The bacterial strains used in this study are listed in Table 1. Pectobacterium carotovorum subsp. carotovorum (Pcc) isolates were kindly provided by the Rural Development Administration (RDA) at Wanju-gun, Korea. All the Pcc isolates were incubated in tryptic soy broth (TSB) and tryptic soy agar (TSA) [1.5% (w/v) agar] at 28°C. Escherichia coli MFDpir was grown in Luria-Bertani (LB) broth and plates supplemented with 0.3 mM diaminopimelic acid (DAP). Antibiotics were added to the media at the following concentrations if necessary: ampicillin (Amp), 50 μg mL⁻¹; carbenicillin (Car), 100 μg mL⁻¹; kanamycin (Kan), 50 μg mL⁻¹; rifampicin (Rif), 50 μg mL⁻¹, while isopropyl β-D-1-thiogalactopyranoside (IPTG) was added at a concentration of 50–100 μM.

Bacterial susceptibility against the phage was confirmed with the spot assay as described previously (Kim and Ryu, 2012). Briefly, the host bacterial lawn on the TSA plate was prepared by pouring 5 mL of soft TSB agar [0.4% agar (supplemented with antibiotics and IPTG, if necessary)] inoculated with the host bacteria. After 30 min of solidification, 10 μL of serially diluted (10-fold) phage lysates were spotted on the bacterial lawn and dried for 20 min at room temperature. Plates were incubated overnight at 28°C, and the phage plaques formed were counted on the next day.

Purified phage POP72 was morphologically characterized by transmission electron microscopic (TEM) analysis as described by Kim and Ryu (2011). Briefly, a high titer of phage stock (~1 × 10¹⁰ PFU mL⁻¹) was adsorbed onto carbon-coated copper grids and negatively stained with 2% aqueous uranyl acetate (pH 4.0). The phages were observed with TEM (Carl Zeiss LEO 913AB) at 100-kv accelerating voltage at the National Institute of Agricultural Sciences (Wanju-gun, South Korea). The phage was classified morphologically using the International Committee on Taxonomy of Viruses (ICTV) classification (King et al., 2011).

Phage DNA was extracted from the virions with proteinase K and SDS as previously described (Sambrook and Russell, 2001). The purified phage DNA was sequenced with a Genome Sequencer FLX titanium sequencer (Roche, Mannheim, Germany) and assembled with GS de novo assembler software (Roche) at Labgenomics Inc., South Korea. The ORFs of the POP72 genome were predicted using Glimmer3 (Delcher et al., 1999), Genemark.hmm (Lukashin and Borodovsky, 1998), FgenesB software (Softberry, Inc., Mount Kisco, NY), and RAST annotation engine (http:://rast.nmpdr.org/) (Aziz et al., 2008; Overbeek et al., 2014). The assortment and editing of genome-sequencing and annotation data were conducted using Artemis (Carver et al., 2008). The tRNA-coding sequences in the phage genome were scanned with a tRNAscan-SE program (Lowe and Eddy, 1997). The prediction of protein functions was conducted with NCBI BLASTp, CDD (Conserved Domain Database), and InterProscan (Altschul et al., 1990; Jones et al., 2014; Marchler-Bauer et al., 2014). A comparative genomic analysis of POP72 with homologous phages was conducted by BLASTn, and the results were visualized using ACT13 (Carver et al., 2008). Putative promoters and terminators in CA biosynthesis gene cluster were predicted using BPROM and iTerm-PseKNC (Solovyev and Salamov, 2011; Feng et al., 2018). BioCyc database collection was used to predict putative operons from the gene cluster (Karp et al., 2017).

