Phage-Resistant Bacteria Reveal a Role for Potassium in Root Colonization

ABSTRACT Bacteriophage predation is an important factor in bacterial community dynamics and evolution. Phage-bacterium interaction has mainly been studied in lab cultures, while dynamics in natural habitats, and especially in the plant root niche, are underexplored. To better understand this process, we characterized infection of the soil bacterium Bacillus subtilis NCBI 3610 by the lytic phage SPO1 during growth in LB medium and compared it to root colonization. Resistance in vitro was primarily through modification of the phage receptor. However, this type of resistance reduced the ability to colonize the root. From a line that survived phage infection while retaining the ability to colonize the root, we identified a new phage resistance mechanism involving potassium (K+) ion influx modulation and enhanced biofilm formation. Furthermore, we show that potassium serves as a stimulator of root colonization among diverse growth-promoting bacilli species, with implications for plant health.

also cause changes in colony morphology (14), genome-wide mutation rate (15), and lateral spread of genetic material (16). In recent years, insights have been gained into how phage-bacterium interactions affect bacterial growth in ocean (17) and mammalian gut ecosystems (18). However, little is known about the effects of phage-bacterium interactions in other ecosystems, especially in soil and the root rhizosphere (19).
To address this question, we utilized Bacillus subtilis strain NCBI 3610 (henceforth, B. subtilis) (20) and its cognate lytic phage SPO1 (21) to explore their interaction during root colonization of the plant model system Arabidopsis thaliana. B. subtilis is a Gram-positive spore-forming bacterium isolated from soil and is able to colonize plant roots (22). Root colonization by B. subtilis is mediated by formation of a biofilm on the root (23). Biofilms are bacterial communities encased in an extracellular matrix. The B. subtilis matrix is mainly composed of sugar polymers, encoded by the eps operon, and protein fibers encoded by the tapA-sipW-tasA operon (24). B. subtilis defective in biofilm formation exhibits severe defects in root colonization (23). The phage SPO1 is a lytic phage, representing a large and diverse group of Myoviridae bacteriophages, harboring a long contractile tail. SPO1 phage exhibits a complex infection cycle, involving the subversion of the host transcription machinery for its own use (25). SPO1 utilizes the wall teichoic acid polymers (WTAs) as receptors to invade B. subtilis cells (26). WTAs are long sugar polymers that are incorporated into the cell wall and membrane of B. subtilis and other Gram-positive bacteria (27). SPO1 binds these polymers when they are decorated by glucose moieties (gWTAs). Comparison of phage-bacterium evolution upon infection in LB medium and during root colonization revealed that phage infection in vitro and in planta exhibits different evolutionary trajectories. Characterization of bacteria resistant to phage infection during root colonization led to the identification of a novel phage resistance mechanism through modulation of potassium (K 1 ) ion influx and enhanced biofilm formation. Furthermore, we show that potassium serves as a stimulator of root colonization among diverse bacilli species.

RESULTS
Loss of phage receptor results in a fitness cost for root colonization. To explore phage-bacterium interactions during root colonization, we inoculated the roots of A. thaliana with B. subtilis 3610 bacteria together with SPO1 phage at either a high (phage/bacteria, 1:1), or low (phage/bacteria, 1:10) multiplicity of infection (MOI). Measuring bacterial colonization (CFU) after 48 h revealed a reduction in root colonization of ;95% at an MOI of 1 and an ;90% reduction at an MOI of 0.1 (Fig. 1A), indicating that SPO1 can efficiently infect and kill its host bacteria during root colonization. Significant levels of PFU were recovered from the root after SPO1 infection (5.56 Â 10 4 6 4.67 Â 10 4 , for MOI of 1, n = 6). However, no increase in the amount of the phage was observed in comparison to the input (PFU = 10 6 ), probably reflecting the fact that phage cannot adhere to the root without the presence of the host bacteria. Selection pressure by phage drives bacterial evolution of resistance mechanisms (28). To explore the evolution of phage resistance mechanisms in vitro versus those in planta, we isolated 300 bacteria (screen 1) that survived phage infection on the root and restreaked them on agar plates containing SPO1 phage, (maximum of 10 bacteria from each root). We found 20 in planta survivors that became resistant to phage infection in vitro. On the other hand, 100% of 70 bacteria isolated after surviving phage infection in LB medium (taken from 3 independent LB plates) became immune to further infection. Twenty randomly