Comparative characterization reveals conserved and divergent ecological traits of oral corynebacteria

TL;DR

This study explores how different types of oral corynebacteria form biofilms and interact with other microbes, revealing some are key biofilm organizers while others are less specialized.

Contribution

The study introduces two new corynebacterial strains and reveals conserved and divergent traits in biofilm formation and interspecies communication.

Findings

C. durum strains form thick biofilms with extensive EPS networks, while C. argentoratense forms thinner biofilms with minimal EPS.

All tested strains secrete extracellular membrane vesicles that induce chain elongation in Streptococcus sanguinis.

C. argentoratense lacks nitrate reductase genes and cannot grow anaerobically, suggesting a less specialized oral niche.

Abstract

Corynebacteria are abundant members of the oral microbiome and increasingly recognized as key structural organizers of supragingival biofilms. Despite their prevalence, the ecological roles and phenotypic traits of many oral corynebacterial species remain poorly defined. Here, we isolated and characterized two new strains, Corynebacterium durum JJ2 and Corynebacterium argentoratense MB1, and compared them with previously characterized and published Corynebacterium durum JJ1 and reference strain Corynebacterium matruchotii ATCC 14266. Phenotypic assays revealed that C. durum strains displayed robust aggregation, thick biofilm formation, and extensive extracellular polymeric substance (EPS) networks, whereas C. argentoratense MB1 and C. matruchotii ATCC 14266 formed thinner biofilms with minimal EPS production. All four strains secreted extracellular membrane vesicles capable of inducing chain elongation in Streptococcus sanguinis, underscoring a conserved interspecies signaling function. Genomic analysis demonstrated close relatedness between C. durum and C. matruchotii, while C. argentoratense MB1 was more distantly related, with a reduced genome, fewer metabolic pathways, and the absence of nitrate reductase genes, consistent with its inability to grow under anaerobic conditions. These findings suggest that C. argentoratense MB1 may represent a less specialized or transient inhabitant of the oral cavity, whereas C. durum and C. matruchotii are well adapted to the oral niche. Together, this study expands our understanding of phenotypic diversity, metabolic capacity, and interspecies interactions among selected oral corynebacteria, highlighting their potential importance as biofilm organizers and contributors to oral microbial ecology.

Oral corynebacteria contribute to the structural and ecological stability of supragingival communities. Yet, their species-level functions remain poorly defined. By isolating and characterizing new strains of Corynebacterium durum and Corynebacterium argentoratense, and comparing them with reference strains including Corynebacterium matruchotii, we provide new insight into their phenotypic diversity, metabolic capacity, and ecological roles. Our results demonstrate that C. durum strains form robust biofilms enriched in extracellular polymeric substances, while C. argentoratense produces thinner biofilms and lacks the genomic features required for anaerobic growth, suggesting a less specialized or transient role in the oral cavity. Importantly, we show that extracellular membrane vesicles secreted by all tested strains promote chain elongation in Streptococcus sanguinis, highlighting a conserved mechanism of interspecies communication. These findings advance our understanding of how oral corynebacteria contribute to biofilm organization and microbial homeostasis and position them as critical but understudied players in oral microbial ecology.

Figures

(A) Transwell assays for S. sanguinis SK36 alone, SK36 with Corynebacterium argentoratense MB1, and SK36 with Corynebacterium durum JJ2. (B) Bar graph showing the concentration of EMVs produced by Corynebacterium species grown in ASM with either glucose or sucrose supplemented (n = 3, standard deviation). (C) Graph showing the varying diameter of EMVs produced by Corynebacterium species grown in ASM with either glucose or sucrose supplemented (n = 3, standard deviation).

(A) Average autoaggregation percentage of Corynebacterium species after 120 min of static incubation (n = 3, standard deviation). (B) Bacterial species after 120 min of static incubation. 1 = C. durum JJ1, 2 = C. durum JJ2, 3 = C. argentoratense MB1, 4 = C. matruchotii ATCC 14266, 5 = S. sanguinis SK36, 6 = PBS control (representative picture of n = 3).

(A) Average absorbance of CV-stained biofilms read at absorbance 590 nm. (n = 3, standard deviation) (B) Confocal laser scanning microscopy (CLSM) of the four Corynebacterium strains. 3D rendering performed using Imaris with a maximum intensity projection performed. *Ordinary one-way ANOVA statistical analysis showed a significant P-value <0.001.

SEM of four Corynebacterium isolates with fructose-, glucose-, and mucin-supplemented growth media at 20,000× magnification.

BHI agar plates with spotted serially diluted Corynebacterium species grown for 2 days in either aerobic (+O2) or anaerobic (−O2) chambers.

(A) A chart displaying unique characteristics to each genome. (B) Phylogenomic Tree of 6 Corynebacterial genomes, specifically the evolutionary differences between the shared single-copy core genes present within all genomes. (C) Pangenome assembly of five Corynebacterium genomes using the bioinformatics tool Anvio.

(A) Venn diagram showing the gene clusters in the five genomes used for the pangenome. (B) Number of hits for each COG category within the pangenome. (C) Chart showing unique nitrate genes in four out of five genomes (no homologs present in C. argentoratense MB1).

