to changes in capsule production, there were also changes in cell wall architecture, as evidenced by the formation of unusually long chains of cocci in these strains (37). The addition of the CpsA peptide to wild-type cultures changed the normal chain length from 2 to 4 cocci per chain (Fig. 6A) to most chains being >10 cocci per chain (Fig. 6B). To visualize the interaction of the peptide with the bacteria and any cell wall changes that may be occurring, the peptide was labeled with FITC and examined by fluorescence microscopy (Fig. 7). When GBS cultures were grown in the presence of the FITC-labeled peptide, nearly identical staining patterns were apparent in both wild-type (Fig. 7A) and ΔcpsA (Fig. 7B) cultures, indicating that the peptide is binding to a factor other than the CpsA protein. Both wild-type and cpsA deletion strain cultures exhibited a long-chain phenotype, indicating that the peptide is able to exert a dominant-negative effect on the cell wall, similar to the episomal expression of CpsA-245, leading to the bacteria forming long chains of cocci (37). Staining was most intense at the midcell of the bacteria and the region of the division septae, suggesting that peptide interactions are located at the division septum, potentially as part of the cell division machinery. When FITC-peptide was incubated with cells in the absence of growth, no binding was visualized (data not shown), indicating that actively growing cells are required for peptide binding.

FIG 6.

Effect of peptide on chain length. Phase-contrast microscopy of overnight cultures of wild-type bacteria grown in the presence of protease inhibitors in the absence of CpsA peptide (A) or with 200 pmol of CpsA peptide/ml (B) was performed. Scale bar, 10 μm.

FIG 7.

Localization of FITC-labeled CpsA peptide. Fluorescent micrographs show wild-type (A) and ΔcpsA mutant (B) bacteria grown overnight in the presence of FITC-labeled peptide. Scale bar, 10 μm.

To determine whether the binding of FITC-peptide to the bacteria was specific, competition experiments were performed. Bacteria were grown in the presence of the FITC-labeled peptide, or labeled peptide plus an excess of unlabeled peptide or unlabeled scrambled peptide. Relative fluorescence units (RLU) were determined using a 96-well fluorometer and compared between the two conditions. The wild-type culture and the cpsA deletion culture with both labeled peptide and a 25-fold excess of unlabeled peptide exhibited (3.05 ± 0.21)-fold and 3.42 ± 1.22)-fold less fluorescence (expressed as means ± the standard errors of the mean [SEM]), respectively, compared to the fluorescence of the cultures grown with only the labeled peptide (P < 0.01). When grown with a 25-fold excess of unlabeled scrambled peptide, the fluorescence levels were changed by (1.12 ± 0.26)-fold and (1.07 ± 0.33)-fold, respectively, compared to the fluorescence of cultures grown with labeled peptide alone, demonstrating that the scrambled peptide (same amino acids in a different order) does not compete with the FITC-labeled peptide.

The role of the CpsA protein in virulence in a zebrafish sepsis-meningoencephalitis model was also examined. Fish were injected intramuscularly and assayed for bacterial dissemination to the heart, spleen, and brain at 24 h postinfection (Fig. 8). The cpsA deletion strain was attenuated in the ability to disseminate to the heart and brain, while the cpsA insert strain was attenuated in the ability to disseminate to all three major organs. Moreover, the ability of the cpsA insert strain to disseminate to the heart and brain was reduced from that shown for the cpsA deletion, further supporting the key detrimental effect of the aberrant function of the accessory domain of CpsA in the virulence of GBS, allowing dissemination away from the initial site of infection into the bloodstream and crossing the blood-brain barrier.

FIG 8.

GBS dissemination to zebrafish major organs. Viable CFU in the spleens, hearts, and brains of individual zebrafish were measured 24 h after intramuscular injection of 10⁶ CFU of GBS. Each point represents one fish; each bar indicates the mean. Error bars represent the standard deviations. *, P < 0.05; **, P < 0.01 (different from wild type). n = 14 to 18 fish per strain.

DISCUSSION

LCP (LytR-CpsA-Psr) family members are found in a variety of Gram-positive bacteria and have been credited with a wide range of functions and phenotypes. However, functional mechanisms for these proteins are still as yet undetermined. Recently, LCP proteins were implicated in the attachment of cell wall carbohydrates, capsule, and teichoic acids to the cell wall (36, 40, 41). Cell wall carbohydrate precursors are made processively on the inner and then the outer leaflet of the bacterial membrane. The carbohydrate precursors are anchored to lipid carriers during this process. LCP family members are hypothesized to act enzymatically to transfer the precursors off the lipid carrier and onto the peptidoglycan cell wall with some substrate specificity (36, 40, 41). A Staphylococcus aureus triple mutant in three LCP family members has no wall teichoic acids, releasing teichoic acid instead into the culture supernatant (42). A double mutant of Streptococcus pneumoniae in two LCP family members releases capsule into the culture supernatant (40). Mutation of multiple LCP family members can be lethal (43) or cause growth deficits unless compensatory mutations of other genes also occur (36, 40, 44, 45).

