Live cell imaging of Plasmodiophora brassicae—host plant interactions based on a two‐step axenic culture system

Abstract Plasmodiophora brassicae, a parasitic protist, induces club‐shaped tumor‐like growth of host Brassicas roots and hypocotyls after infection. Due to its soil‐borne nature and intracellular, biotrophic parasitism the infection biology and early pathogenesis remains in doubt. In this study, we have established a new protocol, based on a two‐step axenic culture of P. brassicae with its host tissues, for easy and in planta observation of cellular interactions between P. brassicae and host plants: first, coculture of P. brassicae with infected canola root tissues, on growth‐medium plates, enables the propagation of P. brassicae that serves as pure inoculum for pathogenicity assays, and second, the pure inoculum is subsequently used for pathogenicity tests on both canola and Arabidopsis seedlings grown on medium plates in Petri dishes. During the first axenic culture, we established a staining protocol by which the pathogen was fluorescently labeled with Nile red and calcofluor white, thus allowing in planta observation of pathogen development. In the pathogenicity assays, our results showed that axenic cultures of P. brassicae, in calli, remains fully virulent and completes its life cycle in both canola and Arabidopsis roots grown in Petri dishes. Combining visualization of fluorescent probe‐labeled P. brassicae structures with fluorescent protein tagging of Arabidopsis cellular components, further revealed dynamic responses of host cells at the early stages of P. brassicae infection. Thus, established protocols for in planta detection of P. brassicae structures and the live cell imaging of P. brassicae—Arabidopsis interactions provide a novel strategy for improving our detailed knowledge of P. brassicae infection in host tissues.

resting spores, primary infection of root hairs and epidermal cells where primary plasmodia, zoosporangia, and secondary zoospores are formed, and secondary infection resulting from released secondary zoospores that develop into multinucleate secondary plasmodia in the cortex and stele, resulting in cell hypertrophy and hyperplasy of infected roots (Kageyama & Asano, 2009). Finally, cleavage of secondary plasmodia produces numerous resting spores that are released into the soil, through infection tissue decay, where they can survive up to 20 years (Kageyama & Asano, 2009). Plasmodiophora brassicae displays a complex intracellular infection process involving various developmental stages during pathogenesis. Apart from resting spores released into the soil, the majority of P. brassicae life stages occur obligately inside distinct host tissues. Although much effort has been expended to understand the infection process, the intracellular, biotrophic life history of P. brassicae renders significant challenges to elucidating morphological and cellular characteristics of P. brassicae development in host tissues during clubroot disease initiation and development.
Many studies on host plant-P. brassicae interactions have been performed under artificial inoculation conditions in growth chambers and greenhouses, that is, flooding potting mix with resting spore suspensions (Sharma, Gossen, & McDonald, 2011a), growing a host plant in slurries of resting spores , and using sand-liquid culture (Agarwal, Kaul, Faggian, & Cahill, 2009;Deora, Gossen, & McDonald, 2013). However, observations of the infection process, especially during the early stages of infection, have proven difficult due to adherent soil and sand particles not easily removed by washing, subsequent root hair or root surface damage during root sample preparation and the presence of soil-borne bacteria, fungi and oomycetes on and/or inside host tissues, increasing the complexity of this system for elucidating the P. brassicae life cycle.
In addition to the inoculation in soil and sand, several dual culture systems such as callus culture (Takahashi et al., 2001), hairy root culture induced by Agrobacterium (Asano, Kageyama, & Hyakumachi, 2000;Asano, Kodama, & Kageyama, 2006), cell suspension culture , and adventitious root culture (Takahashi et al., 2006) have been developed for maintaining an axenic culture of P. brassicae within infected host tissues. Many of these studies have used Brassica napus or Brassica rapa as host species and provided adequate systems for studying P. brassicae-host cell interactions avoiding the influence of other microbe contaminants.
Various histological methods have been used to study P. brassicae-host plant interactions (Schuller & Ludwig-Müller, 2016), but routine techniques for selective labeling of pathogen structures have not been achieved. As we often see in reported studies on the cytology of clubroot, there is a striking resemblance between the primary plasmodium and the zoosporangium (sporulating plasmodium) of P. brassicae and host cellular compartments. In some instances, the appropriate identification of pathogen structures, using histological stains, has been rather ambiguous, especially during primary infection (Donald & Porter, 2004;Fei et al., 2016;Hwang et al., 2011;Sharma, Gossen, & Mcdonald, 2011b;Takahashi et al., 2006). Due to its obligate biotrophic nature, P. brassicae is not amenable for transgenic research, thus impeding the availability of fluorescent protein-tagged strains for pathogen visualization during the infection process. Therefore, there is an urgent need to develop a methodology that will enable the in planta detection and observation of

