Methylation of foreign DNA overcomes the restriction barrier of Flavobacterium psychrophilum and allows efficient genetic manipulation

Abstract

Flavobacterium psychrophilum causes bacterial cold-water disease (BCWD) in salmonids and other fish, resulting in substantial economic losses in aquaculture worldwide. The mechanisms F. psychrophilum uses to cause disease are poorly understood. Despite considerable effort, most strains of F. psychrophilum have resisted attempts at genetic manipulation. F. psychrophilum restriction-modification (R-M) systems may contribute to this resistance. Restriction endonucleases (REases) rapidly degrade nonself DNA if it is not properly methylated by their cognate DNA methyltransferases (MTases). We used comparative genomics to show that R-M systems are abundant in F. psychrophilum and that strain-specific variations partially align with phylogeny. We identified two critical type II R-M systems, HpaII-like (FpsJI) and ScrFI-like (FpsJVI), that are conserved in most of the sequenced strains. Protection of foreign DNA against HpaII and ScrFI was achieved by expression of the MTases M.FpsJI and M.FpsJVI in Escherichia coli. Furthermore, deleting the two REase genes from F. psychrophilum resulted in efficient conjugative DNA transfer from E. coli into the otherwise genetically intractable F. psychrophilum strain CSF259-93. This allowed us to construct a CSF259-93 mutant lacking gldN, a core component of the type IX protein secretion system. The pre-methylation system developed in this study functions in all tested F. psychrophilum strains harboring HpaII-like and ScrFI-like REases. These newly developed genetic tools may allow the identification of key virulence factors and facilitate the development of live attenuated vaccines or other measures to control BCWD.

Importance: Bacterial cold-water disease (BCWD) caused by Flavobacterium psychrophilum is a problem for salmonid aquaculture worldwide, and current control measures are inadequate. An obstacle in understanding and controlling BCWD is that most F. psychrophilum strains resist DNA transfer, thus limiting genetic studies of their virulence mechanisms. F. psychrophilum restriction enzymes that destroy foreign DNA were suspected to contribute to this problem. Here, we used F. psychrophilum DNA methyltransferases to modify and protect foreign DNA from digestion. This allowed efficient conjugative DNA transfer into nine diverse F. psychrophilum strains that had previously resisted DNA transfer. Using this approach, we constructed a gene deletion mutant that failed to cause disease in rainbow trout. Further genetic studies could help determine the molecular factors involved in pathogenesis and may aid development of innovative BCWD control strategies.

Keywords: DNA methyltransferases; DNA transfer; Flavobacterium psychrophilum; bacterial cold-water disease; genetic manipulation; restriction-modification systems; type IX secretion system.

Fig 1

Distribution of MTase-encoding genes in F. psychrophilum genomes. (A) Conservation of MTase-encoding genes of R-M systems in a collection of 17 strains retrieved from various fish species and geographic origins: blue, orthologous gene present (>90% identity in protein sequence); grey, genes sharing between 50% and 90% identity in protein sequence (each group of homologs is surrounded by a black line); white, gene not conserved (<50% identity in protein sequence). The numbers “2” and “3” inside some squares indicate the number of gene copies present in the genome. All other squares (with no number) indicate that a single copy is present. MLST, multi-locus sequence type data from F. psychrophilum PubMLST database (71); CC-ST, clonal complex-sequence type. Fish host: RbT, Oncorhynchus mykiss; Ayu, Plecoglossus altivelis; CoS, Oncorhynchus kitsutch. MLST data are not available for CN38. (B) R-M system details from the REBASE database (69): R-M type, MTase name, enzymatic activity (methylated bases and motif specificity), and corresponding REase prototype for isoschizomers. Information is provided for enzymes with demonstrated activity based on PacBio DNA methylation data, available respectively for strains JIP02/86 and NCIMB 1947(T) in references and . (C) Locus tag of R-M system genes: MTase, DNA methyltransferase encoding gene; REase, restriction endonuclease encoding gene; n.p., gene not present.

Fig 2

Maps of pACYC184 and the representative methylation plasmids. M.FpsJI and M.FpsJVI-encoding genes were cloned between the BamHI and SalI sites of pACYC184 to generate pSS01 and pSS02, respectively. M.FpsJVI gene was cloned between the SalI and NarI sites of pSS01 to generate pSS05. Other methylation plasmids listed in Table S1 (pSS07, pSS08, pSS11, and pCB01) were constructed by inserting the corresponding MTase gene between the BamHI and NarI sites of pSS01. Numbers immediately inside of the ring refer to base pairs of sequence. p15A ori refers to the origin of replication that functions in E. coli, but not in F. psychrophilum. camR and tetR confer chloramphenicol and tetracycline resistance on E. coli, respectively, but not on F. psychrophilum. Binding sites for primers used in PCR reactions to clone the MTase genes are shown by the arrows inside and outside of the rings, with the perpendicular ends indicating the actual binding sites.

Fig 3

Improvement of conjugation efficiency in F. psychrophilum CSF259-93 using different methylation plasmids. The E. coli S17-1 λ pir strains, carrying both pCP11 and one of the different methylation plasmids, were used as the donor, and wild-type F. psychrophilum CSF259-93 was used as the recipient in the conjugations. E. coli S17-1 λ pir carrying pCP11 and pACYC184 was used as a negative control. Equal amounts of F. psychrophilum and E. coli cells were used in all the conjugation experiments. Erythromycin-resistant CFUs were counted from four plates in each conjugation experiment. Error bars represent standard deviation.

