Establishment of a HicA toxin-based counterselection system for markerless genetic engineering in Veillonella atypica OK5

TL;DR

A new genetic tool using the HicA toxin was developed for Veillonella atypica, enabling markerless gene deletions and promoter analysis to study oral biofilm interactions.

Contribution

Adaptation of the HicA toxin-based counterselection system from Fusobacterium nucleatum for use in Veillonella atypica OK5.

Findings

The HicA toxin system enables robust negative selection in V. atypica OK5.

A markerless deletion of the hemagglutinin gene hag1 was successfully generated.

Hag1 is essential for coaggregation with Streptococcus gordonii and is growth-phase-dependent.

Abstract

Genetic manipulation in Veillonella atypica remains limited due to the scarcity of robust counterselection systems, hindering mechanistic studies of its role in oral biofilm ecology and host interactions. In this study, we adapted a highly efficient, HicA toxin–based counterselection system, originally developed for Fusobacterium nucleatum, for use in V. atypica OK5. We demonstrated that the inducible expression of a truncated Fusobacterium periodonticum HicA toxin, controlled by a theophylline-responsive riboswitch E, provides robust and reliable negative selection. The utility of this system was validated by generating a markerless deletion of the hemagglutinin gene hag1. The resulting Δhag1 mutant exhibited a complete loss of coaggregation with the early colonizer Streptococcus gordonii, confirming the essential role of Hag1 in interspecies adhesion. Furthermore, we constructed a hag1-luciferase transcriptional fusion reporter, which revealed that hag1 expression is growth-phase-dependent, peaking during early to mid-log phase. This work establishes the HicA system as a powerful genetic tool for Veillonella, enabling scarless gene deletions and promoter analysis, thereby facilitating future mechanistic studies of its role in oral biofilm development and microbial community dynamics.

The online version contains supplementary material available at 10.1186/s12866-026-04809-y.

Figures

Construction and testing of *E. coli*-*Veillonella* shuttle plasmid pVZP01 expressing the FpHicA toxin under the control of *rpsJ*_*Fp*_ promoter and a riboswitch E-based inducible system. A Construction of pVZP01. The plasmid pBCG02 containing the *rpsJFp-E-hicA* expression cassette was previously reported to be functional in *Fusobacterium*. The linearized pBSJL2 backbone and *rpsJFp-E-hicA* cassette were fused by Gibson assembly to generate pVZP01. B Overnight *E. coli* harboring pVZP01 was sub-cultured 1:1,000 into 6 mL LB broth supplemented with the indicated concentration of theophylline

Deletion of the *hag1* gene in *V. atypica* OK5. A Construction of the *Veillonella* suicide plasmid pVZP02. The *rpsJFp-E-hicA* expressiin cassette was amplified from pVZP01 and cloned into the linearized *Veillonella* suicide vector pBST via Gibson assembly to yield pVZP02. B Construction of the *hag1* deletion plasmid pVZP02-*Δhag1*. Approximately 1.0 kb fragments upstream and downstream of *hag1* gene were PCR amplified and fused by overlapping PCR. The fused amplicon was cloned into linearized pVZP02 via Gibson assembly, yielding pVZP02-*Δhag1*. C Schematic representation of the strategy

Luciferase gene knock-in fusion with *hag1* in *V. atypica* OK5. A Construction of plasmid pVZP02-*hag1-luc* for luciferase gene transcriptional fusion with *hag1*. The luciferase (*luc*) gene, PCR amplified from pFW5-*luc* vector, was assembled by overlapping PCR between two 1.0 kb PCR fragments flanking the 3′ end of *hag1* (internal coding region) and the immediate downstream intergenic sequence, generating a chromosomal transcriptional reporter in which *luc* is expressed from its own RBS/start codon. The fusion product was cloned into linearized pVZP02 via Gibson assembly to generate pVZP