selected in vitro survivors became phage resistant through loss of the phage receptor, as judged by a lack of staining with concanavalin A, Alexa Fluor 488 conjugate (ConA 488 ), a lectin that binds specifically glycosylated WTA (gWTA) (Fig. 1B) (see also reference 26). Of the 20 SPO1-resistant bacteria isolated from the plant all but one (m28) had lost their ability to recolonize the root (Fig. 1C). Nineteen of these isolates concomitantly lost their ConA 488 staining (see Fig. S1A in the supplemental material). One of the isolates, m28, was completely phage resistant in vitro when infected in liquid culture ( Fig. S1B) but still exhibited faint ConA 488 staining ( Fig. S1A and C) and partial sensitivity when infected on roots ( Fig. 2A). The correlation between SPO1 sensitivity and root colonization ability suggests that gWTA is important for efficient root colonization. To test this hypothesis, we inoculated roots with DtagE bacteria, which lack gWTA (29), and found a significant reduction in root colonization by CFU measurement and fluorescent bacteria ( Fig. 1D and E). Because biofilm formation has been shown to be necessary for root colonization by B. subtilis (30), we tested the ability of DtagE bacteria to form a biofilm in vitro. DtagE cells exhibit normal biofilm formation both on agar plates and in liquid medium (Fig. 1F). Of note, gWTA is required for nasal epithelium colonization by Staphylococcus aureus (31). Our results suggest that gWTA in B. subtilis is similarly important for plant surface adhesion, irrespective of biofilm formation.
Bacteria evolved a novel phage resistance mechanism. Our results indicate that loss of the phage receptor results in a fitness cost to bacteria during root colonization. To identify alternative pathways utilized by bacteria to resist phage infection during root colonization, we utilized the bacteria that survived phage infection on the root (100 of the 300 bacteria from screen 1) and infected them with SPO1, this time during root colonization (screen 2). Most of the bacteria exhibited phage sensitivity similar to that of the parental bacteria. However, of 100 bacterial strains tested, we found 4 that survived SPO1 infection on roots better than wild-type (WT) cells ( Fig. 2A). To explore the mechanism of bacterial survival, we sequenced the genomes of m5, m11, m54, and m93, the 4 bacteria isolated during screen 2, along with m28 isolated from screen 1. Table 1 presents the mutations in each of the bacteria.
One mutant, m5, exhibited enhanced phage resistance when infected in planta at a medium MOI (1 to 0.1) (Fig. 2C, CFU, and Fig. 2D, fluorescent bacteria), phage sensitivity in vitro, and ConA 488 staining similar to that of WT ( Fig. 2B and E and Fig. S2B), indicating that it harbors a novel phage resistance mechanism. The m5 mutant also exhibited enhanced resistance to infection by SPP1, a lytic phage from a different bacteriophage family (Siphoviridae) (Fig. 2F). SPP1 utilizes YueB, a membrane protein, as a receptor. However, gWTA also plays an important role for SPP1 infection as well (32). Nevertheless, m5 has a resistance mechanism that is not phage-family specific. m5 harbors a single point mutation in the ktrC gene, encoding a low-affinity potassium channel. This missense mutation changes proline 189 into threonine (KtrC P189T) and resides in a conserved RCK domain, known to bind the regulatory molecule ci-di-AMP (33,34). ci-di-AMP negatively regulates potassium uptake (35). We hypothesized that KtrC P189T is a gain of function mutation, affecting the regulation of potassium uptake.
Potassium enhances root colonization and phage resistance through modulation of biofilm formation. Several potassium channels are encoded in the B. subtilis genome: ktrAB and kimA encode high-affinity channels (36,37), while the ktrCD operon encodes a low-affinity channel (36). To understand the effect of the KtrC P189T mutation and potassium uptake on bacterial root colonization, we examined the colonization efficiency of bacteria lacking these channels. While DktrC and DkimA mutants colonized the root with similar efficiency to WT bacteria (Fig. 3A), DktrA bacteria exhibited reduced root colonization (Fig. 3A). The KtrC P189T mutation is able to restore the colonization of DktrA bacteria (Fig. 3A, M5 DktrA), suggesting that this mutation enhances potassium uptake through the KtrC channel. Growing plants on media with different levels of potassium revealed a positive correlation between potassium concentration and root colonization   (Fig. 3C). Stimulation of root colonization was specific to potassium ions, as neither sodium nor nitrogen had a similar effect (Fig. S2A). Simply adding extra potassium to the plant growth medium was sufficient to enhance survival of WT cells upon phage infection ( Fig. 3C and D), reinforcing the idea that the KtrC P189T mutation affects phage infection through increased potassium uptake.