Tables

TABLE 1. Homology of selected open reading frames

C. argentoratense	C. durum	C. matruchotii
GC00003614Enoyl-CoA hydratase	30.81%	27.64%
GC00003643Branched-chain amino acid efflux pump	47.22%	37.02%
GC00003689Glutamyl-tRNA reductase	49%	43.64%
GC00004088Amino acid permease	No significant similarity
GC00004161Septum formation Initiator FtsB	40.67%	36%
GC00004206DNA repair protein RecO	46.18%	41.86%
GC00004250Unknown function	52.43%	No significant similarity
GC00004307Maltose 6’-phosphate phosphatase	No significant similarity
GC00004682N-acetylmuramoyl C-alanine amidase	No significant similarity
GC00004696Sensor histidine kinase	50.81%	45.49%

Full text

INTRODUCTION

The human oral microbiome is one of the most densely populated and taxonomically diverse microbial communities in the body (1). The expanded Human Oral Microbiome Database (eHOMD) currently lists over 800 microbial taxa associated with the oral cavity (2). While each harbors a subset of approximately 150–300 species, the composition of this community plays a pivotal role in determining the status of oral health and disease (3). Importantly, many oral disease-associated microorganisms do not meet the classical criteria of pathogens as defined by Koch’s postulates. Instead, these organisms, referred to as pathobionts, can transition from harmless to harmful depending upon the ecological and environmental contexts within the oral niche (4–7). For instance, Streptococcus mutans, a key contributor to dental caries, exerts its pathogenic potential primarily under conditions of frequent dietary sugar intake, which leads to acid production and subsequent enamel demineralization (8).

In contrast, commensal species are generally regarded as harmless, and in many cases, beneficial to the host. These organisms contribute to oral health through mechanisms such as colonization resistance, where a stable and diverse microbiota prevents the establishment of invading pathogens via nutrient competition, physical exclusion, and production of antimicrobial compounds (7). The classification of a species as a commensal is context-dependent. A prime example is Streptococcus sanguinis, which is typically abundant in health-associated oral biofilms where it can inhibit S. mutans and other pathobionts through hydrogen peroxide production. However, when it gains access to the bloodstream, S. sanguinis can cause serious extraoral infections, such as infective endocarditis or, more rarely, brain abscesses in immunocompromised individuals (9).

Despite their dominance in the health-associated oral biofilm and their ecological importance, oral commensals remain underexplored compared to pathobionts and other human-associated pathogens (7). The molecular strategies employed by commensals to maintain microbial homeostasis, a concept we define as molecular commensalism (7, 10, 11), have been most extensively studied in oral streptococci. Recent investigations have highlighted the cooperative and spatial interactions between streptococci and Corynebacterium spp., two of the most abundant genera in health-associated oral biofilms (11–15). Notably, Corynebacterium spp., long known primarily for pathogenic species such as C. diphtheriae, have gained attention for their role as a commensal in the oral cavity as well as the nasopharynx (11, 16, 17). For example, Corynebacterium spp. form striking “corncob” structures with streptococci in vivo, in which a central Corynebacterium rod is surrounded by chains of cocci as demonstrated with fluorescence in situ hybridization (FISH) using native human biofilm samples (18). Coculture experiments investigating the corynebacterial-streptococcal interaction in vitro have shown that C. durum induces up to 10-fold elongation of S. sanguinis chains. This effect persists even in transwell setups, indicating that direct contact is not required; instead, a diffusible factor mediates the response (15). The study also identified specific fatty acid cargo within Corynebacterium extracellular membrane vesicles (EMVs) as the responsible agents for chain elongation. These findings suggest a novel and perhaps widespread form of interspecies communication in the oral biofilm via diffusible EMVs.

The eHOMD lists 20 Corynebacterium species inhabiting oral sites, with C. durum and C. matruchotii among the most abundant (18, 19). While recent studies have begun to explore their roles in oral microbial ecology, information about phenotypes and genomic arrangements as well as cultured representatives remains limited. To address this gap, we have isolated and characterized new oral Corynebacterium strains and compared them to previously characterized and published Corynebacterium durum JJ1 (15) and reference strain Corynebacterium matruchotii ATCC 14266. Here, we present their phenotypic properties, compare their genomes to published strains, and discuss their potential ecological roles within the oral microbiome.

MATERIALS AND METHODS

Bacterial strains and media

Strains are described in Table S1. Media used for bacterial growth include Brain Heart Infusion (BHI; Becton Dickinson & Co., MD, USA) and Artificial Saliva Medium (ASM) (20). Bacteria were grown in an ambient atmosphere chamber with 5% CO_2_ at 37°C unless otherwise specified. Strain isolation: Corynebacterial strains were isolated as previously described (15, 21). Briefly, after saliva collection from volunteers, samples were placed on ice to rest prior to being centrifuged (10 min, 4,000 rpm, A-4-62 Rotor; Eppendorf 5425 Centrifuge) at 4°C. The pellets were then plated on a selective medium for oral Corynebacterium species (OCM). BHI was used as the OCM base medium supplemented with galactose and bromocresol purple to distinguish C. durum from C. matruchotii. C. matruchotii is unable to catabolize galactose (22). Two antibiotics, fosfomycin and amphotericin B, were included to inhibit the growth of other oral microbes, in particular, oral streptococci and fungi. Two new isolates were chosen for comprehensive characterizations. C. durum JJ2 and C. argentoratense MB1. The protocol for corynebacterial isolation was approved by the OHSU institutional IRB, study number STUDY00016426.