Full text

Introduction

The human oral microbiome is a multispecies niche colonized by more than 1,000 microbial species, with 100–200 typically residing in an individual mouth [1, 2]. Formation of this community involves a series of sequential steps [3]: initial colonizers consisting of the mitis streptococci (i.e., Streptococcus gordonii) attach to the teeth surface and create a favorable environment conducive to the colonization by the bridging species, such as members of the genus Veillonella [4]. Veillonella, as one of the most predominant bacteria in oral microbiota [4–7], encompasses anaerobic, gram-negative cocci frequently encountered in human oral and gut microbiomes, where they primarily metabolize lactate into short-chain fatty acids like propionate and acetate, contributing to biofilm maturation, metabolic cross-feeding and shaping community structure [4, 8, 9]. These early biofilm colonizers act as metabolic cross-feeders and bridging species. They also reduce oxygen-rich micro-niche and generate nutrients (e.g., heme), thereby promoting the growth of later colonizers - many of which are obligate anaerobic periodontopathogens (e.g. Fusobacterium nucleatum and Porphyromonas gingivalis) - and shaping oral health versus dysbiosis [4, 10, 11].

Despite their ecological importance, Veillonella species have historically faced limitations due to the lack of robust genetic tools. The strain Veillonella atypica OK5 stands out as a pioneering model because it is genetically transformable and has a fully sequenced genome, enabling emerging studies into its genetics and ecology [12–14]. Our group developed a counterselectable markerless mutagenesis system (pheS-based), which allows clean and scarless deletions of genes in V. atypica OK5 [13]. Additionally, the Merritt group has demonstrated that Veillonella parvula clinical isolates are naturally competent, allowing allelic exchange mutagenesis and transformation via genomic DNA or PCR products under optimized medium conditions [15]. Recently, a small shuttle plasmid for fluorescence-based applications has been created (pCF1135), which has been successfully used in V. parvula and V. atypica, enabling expression of oxygen-independent fluorescent proteins [16]. Beyond these advances, a suite of additional genetic tools has been developed in V. parvula that further enables functional genomics. These include adaptation of a mariner transposon mutagenesis system (pRPF215) for generating transposon mutant libraries, plasmid-based complementation using pRPF185 or the chloramphenicol-resistant shuttle vector pBSJL2-cat, conjugational transfer of mobilizable plasmids from E. coli into V. parvula, and chromosomal promoter-replacement strategies that place genes under an anhydrotetracycline-inducible pTet promoter at their native loci [17–20]. Collectively, these tools have powered recent mechanistic studies of V. parvula outer-membrance biology and diderm organization, as well as identification and functional analysis of adhesins mediating interspecies co-aggregation and mixed-species biofilm architecture.

A recent innovation in Fusobacterium genetics introduced a hicA-based counterselectable marker: a truncated Fusobacterium periodonticum HicA (FpHicA) toxin under a constitutive promoter and a theophylline-responsive riboswitch [21]. Mechanistically, HicA is a type II toxin–antitoxin (HicAB) toxin that functions as an RNase; when expressed without sufficient cognate antitoxin (HicB), HicA degrades cellular mRNA and tmRNA, leading to translational arrest and growth inhibition [21, 22]. In our prior study in Fusobacterium, the active counterselection allele corresponded to an N-terminally truncated FpHicA, whereas the full-length form showed little/no toxicity - suggesting that N-terminal sequences can inhibit activity in that context. Notably, the truncated FpHicA arrested growth in multiple bacteria (including F. nucleatum, F. periodonticum, and E. coli), consistent with a broadly portable RNase-based mechanism when expression is tightly controlled [21]. This system demonstrated high efficiency in generating markerless in-frame deletions in F. nucleatum. However, to date, no such system has been applied to Veillonella.

In this study, we adapted the FpHicA-riboswitch-based counterselection system for Veillonella. We hypothesized that FpHicA expression could act as an effective toxin in V. atypica OK5 when regulated by theophylline, enabling markerless mutagenesis. To test this, we engineered the rpsJFp-E-hicA cassette into OK5 to (i) test its toxicity and utility as a counterselectable marker, (ii) generate a knock-out deletion of the adhesin gene hag1, and (iii) create a transcriptional knock-in fusion of hag1 with luciferase gene for dynamic expression profiling. Our results show that the hicA system is highly functional in V. atypica and can enable precise genetic manipulation, significantly enriching the molecular toolkit for Veillonella research.