A Tn5 transposon mutant library was constructed as previously described (Larsen et al., 2002; Ferrieres et al., 2010) with some modifications. Briefly, a donor E. coli MFDpir containing a suicide vector pRL27 and a recipient Pcc27 at early log phase (optical density at 600 nm, ~0.8) was harvested and washed three times with 10 mM MgSO₄. A mixture of donor and recipient cells (3:1, v/v) was spotted onto the TSA plate supplemented with DAP and incubated for 24 h at 28°C for the conjugation. To screen the phage-resistant mutants, pools of transconjugants were mixed with POP72 (10⁹ PFU mL⁻¹) in 1 mL of TSB broth and incubated at room temperature for 15 min prior to plating on TSA/Kan plates for 24 h at 28°C. Surviving colonies were isolated by streaking three times on TSA plates, and the absence of the remaining phages in the cells was verified by spotting culture supernatants on the bacterial lawn. Isolated phage-resistant Tn5 mutants were further characterized.

Genomic DNA extracted from the selected transposon mutants was digested with the restriction enzyme BamHI, and purified DNA fragments were self-ligated with T4 ligase (Roche). E. coli DH5α λpir was transformed with the circularized DNA fragments by heat shock and plated on LBA/Kan. Rescued plasmid DNA harboring the R6Kγ origin with the Tn5 transposon was extracted from the selected clones, and a locus of transposon insertion was identified by DNA sequencing using a pair of oligonucleotide primers tpnRL17-1 and tpnRL 13-2 (Table 2) (Larsen et al., 2002). The nucleotide sequence of Pectobacterium carotovorum subsp. carotovorum strain Pcc21 (GenBank accession number CP003776) that was registered in the NCBI database was used as a reference for the sequence analysis.

Site-directed mutagenesis of Pcc27 was performed by the lambda red recombination system with some modifications (Datsenko and Wanner, 2000). Fifty milliliter of Pcc27 cells harboring plasmid pKD46 were incubated at OD₆₀₀ = 0.5 and made electrocompetent by three time washes with ice-cold 10% glycerol. The kanamycin resistant (Kan^R) cassette from plasmid pKD13 was amplified using primers that contains 40-mer homologous sequences of wzc gene and 20-mer priming sequence of pKD13. Fifty microliter of the competent cells were transformed with 3.5 μg of the concentrated PCR product by electroporation, and recovered in 1 ml SOC media for 3 h at 30°C. Two-hundred microliter of the culture were plated out on LBA/Kan. The remaining cell cultures were kept at room temperature overnight and plated out the next day. The Kan^R cassette of the positive transformants was removed by introducing the plasmid pCP20. Gene deletion was confirmed by PCR and subsequent DNA sequencing.

To isolate spontaneous POP72-resistant mutants, a high-titer overlay assay was performed as previously described (Kim and Ryu, 2011). Briefly, 5 mL of soft TSB agar (0.4% agar) inoculated with a mixture of serially diluted (10-fold) Pcc27 culture at stationary phase and 100 μL of POP72 suspension (~10⁹ PFU mL⁻¹) were poured on TSA plates. The solidified plates were incubated for 24 h at 28°C. Randomly selected single colonies from each plate were isolated by repeated subcultures on TSA plates. The phage resistance and the absence of phages from the isolates were verified using a spot assay.

The plasmids used in this study are listed in Table 1. Plasmid pCpsG, which expresses the putative cpsG gene under the control of the lac promoter, was constructed as follows. The putative cpsG gene from Pcc27 was amplified by PCR with oligonucleotides Pcc27_cpsG_F_BamHI and Pcc27_cpsG_R_HindIII. The purified PCR product was digested with the restriction enzymes BamHI and HindIII and ligated with the same restriction enzyme-digested pUHE21-2 lacI^q plasmid vector (Soncini et al., 1995). The sequence of the cloned gene was verified by DNA sequencing. All the other plasmids were constructed with similar experimental steps and corresponding oligonucleotides (Table 2).

