Natural Transformation of Riemerella columbina and Its Determinants

In a previous study, it was shown that Riemerella anatipestifer, a member of Flavobacteriaceae, is naturally competent. However, whether natural competence is universal in Flavobacteriaceae remains unknown. In this study, it was shown for the first time that Riemerella columbina was naturally competent in the laboratory condition; however, Flavobacterium johnsoniae was not naturally competent under the same conditions. The competence of R. columbina was maintained throughout the growth phases, and the transformation frequency was highest during the logarithmic phase. A competition assay revealed that R. columbina preferentially took up its own genomic DNA over heterologous DNA. The natural transformation frequency of R. columbina was significantly increased in GCB medium without peptone or phosphate. Furthermore, natural transformation of R. columbina was inhibited by 0.5 mM EDTA, but could be restored by the addition of CaCl2, MgCl2, ZnCl2, and MnCl2, suggesting that these divalent cations promote the natural transformation of R. columbina. Overall, this study revealed that natural competence is not universal in Flavobacteriaceae members and triggering of competence differs from species to species.


INTRODUCTION
Naturally competent bacteria can actively take up naked DNA from their environment and integrate it into the genome, which is called natural transformation (Mell and Redfield, 2014). As one of the three horizontal gene transfer mechanisms, natural transformation facilitates bacterial acquisition of virulence genes and antibiotic-resistant cassettes to help bacteria adapt to the environment (Wiedenbeck and Cohan, 2011;Seitz and Blokesch, 2013a). Natural transformation was first discovered in Streptococcus pneumoniae in 1928 (Griffith, 1928). Currently, at least 83 species have been found to have natural competence (Johnston et al., 2014;Liu et al., 2017).
In Gram-positive and Gram-negative bacteria, there are different mechanism to take up DNA. Naturally competent Gram-negative bacteria, such as Neisseria species and Haemophilus influenzae, use type IV pili (T4P) to take up exogenous double-stranded DNA (dsDNA), in contrast to Helicobacter pylori, which uses a type IV secretion system (T4SS) (Hofreuter et al., 2000), and Campylobacter jejuni, which uses a type II secretion system (T2SS) (Wiesner et al., 2003) to take up exogenous dsDNA. Gram-positive bacteria, such as S. pneumoniae and Bacillus subtilis, use a competence pseudopilus, a structure similar to T4P, to take up dsDNA (Hahn et al., 2005). Once dsDNA is transported across the outer membrane in Gram-negative bacteria or the peptidoglycan layer in Gram-positive bacteria, dsDNA is degraded to single-stranded DNA (ssDNA) and transported through the pore protein ComEC into the cytoplasm (Johnston et al., 2014). Internalized ssDNA is presumably bound by DNA-processing protein A (DprA), which recruits the recombinase RecA to mediate homologous recombination by facilitating strand exchange (Johnston et al., 2014;Huang et al., 2019). At present, the natural transformation of H. influenzae in Pasteurellaceae, Vibrio cholerae in Vibrionaceae, and S. pneumoniae in Streptococcaceae are well studied (Redfield et al., 2006;Blokesch, 2012;Johnston et al., 2014). Among the six genera of Pasteurellaceae, only the genera Actinobacillus and Haemophilus are naturally competent (Redfield et al., 2006). In Streptococcaceae, both Streptococcus and Lactococcus show natural competence (Griffith, 1928;Dalia et al., 2017). Within the genus Haemophilus, when H. influenzae is transferred from rich medium to defined competence medium (M-IV) or the cell culture reaches stationary phase, it becomes naturally competent (Herriott et al., 1970;Redfield, 1991). However, natural transformation of Haemophilus parasuis was readily induced by nutrient-rich medium (Zhang et al., 2015;Li et al., 2016). All information suggests that the occurrence of natural transformation is different among bacteria, even within the same genus.
In a previous study, it was shown that one member of the Flavobacteriaceae family, Riemerella anatipestifer (R. anatipestifer, RA), which causes septicemic diseases in ducks, geese, turkeys, and other birds , is naturally competent . However, whether other Flavobacteriaceae species are also naturally competent and under which condition natural transformation is induced remains unknown. Here, Flavobacterium johnsoniae (F. johnsoniae), a common soil and aquatic bacterium (McBride, 2004), and Riemerella columbina (R. columbina), widely distributed species among pigeon populations (Rubbenstroth et al., 2013), were selected as models to explore the occurrence of natural transformation and its influencing factors.