Materials and methods

Bacterial strains and culture conditions

The bacterial strains and plasmids used or constructed in this study are listed in Table 1. V. atypica OK5 was cultured in brain heart infusion broth (BHI; Difco) supplemented with 0.6% sodium lactate (BHIL), or on BHIL agar plates. S. gordonii DL1 was grown in BHI broth. All bacterial strains were cultured anaerobically (85% N_2_, 10% CO_2_, 5% H_2_) at 37 °C. The plasmids were transformed into V. atypica OK5 by using electroporation, as described previously [23]. For transformant selection, the bacteria were grown in Todd-Hewitt (TH) broth (Difco) with 0.6% sodium lactate (THL) plus tetracycline (2.5 µg/mL; Sigma-Aldrich). Plasmids maintenance in Escherichia coli was performed using DH5α in Lysogeny Broth (LB; Difco) medium with tetracycline (10 µg/mL) at 37 °C.

Table 1. Bacterial strains and plasmids used in this studyStrain or plasmidDescriptionReference or sourcestrains* E. coli* C2987Cloning strainNEB* V. atypica* OK5Wild type[12]* S. gordonii* DL1Wild type[24] Δhag1OK5 hag1 gene deletion mutantThis work OK5-hag1-lucOK5 hag1-luciferase transcriptional fusion reporterThis workPlasmids pBCG02E. coli-Fusobacterium suicide vector containing expressive cassette rpsJFp-E-hicA, Chloramphenicol resistance; cm^R^[21] pBSJL2Veillonella-E. coli shuttle vector, tetracycline resistance; Tet^R^[12] pVZP01Derivative of pBSJL2, A**mp^R^ gene was replaced by rpsJFp-E-hicA, tetracycline resistance; Tet^R^This work pBSTVeillonella-E. coli suicide vector,tetracycline resistance; T**et^R^[13] pVZP02Derivative of pBST, cassette rpsJFp-E-hicA was inserted in, tetracycline resistance; Te**t^R^This work pVZP02-Δhag1Derivative of pVZP02, hag1 deletion plasmidThis work pVZP02-hag1-lucDerivative of pVZP02, transcriptional fusion luc gene with hag1 geneThis work

Plasmid construction

All plasmids used in this study were constructed via Gibson assembly cloning according to the manufacturer’s instructions with NEBuilder HiFi DNA Assembly master mix. A 2 x Phanta Max Master Mix from Vazyme was used for all high-fidelity PCR amplification. The primers used in this study were listed in Table 2.

Table 2. Primers used in this studyPrimerSequence (5’ to 3’)^a^PurposepVZP01-FCACATTTCCCCGAAAAGTGCCACReverse PCR for pBSJL2 to generating pVZP01 backbonepVZP01-RCTGTCAGACCAAGTTTACTCATATATACReverse PCR for pBSJL2 to generating pVZP01 backboneFpHicA-FGTGGCACTTTTCGGGGAAATGTGAGCACCAGGTTTAGGGGCTGCpBCG02 as template for rpsJFp-E-hicA amplificationFpHicA-RATATGAGTAAACTTGGTCTGACAGATTAAAAGTATGAGGAGAACAAGCpBCG02 as template for rpsJFp-E-hicA amplificationpVZP02-FATAGCCATGAGCGGATACATATTTGReverse PCR for pBST to generating pVZP02 backbonepVZP02-RTACAAATATGCTCTTACGTGCTATTATTTAReverse PCR for pBST to generating pVZP02 backbonepVZP02-FphicA-FCAAATATGTATCCGCTCATGGCTATAGCACCAGGTTTAGGGGCTGCpVZP01 as template for rpsJFp-E-hicA amplificationpVZP02-FphicA-RTAAATAATAGCACGTAAGAGCATATTTGTAATTAAAAGTATGAGGAGAACAAGCpVZP01 as template for rpsJFp-E-hicA amplificationhag1-up-FGGTACCGGGCCCCCCCTCGAGGTCGCGTCAAGAAATTAGCAAGGTTGCpVZP02-Δhag1**hag1-up-RGATTTTTATTATCAATCATATATTACTCGTTTTCCTTTTACTAACTATCpVZP02-Δhag1**hag1-dn-FGATAGTTAGTAAAAGGAAAACGAGTAATATATGATTGATAATAAAAATCpVZP02-Δhag1**hag1-dn-RGAGCTCCACCGCGGTGGCGGCCGCTCGGATGCAATAAGGTTAGACCpVZP02-Δhag1pVZP02-rev-FCGACCTCGAGGGGGGGCCCGGTACCReverse PCR for pVZP02 to generating gene deletion backbonepVZP02-rev-RAGCGGCCGCCACCGCGGTGGAGCTCReverse PCR for pVZP02 to generating gene deletion backbonehag1-check-FGAATACAAGTTAACAATTCGConfirmation of hag1 deletionhag1-check-RGTATCTACCGGCCCCAGTCGConfirmation of hag1 deletionhag1-luc-up-FGGTACCGGGCCCCCCCTCGAGGTCGCTACTGGCACGGTAACAGGCpVZP02-hag1-luc**hag1-luc-up-RCATATTTACCTCCTCGATTAGCCTTTGGATGCTAACAATTTATCpVZP02-hag1-lucluc-FCAAAGGCTAATCGAGGAGGTAAATATGGAAGACGCCAAAAACpVZP02-hag1-lucluc-RATTTTTATTATCAATCATATATTACTTACAATTTGGACTTTCCGCpVZP02-hag1-luc**hag1-luc-dn-FGTAATATATGATTGATAATAAAAATCAACTTGTAAAGpVZP02-hag1-luc**hag1-luc-dn-RGAGCTCCACCGCGGTGGCGGCCGCTCAAAGCGCCATTTTGACGGATGpVZP02-hag1-luc**hag1-luc-check-FCCTATTAGCTCTGTATATGTAATGCConfirmation of OK5-hag1-luc**hag1-luc-check-FCGGTAATATGGTATCTACTGAGAACConfirmation of OK5-hag1-luc^a^underlined are primer sequences specific to the ends of the corresponding assembly