The phage adsorption assay was performed as previously described with some modifications (Kim and Ryu, 2012). Briefly, Pcc cells were harvested and suspended in TSB media (approximately 2~3 × 10⁸ CFU mL⁻¹). Phage POP72 was added to the bacterial suspension at an MOI of 0.01 and aliquoted in an equal volume to five microtubes. During the phage adsorption for 15 min at 28°C, each tube was centrifuged (16,000 × g, 4°C for 1 min) at the indicated time points, and the supernatants were immediately filtered (0.22 μm-pore-size filter). The number of remaining free phage particles in the filtrates was determined by the spot assay using Pcc27 as an indicator strain.

The CA of Pcc was extracted as previously described with some modifications (Dische and Shettles, 1951). EPS-degrading enzymes in the Pcc cell culture were inactivated by heating at 100°C for 15 min. After cooling, 50 mL of the culture were centrifuged (13,000 × g, 4°C for 30 min), and the supernatant was mixed with three volumes of ethanol. CA-containing precipitates were harvested by centrifugation as described above after overnight incubation at 4°C. The pellet was dissolved in 4 mL of distilled water and dialyzed for 48 h against distilled water (Spectra/Por 3 regenerated cellulose membrane, MWCO 3.5 kDa, SPECTRUM). Residual polypeptides were removed by precipitation with 5 mL of 10% (v/v) trichloroacetic acid for 1 h and centrifuged (13,000 × g, 4°C for 30 min). The supernatant was dialyzed again for 120 h against distilled water. The resulting preparation was stored at 4°C until quantification.

The quantification of CA was conducted by measuring the amount of non-dialyzable methyl pentose (i.e., fucose) which is a specific component of the CA (Bechet et al., 2010). One milliliter of the CA preparation was mixed with 4.5 mL of H₂SO₄/H₂O (6:1, v/v), and the mixture was heated at 100°C for 20 min and cooled to room temperature. For each sample, the absorbance at 396 nm (A₃₉₆) and 427 nm (A₄₂₇) was measured either directly (control sample, A-co) or after the addition of 100 mL of cysteine hydrochloride (cysteine sample, A-cy) (Bechet et al., 2010). Biological extracts contained various compounds, which following reaction with H₂SO₄, yield products absorbing between 396 and 427 nm. Nonspecific reactions were subtracted from the total absorption of the sample: the value of A_396−co and A_427−co were deducted from A_396−cy and A_427−cy, respectively, to acquire ΔA₃₉₆ and ΔA₄₂₇. (ΔA₃₉₆-ΔA₄₂₇) were associated with the concentration of methylpentose using a diverse concentration of fucose standard curve ranging from 5 to 100 μg mL⁻¹ (Obadia et al., 2007).

Inactivation of the phage particles by Pcc27 CA was examined as previously described with modifications (Heller and Braun, 1979). Briefly, dilutions of phage POP72 (5 × 10⁶ PFU mL⁻¹) were mixed with equal volumes of extracted Pcc27 CA (36 μg mL⁻¹) and incubated at 28°C for 30 min. The mixtures were diluted 10-fold with ice-cold PBS, and the number of remaining active phage particles was counted using the double-agar layer overlay assay. The number of phages from a parallel assay conducted with distilled deionized water (ddH₂O) was used as a negative control.

Rifampicin-resistant mutants were selected as previously described (Glandorf et al., 1992). Spontaneous rifampicin-resistant mutants of Pcc27 were obtained by transferring colonies of this strain to TSA agar plates containing increasing concentrations (12.5, 25, 37.5, and 50 μg mL⁻¹) of rifampicin (Sigma). The stability of these rifampicin-resistant mutants was tested by three subcultures of the rifampicin-resistant mutants on TSA agar plates containing rifampicin (50 μg mL⁻¹).