Transwell assay

To test whether the newly isolated corynebacterial strains induce chain elongation in S. sanguinis as reported before (15), transwell assays were performed. Briefly, fresh overnight cultures of bacterial cells were inoculated in the upper (C. durum JJ2, C. argentoratense MB1) and lower (S. sanguinis SK36) chambers of the well and incubated overnight at 37°C in a 5% CO_2_ atmosphere. The BHI medium for growth was shared between both bacterial strains, while the cells could not pass the 0.4 µm membrane barrier of the transwell inserts (Transwell Clear Inserts, Corning, AZ, USA). Cell morphology was then examined in comparison to SK36 in the BHI medium control using an Olympus IX73 inverted microscope. Images were acquired using the imaging software platform cellSens (Olympus) to document the elongation phenotype.

Autoaggregation assay

An autoaggregation protocol was adapted from Raju et al. (23). Briefly, bacterial cells were grown aerobically (5% CO_2_) at 37°C in BHI media for 16 h. Overnight cultures (OD_600_ >1.00) were centrifuged (10 min, 3,000 rpm, A-462 Rotor; Eppendorf 5425 Centrifuge), and the medium was removed. Next, the pellet was resuspended in 1 mL aggregation buffer (1 mM Tris, 2 mM CaCl_2_, 3 mM MgCl_2_, 150 mM NaCl, buffered to pH 7.4 with HCl) and centrifuged again. The buffer was removed, and cells were resuspended in 1 mL aggregation buffer and diluted to OD_600_ ~1.0 in 1 mL aggregation buffer in polyurethane cuvettes. Samples were incubated at room temperature for 120 min and the OD_600_ was measured in increasing intervals up to 120 min using a BioPhotometer Plus UV/Vis Photometer (Eppendorf, Germany).

Crystal violet biofilm assay

Bacteria were grown overnight in BHI and diluted to OD_600_ = 0.1 in 5 mL BHI. Using a 24-well plate, 500 µL of each culture was added to each well and left to grow for 2 days. Media was removed and washed 2× with sterile dH_2_O. 600 µL of 0.01% crystal violet (CV) was added to each well and left for 30 min to stain while rocking on a plate shaker. Next, the CV solution was removed and washed once with dH_2_O and left to dry. To solubilize, 1 mL of 70% ethanol was added to each well and left to sit for 30 min while rocking. The absorbance of each well was read at OD_590_ using BioTek Cytation 5 (Agilent, Santa Clara, CA).

Evaluation of oxygen-tension-dependent growth conditions

To determine growth under different oxygen tensions, cultures were grown overnight in BHI and diluted to an OD_600_ of 0.1. Next, the cultures were spotted in a 10-fold serial dilution on BHI agar plates and incubated at 37°C either in an ambient atmosphere chamber with 5% CO_2_ or an anaerobic chamber containing an atmosphere of 90% N_2_, 5% CO_2_, and 5% H_2_. Bacterial growth was assessed after 2 days of incubation and photographed for documentation.

Scanning electron microscopy

Cultures were grown in ASM supplemented with glucose, fructose, or mucin, at 1% wt/vol, and diluted to OD_600_ = 0.3 in ASM. 1 mL of culture was added to 13 mm Thermanox discs situated in a 24-well plate. Biofilms were grown overnight at 37°C with 5% CO_2_ supplementation. The media was removed, and biofilms were fixed with 2% glutaraldehyde in Sorensen’s buffer for 24 h at 4°C.

Biofilms were prepared and imaged at the OHSU Multiscale Microscopy Core, a member of the OHSU University Shared Resource Cores (RRID:SCR_009969). Samples were sputter coated with 10 nm thick carbon (ACE600 coater). Imaging was then performed using a Helios Nanolab 660 dual-beam scanning electron microscope (SEM; FEI).

Confocal laser scanning microscopy

Cultures were grown in BHI and diluted to 0.1 in fresh BHI. 700 µL of diluted culture was added to ibidi µ-slide 4-well chambered coverslips. After 1 day of growth, cultures were supplemented with 5 µL of 5 mM Syto 9 stain solution to result in a 0.0357 mM solution and taken to the Advanced Light Microscopy Core at Oregon Health Science University for microscopy. Subsequent images were analyzed using Imaris 3D rendering with Blend and MIP (Maximum Intensity Projection) thumbnails.

Corynebacterial API test

Newly isolated strains were characterized using the commercial API Coryne system (BioMérieux, France). In brief, this biochemical reaction system consists of dehydrated substrates for 11 enzymatic activities (nitrate reductase, pyrazinamidase, pyrrolidonyl arylamidase, alkaline phosphatase, glucuronidase, β-galactosidase, α-glucosidase, N-acetyl-β-glucosaminidase, glucosaminidase, esculin, urease, and hydrolysis of gelatin) and eight sugar fermentations (glucose, ribose, xylose, mannitol, maltose, lactose, sucrose, and glycogen). Bacterial samples were prepared and transferred to API Coryne strips according to the manufacturer’s instructions. The strips were then incubated at 37°C for 24 h. The readings, except for the esculin, urease, and gelatin tests, were performed after adding the appropriate reagents. The fermentation reactions were considered positive when they turned yellow. Identification and interpretation were conducted according to the manufacturer’s instructions. Catalase activity was determined by adding a drop of hydrogen peroxide (3%) to the esculin test after 24 h.

