PBP1a-Deficiency Causes Major Defects in Cell Division, Growth and Biofilm Formation by Streptococcus mutans

Streptococcus mutans, a key etiological agent of human dental caries, lives almost exclusively on the tooth surface in plaque biofilms and is known for its ability to survive and respond to various environmental insults, including low pH, and antimicrobial agents from other microbes and oral care products. In this study, a penicillin-binding protein (PBP1a)-deficient mutant, strain JB467, was generated by allelic replacement mutagenesis and analyzed for the effects of such a deficiency on S. mutans’ stress tolerance response and biofilm formation. Our results so far have shown that PBP1a-deficiency in S. mutans affects growth of the deficient mutant, especially at acidic and alkaline pHs. As compared to the wild-type, UA159, the PBP1a-deficient mutant, JB467, had a reduced growth rate at pH 6.2 and did not grow at all at pH 8.2. Unlike the wild-type, the inclusion of paraquat in growth medium, especially at 2 mM or above, significantly reduced the growth rate of the mutant. Acid killing assays showed that the mutant was 15-fold more sensitive to pH 2.8 than the wild-type after 30 minutes. In a hydrogen peroxide killing assay, the mutant was 16-fold more susceptible to hydrogen peroxide (0.2%, w/v) after 90 minutes than the wild-type. Relative to the wild-type, the mutant also had an aberrant autolysis rate, indicative of compromises in cell envelope integrity. As analyzed using on 96-well plate model and spectrophotometry, biofilm formation by the mutant was decreased significantly, as compared to the wild-type. Consistently, Field Emission-SEM analysis also showed that the PBP1a-deficient mutant had limited capacity to form biofilms. TEM analysis showed that PBP1a mutant existed primarily in long rod-like cells and cells with multiple septa, as compared to the coccal wild-type. The results presented here highlight the importance of pbp1a in cell morphology, stress tolerance, and biofilm formation in S. mutans.

Introduction actively growing cultures were diluted 1:100 in fresh medium, and growth in culture optical density (OD 600nm ) was continuously monitored using a Bioscreen C (Oy Growth Curves AB Ltd, Finland) at 37°C with and without sterile mineral oil overlay.

Construction of deficient mutant and complement strain
To construct a PBP1a-deficient mutant, a PCR-ligation-mutagenesis strategy was used as previously described to remove nucleotides 60-2066 within the coding sequence [29,30]. Briefly, the 5' and 3' regions flanking pbp1a were amplified by PCR using Phusion high-fidelity DNA polymerase (New England Biolabs, Ipswich, MA) with gene-specific primers shown in Table 1. Following proper restriction enzyme digestions, the flanking regions were ligated to a nonpolar kanamycin resistant element (aphA encoding aminoglycoside 3'-phosphotransferase) [31] that was digested similarly. The resulting ligation mixtures were used to directly transform S. mutans UA159 in the presence of competence stimulating peptide (CSP) [32]. An allelic replacement mutant, JB467 with pbp1a deficiency, was isolated on BHI-Kan plates and further confirmed using PCR and sequencing. For complementation of the JB467, the pbp1a-coding sequence plus its cognate promoter region were amplified by PCR, digested, and cloned into shuttle vector pDL278. After sequence confirmation, the resulting construct, pDL-pbp1a, was used to transform JB467, generating complement strain, JB467C.