P. brassicae.
Abundant electron-dense lipid droplets have been identified in P. brassicae plasmodia and resting spores (Deora et al., 2013;Williams & McNabola, 1967). The presence of enriched lipid droplets within pathogen structures allows for the pathogen to be distinguished from surrounding host cells of infected root tissues and may therefore serve as targets for histological staining. Lipid droplets have a core of neutral lipids, mainly triacylglycerides and sterol esters. In many studies, 9-diethylamino-5H-benzo[a]phenoxazine-5-one (Nile red), a hydrophobic and metachromatic dye with poor solubility and fluorescence in water (Rumin et al., 2015), has been reported as a general lipid dye for both polar and nonpolar lipids using different excitation and emission wavelength ranges (Diaz, Melis, Batetta, Angius, & Falchi, 2008) in a wide range of biological systems (Guzmán, Jara, Duarte, & Presmanes, 2010;Kimura, Yamaoka, & Kamisaka, 2004;Kopischke et al., 2013;Miquel et al., 2014;Rumin et al., 2015). Detection of nonpolar, neutral lipids by Nile red staining is coupled with a short excitation wavelength such as EX/EM wavelengths of 488/550-650 nm, whereas detection of polar lipids can be achieved by Nile red staining using a long excitation wavelength of 530-560 nm (Guzmán et al., 2010;Li, Sun, Hicks, & Raikhel, 2015;Rumin et al., 2015).
Nevertheless, Nile red staining for intracellular lipid droplets for identifying P. brassicae structures within clubroot tissues warrants further investigation and potential procedure optimization.
In this study, we have developed a two-step dual culture protocol to obtain axenic cultures of P. brassicae from canola galls and to conduct pathogenicity assays on canola and Arabidopsis seedlings grown on medium plates in Petri dishes using pure inoculum from the axenic cultures. In addition, we have established and optimized live cell imaging techniques, by specific labeling of intracellular lipid droplets, the nucleus, and cell walls of the pathogen, that enable the identification of initial stages of the P. brassicae infection cycle. Combining visualization of fluorescent probe-labeled P. brassicae structures and fluorescent protein tagging of Arabidopsis cellular components further revealed dynamic responses of host cells to P. brassicae infection. The live cell imaging based on an axenic culture system has proved to be a valuable tool for investigating the cellular and molecular mechanisms underlying P. brassicae-host interactions.

| Pathogen inoculum
Canola, B. napus cv. "Westar," was used to maintain the pathogen P. brassicae Woron. P. brassicae inoculum used in this study may represent a mixture of pathotypes, but pathotype 3 would certainly predominate . To prepare resting spores, 5 g of airdried clubroot galls were soaked in 200 ml distilled water for 2 hr to soften the tissue before grinding in a mortar and pestle. After filtering through seven layers of cheesecloth, the spore concentration was estimated using a hemocytometer and adjusted to 5.0 × 10 7 spores/ml. Canola seeds were sown in a soilless potting mix (Sunshine Mix #3/LG 3; Sun Gro Horticulture Canada Ltd., Canada) and plants were grown under 22°C, a 16 hr/8 hr light/dark photoperiod in a growth chamber (Conviron, USA). Five-day-old seedlings of canola were inoculated by pipetting 5 ml of resting spore suspension (5.0 × 10 7 spores/ml) around the base of the seedlings shoots, where it meets the soil.

