Bacterial strains are listed in Supplementary Table S1. P. gingivalis strains were cultured at 37°C in 5 mL of Brain-Heart Infusion medium (Becton, Dickinson and Company, Maryland, United States) supplemented with 5 μg/mL hemin and 1 μg/mL menadione (referred to as BHI-HM) under anaerobic conditions: 10% H₂, 10% CO₂, and 80% N₂ (Nakao et al., 2006). BHI-HM blood-agar plates were prepared by adding 5% defibrinated sheep blood and 1.5% (w/v) agar into BHI-HM. To select transformants, the medium was supplemented with 10 μg/mL erythromycin or 1 μg/mL ampicillin.

Primers used in this study are listed in Supplementary Table S2. An erythromycin resistant gene, ermF, was generated by PCR using the PKB-301/302 primer set and pHS17 (Haake et al., 2000) as the template DNA. DNA fragments of 0.5 kb corresponding to the upstream and downstream regions of PGN_0032–PGN_0033 were amplified with primer sets PKB-440/418 (500-bp upstream homology arm) and PKB-419/441 (500-bp downstream), respectively. These fragments were assembled into the ermF cassette by overlap extension PCR with primers PKB-440/441. The upstream and downstream homology arms ranging from 50 to 2,000 bp were generated using the following primer sets: PKB-446/418 (50-bp arm), PKB-444/418 (100-bp arm), PKB-442/418 (250-bp arm), PKB-428/418 (1,000-bp arm), and PKB-417/418 (2,000-bp arm) for the upstream homology arms; PKB-419/447 (50-bp arm), PKB-419/445 (100-bp arm), PKB-419/443 (250-bp arm), PKB-419/429 (1,000-bp arm), and PKB-419/420 (2,000-bp arm) for the downstream homology arms.

Schematic of the transformation assay is shown in Figure 1. The optimized procedure and conditions are outlined here, with different conditions indicated in the results section or figure legends. P. gingivalis strains were anaerobically cultured at 37°C in 5 mL of BHI-HM until reaching the early-mid exponential phase. The cells were harvested by brief centrifugation at 8,000 g for 4 min at room temperature. After removing the supernatant, the cell pellet was suspended in 500 μL of fresh prewarmed BHI-HM. Twenty microliters of the cell suspension were mixed with 100 ng of donor DNA and spotted onto a BHI-HM blood-agar plate lacking erythromycin or ampicillin. The plate was incubated anaerobically at 37°C for 24 h, after which a colony biofilm was scraped with a disposable inoculating 10-mm loop from the plate and suspended in 500 μL of fresh pre-warmed BHI-HM by the stirring the loop (the biofilm is non-sticky and can be easily removed from the loop and suspended in the liquid medium). The colony suspension was serially diluted in BHI-HM and spread onto BHI-HM blood-agar plates containing the appropriate antibiotics. Then, the plates were incubated at 37°C for 5 days under anaerobic conditions. Transformation efficiency was assessed by counting the number of the antibiotic-resistant colonies.

To create the deletion mutant for comEA (PGN_0421), DNA fragments corresponding to the 5′- and 3′-flanking regions of the gene were amplified from the genome of P. gingivalis ATCC 33277 by PCR using the primer sets PKB-462/455 and PKB-456/463, respectively. The ampicillin resistant gene, cepA, was amplified from P. gingivalis KDP501 (Sato et al., 2018) using the primer set PKB-421/422. The DNA fragments were assembled to a ΔcomEA::cepA cassette by overlap extension PCR with primers PKB-462/463. The resulting PCR product (100 ng) was introduced into P. gingivalis ATCC33277 via natural competence, as described above. Transformants were selected using 1 μg/mL ampicillin.

For the comEC (PGN_0519) deletion mutant, the 5′- and 3′-flanking regions were amplified using the primer sets PKB-460/451 and PKB-452/461, respectively. Preparation of the cepA cassette and transformation were carried out following the same procedure as used for the comEA mutant.

A previous study reported the natural competence of P. gingivalis mentioning that biofilm conditions appear to enhance extracellular DNA uptake (Tribble et al., 2012). Based on the study, we designed a transformation protocol for P. gingivalis, as shown in Figure 1. Recipient cells were propagated by cultivating them in a liquid BHI-HM medium. Donor DNA was added to the recipient cells and then spotted onto a BHI-HM blood-agar plate without antibiotics to facilitate colony biofilm formation and DNA uptake. The colony biofilm was suspended in liquid BHI-HM and spread onto an antibiotic-containing BHI-HM blood-agar plate to select transformants.