A Pcc virulence assay was performed as previously described with modifications (Marquez-Villavicencio et al., 2011). Chinese cabbages purchased in a local market in Seoul, South Korea were cut to an equal size (~10 by 7 cm) and surface sanitized for 5 min with 1% sodium hypochlorite. After rinsing with sterilized water three times for 5 min and air-drying, Chinese cabbage samples were stabbed with a sterilized pipette tip, and 10 μL of Pcc bacterial culture (10⁹ CFU mL⁻¹) was added to the wound. The negative control was inoculated with sterile distilled water. For the phage-treated group, POP72 lysate was placed on the Pcc-inoculated wound at an MOI of 1, 10, and 100. The Chinese cabbages were placed in plastic boxes to maintain high humidity and incubated at 28°C for 24 h. The samples were examined periodically for soft rot disease symptoms. The number of Pcc recovered from the Chinese cabbage was measured as described by Lee et al. (2016). Rifampicin-resistant Pcc27 (Pcc27^RifR) cells from the inoculated Chinese cabbage samples were separated and collected using stomaching with 100 mL of sterilized buffered peptone water (B.P.W) for 1 min. After the stomaching, the Chinese cabbage debris was removed using a sterile bag (FILTRA-BAG with Open Top, SCTO7012A, LABPLAS), and the supernatants were transferred to new conical tubes. After centrifugation at 13,000 × g, 4°C for 1 min, the bacterial cell pellets were resuspended with PBS buffer. Serially diluted (10-fold) cell suspensions were spread onto LB Rif agar plates, and the colonies formed were counted after overnight incubation at 28°C.

Pectobacterium carotovorum phages were screened by a plaque assay from environmental samples (i.e., soil from cultivated land, agricultural water, and irrigation water) collected from various regions in South Korea. Nine phages were isolated, and the ability of these phages to infect bacteria was confirmed by single plaque formation in a spot assay with serially diluted phage lysate (10-fold; from 10⁹ to 10² PFU/ml). Phage POP72 has the broadest host range compared with the other eight phages and forms clear plaques on several Pcc (20 of 22 strains) and P. carotovorum subsp. brasiliensis strains (Pcb; 13 of 15 strains) (Table S1). Morphologically, POP72 belonged to the family Podoviridae with an icosahedral head (60.09 ± 2.79 nm) and a short tail (15.00 ± 3.80 nm) (Figure S1).

Genomic analysis revealed that POP72 has 44,760 bp of double-stranded DNA with a GC content of 49.7% containing 55 putative open reading frames (ORFs) and no tRNAs (Figure S2 and Table S2). No genes known to be associated with lysogen formation were found, suggesting the lytic nature of the POP72. Comparative genomic analysis using BLAST and ACT13 revealed that phage POP72 highly resembled Pectobacterium carotovorum phage PP1 (Lim et al., 2013). Although the whole genome of phage POP72 shares 99% nucleotide identity with the PP1 genome (Figure S3), an additional 360-bp sequence flanked by ORF28 (endonuclease) and ORF29 (phosphoesterase) exists in POP72. With the exception of phage PP1, phage phD2B that specifically infects the plant pathogen Lelliottia spp. is the phage most similar to POP72 (Figure S3).

To identify the phage receptor for POP72, a random mutant library of Pectobacterium carotovorum subsp. carotovorum isolate Pcc27 was constructed using a Tn5 transposon (Larsen et al., 2002; Ferrieres et al., 2010), and POP72-resistant clones were screened. Twenty-four clones of POP72-resistant Pcc27 were obtained from the library (Table 3). Through partial DNA sequencing, 21 of the POP72-resistant clones were found to have transposon insertions in various genes associated with CA biosynthesis (Figure 1), including operons and genes involved in the exportation of CA repeat unit to periplasm, biosynthesis of polysaccharides, and biosynthesis of precursor sugar of CA. Among these, three putative genes, putative cpsG (homology to Pcc21_RS06735 in Pectobacterium carotovorum subsp. carotovorum strain Pcc21), putative wcaA (Pcc21_RS06680 in Pcc21), and putative wzc (Pcc21_RS06645 in Pcc21), were selected for further investigation. The cpsG gene in the other Enterobacteriaceae bacteria is directly involved in the biosynthesis of GDP-mannose, which is converted into a GDP-fucose via a three-step pathway (Andrianopoulos et al., 1998). Two l-fucoses and one d-glucose constitute a main chain of CA monomer, and a side chain comprised of d-galactose and d-glucuronic acid is attached to the one of l-fucoses, suggesting that sufficient expression of the cpsG gene may be required to construct a basic structure of CA (Stevenson et al., 1996). The putative wcaA and putative wzc genes encode a glycosyltransferase and an exopolysaccharide exporter accessory protein, respectively. Although the exact function of WcaA glycosyltransferase has not been clearly revealed, this enzyme might be associated with the transfer of some saccharide moieties to biosynthesize the CA. The gene product of the putative wzc might play a role in exporting the capsular polysaccharide beyond the outer membrane and regulating the tyrosine phosphorylation of the UDP-glucose dehydrogenase involved in the CA biosynthesis (Table 3) (Stevenson et al., 1996; Grangeasse et al., 2003; Bechet et al., 2010).