| Establishment of callus culture
The callus culture system was established according to Bulman et al. (2011) with minor modifications. Canola clubroot galls were harvested 5 weeks after inoculation. After washing thoroughly in tap water, galls were surface-sterilized with 70% ethanol for 1 min, 30% bleach with 0.05% tween (~1.4% sodium hypochlorite) for 20 min, followed by three washes in sterile distilled water. After washing, galls were cut into sections between 2 and 3 cm in length, which were sterilized with 70% ethanol for 1 min, 2% Chloramine-T (Sigma-Aldrich, Oakville, Canada) for 20 min, followed by three washes in sterile distilled water. After this second sterilization, the sections were cut into ~0.5 cm segments and were placed on a medium consisting of Gamborg B5 basal medium (Sigma-Aldrich, Oakville, Canada) plus 3% sucrose, 300 mg/L timentin (Gold Biotechnology, St. Louis, US), and 1.5% agar. Lateral root segments of non-inoculated plants were used as control, because mature main roots were inflexible and too difficult to cut into small segments.
After 2 week incubation in the dark at 23°C, gall segments showing callus growth on the cut edges were transferred to fresh medium minus timentin. At the same time, a thin section was removed from each cultured segment, divided into two pieces, with one piece placed into Luria-Bertani (LB) liquid medium and one onto Potato Dextrose Agar (PDA) medium to confirm that the gall/callus segment was free of any endophytic bacteria and fungi. Each medium was observed for 3 days to confirm no bacterial or fungal growth.
Endophytic microbe-free calli were used to inoculate host seedlings.

| Pathogenicity assay of canola and Arabidopsis seedlings grown in Petri dishes
Pathogenicity testing of canola and Arabidopsis seedlings grown on medium plates in Petri dishes, used canola cultivar Westar and Arabidopsis thaliana Col-0 ecotype. Westar seeds, surface-sterilized for 1 min in 70% ethanol, followed by 20 min in 30% bleach and three washes in sterile water, were grown in square Petri dishes on half-strength Murashige and Skoog (MS) medium in a growth chamber at 23°C with a 16 hr/8 hr, light/dark photoperiod (light intensity 120 µmol/ms). Approximately 10 canola seeds or 15-20 Arabidopsis seeds were germinated and grown on MS plate in Petri dish. After 4 days, each seedling was inoculated with 400 µl of 5.0 × 10 7 spores/ ml resting spore suspension, freshly prepared from callus, by pipetting onto the roots. Petri dishes were sealed thoroughly with Parafilm to confine the internal environment. The inoculation procedure in Arabidopsis was similar with one difference: The time for bleach sterilization was reduced to 5 min. The infection process was observed under transmission light microscopy (AxioPlan; Zeiss).
The pathogenicity tests have been repeated at least three times on both Arabidopsis and canola seedlings. Arabidopsis transgenic lines, expressing PMA4-GFP (Nottingham Arabidopsis Stock Centre, stock number N799380 (Teh & Moore, 2007)) and AT4G19150-GFP (Arabidopsis Biological Resource Centre, stock number CS84731 (Cutler, Ehrhardt, Griffitts, & Somerville, 2000)) were infected using the same inoculation procedure.

| Live in planta staining of Plasmodiophora brassicae
During tissue culture, Nile red (Sigma-Aldrich, Oakville, Canada) and calcofluor white (CFW, Sigma-Aldrich, Oakville, Canada) doublestaining techniques were used to observe pathogen development in calli. Calli were incubated in a final solution of 1 µg/ml Nile red, diluted in sterile water from a 10 mg/ml stock, for 10 min in the dark.
After a wash with sterile water, calli were incubated in two drops of CFW for 5 min in the dark following the supplier ' s instruction. After a final wash, calli were mounted on glass slides for observations. For live cell imaging of P. brassicae-Arabidopsis interactions, 4,6-Diamidine-2-phenylindole dihydrochloride (DAPI, Sigma-Aldrich, Oakville, Canada) staining of inoculated Arabidopsis roots was carried out in a working solution of 1 μg/ml DAPI, diluted from a 10 mg/ml stock in water, for 20 min in the dark, followed by a wash with sterile water and where appropriate, 1 µg/ml Nile red was added for 5 min. Images of stained roots were viewed and captured with a confocal laser-scanning microscope (Zeiss Meta 510).

| Confocal microscopy
To observe neutral lipids stained with Nile red an excitation wavelength of 488 nm, with emission recorded with a 585-615 nm filter set, was used. For polar lipids, an excitation wavelength of 543 nm was used, and emission recorded with a >650 nm filter set. For both CFW and DAPI stained samples, the excitation wavelength was 405 nm with emission measured between 420 and 480 nm. For imaging GFP, a 488 nm excitation wavelength was used with GFP fluorescence detected with a 505-530 nm band pass. All observations were obtained from at least three independent experiments. After acquisition, images were analyzed and processed with LSM Image Browser and Adobe Photoshop (Adobe Systems). Contrast and brightness levels were optimized.