Biofilm analysis
S. mutans biofilms were grown in modified biofilm medium with glucose (20 mM, BMG), sucrose (10 mM, BMS), or glucose (18 mM) and sucrose (2 mM) (BMGS) as the supplemental carbohydrate sources as previously described [33,34]. Overnight cultures were inoculated 1:100 into fresh BHI and allowed to grow until ODffi 0.5, and then the biofilm media was inoculated by 1:100 dilutions. Biofilm mass on 96-well plates (Corning, NY) was measured spectrophotometrically after crystal violet staining [35]. For biofilm structural analysis, hydroxylapatite (HA) disks, commonly used as in vitro tooth model, were also used and placed vertically in fabricated orthodontic wire holders fit for 24-well plates [36], and biofilms were analyzed using scanning electron microscopy (SEM). Glycolytic pH drop, acid killing and hydrogen peroxide challenge assays The impact of pbp1a deficiency on the ability of S. mutans to withstand acid and oxidative stress were determined by using procedures described elsewhere [29,37,38]. The glycolytic pH drop assay was carried out as described by Belli et al. [38]. Briefly, S. mutans strains were grown in BHI until mid-exponential phase (OD 600nm ffi0.5), washed twice with ice cold deionized water by centrifugation at 4,000 rpm at 4°C for 10 minutes, and the cells were then resuspended in 50 mM KCl and 1 mM MgCl 2 . The pH was adjusted to 7.2 with 0.1 M KOH before the addition of 50 mM glucose. The pH drop was monitored continuously for a period of 30 minutes. For acid tolerance response, middle-exponential phase culture (ODffi0.3) were properly washed and then incubated in glycine buffer, 0.1 M, pH 2.8 for periods of 30, 45 and 60 minutes [37,39]. To evaluate superoxide stress susceptibility, paraquat (also, methyl viologen, Sigma, St. Louis, MO) was prepared freshly and added to the growth medium, BHI, and the growth impact was monitored using Bioscreen C [29]. For hydrogen peroxide killing, bacterial cells were prepared similarly as above, and then incubated in glycine buffer, 0.1 M, pH 7.0 containing 0.2% (w/v, or 58 mM final conc.) hydrogen peroxide for periods of 90 and 110 minutes as detailed elsewhere [37,40].

Field Emission-SEM
For field emission-scanning electron microscopic (FE-SEM) analysis, S. mutans overnight cultures were diluted 1:100 in either BMG, BMGS or BMS on HA discs placed horizontally in 24-well plates as previously described [29,36]. Briefly, after 24 hours of growth, the HA disks were rinsed briefly with PBS, pH 7.4, and then fixed in 2.5% glutaraldehyde (Polysciences, Warrington, PA) overnight at 4°C. The fixed samples were dehydrated using increasing concentrations of ethanol, and then dried at the critical point of CO 2 (Electron Microscopy Sciences, Hatfield, PA). A small amount of carbon was sputtered onto the samples to avoid charging in the microscope. Microscopy was performed with a Hitachi S-4800 High-resolution microscope (Tulane University, New Orleans, Louisiana).

TEM analysis
Transmission electron microscopic (TEM) analysis was carried out similarly as described previously [29]. Briefly, S. mutans strains were grown in BHI until mid-exponential phase, harvested by centrifugation at 4,000 rpm, 4°C for 10 minutes, washed twice with PBS, and then fixed in 2% paraformaldehyde/2.5% glutaraldehyde (Polysciences, Warrington, PA) in PBS for 1 h at room temperature. Cells were washed in PBS and post-fixed in 1% osmium tetroxide (Polysciences) for 1 h. Samples were then rinsed extensively in dH 2 0 prior to en bloc staining with 1% aqueous uranyl acetate (Ted Pella Inc., Redding, CA) for 1 hour. The cells were then washed with dH 2 0 and dehydrated in a graded series of ethanol and embedded in Eponate 12 resin (Ted Pella Inc., Redding CA). Sections of 90-100 nm were prepared, stained with uranyl acetate and lead citrate, and viewed under a JEOL 1200 EX transmission electron microscope (JEOL USA Inc., Peabody, MA).

Autolysis assay
For autolysis assay, S. mutans strains were grown in BHI overnight. Bacterial cells were harvested by centrifugation at 4,000 rpm, 4°C for 10 minutes, washed once with phosphate buffered saline (PBS, pH 7.2), and then resuspended in PBS with and without inclusion of Triton X-100 (0.2%, v/v, Sigma) and adjusted to similar OD 600nm [35]. The cell suspension was then incubated at 37°C, and autolysis in reduction of OD 600nm was monitored automatically using Bioscreen C every 30 minutes with moderate shaking (10 seconds) before measurement.