• (i)pVZP01. To create this plasmid, pBCG02 [21] was used as a DNA template to amplify the expressive cassette rpsJFp-E-hicA, in which the expression of hicA was driven by the rpsJFp promoter and regulated by theophylline-responsive synthetic riboswitch E. This riboswitch has been reported to function in both Gram-positive and Gram-negative bacteria [25–29]. The Veillonella-E. coli shuttle vector pBSJL2 [12] was used as a DNA template to perform inverse PCR to remove ampR region (β-lactamase) and then generate the linearized plasmid backbone using the primer pair pVZP01-F/pVZP01-R. Next, the rpsJFp-E-hicA cassette was cloned into plasmid backbone by Gibson assembly (Fig. 1A) and the recombinant plasmid pVZP01 was transformed into E. coli C2987 cells and confirmed by sequencing.

Fig. 1. Construction and testing of E. coli-Veillonella shuttle plasmid pVZP01 expressing the FpHicA toxin under the control of rpsJFp promoter and a riboswitch E-based inducible system. A Construction of pVZP01. The plasmid pBCG02 containing the rpsJFp-E-hicA expression cassette was previously reported to be functional in Fusobacterium. The linearized pBSJL2 backbone and rpsJFp-E-hicA cassette were fused by Gibson assembly to generate pVZP01. B Overnight E. coli harboring pVZP01 was sub-cultured 1:1,000 into 6 mL LB broth supplemented with the indicated concentration of theophylline. Cultures were incubated aerobically at 37 °C with rotation for 18 h. C V. atypica OK5 containing either pVZP01 or the parental plasmid pBSJL2 was streaked on BHIL agar plates supplemented with 0 or 2 mM theophylline and incubated anaerobically at 37 °C for 24 h. D Overnight V. atypica OK5 carrying pVZP01 or pBSJL2 was diluted 1:100 into 8 mL BHIL broth with the indicated concentration of theophylline. The cultures were incubated in the anaerobic chamber at 37 °C for 24 h. All experiments were performed at least three times, and representative results are presented

• (ii)pVZP02. After confirming the reliability of FpHicA in V. atypica OK5, pVZP01 was used as a DNA template to amplify the rpsJFp-E-hicA cassette using the primer pair pVZP02-FphicA-F/pVZP02-FphicA-R. This amplicon was cloned into linearized Veillonella suicide vector pBST [13] via Gibson assembly, yielding pVZP02 (Fig. 2A). pVZP02 was derived from the previously published vector pBST and retains the parental lacZ module shown in Fig. 2. The cloning region in pVZP02 is compatible with Gibson assembly; in this study, the backbone was linearized across the cloning region and inserts were assembled using overlapping ends, followed by colony PCR and Sanger sequencing confirmation. The same region contains an MCS site, so pVZP02 can also be used for standard restriction–ligation cloning if desired. Blue–white screening is technically feasible in appropriate E. coli hosts but was not required for our Gibson-based workflow. This recombinant plasmid was transformed into E. coli C2987 cells and confirmed by sequencing.