Bacterial Strains, Primers, and Growth Conditions
Riemerella columbina and Flavobacterium johnsoniae were purchased from the Culture Collection of the University of Gothenburg (CCUG) and the China General Microbiological Culture Collection Center (CGMCC), respectively. The bacterial strains and primers used in this study are listed in Table 1. The culture conditions for R. columbina and F. johnsoniae were identical to those used for R. anatipestifer described in a previous study . Briefly, R. columbina was cultured in GC broth (GCB) medium with shaking or GCB agar plates and LB plates supplemented with 5% sheep blood (blood plates) at 37 • C, however, F. johnsoniae was cultured in GCB medium with shaking or GCB plates at 25 • C. When required, erythromycin was added into the medium at a final concentration of 1 µg/ml for R. columbina and 50 µg/ml for F. johnsoniae.

Preparation of Donor DNA
The homologous gene of dprA (C237_RS0105470) in R. columbina, which protects ssDNA and loads RecA to facilitate homologous recombination (Mirouze et al., 2013), and the gliding motility gene gldH in F. johnsoniae were selected as targeted deletion gene, since they are not essential for the growth of bacteria and can be deleted (McBride et al., 2003;Hovland et al., 2017;Huang et al., 2019). Donor DNA was composed of upstream of target gene, an antibiotic resistance cassette and downstream of target gene. Briefly, the ∼620 bp upstream sequence and ∼620 bp downstream sequence of C237_RS0105470 were amplified from the genome of R. columbina using the primers RC-Up P1 and RC-Up P2, RC-Down P1 and RC-Down P2, respectively. The ∼620 bp upstream sequence and ∼620 bp downstream sequence of gldH were amplified from the genome of F. johnsoniae using the primers Up(gldH) P1 and Up(gldH) P2, Down(gldH) P1 and Down(gldH) P2, respectively. An erythromycin resistance cassette was amplified from the genome of R. anatipestifer CH-1 using the primers RC-Erm P1 and RC-Erm P2 or Erm(gldH) P1 and Erm(gldH) P2, respectively (Liao et al., 2015;Luo et al., 2015). The three fragments were fused using overlapping PCR (Xiong et al., 2006;Huang et al., 2017Huang et al., , 2019. The fused fragments served as donor DNA for natural transformation.

Natural Transformation Procedure
The procedure of natural transformation was similar to that used for R. anatipestifer described in a previous study Huang et al., 2019). Briefly, R. columbina and F. johnsoniae were cultured in GCB liquid with shaking at 37 • C for R. columbina and at 25 • C for F. johnsoniae. The bacteria were collected during the logarithmic phase (OD 600 = 3-4 for R. columbina, OD 600 = 1-1.5 for F. johnsoniae) and adjusted to an optical density (OD) of 1. The growth curve of F. johnsoniae in GCB was shown in Supplementary Figure 1. The donor DNA was added to the bacterial cells and incubated for 1 h at 37 • C for R. columbina and at 25 • C for F. johnsoniae. Then, 100 µl of cells were plated on GCB agar plates supplemented with erythromycin (1 µg/ml for R. columbina; 50 µg/ml for F. johnsoniae) to count transformants. Then, 10 µl of cells were serially diluted with PBS and plated on GCB agar plates to count viable bacteria. The transformation frequency (TF) was calculated as transformants divided by viable bacteria. Then, 100 µl of cells were plated on GCB supplemented with the corresponding concentration of erythromycin to check for spontaneous mutants.

Determination of Growth Curves
The bacteria were streaked on blood plates or GCB agar plates. A single colony was cultured in 5 ml of GCB liquid medium with shaking at 37 • C for 14 h. The bacterial cells were transferred into 20 ml of GCB with or without peptone, phosphate or iron at an OD of 0.05 and cultured at 37 • C with shaking. The OD 600 was

The Effect of Components of GCB on Natural Transformation in R. columbina
Bacterial cells were cultured to the logarithmic phase (OD 600 = 3-4) and adjusted to an OD 600 of 1. The bacterial cells were collected and resuspended in GCB medium depleted of vitamin B1 (VB 1 ), glucose, L-glutamine, NaCl, peptone, or phosphate. After the bacteria were incubated at 37 • C for 30 min, donor DNA was added to the cultured cells, and natural transformation was performed. Iron is essential for the growth of most bacteria (Liao et al., 2016). To investigate whether iron affects the growth and natural transformation of R. columbina, the growth curve of R. columbina in GCB supplemented with different concentrations of iron chelator ethylenediamine-N,N'-bis(2hydroxyphenylacetic acid) (EDDHA) according to the method mentioned previously (Press et al., 2001;Liu et al., 2016Liu et al., , 2019 and natural transformation were performed after the bacteria were incubated into GCB supplemented with the corresponding concentration of EDDHA at 37 • C for 30 min. The viable bacteria and transformants were counted, and the TF was calculated.