EMV concentration

EMV isolation and characterization followed a previously published protocol (24). Bacterial cells were grown in 25 mL ASM overnight and diluted to 1.0 in 250 mL ASM media supplemented with glucose or sucrose at 0.6% as relevant carbohydrates known to influence the oral biofilm. Cultures were left to grow overnight on a shaking incubator at 37°C and 150 rpm. Samples were centrifuged (15 min, 3,750 rpm, A-4-62 Rotor; Eppendorf 5425 Centrifuge) at 4°C. The resulting supernatant was filter sterilized with a 0.45 µm pore and then divided among Vivaspin 20 ultracentrifugation units (100 kDa MWCO, GE Healthcare) columns for further concentration. The columns were centrifuged at 4°C at increasing times of 15–45 min and 4,000 rpm (A-4-62 Rotor; Eppendorf 5425 Centrifuge). Once the supernatant was concentrated to <5 mL, the resulting solution was ultracentrifuged for 2 h at 37,000 rpm and 4°C (Beckman Coulter, Optima XL100K Ultracentrifuge; rotor type 50.3TI). The final precipitate was resuspended in 1 mL PBS and stored at −80°C for long-term storage. EMVs were quantified with Nanoparticle Tracking Analysis (NTA) using ZetaView (Particle Metrix, Germany), scanning 11 cell positions with 60 frames per position for every measurement. These positions were analyzed by ZetaView software version 8.05.12 with the following parameters: laser wavelength (488 nm), filter wavelength (scatter), maximum particle size (1,000), minimum particle size (10), and minimum particle brightness (20).

Corynebacterium pangenome and phylogenomic tree

To compare the pangenome and phylogeny of the strains examined here, Anvi'o (development version) was used to generate a pangenome of the four strains examined here, plus the additional C. matruchotii reference strain NCTC10206. All files and code used here are available at https://github.com/jonbakerlab/Corynebacterium-pangenome. The pangenome was used to select 14 single-copy core amino acid sequences with maximum sequence heterogeneity (but no gaps in the alignment) with which to perform phylogenomic analysis, also performed using Anvi'o. A C. glutamicum genome was added to the phylogenomic analysis as a non-oral outgroup; however, the analysis unexpectedly showed that C. glutamicum was more closely related to C. matruchotii than the other strains. NCBI accession numbers are as follows: C. durum JJ2: CP198957; C. durum JJ1: CP199749; and C. argentoratense MB1: CP199748.

Statistical analysis

Data were analyzed in GraphPad Prism (10.5.0). Results are shown as mean ± SD from at least three independent experiments. Group differences were assessed by one-way ANOVA with Tukey’s post hoc test. A P-value < 0.05 was considered statistically significant.

RESULTS

Isolation of oral corynebacterial species

Using an established protocol for the selective isolation of oral Corynebacterium species (15, 21), six putative isolates were obtained. 16S rRNA gene sequencing identified one isolate as C. durum with 99.25% sequence identity based on NCBI BLASTn analysis (standard database, default parameters). The remaining five isolates were identified as Corynebacterium argentoratense, showing 96%–99% sequence identity (data not shown). The isolation of C. argentoratense was unexpected, as this species is typically associated with the human skin microbiota; however, it has also been reported in the throats of patients with tonsillitis, suggesting a broader ecological niche (25). For subsequent characterization, the C. durum isolate designated as strain JJ2 and one C. argentoratense isolate designated as strain MB1 were selected.

Induction of cell chain elongation in S. sanguinis

One of the few well-characterized in vitro interspecies interactions involving oral Corynebacterium species is the induction of cell chain elongation in the commensal S. sanguinis, mediated by EMVs (15). This phenotype appears to be specific to S. sanguinis and has not been observed in other tested oral streptococci. The effect has previously been demonstrated for both C. durum and C. matruchotii (15). To determine whether the newly isolated strains also elicit this response, transwell assays were conducted with S. sanguinis SK36 and physically separated from corynebacterial species via transwell inserts. Consistent with previous findings, both C. durum JJ2 and C. argentoratense MB1 induced marked chain elongation in S. sanguinis SK36 following overnight incubation (Fig. 1A). Of note, all newly C. argentoratense strains induced cell chain elongation (Fig. S1). These results suggest that the new isolates likely produce EMVs capable of diffusing through the transwell membrane to trigger the observed phenotypic change in S. sanguinis chain formation.

(A) Transwell assays for S. sanguinis SK36 alone, SK36 with Corynebacterium argentoratense MB1, and SK36 with Corynebacterium durum JJ2. (B) Bar graph showing the concentration of EMVs produced by Corynebacterium species grown in ASM with either glucose or sucrose supplemented (n = 3, standard deviation). (C) Graph showing the varying diameter of EMVs produced by Corynebacterium species grown in ASM with either glucose or sucrose supplemented (n = 3, standard deviation).