Statistical Analysis
Quantitative data were analyzed using the paired Student's t-test. A difference at P value equal to or less than 0.05 is considered statistically significant.

PBP1a deficiency causes major defects in growth
A PBP1a-deficient S. mutans strain, JB467, was constructed by replacing nucleotides 59 to 2077 relative to the translational start site of pbp1a with a nonpolar kanamycin-resistance cassette by allelic replacement. When compared to the wild-type, UA159 during growth in BHI, the PBP1a-defcient mutant, JB467 had a reduced growth rate with an averaging doubling time of 135(±13) minutes, compared to 93(±3) minutes for the wild-type (P<0.01). Relative to the wild-type, the PBP1a mutant also had a reduced culture OD 600nm , with the average maximal OD 600nm of the overnight wild-type being 0.779 vs 0.586 for the PBP1a mutant (P<0.005) ( Fig 1A). Besides, the mutant also had higher tendency to form aggregates at the bottom of the culture tube leaving a clear growth medium broth (Data not shown). A PBP1a-complement strain, JB467C, which carries the wild-type copy of the pbp1a gene plus its cognate promoter region, was constructed using shuttle vector pDL278 (Table 1). Complementation in JB467C was able to restore the slow-growth phenotype of PBP1a-deficient mutant to that similar to wild-type UA159 with an averaging doubling time of 92±2.5 minutes and an OD 600nm maximum of 0.781 when grown under similar conditions ( Fig 1A). The complementation in trans also restored the aggregation characteristics to the wild-type.

PBP1a-deficient mutant is more susceptible to acid and oxidative stresses
In an effort to discern the effects of PBP1a-deficiency on the ability to tolerate stresses, the PBP1a-deficeint mutant, along with the wild-type and the complement strain, were grown in BHI adjusted to different pHs. At pH 6.2, S. mutans wild-type UA159 grew at a slower rate than that at neutral pH, as expected ( Fig 1B). However, relative to the wild-type, the PBP1adeficient mutant further reduced its growth rate, with an averaging doubling time of 232   1B). The PBP1a mutant completely stopped to grow when incubated in BHI adjusted to pH 8.5 (Fig 1C). Similarly, the complementation in trans was able to partly restore the phenotypes to those similar to the wild-type. Reduction of growth rate during growth in BHI adjusted to lower pH indicates a reduced glycolytic activity and/or weakened acid tolerance. When analyzed using glycolytic pH drop experiments, an assay commonly used to analyze the proton permeability [38], it was shown that relative to wild-type, the rate (slope) of glucose-induced pH drop of the PBP1a-deficient mutant was significantly reduced and the resting pH of the mutant was significantly higher (P<0.01) (Fig 2). As expected, complementation of the mutant with the wild-type coding sequence plus its cognate promoter region in JB467C was able to restore these phenotypes to levels similar to the wild-type. When incubated in glycine buffer of pH 2.8, the PBP1a-deficient mutant, JB467, showed an increased sensitivity of more than 1-log than the parent strain, UA159 (P<0.001) after 30 minutes (Fig 3A). The complement strain showed a level of sensitivity that is very similar to the wild-type after 30 and 45 minutes, although had a slightly higher survivability than the parent strain after 60 minutes.
To assess the impact of the PBP1a-deficiency on the ability to tolerate oxidative stress, paraquat, a redox cycling agent, was included in the growth medium to generate intracellular superoxide. Relative to the wild-type, the PBP1a-deficient mutant, JB467 displayed apparent effects in both the growth rate and its cell yield in culture OD 600nm during growth in medium with paraquat a level as low as 0.1 mM (Fig 4A). As the concentration of paraquat was increased, the PBP1a-deficient mutant was further exacerbated (Data not shown) and almost completely eliminated growth 5 mM (Fig 4B). Complementation of pbp1a in trans in JB467C was able to restore the paraquat-sensitive growth phenotype of JB467 to the level very similar to the wildtype, UA159. When challenged with hydrogen peroxide, the PBP1a-deficient mutant was about 1-log more sensitive than the wild-type (P<0.001) after 90 minutes of incubation ( Fig 3B). Complementation of the deficient mutant with the wild-type copy of pbp1a plus its cognate promoter region in multiple copy shuttle vector, pDL278, was able to only slightly increase the resistance.