EDTA Treatment
Bacterial cells were cultured until the logarithmic phase (OD600 = 3-4) at 37 • C with shaking and adjusted to an OD 600 of 1. Three hundred microliters of bacteria were collected and resuspended in GCB medium supplemented with 0.5 mM EDTA. Natural transformation was performed after the bacteria were incubated at 37 • C for 30 min. The TF was calculated according to the method described previously. To investigate which divalent cation affects the natural transformation of R. columbina, different concentrations of CaCl 2 , MgCl 2 , ZnCl 2 , MnCl 2 , or CuCl 2 were added into the GCB medium supplemented with the corresponding concentration of EDTA. The bacterial cells were first incubated in the above medium at 37 • C for 30 min, and natural transformation was then performed. The TF was calculated as described previously.

Statistics
Statistical analysis was performed using GraphPad Prism 8.0 (GraphPad Software Inc., La Jolla, United States). An unpaired two-tailed Student's t-test was used to compare two groups, and a value of P < 0.05 was considered significant. Data represent the mean and standard deviation (SD) from at least three independent experiments.

R. columbina, but Not F. johnsoniae, Is Naturally Competent Under the Same Conditions
To assay whether other members of Flavobacteriaceae were able to undergo natural transformation, R. columbina and F. johnsoniae were selected. We used the same method as described in a previous study for R. anatipestifer to determine the natural competence of these two species . After R. columbina incubated with donor DNA which contains the upstream sequence of C237_RS0105470, an erythromycin resistance cassette and the downstream sequence of C237_RS0105470 (Figure 1A), many resistant colonies grew on the plate with erythromycin. However, no resistant colonies appeared in the control group without donor DNA (the spontaneous mutation rate of erythromycin resistance was lower than the detection limitation). Random single colonies were verified using PCR to ensure that the target gene was replaced by the erythromycin resistance cassette through homologous recombination. As shown in Figure 1B, compared to the wild-type strain, the resistant colonies contained an erythromycin resistance gene but not a target gene. It was suggested that the target gene has been replaced by the erythromycin resistance cassette and that the target sequence of the R. columbina strain was lost. It was strongly supported that R. columbina was naturally competent and that natural transformation could be used to efficiently generate targeted gene disruptions in R. columbina, with a TF of 4.14 (±0.5) × 10 −6 .
After F. johnsoniae incubated with donor DNA containing the upstream sequence of gldH, an erythromycin resistance cassette and the downstream sequence of gldH, the transformants were selected on GCB plates supplemented with 50 µg/ml erythromycin (the MIC of erythromycin for F. johnsoniae is 16 µg/ml). However, no resistant colony appeared with or without donor DNA. It has been shown that gldH can be deleted in F. johnsoniae through other methods (McBride et al., 2003), suggesting that this gene is not essential for the survival of the bacteria. Overall, it was suggested that F. johnsoniae could not perform natural transformation using the same method as R. columbina under the same conditions.
Searching for the Components of the Natural Transformation Machinery in the Genome of F. johnsoniae To investigate whether F. johnsoniae contains the homologous proteins that involved in natural transformation. We aligned the amino acids sequences of T4SS from H. pylori and Agrobacterium . Lane 1, the 16S rRNA was amplified from wild-type R. columbina and the mutant strain R. columbina C237_RS0105470::Erm using the primers RC-16S rRNA P1 and RC-16S rRNA P2. Lane 2, the erythromycin resistance cassette was amplified from wild-type R. columbina and the mutant strain R. columbina C237_RS0105470::Erm using the primers RC-Erm P1 and RC-Erm P2. Lane 3, the target gene C237_RS0105470 was amplified from wild-type R. columbina and the mutant strain R. columbina C237_RS0105470::Erm using the primers C237_RS0105470 P1 and C237_RS0105470 P2. Lane 4, the upstream sequence and erythromycin resistance cassette was amplified from wild-type R. columbina and the mutant strain R. columbina C237_RS0105470::Erm using the primers RC-Up P1 and RC-Erm P2. Lane 5, the erythromycin resistance cassette and downstream sequence was amplified from wild-type R. columbina and the mutant strain R. columbina C237_RS0105470::Erm using the primers RC-Erm P1 and RC-Down P2.
Frontiers in Microbiology | www.frontiersin.org tumefaciens, T4P from V. cholerae, T2SS from C. jejuni, and other hypothetical competence proteins from R. anatipestifer with genome of F. johnsoniae. As shown in Table 2, only the homolog of ComB11 in T4SS, which is a putative VirB11-homologous ATPase (Karnholz et al., 2006), was found in F. johnsoniae and showed 40.8% identity with the ComB11 of H. pylori. Based on the T4P of V. cholerae (Seitz and Blokesch, 2013b), only the homologs of PilB, PilC, PilF, PilQ and PilT were discovered in F. johnsoniae and shared 48.41, 29.62, 25.98, 27.43, and 43.33% with each relative protein of V. cholerae, respectively. PilB and PilT are polymerization and depolymerization ATPases, respectively (Seitz and Blokesch, 2013b). PilC was an inner membrane platform protein which interacts with PilB and PilT to control both pilus assembly and disassembly (Takhar et al., 2013). PilF is pilolin protein which is essential for pilus biogenesis (Matthey and Blokesch, 2016). PilQ is a secretion pore, which plays a role in translocating pilus on the cell surface (Wolfgang et al., 2000). Furthermore, only the homologs of CtsD and CtsF were found in F. johnsoniae based on the T2SS of C. jejuni (Wiesner et al., 2003). CtsD is an outer membrane protein which has homology to the PilQ protein (Wiesner et al., 2003). CtsF is an inner membrane protein and shares similarity to PilG of N. gonorrhoeae which has homology to the PilC of V. cholerae (Tønjum et al., 1995). Other hypothetical competence protein of R. anatipestifer, like DprA, ComEC, RecA, Ssb, ComM and RadC is also present in F. johnsoniae . These results indicated that these homologs of F. johnsoniae may be sufficient to encode a T4P-type DNA uptake system in addition to the proteins usually needed for DNA translocation and cytoplasmic processing.