The CpsA protein is a member of the LCP protein family and is highly conserved in all GBS strains and serotypes. As mentioned above, the CpsA protein contains an additional extracellular domain, called the accessory domain, that is not found in the other LCP family members. Notations in the genomic databases identify this domain by its homology to the DNA polymerase processivity factor (DNA_PPF) of phage and eukaryotic DNA polymerases. However, the homology is quite limited, and this domain is found on the outside the cell and thus not likely to play a role in DNA replication. Analysis of the crystal structure of the extracellular domains of the S. pneumoniae Cps2A protein reveals that the accessory domain folds independently of the LCP domain (36). There is high homology between the pneumococcal and GBS CpsA proteins (50% identical and 69% similarity at the amino acid level), suggesting a conserved topology.

Previous work with ectopic expression of extracellularly truncated GBS CpsA protein identified a dominant-negative effect on capsule expression when the CpsA-245 protein (containing only the accessory domain in the absence of the LCP domain) was expressed in the wild-type strain (37). In addition, ectopic expression of this truncation caused abnormal cell wall phenotypes, leading to the production of long chains of cocci. However, episomal expression of the CpsA-full protein or the CpsA-117 protein did not cause a dominant-negative effect on capsule or cell wall phenotypes. These results indicate that the accessory domain, in the absence of the LCP domain, resulted in aberrant function with a decrease in capsule production and cell wall integrity. Serial truncations constructed in the present study identified a region between amino acids 210 and 245 of the accessory domain that are key to the dominant-negative effect, with another possible region between amino acids 132 and 153. The CpsA-210 and CpsA-187 intermediate truncations that do not have a dominant-negative effect when ectopically expressed in the wild-type strain could indicate that these regions are key to the improper folding and malfunctioning of the accessory domain leading to the aberrant interactions with the binding partners of CpsA and dominant-negative effect.

A dominant-negative effect can occur when a mutant copy of a protein interferes with the function of the wild-type protein. This can be accomplished through direct interaction of the mutant copy with the wild-type protein, thereby blocking function, or through the interaction of the mutant copy of the protein with the normal binding partners or substrates required for proper function of the wild-type protein. In the case of the CpsA protein, we demonstrated that the dominant-negative effect occurs through expression of the protein episomally from within the cell or exogenously by adding a purified truncated CpsA protein or synthetic CpsA peptide to the extracellular environment. The demonstration of the fluorescent peptide binding to both the wild-type and the cpsA deletion bacteria indicates that the dominant-negative effect is not produced by direct interaction with wild-type CpsA protein. The ability of the dominant-negative effect to act extracellularly, without requiring cellular expression of the mutant protein supports a scenario in which the accessory domain/peptide interferes with mechanisms required for the normal cell wall integrity and capsule production.

Disruption of LCP family proteins leads to cell wall and cell division abnormalities in multiple bacterial species. In Streptococcus mutans, mutation of the LCP family member brpA led to abnormal cell division events and atypical cell wall morphology (43, 46). S. mutans lytR mutants also have dysregulated cell division, leading to the formation of long chains of cocci (47). Staphylococcus aureus triple lcp mutants have unusually thin cell walls and aberrantly located division septae (42, 45) and are hypersensitive to the cell wall targeting antimicrobials bacitracin (44) and oxacillin (48). Mutation of the S. aureus msrR gene, encoding an LCP family member, causes cells to be unusually large and have misplaced division septae (49). Double LCP mutants in S. pneumoniae grew more slowly and had cell division septae placed in aberrant locations, leading to cocci of uneven size (40), as did mutations in the pneumococcal lytR protein alone (50). Mutation of the Enterococcus hirae psr gene led to increased sensitivity to the cell wall targeting antimicrobial, lysozyme (51). Together, these studies indicate that LCP family members play key roles in stability of the Gram-positive cell wall and in proper cell division.