Fig. 2. Deletion of the hag1 gene in V. atypica OK5. A Construction of the Veillonella suicide plasmid pVZP02. The rpsJFp-E-hicA expressiin cassette was amplified from pVZP01 and cloned into the linearized Veillonella suicide vector pBST via Gibson assembly to yield pVZP02. B Construction of the hag1 deletion plasmid pVZP02-Δhag1. Approximately 1.0 kb fragments upstream and downstream of hag1 gene were PCR amplified and fused by overlapping PCR. The fused amplicon was cloned into linearized pVZP02 via Gibson assembly, yielding pVZP02-Δhag1. C Schematic representation of the strategy for generating hag1 deletion mutants. Plasmid pVZP02-Δhag1 was transformed into V. atypica OK5 by electroporation, and mutants were selected by a two-step allelic exchange. D Screening of V. atypica OK5 (pVZP02-Δhag1) on theophylline agar plates. A 100-µL aliquot of a 10,000-fold diluted overnight culture was plated on BHIL agar plates containing 0 or 2 mM theophylline to select cells that had excised and lost the plasmid. E Confirmation of Δhag1 deletion mutants by colony PCR using primer pair hag1-check-F/hag1-check-R (Table 2). Ten colonies were randomly selected for PCR, and seven (clones 1, 2, 4, 5, 6, 8 and 10) were deletion mutants, while three reverted to wild type after the second recombination. V. atypica OK5 wild type strain was used as a control. F Coaggregation assays to assess Δhag1 deletion mutants. V. atypica OK5 wild type strain and ten PCR-confirmed strains were tested for coaggregation with S. gordonii DL1. Results shown are representative of three independent experiments

• (iii)pVZP02-Δhag1. To construct the plasmid for hag1 deletion, 1.0-kb fragments of upstream and downstream of hag1 gene were amplified by PCR using primer pairs hag1-up-F/hag1-up-R and hag1-dn-F/hag1-dn-R. Two PCR amplicons were fused by overlapping PCR. The fused product was then mixed with 10 µL Gibson assembly master mix together with an equal amount of pVZP02 vector backbone created by inverse PCR with primer pair pVZP02-rev-F/pVZP02-rev-R. (Fig. 2B). The resulting plasmid pVZP02-Δhag1 was transformed into E. coli C2987 cells and confirmed by sequencing.
• (iv)pVZP02-hag1-luc. To construct the plasmid for luciferase gene (luc) knock-in fusion with hag1 in OK5, 1.0-kb fragments flanking the 3’ end of the hag1 gene (internal coding region) and the immediate downstream region were amplified by PCR using primer pairs hag1-luc-up-F/hag1-luc-up-R and hag1-luc-dn-F/hag1-luc-dn-R. Luciferase gene was amplified by PCR using plasmid pFW5-luc [30, 31] as the template and primer pair luc-F/luc-R. The three PCR amplicons were fused by overlapping PCR. The fused product was then mixed with 10 µL Gibson assembly master mix together with an equal amount of pVZP02 vector backbone created by inverse PCR with primer pair pVZP02-rev-F/pVZP02-rev-R. (Fig. 3A). The resulting plasmid pVZP02-hag1-luc was transformed into E. coli C2987 cells and confirmed by sequencing.