First, we prepared donor DNA containing an erythromycin-or ampicillin-resistant genes (ermF or cepA) for transformation. The intergenic region between PGN_0032 and PGN_0033 in the P. gingivalis ATCC 33277 genome was designated as the donor integration site to minimize the impact on cell viability and growth. The donor DNA fragments were flanked by 500-bp homology arms at their 5′ and 3′ ends, corresponding to the upstream and downstream regions of the integration site (Figure 2A).

To test our transformation method, P. gingivalis ATCC 33277 recipient cells were cultured at 37°C for 1 day in liquid BHI-HM. The recipient cells were mixed with 0.5 μg of ermF donor DNA, and a colony biofilm was developed for 24 h on a BHI-HM blood-agar plate. Afterward, the colony biofilm suspension was spread onto an erythromycin-containing BHI-HM blood-agar plate. Tiny colonies appeared on the plate 3 days after incubation. Five days after incubation, when the colonies had grown larger, colony PCR was performed to verify donor integration accuracy. In this PCR genotyping, a 3.2-kb band signal was detected when ermF was inserted into the intergenic region between PGN_0032 and PGN_0033, while a 2.0-kb band was detected in the wild-type strain. All nine transformant colonies examined confirmed that ermF was correctly inserted into the intergenic region between PGN_0032 and PGN_0033 (Figure 2B, Em^r transformants No. 1–9). The integration had no significant effect on the cell growth, as expected (Supplementary Figure S1). Additionally, instead of ermF, cepA donor DNA was also successfully integrated into the intergenic region (Supplementary Figure S2). We attempted transformation by electroporation following the method described by Bélanger et al. (2007). However, we could not obtain any electroporation-mediated transformants, whereas the natural competence-mediated transformation produced >1 × 10³ Em^r transformants under the same conditions for donor DNA amount and cultivation. These results demonstrate the efficacy of our transformation method.

In general, the expression of natural competence is tightly regulated and temporal (Salvadori et al., 2019). Therefore, we investigated the relationship between transformation efficiency and the growth phase of the recipient cells. Under our culture conditions, the lag phase lasted 6 h after inoculation. The cell culture then entered the exponential phase by 10 h and reached the stationary phase at 24 h (Figure 3A). P. gingivalis recipient cells were harvested at 10, 16, 24, 34, and 42 h post-inoculation (Figure 3A, red arrows). For this transformation assay, we used 100 ng of ermF donor DNA with 1,000-bp homology arms. As shown in Figure 3B, the highest transformation efficiency was 7.7 × 10⁵ CFU/mL of the biofilm suspension (7.7 × 10⁶ CFU/μg of DNA) during the early-exponential phase. Given that the total viable cell count of the biofilm suspension was 3.9 × 10⁹ CFU/mL, the transformation frequency (the ratio of erythromycin-resistant colonies to total colonies) was 2.0 × 10⁻⁴. In contrast, transformation efficiency drastically declined upon entering the stationary phase (24 h and later time points). These results indicate that the natural competence of P. gingivalis ATCC 33277 is activated during the early growth phase in BHI-HM liquid culture.

Biofilm development is a crucial step for DNA uptake and the expression of ErmF, which confers antibiotic resistance. We investigated the effects of incubation duration of the P. gingivalis colony biofilm on the DNA competency (Figure 3C). Although a small number of transformants were generated by 0 to 3 h incubation for the biofilm development (3.8–8.2 × 10¹ CFU/mL), it seemed that at least 6 h of incubation was necessary for stable performance in routine transformation experiments (Figure 3C, 6 h, 1.3 × 10³ CFU/mL). Transformation efficiency plateaued at 24 h incubation, with only a slight increase observed at 48 h. Considering the balance between transformation efficiency and the time required for the experiment, we concluded a 24-h culture as optimal for biofilm development.