The susceptibility of these Tn5 transposon mutants, as well as the Pcc27 wild-type (WT) strain, against POP72 was examined further using the spot assay. As expected, phage POP72 could not form any plaques on the lawns of the three transposon mutants (Figure 2A). When the Tn5-inactivated genes were complemented by a pUHE21-2 lacI^q vector harboring each corresponding gene, the sensitivity of Pcc27 against POP72 was partially or completely restored even without IPTG induction (Figure 3A). Complementation without IPTG induction, as well as the partial restoration, might be due to the improper levels of each protein within a bacterial cell because it is difficult to fine-tune the expression levels of genes with the leaky inducible lac promoter and IPTG (Jacob and Monod, 1978). The in-frame wzc gene deletion mutant was also constructed and tested for the phage susceptibility. As expected, the Δwzc mutant was resistant to POP72 and the wzc complemented strain was sensitive to POP72 (Figure S4), suggesting that insensitivity of transposon mutants against POP72 was not due to the polar effects of the Tn5 insertions on adjacent gene expression. To test whether the inactivated genes played a role in the initial step of POP72 infection, phage adsorption was assayed with the WT and Tn5 mutants. As shown in Figure 2B, most of phage POP72 efficiently bind to the WT Pcc27 cells within 1 min, and no infectious particles were remained in the supernatant after 5 min. However, the phage POP72 adsorption to the Pcc27 cells was completely abolished in the cpsG::Tn5, wcaA::Tn5, and wzc::Tn5 mutants. The complementation of these genes restored the POP72 adsorption in the mutants (Figure 3B). In addition, the specificity of POP72 to the CA was verified further by phage inactivation assays using the CA extracted from the WT Pcc27 strain. As shown in Figure 3C, POP72 directly interacts with the CA. Approximately 80% of the POP72 particles were inactivated by 36 μg mL⁻¹ of Pcc27 CA at 28°C within 30 min. The inactivation might be due to a triggering of DNA ejection from the POP72 head upon CA binding, similar to the case of Podoviridae phage P22 that ejects its DNA after binding to the extracted phage receptor Salmonella LPS (Andres et al., 2010). These results confirm that the CA is the receptor for phage POP72. It is interesting to note that eight phages from the nine isolated phages including POP72 use CA as a receptor (Figure S4), implying a wide involvement of CA as a receptor among Pcc phages.