It was previously shown that B. subtilis bacteria sense potassium influx through the cell membrane, utilizing KinC and KinB protein kinases to induce biofilm formation Phage-Bacterium Interactions during Root Colonization ® genes and sliding motility, respectively (38,39). Both biofilm formation and sliding motility play important roles during root colonization (23,40). We found that m5 DkinC bacteria, but not m5 DkinB bacteria, lost SPO1 resistance (Fig. 3E). Addition of potassium to the plant growth medium failed to stimulate phage resistance for DkinC bacteria (Fig. S2C). Thus, we conclude that KinC is involved in the effect of potassium on root colonization and phage resistance.
KinC activation induces a phosphorylation cascade that culminates in induction of biofilm matrix genes (38). Disruption of the biofilm matrix operons epsA-O and tapA-sipW-tasA abolished the effect of potassium on root colonization (Fig. S2C). Interestingly, disruption of the tapA-sipW-tasA operon, but not epsA-O, reduced the phage resistance phenotype of m5 cells ( Fig. 3F and G), suggesting that an increase in the protein fiber component was responsible for the phage resistance effect of increased potassium influx. Thus, our analysis revealed a novel phage adaptation mechanism that works through enhanced potassium influx by stimulating biofilm matrix formation.
To further characterize the role of biofilm matrix formation in phage resistance, we monitored SPO1 infection during biofilm formation in vitro (20). Measurement of biofilm diameter on MSgg agar plates revealed significant reduction in colony diameter on plates containing SPO1 phage (Fig. 4A). Of note, DktrA cells, although able to form normal biofilm in the absence of phage, exhibited a significant decrease in biofilm diameter in the presence of phage in comparison to that of WT infected cells (Fig. 4A  and B). Similar to what was observed in plants, the KtrC P189T mutation was able to compensate for the increased sensitivity of DktrA cells (Fig. 4A and B). Thus, our in vitro analysis provides further support for the hypothesis that potassium influx modulates biofilm formation to enhance phage resistance. Equivalent results were obtained for pellicle biofilm formation on liquid MSgg medium (Fig. S3), where enhanced potassium influx in m5 bacteria, or addition of extra potassium to the medium of WT bacteria, increased the chance of surviving SPO1 infection, while decreased influx, due to the DktrA mutation, increased phage sensitivity.
Potassium enhances root colonization by diverse bacilli species. To determine if potassium is able to enhance colonization by other bacilli species, we analyzed its effect on Bacillus velezensis FZB42, a plant-associated bacterium (41), shown to enhance plant health through growth stimulation and pathogen inhibition. We found that potassium enhances root colonization of B. velezensis FZB42 (Fig. 5A), and this phenomenon was abolished in B. velezensis FZB42 DkinC cells (Fig. 5B). A similar phenomenon was observed for B. subtilis GB03 and B. pumilus SE34, two other plant growth-promoting bacteria (Fig. 5A). Thus, potassium serves as a wide-spread signal, stimulating root colonization by diverse bacilli species. B. velezensis FZB42 inhibits the growth of the plant fungal pathogen Rhizoctonia solani (42), raising the possibility that simply adding potassium to the growth medium could enhance B. velezensis FZB42 colonization and fungal protection. Indeed, potassium enhanced plant growth when inoculated with Rhizoctonia solani in the presence of B. velezensis FZB42 in comparison to growth of plants inoculated with fungi but no bacteria (Fig. 5C), and this effect of plant protection was abolished in DkinC cells, which are unable to sense potassium influx (Fig. 5C).