PBP1a-deficiency increases autolysis rate
When resuspended in PBS with inclusion of Triton X-100, the PBP1a-deficeint mutant, JB467 reduced its density a lot faster than the parent strain, UA159, and after 24 hours, reached a density level significantly lower than the parent strain (P<0.01) (Fig 5). Similar results were also observed when resuspended in PBS without Triton X-100 (Data not shown).

PBP1a deficiency abolishes biofilm formation
Unlike the parent strain, PBP1a-deficient JB467 was unable to form any biofilms when grown vertically on glass slides (data not shown). Therefore, 96-well plates were used to grow S. mutans strains UA159, JB467, and JB467C in biofilm medium with glucose (BMG), sucrose  (BMS), and glucose plus sucrose (BMGS), and biofilms accumulated at 24 and 48 hours (Fig 6) were measured using a spectrophotometer following crystal violet staining. As expected, wildtype UA159 developed robust biofilms under all different growth conditions tested, especially when grown in medium containing sucrose, which is known to yield copious adhesive glucans [4]. In contrast, however, the PBP1a-deficient mutant, JB467 failed to grow and accumulate any sufficient biofilms in BMG, BMS, or BMGS in 96-well plates (Fig 6A). The most significant differences between UA159 and JB467 were measured when sucrose was included in the growth medium. As shown by FE-SEM analysis with biofilms grown on HA disks (Fig 6B), drastic differences in both quantity and structure of the biofilms were observed between the wild-type and the PBP1a mutant. The wild-type biofilms appeared multi-layered and evenly distributed, while PBP1a mutant had only sparse cell clusters with little or no accumulation of biofilms ( Fig 6B). When compared to the wild-type biofilms, the biofilms of the PBP1a mutant also possessed significantly bigger and seemingly more eDNA nanofibers [2] (Fig 6B). Besides, alterations in cellular morphology of the PBP1a-deficient mutant were also apparent, when compared to wild-type UA159 and the complemented strain JB467C (see below for more details).

PBP1a deficiency causes morphological alterations
As seen by TEM analysis in Fig 7, PBP1a-deficiency in JB467 caused some major alterations in cellular architecture of the deficient mutant. Unlike the coccoid cells of the wild-type, cells of the PBP1a-deficient mutant were mostly elongated, long rods and unevenly sized, and were characterized with increased interseptal distance, incomplete septum, and presence of the multipartite septa (Fig 7A). Similar morphological features were also observed in biofilms under FE-SEM (Fig 6B). FE-SEM showed that the wild-type was present mostly in short chains of coccoid cells, while the PBP1a mutant displayed mostly in chains of unevenly sized long rods (Fig 6B). In addition, the deficient mutant also appeared to have a significantly thinner PG  PBP1a-deficiency significantly decreased biofilm formation by the deficient mutant (* indicates differences at P<0.001 when compared to the wild-type, UA159 under the same conditions). Unlike the robust, evenly distributed biofilms of the wild-type, UA159, biofilms of PBP1a mutant, JB467 were sparse and featured with elongated giant cells. In addition, the biofilms of the PBP1a mutant also possessed significantly bigger and seemingly more eDNA nanofibers (indicated by arrows). FE-SEM micrographs were taken at 20k× magnification. PBP1a in Streptococcus mutans Biofilm Formation layer when compared to the parent strain with an average thickness of 21.2±1.12 nm for the mutant vs 28.4±3.2 nm for the wild-type (P<0.01) (Fig 7B). Complementation with pbp1a in trans in JB467C was able to partially restore the phenotypes of JB467 to those similar to the wild-type. The complement strain had an average PG layer of 35.52±4.23 nm.