Natural Transformation of R. columbina Increases During the Logarithmic Phase
We were wondering whether natural transformation was able to occur in all growth phases in R. columbina, the TF was assayed. Natural transformation was performed at each time point by adding the same amount of donor DNA. As shown in was cultured to the logarithmic phase and mixed with different amounts of donor DNA (0.1, 1, 10, 100, 200, 500, 1,000, 2,000, or 4,000 ng). As shown in Figure 3, the TF increased with increasing DNA concentration when the amount was lower than 1,000 ng. However, when the DNA amount was higher than 1,000 ng, the TF no longer increased. The number of transformants were included in the Supplementary Data Sheet 2. These results suggested that 1,000 ng of donor DNA was saturating for natural transformation in R. columbina.

R. columbina Preferentially Takes Up Its
Own DNA Over Heterologous DNA It has been reported that some bacteria, such as H. influenzae and Neisseria, preferentially take up DNA containing short motifs known as uptake signal sequences (USSs) or DNA uptake sequences (DUSs) (Scocca et al., 1974;Sisco and Smith, 1979). These short motifs have accumulated in the genome to high densities over evolutionary time (Mell and Redfield, 2014). To determine whether R. columbina also preferentially takes up its own DNA, a natural transformation competition experiment was performed. In this experiment, the genome of R. columbina C237_RS0105470 was used as the donor DNA. As shown in Figure 4, when 1 µg of donor DNA was mixed with 1 µg of genomic DNA of R. columbina, the TF was decreased two-fold compared to the control in which without competition DNA was added; Moreover, the TF was decreased with the increase of competing DNA. However, the TF showed no significant changes when 1 µg donor of DNA was mixed with 1 µg of R. anatipestifer or E. coli genomic DNA compared to that when only 1 µg of donor DNA was added. The TF was decreased significantly only when the R. anatipestifer or E. coli genomic DNA was increased to 10 µg, which can be considered as unspecific effect. The   (3-4), and the OD 600 value was determined. The cell culture was adjusted to an OD 600 of 1. A 300 µl cell culture was transformed with 1 µg the genome of R. columbina C237_RS0105470 (control) or 1 µg of donor DNA mixed with different quantity of competing genomic DNA of R. columbina, R. anatipestifer ATCC11845 or E. coli XL1-Blue and incubated at 37 • C for 1 h. The natural TF was calculated as transformants divided by viable bacteria. The P-value refers to the difference between the no-competition and competition measurement, respectively. Insignificance p-value (P > 0.05) was not shown. All the results are representative of three independent experiments. Error bars denote standard deviation.
number of transformants were included in the Supplementary Data Sheet 2. To further investigate whether R. columbina, F. johnsoniae or R. anatipestifer contain putative DUSs or USSs, Jellyfish 1 was to be used to count the numbers of occurrences of individual kmers in both strands of their genome, respectively, with a parameter that limited the length of kmers to less than 10 bp. As shown in Table 3, sequences with the top three repeats for 10 bp, 9 bp, 8 bp, and 7 bp were listed, respectively. It was found that hundreds of repeat sequences or its complement were present in their genomes. The frequency of the 9-bp repeat sequence is 0.6/kb for R. anatipestifer, 0.5/kb for R. columbina and 0.5/kb for F. johnsoniae, respectively, which is much higher than the frequency of 0.1/kb expected for a random sequence of this base composition for them. Whether this sequence has the function of DUSs or USSs needs to be further investigated.