DISCUSSION
Phage infection has been extensively characterized in lab cultures. Recent work revealed that phage-bacterium interactions in natural environments can differ significantly from those  (43) due to the fitness constraints encountered by bacteria in their natural environment. This can impede the evolution of phage receptors, the main resistance pathway observed in the lab (44). We found that gWTA, the SPO1 phage receptor, is essential for root adhesion; thus, receptor modification has a fitness cost in the root environment. Of note, it has been shown that gWTA is essential for adhesion of Staphylococcus aureus to the nasal epithelium (31), while gWTA also serves as phage receptor for several S. aureus-infecting phages (45). It will be interesting to determine if phage-bacterium evolution is constrained in this niche as well. Phage-Bacterium Interactions during Root Colonization ® While evolution of receptor modification is not favored, we found that bacteria in the root adapt to phage infection through evolution of enhanced biofilm formation. Consistent with previous in vitro work (38), we show that during plant colonization, altering potassium efflux through KtrCD channels enhances B. subtilis biofilm formation and thus promotes survival upon phage infection (Fig. 6). Our results indicate that protein fiber density, encoded by the tapA-sipW-tasA operon, is the main determinant of SPO1 resistance. Interestingly, it has been shown that curli protein fibers are important for T7 resistance in a biofilm of Escherichia coli bacteria (46). Phage-bacterium interaction in biofilm is an emerging area of research (47). Biofilms are an important mode of life for bacteria colonizing natural environments, such as plant roots, as well as for pathogenic bacteria (48). Thus, in order to understand bacterial community dynamics in nature, and to deal with biofilm pathogenicity, it is important to explore the interaction of phage and bacteria in this specific niche.
Phage utilize bacterial appendages decorating the cell surface as receptors. Appendages such as flagella, type IV pili, exopolysaccharides, and transporters of nutrients, are important for bacterial survival in the natural environment and specifically, in the plant niche (9). Evolution of phage resistance has been associated with reduced host fitness on plants. An example is phage utilizing type IV pili (49), where the evolved mutant bacteria growly similar to the parental strain in planktonic culture but suffer from reduced biofilm formation and plant leaf colonization. Thus, it is important to explore phage-bacterium interactions in their natural context to gain insight into their coevolutionary dynamics.
Our genetic screen opens the way to discovering new phage resistance strategies that are not accessible during in vitro growth. An interesting future direction derives from the observation that among in planta phage-resistant bacteria, we found 4 of 100 with genetically inherited enhanced phage resistance. This suggests that 96% of the bacteria survived infection in a nongenetic manner. It would be interesting to determine how nongenetic mechanisms enable bacteria to survive phage infection. One possibility is by colonizing root niches that are less accessible to phage, a strategy that was recently demonstrated for phage infection of bacteria colonizing the gut ecosystem (50).
Potassium is an essential macronutrient for bacteria and plants. Unlike nitrogen and phosphate, it is not incorporated into cellular macromolecules but remains a soluble ion (51). Potassium is involved in the regulation of many processes, including osmotic Phage-Bacterium Interactions during Root Colonization ® adaptation, membrane potential, phloem transport, regulation of metabolic enzymes, and biotic and abiotic stress responses. Potassium, nitrogen, and phosphate, are the three main ions consumed by plants and are heavily utilized for fertilization in modern agriculture. (52). The effect of nitrogen and phosphorous on root microbial communities has been characterized (6,53,54). However, the role of potassium in bacterial colonization is less well explored. Our analysis suggests that addition of potassium is a simple method to enhance bacillus colonization. Many bacillus species exert positive effects on their plant host by enhancing plant growth and inhibiting pathogens. Our results suggest that adding potassium could enhance the ability of B. velezensis to protect seedlings from R. solani infection.

MATERIALS AND METHODS
Bacterial strains and growth conditions. WT B. subtilis NCBI 3610 and AR16 (55) (B. subtilis PY79) were kindly provided by Sigal Ben-Yehuda (Hebrew University); DtagE (BKK35730), DktrA (BKK31090), DktrC (BKK14510), and DkimA (BKK04320) strains (all in B. subtilis 168) were purchased from the Bacillus Genetic Stock Center (http://www.bgsc.org/). Genomic DNA was extracted from AR16 and mutant strains using a Wizard genomic DNA purification kit (Promega) and transformed into B. subtilis NCBI 3610. The media and growth conditions used for DNA transformation of B. subtilis NCBI 3610 are described at http://2013.igem.org/Team:Groningen/protocols/Transformation. The bacteria were cultivated routinely on Luria broth (LB) medium. When needed, the medium was solidified with 1.5% agar. For biofilm formation, bacteria grown overnight were inoculated into MSgg medium and incubated without shaking for 2 days at 30°C as described in reference 20. This medium was also solidified with 1.5% agar.