Characterization of corynebacterial EMV production

EMVs play a critical role in mediating interactions between oral Corynebacterium species and other members of the oral microbiome. Notably, EMVs have been shown to induce chain elongation in several isolates of S. sanguinis and to inhibit hyphal formation in Candida albicans (15, 26). To further investigate whether relevant ecological factors implicated in caries development are influencing EMV production in Corynebacterium, we examined the impact of glucose and sucrose supplementation on EMV biogenesis. All four tested Corynebacterium strains produced EMVs when cultured in ASM supplemented with either glucose or sucrose. EMV diameters and concentrations were comparable between the two carbohydrate conditions (Fig. 1B and C), suggesting that the type of sugar does not significantly influence vesicle production. Among the carbohydrate conditions, C. durum JJ2 produced the largest vesicles in glucose, and C. matruchotii ATCC 14266 produced the largest vesicles in sucrose. C. matruchotii ATCC 14266 yielded the highest EMV concentrations, reaching up to 10¹³ particles per mL, which correlates with its enhanced growth and higher final cell density following overnight incubation. C. durum JJ1 and C. argentoratense MB1 yielded lower EMV concentrations and diameters across both carbohydrate conditions. Visible differences in EMV diameter are seen for C. durum JJ2 in glucose in contrast to sucrose, yet the data are not significant (Fig. 1B and C). Overall, all four Corynebacterium strains produced substantial amounts of EMVs under both carbohydrate conditions, and EMV production appeared largely unaffected by the specific sugar source. Statistical analysis using one-way ANOVA indicated that differences in EMV concentration and vesicle diameter were not statistically significant.

Comparison of Corynebacterium autoaggregation

Bacterial aggregation and autoaggregation are commonly mediated by adhesins, including fimbriae or pili, as well as specific surface proteins, as demonstrated for the streptococcal antigen I/II family and CshA/B (27, 28). In addition, secreted factors like polysaccharides and extracellular DNA can contribute to aggregation behavior (29). However, in oral Corynebacterium species, no specific proteins or molecular factors have yet been identified that mediate aggregation or autoaggregation. To assess their potential for autoaggregation, we compared the autoaggregation behavior of all four oral Corynebacterium isolates (Fig. 2A). Both C. durum strains exhibited rapid autoaggregation, with visible clumping occurring within 15 min of incubation in aggregation buffer (Fig. 2B). In contrast, C. matruchotii ATCC 14266 and C. argentoratense MB1 showed minimal aggregation, even after 60–90 min. S. sanguinis SK36 was included as a negative control and did not aggregate during the 120-min observation period. Notably, the time-dependent autoaggregation rate of C. durum JJ1 and JJ2 was approximately fivefold higher than that of the other two Corynebacterium species (Fig. 2A). We recognize that this in vitro assay may not accurately represent in vivo ecological associations and that the outcome may be media-dependent.

(A) Average autoaggregation percentage of Corynebacterium species after 120 min of static incubation (n = 3, standard deviation). (B) Bacterial species after 120 min of static incubation. 1 = C. durum JJ1, 2 = C. durum JJ2, 3 = C. argentoratense MB1, 4 = C. matruchotii ATCC 14266, 5 = S. sanguinis SK36, 6 = PBS control (representative picture of n = 3).

Biofilm quantification

Bacterial aggregation is a key mechanism that can directly impact biofilm formation (23). To assess biofilm production among the corynebacterial strains, we employed the standard CV staining assay. As shown in Fig. 3A, both C. durum strains retained more CV compared to the other species, indicating the formation of thicker biofilms. In contrast, C. argentoratense MB1 and C. matruchotii ATCC 14266 retained less CV, suggesting the formation of thinner biofilms. The observed differences in autoaggregation, with strains JJ1 and JJ2 exhibiting enhanced aggregation capacity, along with their increased biofilm formation, suggest a possible correlation between these two processes. One-way ANOVA revealed statistically significant differences when comparing both C. durum strains to C. argentoratense and C. matruchotii.

*(A) Average absorbance of CV-stained biofilms read at absorbance 590 nm. (n = 3, standard deviation) (B) Confocal laser scanning microscopy (CLSM) of the four Corynebacterium strains. 3D rendering performed using Imaris with a maximum intensity projection performed. **Ordinary one-way ANOVA statistical analysis showed a significant P-value <0.001.

CLSM imaging of 3D biofilm structure

Although CV staining enables a general quantification of biofilm biomass, it does not provide information regarding the structural organization or architecture of the biofilm. To assess the three-dimensional structure of the biofilms, CLSM was performed on all four isolates (Fig. 3B). Consistent with the CV assay, CLSM revealed striking differences in biofilm morphology: the C. durum strains formed thick, highly aggregated biofilms, whereas C. argentoratense MB1 and C. matruchotii ATCC 14266 produced thin, uniform biofilm mats. These findings indicate a consistent phenotypic similarity between the two C. durum strains across multiple assays, CV staining, autoaggregation, and CLSM, highlighting their distinct ability to form robust, structured single-species biofilms compared to C. argentoratense MB1 and C. matruchotii ATCC 14266.