We investigated how transformation efficiency depends on the dose and length of homology arms in the donor DNA. Recipient cells in the exponential phase were mixed with 0, 10, 50, 100, 500, and 1,000 ng of donor DNA, each containing 1,000-bp homology arms. The transformation assay revealed a dose-dependent increase in transformation efficiency, peaking with 100 or 500 ng of donor DNA (Figure 4A). Even as little as 10 ng of donor DNA was sufficient to achieve 10⁴ CFU/mL, highlighting the high efficiency of our transformation method. However, adding 1,000 ng of donor DNA resulted in a slight decrease in efficiency, likely due to the dilution of recipient cells caused by the large volume of the DNA solution when spotted onto the agar plate.

Subsequently, we determined the minimal length of homology arms required for effective homologous recombination with the genomic DNA. Donor DNA with 50-, 100-, 250-, 500-, 1,000-, and 2,000-bp homology arms was tested. Transformation efficiency increased proportionally with the length of the homology arms (Figure 4B), showing more than a 10-fold increase as the length doubled. Unexpectedly, even 50-bp homology arms still produced some transformant colonies. Donor DNA with 2,000-bp homology arms yielded 2.6 × 10⁶ CFU/mL of the transformants. Although this was the highest transformation efficiency in this study, excessively long homology arms can lead to issues such as incomplete amplification or PCR errors. Therefore, 500–1,000-bp homology arms are more manageable and practical for P. gingivalis mutagenesis.

Natural DNA competence involves a species-specific DNA uptake machinery that traverses the cellular envelope (Seitz and Blokesch, 2013). While most components of this machinery in P. gingivalis remain unknown, ComF (locus tag, PG0158 in W83; PGN_0270 in ATCC 33277), a cytoplasmic ATP-dependent DNA helicase, is the only component experimentally confirmed to be necessary for natural competence-mediated genetic transformation (Tribble et al., 2012). Based on P. gingivalis gene annotation by Kyoto Encyclopedia of Genes (KEGG; Kanehisa et al., 2023), we identified two additional proteins potentially involved in natural competence: PGN_0421 and PGN_0519. PGN_0421 contains a ComEA domain (COG accession no. COG1555) spanning positions 94–165 (E-value 6.03 × 10⁻²¹). ComEA is a DNA-binding protein with HhH motifs that facilitates DNA transport and is located in the periplasmic compartment in Gram-negative bacteria such as Vibrio cholerae (Matthey and Blokesch, 2016). On the other hand, PGN_0519 contains a ComEC domain (COG accession no. COG0658) spanning positions 213–488 (E-value 2.05 × 10⁻⁴¹). ComEC is a membrane protein that forms a channel in the cytoplasmic membrane to import extracellular DNA into the cytosol (Pimentel and Zhang, 2018).

To investigate their involvement in natural competence, we constructed deletion mutants of P. gingivalis comEA (PGN_0421) and comEC (PGN_0519). As the donor DNA used in this study carried ermF, we replaced comEA and comEC with cepA (Figure 5A). Using our transformation method, we successfully generated the ΔcomEA and ΔcomEC mutants. Gene replacement with cepA was confirmed by PCR (Figure 5B). Transformation attempts with ermF donor DNA on the ΔcomEA and ΔcomEC mutants resulted in no colonies (Figure 5C), indicating that the transformation efficiency of these strains was below detectable levels. This finding suggests that ComEA and ComEC are essential for the natural competence of P. gingivalis.

This study exclusively used the ATCC 33277 strain; however, other type strains, such as W83 (Nelson et al., 2003) and TDC60 (Watanabe et al., 2011), are also commonly used in P. gingivalis research. To verify the DNA competency of these strains, we employed the strain-specific homology arms (1,000 bp each) for the ermF donor DNA to avoid nucleotide mismatches with ATCC 33277. P. gingivalis ATCC 33277, W83, and TDC60 were cultured to the exponential phase, mixed with 100 ng of donor DNA and allowed to develop biofilms for 24 h. The transformation assay revealed that W83 and TDC60 exhibited high DNA competency comparable to that of ATCC 33277, with no significant differences observed between the strains (Figure 6). In this assay, we used the same transformation conditions that were optimized for ATCC 33277. Since W83 and TDC60 have different characteristics, strain-specific adjustments could further enhance the transformation efficiency for W83 and TDC60. Overall, these results suggest that our transformation method is generally applicable in P. gingivalis.