Phage strains and infection conditions. SPO1 (1P4) and SPP1 (1P7) phage were purchased from the Bacillus Genetic Stock Center. Phage lysate was routinely prepared by adding approximately 10 9 phage to mid-log cells grown in LB medium supplemented with 10 mM MgSO 4 , until the culture was completely cleared. The lysate was then filtered through a 0.22-mm Millipore filter. For lysis dynamics in LB medium ( Fig. 2E and Fig. S1B), approximately 10 9 phage were added to mid-log-phase growing cells at multiplicity of infection (MOI) of 1, and the optical density at 600 nm (OD 600 ) was monitored at 30-min intervals. For SPO1 infection in pellicle biofilm (Fig. S3), bacteria grown overnight were inoculated into MSgg medium, Tzipilevich and Benfey ® and phage was added at a MOI of 1. Bacteria were incubated without shaking for 2 days at 30°C, and the number of cultures with surviving cells was counted by eye. For SPO1 infection in solid biofilm (Fig. 4), bacteria grown overnight were inoculated into MSgg medium supplemented with 10 5 PFU and solidified with 1.6% agar. A spot of 5 ml from the bacterial culture was inoculated onto the plate and incubated for 2 days at 30°C. Plates were scanned and colony diameter was determined using ImageJ.
Monitoring bacterial growth and phage infection on plant roots. Plants were grown on 0.5 MS medium containing 1.1 g Murashige and Skoog basal salts (in 500 ml double-distilled water [ddH2O]), 1% sucrose, 1% agar, and 5 ml (in 500 ml ddH 2 O) morpholineethanesulfonic acid (MES; 50 g/liter, titrated to pH 5.8 with NaOH). Plants were stratified for 2 days in a 4°C dark room and grown vertically for 4 to 10 days under long-day light conditions. Bacteria from fresh colonies were grown in LB medium to an OD 600 of 1.0 and then diluted 1:100 in 1Â phosphate-buffered saline (PBS) for CFU measurements and microscopy, yielding approximately 1 Â 10 6 . Phage were added to the 1Â PBS at an MOI of 0.1 unless otherwise indicated. Six-day-old seedlings were transferred onto square petri dishes containing 0.5 MS but without sucrose. Two microliters of bacterial dilution was deposited immediately above the root tip and left to dry for 2 min. The square plates were kept in a vertical position during the incubation time at 22°C under long-day light conditions (16 h light/8 h darkness) in a plant growth chamber. For bacterial CFU counting and microscopy, plants were incubated with bacteria for 48 h. Then, the inoculated plant roots were cut and washed three times in sterile water. For CFU counting, the seedlings were transferred to a tube with 1 ml of 1Â PBS and vortexed vigorously for 20 s; then, a serial dilution was plated on LB plates. For PFU measurement, roots were treated as described for CFU, but without the washing step.
Microscopy. Roots were observed using a Zeiss LSM 880 laser scanning confocal microscope with 20Â lens objective. ConA 488 (Thermo Fisher C11252) staining was performed as described previously (26). Bacteria were observed using a Zeiss LSM 880 laser scanning confocal microscope with a 40Â (water immersion) lens objective. Fluorescence intensity was measured using ImageJ.
DNA sequence analysis. DNA was extracted from an overnight culture of the mutant bacterial strain along with the WT strain using a DNeasy PowerSoil Pro kit (Qiagen), according to the manufacturer's instructions. DNA sequencing (DNA-seq) libraries were prepared using a Nextera DNA Flex library preparation kit (Illumina), according to the manufacturer's instructions. Illumina NextSeq 500 high-output 50bp single reads were partially assembled into contigs using the SPAdes algorithm (56) with default parameters. The contigs were aligned to the genome of the WT bacteria from our lab stock.
Data analysis and statistics. All graphs and statistical tests were performed in Excel. P values throughout the paper were determined using Student's t test, except for Fig. S3, where a chi-square test was performed.
Data availability. All the data utilized to generate the graph can be found online at https://www .doi.org/10.6084/m9.figshare.14925375. The DNA-seq data set is available in the SRA repository under accession number PRJNA729435.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.