Fig 4

In vitro restriction digestion by HpaII (A) or ScrFI (B) to verify the pre-methylation-mediated protection of pCP11. The digestion pattern of properly methylated pCP11 isolated from the wild-type F. psychrophilum CSF259-93 (grown at 18°C) was used as a control to show the full protection of methylation (lane 2). Lane 1 was loaded with the DNA ladder (Thermo Scientific, SM1163). Lanes 3–7 used plasmids isolated from E. coli S17-1 λ pir grown at different temperatures. The E. coli strains contained pCP11 only, pSS05 only, or pCP11 plus pSS05 were cultivated at 37°C or 18°C as indicated. Samples in panel A were digested by HpaII and in B digested by ScrFI. Intact pCP11 is 9448 bp, and pSS05 is 6745 bp. “Fp” indicates F. psychrophilum CSF259-93. “Ec” indicates E. coli S17-1 λ pir.

Fig 5

Elimination of the Fps.HpaII and Fps.ScrFI REases in F. psychrophilum CSF259-93 resulted in increased conjugation efficiency. The E. coli S17-1 λ pir strains carrying pCP11 plus pSS05 or pCP11 plus pACYC184 were used as the donor, and wild-type F. psychrophilum CSF259-93 or mutants lacking Fps.HpaII (∆FPSM_02393) or Fps.ScrFI (∆FPSM_00613) were used as recipients in the conjugations. Equal amounts of F. psychrophilum and E. coli cells were used in all the conjugation experiments. Erythromycin-resistant CFUs were counted from four plates in each conjugation experiment. Comparison of conjugation efficiencies was performed by t test using GraphPad Prism v10.2.3. Error bars represent standard deviation.

Fig 6

Photomicrographs of F. psychrophilum colonies (row A) and gliding motility of individual cells on agar (row B). WT indicates wild-type CSF259-93, ∆gldN indicates gldN deletion mutant, and ∆gldN + pSS13 indicates gldN mutant complemented with wild-type gldN on pSS13. Colonies grown from single cells were incubated for 8 days at 18°C on 5% TYES solidified with 1% agar (row A). Scale bar beneath row A indicates 0.5 mm and applies to all panels of row A. For individual cell motility assay (row B), cells were grown in TYES at 18°C with shaking for 48 h, spotted on a pad of full-strength TYES solidified with 1% agar on a glass slide, and covered with an O₂-permeable Teflon membrane. A series of images were taken, and individual frames were colored from red (time 0) to yellow, green, cyan, and finally blue (24 s) and integrated into one image, resulting in “rainbow traces” of gliding cells. White cells correspond to cells that exhibited little if any net movement. The scale bar at lower right indicates 10 µm and applies to all panels of row B.

Fig 7

Secreted proteolytic activity of the wild-type F. psychrophilum CSF259-93 (WT), gldN mutant (∆gldN), and complemented strain (∆gldN + pSS13). Equal amounts of cells were spotted on TYES supplemented with 1.5% skim milk and incubated at 18℃ for 7 days. Clear zones surrounding the area of growth indicate hydrolysis of skim milk by the secreted enzymes.

Fig 8

The percent survival of rainbow trout after challenge with F. psychrophilum. For panel A (trial 1), each fish (Troutlodge February spawning line) was injected as indicated in Materials and Methods with 20 µL of PBS, wild-type CSF259-93 (9.17 ± 0.32 × 10⁶ CFU), gldN mutant (11.1 ± 1.08 × 10⁶ CFU), and the complemented strain (7.57 ± 1.23 × 10⁶ CFU). For panel B (trial 2), each fish (Troutlodge May spawning line) was injected with 50 µL of PBS, wild-type CSF259-93 (17.7 ± 2.57 ×10⁶ CFU), gldN mutant (17.8 ± 2.25 × 10⁶ CFU), and the complemented strain (26.7 ± 0.76 × 10⁶ CFU). The percent survival was observed and measured. Comparison of survival curves was performed by Log-rank (Mantel-Cox) test using GraphPad Prism v9.3.1.

Fig 9

Improvement of conjugation efficiency in different strains of F. psychrophilum by the methylation plasmid pSS05. The E. coli S17-1 λ pir strains carrying pCP11 plus pSS05 or pCP11 plus pACYC184 (negative control) were used as the donors, and wild-type F. psychrophilum strains were used as the recipients in the conjugations. Equal amounts of F. psychrophilum and E. coli cells were used in all the conjugation experiments. Erythromycin-resistant CFUs were typically counted from four plates in each conjugation experiment. Data from a total of 12 plates and three independent experiments were used for CN38 and OSU THCO2-09. Comparison of conjugation efficiencies was performed by t test using GraphPad Prism v10.2.3. Error bars represent standard deviation.

Fig 10

Scheme of the pre-methylation system developed in this study. Ap^r, ampicillin resistant; Cm^r, chloramphenicol resistant; Em^r, erythromycin resistant (expressed in F. psychrophilum but not in E. coli); and Em^s, erythromycin sensitive.