Discussion
Like other Gram-positive bacteria, S. mutans possesses a thick layer of peptidoglycan whose homeostasis is known to be crucial in pathophysiology including cell division and growth, stress tolerance, and biofilm formation. Recent studies have shown that biofilm regulatory protein BrpA, a member of the LytR-CpsA-Psr family of proteins widespread in Gram-positive bacteria, plays a major role in cell envelope homeostasis, influencing cell envelope stress response and biofilm formation [35,40,41]. Like BrpA, penicillin-binding protein PBP1a possesses a C-terminal region with rich serine residues that appear to be unique to S. mutans [27]. Here, we present evidence that PBP1a also plays a critical role in S. mutans cellular biology and affects its abilities to tolerate acid and oxidative stress and form biofilms, traits critical to the pathogenicity of this bacterium. PBP1a-deficiency causes severe defects in cell division and cell morphology of the mutant, significantly decreases the growth rate and weakens the ability of the deficient mutant to tolerate acid and oxidative stressors, and reduces biofilm formation. However, it is also worthy noting that not all phenotypes of the mutant analyzed were restored fully to the wild-type level by complementation of the deficient mutant with wild-type coding sequence plus its cognate promoter region. Part of the reason could be related to the nature of the complementation system being multiple-copy that is associated with the shuttle vector, pDL278 used (see more details below).
Like S. pneumoniae, PBP1a-deficiency in S. mutans is viable when grown in rich medium BHI and semi-defined biofilm medium [29,34], but significantly affects its growth rate and significantly reduces cell yield in culture OD 600nm , especially at acidic and alkaline pH environment. The weakened capacity to tolerate low pH, as shown by acid-killing assay, and the compromised abilities to deal with toxic oxygen metabolites such as hydrogen peroxide and superoxide inside and outside the cell, as revealed by paraquat and hydrogen peroxide challenge assays, all are likely contributors to the reduced ability to grow as reflected by the reduced growth rate and culture density.
Unlike other ovococci bacteria, S. mutans displays heterogeneity among its strains in terms of the length to width ratios, which was shown to be largely due to changes in the K + /bicarbonate in the growth medium [42]. Like in S. pneumoniae and other bacteria [15,43], our data also further supports the notion that PBP1a in Gram-positive bacteria plays a role in surfaceassociated cell separation. Unlike the wild-type that mostly exists in chains of ovococci or short rods, the PBP1a-deficient mutant displays mostly in chains of long, mis-shaped rods with incomplete septa and multiple septa. One streptococcal virulence factor that leads to the establishment of biofilms is the innate ability to form chains of bacteria [25]. Likely, the altered cell morphology, especially the long chaining, swelling and aggregation during growth in broth, plays a major role influencing bacterial adherence and biofilm accretion on a surface. PBPs are known to play a crucial role in peptidoglycan synthesis [15,43]. In E. coli, PBP1a is shown to be capable of polymerizing lipid II to form a cross-linked peptidoglycan product that contains glycan strands with an average length of~20 disaccharide units and with~18-26% of peptides present in cross-links [15]. In the presence of purified peptidoglycan sacculi, PBP1a is capable of attaching a fraction of the newly assembled peptidoglycan to the sacculi via transpeptidase reaction. Consistently, deficiency of PBP1a in S. mutans leads to development of cells with thinner layer of cell wall.
S. mutans is known to produce copious extracellular glucose polymers via Gtf enzymes when grown in sucrose rich medium. The Gtf enzymes can also bind to glucans and surfaces of other bacterial species [4,44], and thus play a central role to S. mutans dominance in cariogenic plaque and its cariogenicity [4]. When analyzed on 96-well plates, the most drastic difference in biofilm formation between the PBP1a-deificient mutant and the wild-type was measured during growth in the presence of sucrose. As detailed above, PBP1a-deficiency renders major growth defects of the deficient mutant, which is likely the major factor that underlies the reduced biofilm formation by the mutant. It is also possible, but awaits further investigation, however, if the Gtfs' localization and