Fig. 3. Luciferase gene knock-in fusion with hag1 in V. atypica OK5. A Construction of plasmid pVZP02-hag1-luc for luciferase gene transcriptional fusion with hag1. The luciferase (luc) gene, PCR amplified from pFW5-luc vector, was assembled by overlapping PCR between two 1.0 kb PCR fragments flanking the 3′ end of hag1 (internal coding region) and the immediate downstream intergenic sequence, generating a chromosomal transcriptional reporter in which luc is expressed from its own RBS/start codon. The fusion product was cloned into linearized pVZP02 via Gibson assembly to generate pVZP02-hag1-luc. B Schematic strategy for generating the OK5-hag1-luc reporter strain. Plasmid pVZP02-hag1-luc was transformed into V. atypica OK5 by electroporation. The hag1 luciferase reporter strains OK5-hag1-luc were obtained via two-step allelic exchange. C Confirmation of OK5-hag1-luc strains by colony PCR using primer pair hag1-luc-check-F/hag1-luc-check-R (Table 2). Of ten colonies tested, four (clones 3, 5, 6 and 8) were luciferase reporters, while six reverted to wild type after the second recombination. V. atypica OK5 wild type strain was used as a control. D luciferase expression profile of OK5-hag1-luc. Growth was measured every 2 h by OD_600_, and luciferase activity (relative light units, RLU) was normalized to OD_600_. Results are presented as mean ± SD from three independent experiments

Transformation and counterselection

All plasmids were introduced into V. atypica OK5 via electroporation as previously described [13, 32]. The reactions were electroporated using a Bio-Rad Gene Pulser II set at 25 kV, 25µF and 200 Ω. Plasmid insertion into Veillonella’s chromosome by homologous recombination was selected on BHIL plates supplemented with 2.5 µg/mL tetracycline at 37 °C. A tetracycline-resistant colony was inoculated in BHIL broth without antibiotics for overnight culture. The following day, a 100 µL aliquot of 10,000-fold diluted culture was plated on BHIL agar plates containing 2 mM theophylline for counterselection. Cells growing on this plate all lost the plasmid via recombinative excision, thus lost resistance to tetracycline (Figs. 2C and 3B). Ten colonies were tested by PCR amplification for hag1 deletion or luciferase gene fusion (Figs. 2E and 3C). To minimize the risk of unintended secondary changes, whole-genome sequencing (WGS) of representative independent clones can be performed as an additional validation step.

Coaggregation assay

In vitro coaggregation assays were performed with V. atypica OK5 wild-type, ten colonies obtained from HicA-based counterselection of OK5 and S. gordonii DL1, as previously described [32]. The bacterial cells in the stationary-phase were harvested and washed twice with coaggregation buffer (1 mM Tris buffer [pH 8.0], 0.1 mM CaC1_2_, 0.1 mM MgCl_2_, 150 mM NaCl) at room temperature. Then, cells were resuspended in coaggregation buffer and normalized to OD600 = 1.2. Equal volumes (0.2 mL) of each cell suspension were mixed in a 24-well plate and mixed by vortexing at room temperature until aggregates formed. Wells containing single species cell suspension alone (0.4 mL) were included as controls.

Luciferase assays

Overnight culture of V. atypica OK5-hag1-luc reporter strains were harvested and re-suspended with fresh BHIL media to OD600 ~ 1.0. Suspended cultures then were 1:100 diluted into fresh BHIL media. All cultures were grown in anaerobic condition for 24 h, and the data were measured every 2 h. Briefly, a 100-µL aliquot was withdrawn from the anaerobic culture, mixed with 25 µL of 1 mM D-luciferin (Sigma) solution (prepared in 0.1 M citrate buffer, pH 6.0) by pipetting up and down ~ 10 times; this mixing step also briefly aerated the sample to supply oxygen required for the luciferase reaction. Luciferase activities were immediately measured using a GloMax Navigator Luminometer (Promega). The optical densities at 600 nm were measured with a V-1200 Visible Spectrophotometer (VWR) to express relative light unit (RLU)/OD_600_.

Results

A HicA toxin from F. periodonticum is functional in V. atypica OK5

Our previous study employed a mutant pheS gene as a counterselectable marker to efficiently generate in-frame deletions in V. atypica OK5 [13]. To expand the genetic arsenal for Veillonella, we tested a hicA-based in-frame deletion system recently developed in F. nucleatum [21]; in which a F. periodonticum hicA homolog (named as FpHicA) served as an effective toxin.