SEM analysis of biofilms and EPS structures

We previously demonstrated using SEM that C. durum produces an elaborate network of extracellular polymeric substances (EPS), enmeshing cells within a dense biofilm matrix (13). To further investigate the extracellular matrix and cellular morphology of the four Corynebacterium strains, SEM imaging was performed on cells grown in ASM supplemented with different carbohydrate sources: glucose, fructose, and mucin. As shown in Fig. 4, both C. durum strains exhibited prominent extracellular matrix structures, particularly in the presence of glucose, where a dense, highly branched EPS network surrounded the cells. This network was less developed when cells were grown with fructose. Interestingly, mucin supplementation also led to the formation of extracellular material, though it appeared morphologically distinct. Notably, both C. durum strains (JJ1 and JJ2) showed similar cellular morphology and EPS features under all tested conditions. In contrast, C. argentoratense MB1 and C. matruchotii ATCC 14266 did not form extensive EPS networks in the presence of glucose or fructose. However, mucin-supplemented conditions induced the production of some extracellular material in these species, albeit to a lesser extent than in C. durum. Additionally, marked differences in cell morphology were observed: C. matruchotii ATCC 14266 formed elongated cells, while C. argentoratense MB1 exhibited much shorter, stubbier cells compared to the other three strains (Fig. 4). Together, these findings highlight distinct structural phenotypes among oral Corynebacterium species, with C. durum demonstrating a unique capacity to produce a dense EPS matrix and maintain consistent cell morphology across nutrient conditions, supporting its role in robust biofilm formation.

SEM of four Corynebacterium isolates with fructose-, glucose-, and mucin-supplemented growth media at 20,000× magnification.

Biochemical analyses of new isolates

To further characterize the recently isolated oral Corynebacterium strains, biochemical profiling was performed using the API Coryne system with results recorded after 24-h incubation. Both C. durum isolates (JJ1 and JJ2) displayed largely similar biochemical profiles, with minor differences observed in the β-galactosidase activity (βGAL) with lower activity of JJ1 compared to JJ2 and maltose (MAL) fermentation tests, which seems to be absent in JJ1 (Fig. S2). In contrast, C. argentoratense MB1 showed a distinct biochemical signature, lacking positive reactions for nitrate reduction (NIT) and esculin hydrolysis (ESC), and appeared to utilize glucose (GLU) as the sole metabolizable substrate under the test conditions. The results were largely consistent with an earlier biochemical characterization of clinical C. argentoratense isolates (30); however, the earlier study reported metabolization of ribose after 48 h as well as fructose after 4 h, which was not a carbohydrate included in the API Coryne test used in the present study. Additionally, C. matruchotii ATCC 14266 differed by exhibiting positive reactions for pyrrolidonyl arylamidase (PyrA) and α-glucosidase (αGLU), which were negative in the other isolates (Fig. S2). Overall, the majority of test results were consistent across all four strains, with only a subset showing species-specific differences. These findings agree with previous biochemical profiling conducted by our group (31), where C. durum JJ1 and C. matruchotii ATCC 14266 displayed similar metabolic patterns.

The effect of oxygen tension on growth

A notable difference in the biochemical profiles of the four Corynebacterium strains was the absence of nitrate reduction in C. argentoratense MB1. While corynebacteria are generally considered aerobic bacteria, some species are facultative anaerobes, such as C. glutamicum, which is capable of anaerobic growth through nitrate respiration and fermentation, resulting in mixed acid production (32). In the oral cavity, corynebacteria encounter oxygen-rich conditions during early biofilm formation, but may experience oxygen limitation in more mature, stratified biofilms (33, 34). To assess their growth potential under varying oxygen levels, all four strains were cultured under aerobic (ambient air with 5% CO_₂_) and anaerobic (90% N_₂_, 5% CO_₂_, 5% H_₂_) conditions (Fig. 5). BHI agar plates spotted with Corynebacterium cultures were incubated at 37°C for 48 h. Under aerobic conditions, all four strains displayed robust and consistent growth. In contrast, anaerobic conditions led to a general reduction in growth, with C. argentoratense MB1 showing no detectable growth. These findings suggest that C. argentoratense may have limited adaptive capacity for survival in anoxic environments, potentially influencing its spatial niche within the oral biofilm.

BHI agar plates with spotted serially diluted Corynebacterium species grown for 2 days in either aerobic (+O2) or anaerobic (−O2) chambers.

Comparative genome analysis

The general genome feature analysis showed that the GC content was similar across species (~57%), while genome size and total gene count were comparable between C. matruchotii and C. durum but markedly reduced in C. argentoratense MB1 (Fig. 6A). Pangenomic and phylogenomic analyses were performed to compare the genomes of Corynebacterium isolates. For comparison, C. glutamicum, a prominent non-oral Corynebacterium used for industrial large-scale amino acid production, was also included. Phylogenomic analyses indicated that C. glutamicum was more closely related to C. matruchotii, while C. durum was more distant, and C. argentoratense MB1 was most distantly related. (Fig. 6B). In line with the phylogenomic data, pangenomic analysis of the newly isolated and sequenced C. durum JJ2 and C. argentoratense MB1 with the published sequences of C. durum JJ1 and two C. matruchotii strains revealed considerable pangenomic overlap within C. durum and C. matruchotii, but a striking divergence from C. argentoratense MB1*,* which had 807 unique genes. (Fig. 6C). Analysis of shared genes further supported this divergence: C. matruchotii and C. durum collectively shared 240 clusters, whereas C. argentoratense MB1 shared only 107 with C. durum and 29 with C. matruchotii. Direct comparision of selected genes confirmed relative distance with lower homology to C. durum and C. matruchotii (Table 1). Across all five genomes, 910 core genes were shared out of 11,645 total (Fig. 7A). COG20 functional categorization indicated that the “Translation, ribosomal structure and biogenesis” category was the most abundant in all genomes, with 1.5–2× more hits than other categories (Fig. 7B). Targeted examination of nitrate reduction genes revealed that C. durum and C. matruchotii harbor three of the four narGHIJ operon genes (narG, narI, and narJ) essential for nitrate reduction and anaerobiosis in E. coli and other bacteria. These genes were absent in C. argentoratense MB1, consistent with its lack of nitrate reductase activity and inability to grow under anaerobic conditions (Fig. 7C). In summary, C. argentoratense MB1 exhibits substantial genomic divergence from C. durum and C. matruchotii, reflecting distinct functional capacities and potential adaptation to unique oral ecological niches.