activity are affected as a result of PBP1a-deficeincy. S. mutans also releases eDNA, including active releases during growth in biofilms, especially during early biofilms [2,45] contributing to the extracellular matrices and facilitating bacterial adherence, biofilm formation and structure integrity [2,46]. Our recently studies have also shown that such release is highly regulated, although the exact mechanism of regulation remains unclear. As shown under FE-SEM, both the wild-type and the PBP1a mutant produce an eDNA network when grown on the HA discs [2]. Relative to the parent strain, the mutant appeared to have more and bigger eDNA nanofibers, although the impact of such eDNA presence and the underlying mechanism remain unclear.
Biofilm formation is a sequential process that is initiated with bacterial adherence to a substratum and followed by formation of microcolonies and accumulation of multilayered cell clusters. During the process, cell envelope plays essential role in not only cell division and multiplication, but also in cell-surface and cell-cell interactions that is known to be crucial in biofilm initiation and development. It is apparent that both chain length and cell shape are aberrant in the PBP1a-deficient mutant. Additionally, defects in remodeling and/or maintenance of the peptidoglycan as a result of PBP1a-deficiency will likely compromise the integrity of the cell envelope and render the cells to be more permeable to protons and other toxic metabolites, which could in part explains the elevated susceptibility to low pH and the superoxide-generating reagent paraquat and hydrogen peroxide. Autolysis, also programmed cell death, is a highly regulated process that can be triggered by stress responses, such as starvation, viral infection and DNA mutations [47]. Unlike the wild-type, the PBP1a mutant displayed a quick and major reduction in OD 600nm when incubated in buffered saline with and without Triton X-100, a detergent commonly used in laboratories to permeabilize the membrane of living cells. Such a reduction also suggests that PBP1a mutant may have lost its proper cell envelope protection, a likely result of compromised integrity in response to PBP1a-deficiency. Compromises in cell envelope and cell envelope integrity can be at least in part attributed to the weakened stress tolerance responses observed with the PBP1a-deficient mutant, which in turn influences biofilm formation. In addition, compromises in cell envelope could also influence the localization and function of surface-associated proteins, including the Gtfs, Gbps and SpaP known to play a central role in S. mutans biofilm formation, directly affect bacterial adherence, inter-cellular interactions and biofilm accumulation, contributing to the reduced biofilms of the PPBP1a-deficient mutant.
Similar to many other Gram-positive bacteria, S. mutans possesses at least six genes encoding proteins homologous to PBPs, which besides PBP1a (SMU.467), also include one PBP1b (SMU.1991), one PBP2x (SMU.455), one PBP2a (SMU.1949), and two PBP2b (SMU.597 and SMU.889) [27]. It is plausible that the PBP redundancy might suggest that optimal conditions exist in S. mutans for peptidoglycan biosynthesis and/or homeostasis by different PBPs. As suggested by results of BLAST search, S. mutans pbp1a gene is most closely linked to genes encoding PBP1a of Streptococcus agalactiae (GBS) and other streptococci. In GBS, surface-associated PBP1a, encoded by ponA, plays an essential role in resistance of GBS to phagocytosis in a neonatal rat sepsis model, and PonA-deficiency led to an increased susceptibility to antimicrobial peptides [48]. In S. pneumoniae, alterations of PBP1a are often associated with high level of βlactam resistance, such as Penicillin G and Cefotaxime [16]. However, it awaits further investigation how deficiency of PBP1a in S. mutans affects antibiotic resistance. It is also known that major differences exist between the Gram-positive and Gram-negative PBPs, suggesting compensatory mechanisms among the PBPs are likely different. Therefore, it is also possible that overexpression of PBP1a in the complement strain that carries a wild-type pbp1a in multiplecopy shuttle vector can cause imbalance between the different PBPs, contributing to the failure to fully restore the phenotypes analyzed to the wild-type. Further studies on the functions of the other PBPs and the interplay of the 6 different PBPs in S. mutans should provide insights on the underlying mechanisms and other related aspects.