Our data similarly show disruption of the GBS cell wall and cell division in cpsA mutant strains or through binding of the CpsA peptide to the wild-type strain causing the formation of long chains of cocci. Visualization of the FITC-labeled peptide that contains 24 amino acids from the C-terminal end of the CpsA accessory domain binding to the midcell region of the bacteria, potentially at the division septae, is the first time that GBS CpsA has been shown to interact physically with the cell division process. Midcell septal localization was also recently demonstrated for LCP family members of S. pneumoniae (40), suggesting a common role for these proteins. In GBS we show that this interaction results in wild-type strains forming abnormally long chains of cocci. This phenotype occurs with expression of the accessory domain in the absence of the LCP domain, either through episomal expression of truncated CpsA or in the case of the cpsA insert strain, which has expression of a chromosomal truncated CpsA. These results suggest that either the LCP domain is required for proper function or the accessory domain is participating in aberrant interactions preventing proper function of the protein. This provides an explanation as to why the episomally expressed CpsA protein truncations containing all or part of the accessory domain (cpsA-245, -210, -187, -153, and -132) cannot complement for capsule expression when expressed in the cpsA deletion strain (Fig. 3). However, episomal expression of the CpsA protein truncation that removes the entire extracellular domain (CpsA-117) is able to complement for capsule expression when expressed in the cpsA deletion strain. Collectively, these results demonstrate that the intracellular and transmembrane domains alone of CpsA are able to complement the deletion in vitro, but only in the absence of the accessory domain.

The CpsA protein of S. pneumoniae, as well as the zoonotic fish pathogen Streptococcus iniae, has been implicated in regulation of expression of the cps locus genes and in virulence (38, 39, 52–56). Proper regulation of cps gene expression is also critical for invasive infection by S. pneumoniae. Ectopic expression of the S. pneumoniae cps locus from a constitutively strong or weak promoter caused attenuation of colonization and virulence in animal models (57). S. pneumoniae strains that are more invasive have higher expression of cpsA in in vitro growth, indicating that cpsA has a role in virulence in invasive infections (58). Similarly, an S. iniae strain with an insertion in the cpsA gene had reduced capsule expression and reduced lethality and ability to disseminate to major organs in a zebrafish model of systemic disease (39).

The GBS cpsA insertion strain analyzed in the present study carries a nonpolar chromosomal insertion in the cpsA gene, resulting in expression of a truncated CpsA containing the accessory domain but missing the LCP domain. This is different from the expression of the CpsA-245 construct in the cpsA deletion strain, because it is in multicopy, whereas the cpsA insert is on the chromosome and is expressed from the endogenous promoter. As shown for S. iniae and S. pneumoniae, GBS CpsA mutants are also attenuated for virulence. We demonstrated that the cpsA insertion strain produces less capsule than the cpsA deletion strain and is more attenuated in dissemination to major zebrafish organs than the cpsA deletion strain, indicating expression of a truncated CpsA is even more detrimental to GBS virulence than the complete absence of CpsA. This suggests that the aberrant function of the CpsA accessory domain, either alone or in the presence of the wild-type CpsA protein, plays an important role in virulence. This could be at the level of capsule dysregulation, or cell wall instability could be causing increased bacterial lysis in vivo, or a combination of both.

Although this truncation of the CpsA protein is causing a function that we assume is aberrant for the normal full-length CpsA, conversely, the accessory domain may have a function in vivo of controlling downregulation of capsule production when in particular host environments. Specific conformations of the full-length protein may allow the accessory domain to interact with CpsA binding partners in a negative way, thereby causing inhibition of capsule production and cell wall synthesis/division. The recent crystal structure of the S. pneumoniae Cps2A revealed a region that could not be modeled, containing a disordered loop (36, 40). In comparing this structure to the GBS CpsA, we find that the disordered loop region is in the most C-terminal region of the accessory domain, the region responsible for the dominant-negative effect of the CpsA-245 protein. Therefore, this region of the protein may take on multiple conformations in response to sensing of the local environment, allowing positive or negative interactions with components of the capsule production/cell wall machinery.

The ability to produce the dominant-negative effect from extracellular addition of the CpsA peptide has obvious implications for treatment and/or prevention of GBS infection. Since GBS treated with the dominant-negative factor have less capsule, this would allow clearance by the host immune system. Furthermore, these bacteria also have aberrant cell walls potentially making them more susceptible to lysis in vivo. Research is currently in progress to further exploit and optimize these phenotypes (M. N. Neely and H. M. Rowe, U.S. patent application 62/066,215). Taken together, the present study provides additional information on the roles of the extracellular domains of GBS CpsA in capsule expression, cell wall integrity, and the virulence of GBS. This supports a role of LCP family members as serving key functions in regulation of cell envelope polysaccharide expression and cell wall integrity and as virulence factors of Gram-positive bacteria.