(A) A chart displaying unique characteristics to each genome. (B) Phylogenomic Tree of 6 Corynebacterial genomes, specifically the evolutionary differences between the shared single-copy core genes present within all genomes. (C) Pangenome assembly of five Corynebacterium genomes using the bioinformatics tool Anvio.

(A) Venn diagram showing the gene clusters in the five genomes used for the pangenome. (B) Number of hits for each COG category within the pangenome. (C) Chart showing unique nitrate genes in four out of five genomes (no homologs present in C. argentoratense MB1).

DISCUSSION

Oral microbiology has traditionally focused on a limited number of bacterial species strongly associated with oral diseases, for example, S. mutans in dental caries and P. gingivalis and Treponema denticola in periodontitis (35–38). These studies have greatly advanced our understanding of specific pathogenic mechanisms and contributed to the broader field of microbial virulence. However, despite these insights, the overall prevalence of caries has only slightly improved, and periodontal disease has remained largely unchanged (39). This paradox reflects a shift in our understanding of oral disease etiology. It is now widely accepted that diseases such as caries and periodontitis are polymicrobial in nature, resisting traditional pathogen-centered diagnostic and therapeutic approaches (4, 40). The concept of microbial dysbiosis has emerged to explain how a once-balanced microbial community can shift toward a disease-associated state. Enabled by high-throughput sequencing technologies, studies of the oral microbiome have revealed substantial compositional and functional changes in microbial communities associated with disease development (41–43). Crucially, these changes reflect a transition from a healthy (eubiotic) microbiome to a dysbiotic one and not through the invasion of new pathogens, but through behavioral and metabolic shifts in resident species that adapt to the altered oral environment (6). This paradigm shift has underscored the need for a more holistic understanding of oral disease, one that accounts not only for pathogenic processes but also for the ecology and function of the healthy microbiome. In this context, oral commensals, particularly commensal streptococci, have received renewed attention (7). These organisms contribute to oral health through a variety of mechanisms that enhance their fitness and confer colonization resistance. One such mechanism is hydrogen peroxide (H_₂_O_₂_) production via the SpxB pathway, which inhibits the growth of H_₂_O_₂_-sensitive pathobionts such as S. mutans and P. gingivalis (44).

Interest in other abundant yet understudied members of the oral microbiome has also grown. A landmark study using FISH to characterize the spatial organization of oral biofilms highlighted the high abundance and central positioning of corynebacteria within supragingival biofilms (18). This finding catalyzed interest in the ecological role of oral Corynebacterium species, which, despite their prevalence, had long remained functionally obscure (45). This renewed focus coincides with broader interest in corynebacteria colonizing other mucosal surfaces, such as the nasopharynx (16), ocular surface (46), and human skin (17), suggesting that these bacteria may play important and possibly conserved roles across different host-associated microbial communities. The corynebacterial species examined here, a newly isolated C. durum strain and the less commonly studied C. argentoratense, expand our understanding of phenotypic traits within this underappreciated genus.

Overall, their genome features are consistent with other human-associated corynebacteria. The G + C content of C. durum (JJ1 and JJ2) and C. matruchotii (ATCC 14266 and NCTC 10206) is ~57%, whereas C. argentoratense MB1 is slightly higher at 58.9%, matching previously reported values for three C. argentoratense strains isolated from the respiratory tract and from blood. In sum, their G + C content falls within the published range for human-associated corynebacteria (approximately 53%–58%) (22, 25). Similarly, the genome sizes of C. durum and C. matruchotii fall within the expected range for corynebacteria, averaging around 3 Mb. In contrast, C. argentoratense MB1 is considerably smaller, at just over 2 Mb. This reduction is also evident in its gene content, with only 1,889 predicted genes compared to approximately 2,500 in the other two species (25, 47).

The phenotypic differences observed across the range of experiments highlight a strong similarity between the two C. durum strains, JJ1 (previous and published isolate) and JJ2, which is to be expected, while also revealing partial phenotypic overlap between C. argentoratense MB1 and C. matruchotii ATCC 14266. This is particularly evident in biofilm formation and autoaggregation assays, which indicate that certain characteristics are conserved between C. argentoratense and C. matruchotii. However, phylogenomic tree analysis reveals a different evolutionary relationship: C. durum, C. matruchotii, C. argentoratense, and C. glutamicum (used as a non-oral Corynebacterium for comparison) share 14 core genes, yet only C. matruchotii and C. glutamicum cluster on the same branch, suggesting that they come from a recent common ancestor, while the other two species have evolved differently (Fig. 6B). These findings may reflect spatial niche differences within the oral and upper respiratory microbiomes. For instance, C. argentoratense has been isolated from a variety of anatomical sites, including the throat (particularly in patients with tonsillitis), upper respiratory tract, blood cultures, and mucosal biofilms (25, 48). In contrast to C. durum and C. matruchotii, which are well represented in eHOMD with 8 and 11 genome entries, respectively, C. argentoratense is notably absent from the eHOMD genome table. In the present study, we isolated C. argentoratense from saliva. Despite current insights, the role of C. argentoratense in oral ecology remains largely unresolved. Determining whether it is a transient inhabitant or a persistent community member, and whether candidate virulence determinants contribute to disease development, merits further systematic investigation.