To assess the potential toxicity of FpHicA in Veillonella, its expression cassette (rpsJFp-E-hicA) was amplified from pBCG02 [21] and cloned into the Veillonella-E. coli shuttle vector pBSJL2 [12] via Gibson assembly, yielding pVZP01 (Fig. 1A). The expression of hicA was driven by the rpsJFp promoter. To regulate toxin expression, a well-characterized, theophylline-responsive synthetic riboswitch E [28] was inserted between the rpsJFp promoter and the hicA start codon [21]. This riboswitch, previously demonstrated to function in both Gram-positive and Gram-negative bacteria [28, 29], consists of an aptamer domain coupled to a synthetic ribosome-binding site (RBS) [21]. In theory, transcription from the rpsJFp promoter occurs constitutively during growth, but translation is activated only in the presence of theophylline.

We first evaluated the functionality of this system in E. coli harboring pVZP01. As expected, the growth assays confirmed that the presence of 2 mM theophylline completely inhibited the growth of E. coli carrying pVZP01 (Fig. 1B). We next introduced pVZP01 into V. atypica OK5. As a control, OK5 was transformed with the parental vector pBSJL2 lacking the toxin cassette. Growth phenotypes were assessed on agar plates and in liquid broth. As shown in Fig. 1C and D, in the presence of 2 mM theophylline, growth of OK5 (pVZP01) was completely inhibited under both conditions, whereas OK5 (pBSJL2) exhibited no growth defect. These results indicate that FpHicA is highly toxic in V. atypica OK5 and can serve as a robust counterselectable marker for genetic manipulation in Veillonella.

Deletion of hag1 gene in V. atypica OK5

We previously reported that the hemagglutinin gene hag1, which encodes a YadA-like autotransporter Hag1, contributes to adherence to several Streptococci species, P. gingivalis and human oral buccal cells [32]. However, the hag1 mutant in that study was generated by an insertional (single-crossover) mutation. To generate a clean hag1 deletion mutant and test hicA system in Veillonella, the plasmid named pVZP02-Δhag1 was constructed to generate the deletion mutant of the entire hag1 coding region. The rpsJFp-E-hicA expression cassette was amplified from pVZP01 and cloned into linearized Veillonella suicide vector pBST [13] via Gibson assembly, yielding pVZP02 (Fig. 2A). The upstream and downstream regions of hag1 gene were amplified and fused by overlapping PCR. The fused fragment was cloned into linearized pVZP02 to generate pVZP02-Δhag1 (Fig. 2B). The sequenced and confirmed plasmid was introduced into V. atypica OK5 via electroporation as previously described [13, 32]. Tetracycline resistant colonies contain the transforming plasmid integrated either at either upstream or downstream regions of hag1 gene via gene recombination (Fig. 2C). For the counterselection, a randomly selected tetracycline resistant colony was grown overnight in liquid BHIL broth without antibiotics to allow recombinative excision of the plasmid. The overnight culture was then plated and screened on BHIL plates supplemented with 2 mM theophylline. As shown in Fig. 2C, following counterselection, only a few of cells can survive on the theophylline-supplemented plates, thus undergo the second recombination to lose the integrated plasmid (Fig. 2D). PCR and sequencing confirmed deletion of the entire hag1 coding region in a subset of isolates (Fig. 2E). The resulting Δhag1 mutant (clone 1, 2, 4, 5, 6, 8 and 10) displayed complete loss of coaggregation with S. gordonii (Fig. 2F), confirming the phenotype reported previously [32].

Luciferase gene knock-in fusion with hag1 in V. atypica OK5

To study the expression and regulation of hag1 gene, the plasmid pVZP02-hag1-luc was constructed to generate a firefly luciferase reporter in OK5. 1.0-kb fragments flanking the 3’ end of the hag1 gene (internal coding region) and the immediate downstream region were amplified by PCR, while luciferase gene was amplified from vector pFW5-luc [31], and three fragments were fused by overlapping PCR. The fused product was cloned into linearized pVZP02 to generate pVZP02-hag1-luc (Fig. 3A). The sequenced and confirmed plasmid was electroporated into V. atypica OK5. The procedure of tetracycline selection and theophylline counterselection is same as above described (Fig. 3B). PCR and sequencing confirmed that the luciferase gene was successfully inserted at the 3’ end of hag1 (Fig. 3C). In the OK5-hag1-luc strain, the luciferase cassette was integrated at the hag1 chromosomal locus as a transcriptional reporter. The luciferase ORF is expressed as an independent protein from its own RBS/start codon, rather than as a Hag1-luciferase translational fusion. Thus, luminescence primarily reports hag1 transcriptional activity across growth. Expression of OK5-hag1-luc reporter was monitored across the growth curve. As shown in Fig. 3D, hag1 expression peaked during early log phase, declined from mid-log phase, and persisted at a reduced level into stationary phase. The establishment of the knock-in reporter validated the efficiency of this system for generating transcriptional fusions.