Bacteria generally evade phage infections by modifying or eliminating the phage receptor (Labrie et al., 2010). To investigate whether a similar phenomenon could occur in Pcc, spontaneous POP72-resistant Pcc27 mutants were isolated through the high-titer overlay assay as previously described (Kim and Ryu, 2011). Eleven mutants were obtained, and the resistance against POP72 infection was verified by the spot assay (data not shown). PCR amplification and DNA sequencing revealed missense mutations within the putative wzc gene from the two spontaneous mutants (Figure S5). In particular, one POP72-resistant mutant, designated Pcc27^Mu−1, contained a single base substitution at the 1081st nucleotide from the start codon of the putative wzc gene: guanine (G) was substituted by thymine (T). This substitution resulted in a replacement of valine (V) with phenylalanine (F) at amino acid residue 361 of the wzc gene product. A nucleotide adenine (A) was replaced with a cytosine (C) at the 1229th nucleotide from the start codon of putative wzc gene in another POP72-resistant mutant Pcc27^Mu−2, resulting in a substitution from aspartic acid (D) to alanine (A) in amino acid residue 410 of the gene product. Notably, these substitutions occurred near or at a G-rich domain characterized by a highly conserved GNVR sequence motif in the putative tyrosine kinase Wzc, suggesting that the substituted residues (V and D) are important for the normal activity of Wzc. When the spontaneous mutant Pcc27^Mu−1 was complemented with the WT wzc gene, POP72 formed clear plaques on the bacterial lawn similar to the WT level (Figure 4). In contrast, complementation with the mutant wzc gene (wzc^Mu−1) did not restore the susceptibility to POP72 (Figure 4), indicating that the single point mutation can lead to POP72 resistance. These results suggest that the CA biosynthesis genes, at least the putative wzc gene, play a critical role in the POP72 infection, probably biosynthesizing the phage receptor. Unidentified mutations in the genes other than the wzc gene related to the CA biosynthesis might be responsible for the POP72-resistance in the other nine spontaneous mutants.

Based on these results, we hypothesized that CA is an important exopolysaccharide for the initial infection of phage POP72. To determine whether the phage-resistant mutants lost their CA from the bacterial cell surface, the CA was extracted and quantified from a stationary phase culture of each strain as described by Meredith et al. (2007) and Kim et al. (2015). As expected, both Tn5-inserted and spontaneous phage-resistant mutants produced approximately half the amounts of CA produced by the WT Pcc27 (Figure 5A). Complementation of the inactivated genes reversed not only the phage sensitivity (Figure 3A) but also the amounts of CA biosynthesized to the WT Pcc27 level (Figure 5B), further suggesting that CA production is required for phage POP72 infection.

Pectobacterium spp. regulates the PCWDE expression under the control of quorum sensing using the N-acyl homoserine lactones (N-AHL) as signal molecules (Jones et al., 1993; Pollumaa et al., 2012). At high cell density, more than ~5.0 × 10⁶ CFU mL⁻¹, this plant pathogen produces PCWDE and causes soft rot diseases on host crops and vegetables (Liu et al., 2008). To test the potential ability of POP72 as an antimicrobial agent against Pectobacterium spp. infection in vegetables, Chinese cabbage artificially inoculated with Pcc27 was treated with the POP72 phage. Rifampicin-resistant Pcc27 (Pcc27^RifR) was used when it was necessary to quantify the Pcc cells recovered from Chinese cabbage. As expected, Pcc27, as well as Pcc27^RifR, caused a prominent soft rot disease in Chinese cabbage when the initial spiked bacterial cell density was more than 10⁶ CFU mL⁻¹ (data not shown). Pieces of Chinese cabbage (cut ~10 by 7 cm) inoculated with 10⁷ CFU mL⁻¹ of Pcc27 started to show a typical soft rot of cabbage tissues after 12 h of storage at 28°C. In contrast, symptom development, as well as bacterial growth, was comparatively retarded up to 12 h by the POP72 treatment (Figures 6A,B). Although the softening of the cabbage tissues treated with POP72 started after 12 h of Pcc27 infection, the extent of the tissue softening was still less significant compared with the untreated cabbage. In addition, the progress of the symptoms was mitigated with the higher concentration of POP72 (Figure 6A). The recovered Pcc population was also reduced to a significant level by POP72 treatment at MOIs of 10 and 100 compared with the SM buffer-treated control (Figure 6B). These results indicate that phage POP72 has the potential to serve as an antimicrobial agent to control Pectobacterium spp. Further studies to elucidate the optimum phage treatment protocol to control Pcc in Chinese cabbage are required.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.