One of the most distinguishing biochemical features of C. argentoratense MB1 is its inability to reduce nitrate, as demonstrated in API strip assays (Fig. S2). The capacity for nitrate reduction, whether to nitrite or further downstream products, is an ecologically relevant trait among oral bacteria (49) and may enable specific taxa to adapt to anaerobic microenvironments. This suggests that C. argentoratense may occupy a distinct functional niche within the oral mucosal microbiome that has guaranteed oxygen availability. The well-documented biogeographical distribution of C. matruchotii along the supragingival margin, where it serves as a structural scaffold for biofilm development (19, 50), likely selects for its ability to grow even under increasingly anaerobic conditions as the biofilm matures. In addition, C. matruchotii has been shown to promote calcification through the deposition of salivary calcium, contributing to dental calculus formation (51), a process that may further limit oxygen diffusion over time. Although the specific spatial distribution of C. durum in the oral cavity has not been determined, it is conceivable that, similar to C. matruchotii, it is a constituent of the supragingival biofilm.

A notable ecological trait conserved across all oral Corynebacterium species examined to date is their ability to induce a species-specific cell chain elongation phenotype in Streptococcus sanguinis. This interaction was first studied in detail with C. durum and attributed to EMVs, specifically those containing fatty acid cargo (15). C. durum secretes fatty acids, most prominently oleic acid, stearic acid, and palmitic acid, and in vitro reconstitution of these fatty acids reproduces the chain elongation phenotype in S. sanguinis (15). EMV fatty acid content can be modulated by environmental conditions; for instance, growth of C. durum in the absence of glucose results in EMVs with a ~50-fold reduction in fatty acid content, eliminating their chain elongating activity (15). Although the fatty acid content of EMVs was not determined in the present study, we observed that all four Corynebacterium strains produced EMVs of comparable concentration and diameter, suggesting a conserved EMV production capacity responsible for the induction of chain elongation in S. sanguinis. Given their ability to diffuse beyond the immediate vicinity of the producing cells, corynebacterial EMVs may influence other microbial community members and even the host. Indeed, EMV uptake by oral epithelial cells has been demonstrated for several oral bacterial species (52–54), underscoring their potential role in modulating both interspecies interactions and host responses within the oral biofilm ecosystem.

A clear distinction was also observed in the metabolic profiles determined using the API Coryne strip test. While not exhaustive, this assay provides a useful means of differentiating corynebacteria and revealed that C. argentoratense MB1 is unable to metabolize several key carbohydrates that both C. durum (JJ1 and JJ2) and C. matruchotii ATCC 14266 can utilize as energy sources. Interestingly, studies on human nasal-associated corynebacteria, including C. propinquum, C. pseudodiphtheriticum, C. accolens, and C. tuberculostearicum, have shown that their metabolic pathways are largely conserved, consistent with their shared genomic and pangenomic structures (47). These findings support the conclusion that metabolic capacity is closely tied to niche specialization within the nasopharynx.

By comparison, the results presented here suggest that C. argentoratense may be less well adapted to the oral niche, given its reduced metabolic repertoire relative to the other two corynebacterial strains. This raises the possibility that C. argentoratense may occur only transiently in the oral cavity. However, it is well established that metabolic interdependencies are widespread within oral biofilms, for example, Veillonella spp. relies on lactic acid produced by oral streptococci for growth (55). A similar cross-feeding relationship with other oral bacteria could support the persistence of C. argentoratense in this environment. At a minimum, further investigation is needed to clearly define the biogeographical distribution of C. argentoratense as well as C. durum, whose precise localization within the oral cavity has likewise not been established but is evident for multiple oral microbiome sequencing projects.

In conclusion, our comparative analyses of C. durum, C. matruchotii, and C. argentoratense underscore both the conserved and divergent traits within oral corynebacteria. While C. durum and C. matruchotii share genomic and phenotypic similarities consistent with stable adaptation to the oral niche, C. argentoratense displays reduced genome size, limited metabolic capabilities, and distinct biochemical features that may reflect either niche specialization under restricted conditions or a more transient role in the oral cavity. Its absence from eHOMD despite repeated isolation from oral samples highlights how little is known about its ecological significance. The conserved ability of all tested oral corynebacteria to influence S. sanguinis morphology through EMV-associated fatty acids further emphasizes their potential importance as biofilm organizers and modulators of community structure. Collectively, these findings expand our understanding of corynebacterial contributions to oral ecology but also reveal critical gaps, particularly regarding spatial distribution, metabolic dependencies, and potential roles in health and disease, that warrant further investigation.