Discussion

In this study, we successfully established a HicA toxin–based counterselection system for V. atypica OK5, thereby expanding the limited genetic toolbox available for this genus. Our results demonstrate that the F. periodonticum HicA toxin, when controlled by a theophylline-responsive riboswitch, is highly toxic to V. atypica and can be harnessed as a reliable counterselectable marker. Using this system, we generated a markerless deletion of the hag1 adhesin gene (knock-out) and constructed a hag1-luciferase transcriptional fusion reporter (knock-in), confirming both the efficiency and versatility of this method.

The establishment of this counterselection system represents an important advance over earlier approaches, such as the mutant pheS counterselection strategy previously developed for V. atypica OK5 [13]. While pheS has been valuable, its reliance on amino acid analogs introduces media-dependent variability. In contrast, HicA provides a strong, toxin-based selection pressure that is easily controlled with theophylline. Notably, similar HicA systems have been applied in F. nucleatum with high efficiency [21], and our adaptation demonstrates its portability to another oral anaerobe. Together, PheS and HicA now offer complementary strategies, providing researchers with greater flexibility in Veillonella genetics.

Our successful deletion of hag1 validates the utility of this system for creating clean, markerless mutations. The Δhag1 mutant displayed a complete loss of coaggregation with S. gordonii, confirming and extending earlier findings obtained with insertional mutants [32]. This provides a more robust genetic foundation for studying hag1’s role in interspecies adhesion, biofilm formation and oral microbial ecology. Similarly, our construction of a hag1-luciferase knock-in reporter highlights the potential of this platform for generating transcriptional fusions. The observed expression dynamics of hag1 - peaking during early log phase before declining from mid-log phase, and persisting at a reduced level into stationary phase - suggest growth-dependent regulation of adhesion factors, consistent with Veillonella’s ecological roles in early colonization and biofilm development [33, 34].

The development of more powerful genetic systems in Veillonella is timely. Members of this genus are central metabolic cross-feeders in the oral microbiome, converting lactate into propionate and acetate while shaping community structure [4, 8]. Recent pan-genomic and metagenomic analyses of human oral Veillonella species provide ecological context that implies genes regulating adhesion, colonization, and metabolism are likely under selective pressure depending on oral site [35, 36], further underscoring the importance of developing robust genetic tools. Consequently, the ease of genetic manipulation now enables direct testing of these hypotheses, allowing researchers to dissect determinants of oral niche specialization and interspecies adhesion.

Moreover, Veillonella has emerged as a key player in oral–systemic health links. Its interactions with Streptococcus mutans can promote biofilm virulence and cariogenicity in human severe caries [37], while coaggregation with P. gingivalis may influence periodontal disease progression [4, 11, 32, 38]. Beyond the oral cavity, Veillonella has been implicated in systemic infections and studied as a potential probiotic in exercise physiology [39, 40]. With markerless editing now feasible, hypotheses about Veillonella’s contributions to both health and disease can be rigorously tested through targeted mutagenesis.

In conclusion, we present a functional HicA toxin–based counterselection system for V. atypica OK5 that enables scarless gene deletion and transcriptional reporter construction. This system not only validates HicA as a versatile genetic tool in oral anaerobes but also opens the door to mechanistic studies of Veillonella’s ecological functions, metabolic networks, and contributions to oral and systemic health. Despite its advantages, the HicA system also carries considerations. For example, the portability of this system to other Veillonella strains or related genera will require validation, as toxin tolerance and transformation efficiency may vary. Future work could combine HicA with PheS in dual-counterselection workflows to increase editing efficiency.

Supplementary Information

Supplementary Material 1.