The TF of R. columbina Is Increased Under Peptone-Restrictive or Phosphate-Restrictive Conditions
To investigate the effect of the nutrients on natural transformation, natural transformation was conducted in GCB depleted for each component, including vitamin B1 (VB 1 ), glucose, L-glutamine, NaCl, peptone and phosphate. As shown in Figure 5A, the TF of R. columbina was 1.9 (±0.1) × 10 −5 in GCB depleted of peptone, which increased five-fold compared to that in GCB. The TF of R. columbina was 9.05 (±0.5) × 10 −6 in GCB depleted of phosphate, which increased approximately two-fold compared to that in GCB. However, compared to the TF of R. columbina in GCB, there was no significant difference 1 https://github.com/gmarcais/Jellyfish when VB 1 , glucose, L-glutamine or NaCl was removed from GCB ( Figure 5A). The number of transformants were included in the Supplementary Data Sheet 2.
Next, we investigated whether the change in TF was associated with the growth ability of bacteria. Therefore, the growth curve of bacteria was determined when peptone, phosphate, NaCl, glucose, L-glutamine or VB 1 was removed from GCB. The results showed that bacteria did not grow in GCB without peptone or VB 1 , whereas the growth of bacteria was significantly decreased in GCB without phosphate; however, there were no significant differences when NaCl, glucose, or L-glutamine was removed from GCB ( Figure 5B). To investigate whether iron affects the natural transformation of R. columbina, different concentrations of iron chelator EDDHA were supplemented into the GCB medium. As shown in Figure 5A, the TF did not change compared to that of the control (without EDDHA). The number of transformants were included in the Supplementary Data Sheet 2. However, the growth of R. columbina was significantly inhibited in iron-depleted medium (GCB supplemented with 200 µM EDDHA), suggesting that iron is essential for the growth of R. columbina ( Figure 5B). Overall, these results suggested that peptone-restrictive or phosphate-restrictive medium had an effect on the natural transformation and this effect is not directly related to the growth ability.
Ca 2+ , Mg 2+ , Zn 2+ , and Mn 2+ but Not Cu 2+ Promote the Natural Transformation of R. columbina Iron has no effect on natural transformation, and we wondered if other divalent cations influence the natural transformation of  R. columbina. Natural transformation was conducted in GCB medium with 0.5 mM EDTA, which had no effect on the survival of bacteria (Supplementary Figure 2). The results showed that the TF in GCB with 0.5 mM EDTA was 5.75 (±0.75) × 10 −7 , which was decreased approximately 4-fold compared to that in GCB [TF = 2.5(±0.1) × 10 −6 ], suggesting that 0.5 mM EDTA had a significant inhibitory effect on natural transformation in R. columbina ( Figure 6A). To investigate which divalent cation has an effect on natural transformation, different concentrations of CaCl 2 , MgCl 2 , ZnCl 2 , MnCl 2 , or CuCl 2 were supplemented into the cell culture after incubation with EDTA. The addition of 0.5 mM Ca 2+ basically restored transformation, and the TF increased as the concentration of Ca 2+ increased ( Figure 6B). The addition of 0.5 mM Mg 2+ completely restored the TF; however, the frequency did not increase as the concentration of Mg 2+ increased (Figure 6B), suggesting that 0.5 mM was likely a saturating concentration of Mg 2+ for natural transformation in R. columbina. The TF gradually increased as the concentration of Zn 2+ increased from 0.1 mM to 0.5 mM. The TF was the highest at 0.5 mM Mn 2+ (Figure 6B). However, the addition of different concentrations of Cu 2+ did not restore the natural transformation but instead inhibited natural transformation ( Figure 6B). The number of transformants were included in the Supplementary Data Sheet 2. Therefore, it was shown that Ca 2+ , Mg 2+ , Zn 2+ , and Mn 2+ were required for the natural transformation of R. columbina, but Cu 2+ was not.

DISCUSSION
Riemerella anatipestifer is the first bacterium of Flavobacteriaceae to be reported to have natural competence .
To check whether other bacteria in Flavobacteriaceae are also naturally competent, F. johnsoniae and R. columbina were selected. The results showed that R. columbina was able to undergo natural transformation; however, F. johnsoniae was not competent under the same conditions. One possibility is that the natural transformation of F. johnsoniae does not occur at all growth phases but only at a certain time point, for example natural transformation happens to S. pneumoniae, in which natural transformation is not constitutive, as synthesis and assembly of the uptake apparatus is a transient and regulated process (Mirouze et al., 2013). Another possibility is that the PCR fragments are not suitable substrates, such as occurs with C. jejuni, which takes up only methylated DNA but not PCR fragments (Beauchamp et al., 2017), and H. influenzae and Neisseria gonorrhoeae, which preferentially take up DNA with an USS or DUS over other sources of DNA (Mell et al., 2012;Berry et al., 2013;Frye et al., 2013;Mell and Redfield, 2014). The third possibility is that natural transformation in F. johnsoniae must be induced by special substrates, such as occurs with natural transformation of Vibrio cholerae, which is induced by chitin (Meibom et al., 2005). The fourth possibility is that we did not choose the correct isolates. It has been reported that even for the competent bacteria, some isolates are non-transformable (Evans and Rozen, 2013;Dalia et al., 2015). The last possibility is that F. johnsoniae does not undergo natural transformation because of the lack of some essential genes for natural transformation. Consistent with the natural transformation of R. anatipestifer , the natural transformation of R. columbina is also constitutive, although the TF is different at the different growth phases. This phenomenon might occur because the expression of genes involved in natural transformation in R. columbina is different at the different growth phases. Similar to R. anatipestifer, R. columbina preferentially takes up selfsourced genomic DNA, suggesting that each bacterium might use a certain mechanism, such as a restriction modification (R-M) system (Aras et al., 2002;Zhang and Blaser, 2012) or other systems, to prevent the uptake of excessive extracellular DNA that may overload the bacteria, subverting the bacterial genome with extracellular DNA from competing strains.
Originally, the function of natural transformation was hypothesized as "DNA for food" (Redfield, 2001), because the natural competence of H. influenzae and B. subtilis was activated under nutrient-limited condition (Bobb, 1963;Herriott et al., 1970). However, this hypothesis is questionable, since the natural competence of some other bacteria, such as A. baumannii, requires a nutrient-rich condition (Traglia et al., 2016). In the case of R. columbina, we showed that the TF of R. columbina was significantly increased under peptonerestrictive or phosphate-restrictive conditions, suggesting that the uptake of DNA may be "food" for R. columbina to supplement the nitrogen and phosphorus.
A more plausible hypothesis for the function of natural transformation is "DNA for repair" (Michod et al., 2008;Engelmoer et al., 2013), since the natural transformation of some bacteria, such as H. pylori (Dorer et al., 2010), S. pneumoniae (Prudhomme et al., 2006) and B. subtilis (Zhang et al., 2018), was activated by antibiotics or DNA damage reagent. Here, we also investigated the effects of antibiotics such as ampicillin (inhibitor of cell wall biosynthesis), kanamycin (inhibitor of protein biosynthesis), nalidixic acid (inhibitor of DNA replication) and mitomycin C (intercalation with DNA) during the natural transformation of R. columbina. We showed that none of the antibiotics affected natural TF of R. columbina did not change after treatment with antibiotics (Supplementary Figure 3), suggesting that the antibiotics used here cannot trigger natural transformation of R. columbina.
Our systematic investigation of natural transformation in the Flavobacteriaceae family shows that it is widely distributed. However, the environmental conditions that trigger natural transformation vary from species to species. In this family, natural transformation appears to play a major role in HGT. The discovery of natural transformation in R. columbina represents the basis for the establishment of gene editing and cloning system in this bacterium.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.