| Establishment of dual culture of calli harboring Plasmodiophora brassicae
Tissue culture of clubroot gall segments of P. brassicae-infected canola roots at 35 days post inoculation (dpi) was established on Gamborg B5 medium plates without the addition of any plant growth regulators. To prevent other soil-borne microorganisms from confusing our observation of the P. brassicae life cycle in host B. napus, we established a modified Bulman dual culture system, taking care to remove all possible contaminating microorganisms (Bulman et al., 2011). Five days after incubation on B5 plates, even in the presence of 300 mg/L timentin, some of the surfacesterilized segments of B. napus clubroot galls frequently displayed either bacterial or fungal contaminations. Subsequently, uncontaminated segments were transferred to fresh medium plates with 300 mg/L timentin. After further incubation for 2 weeks, it became apparent that these transferred gall segments were free of obvious microbial contamination, and outgrowth (callus-like structures) of plant tissues/cells on the peripheral surface of the segments frequently appeared, whereas the non-inoculated B. napus root tissue segments did not produce visible calli on B5 medium plates in the absence of plant growth regulators ( Figure 1).
After 2 weeks incubation on B5 medium plates with antibiotics, uncontaminated calli were transferred to fresh B5 plates in the absence of antibiotics, and at the same time thin sections from these calli were transferred to nutrient-rich liquid LB medium and solid PDA medium plates to double check whether calli were indeed free of contaminating bacteria and fungi. Calli free of contamination were used in further examinations.

| Optimization of differential visualization of intracellular lipid droplets and membrane lipids stained with Nile red
Previous studies had demonstrated that secondary plasmodia and resting spores of P. brassicae contain abundant intracellular lipid droplets (Deora et al., 2013;Williams & McNabola, 1967) and that the mature resting spores contain chitin in cell walls (Moxham & Buczacki, 1983;Schwelm et al., 2015). To facilitate the examination of structures and development of P. brassicae in B. napus callus tissues, we employed two fluorescence probes, Nile red and CFW, for detection of lipid droplets and chitin cell walls, respectively.
Nile red, a lipophilic fluorescent marker, has been used previously to stain intracellular lipid droplets, abundant in secondary plasmodia and resting spores of P. brassicae (Deora et al., 2013;Williams & McNabola, 1967). CFW, selectively binding to β1-3 and β1-4 polysaccharides such as those found in cellulose and chitin, was used to label chitin in the cell walls of mature resting spores of P. brassicae (Moxham & Buczacki, 1983;Schwelm et al., 2015). To develop an applicable staining method for detection of intracellular P. brassicae in host tissues, the optimization of Nile red staining for cellular li-

| Pathogenicity of Plasmodiophora brassicae inoculum derived from axenic culture on host seedlings grown in Petri dishes
Resting spore suspensions of P. brassicae from calli were freshly prepared and used to inoculate both Arabidopsis and canola seedlings grown in Petri dishes, and proliferation of the pathogen within the At 60 dpi, sections from these small galls were stained with CFW and examined under confocal microscopy. Resting spores stained with CFW for cell wall depositions were clearly visible (Figure 4c).
Similarly, at 30 dpi with the resting spore suspension freshly prepared from callus tissues, small galls were observed under light microscopy, on the surface of the elongation zone of canola roots grown in Petri dishes (Figure 4d)

| Live cell imaging of Plasmodiophora brassicae-Arabidopsis interactions during early infection stages
Having shown that P. brassicae derived from callus tissue retains full virulence with the production of mature resting spores in Arabidopsis root tissues grown on medium in Petri dishes, we wished to establish a live cell imaging system to study P. brassicae-Arabidopsis interactions and pathogenesis during early infection stages, that remain largely unknown in the infection cycle (Schuller & Ludwig-Müller, 2016). With a high inoculation pressure (5 × 10 7 -10 8 spores/ml), Similarly, an epidermal cell of the PMA4-GFP-expressing plant with a primary plasmodium at the cytoplasmic periphery showed the green fluorescence of the PM marker (Figure 6b). These images confirmed that Arabidopsis cells hosting primary plasmodia at the initial infection stage maintained cellular integrity. At 3 dpi, intracellular primary plasmodia labeled by Nile red were observed in infected root epidermal cells of the PMA4-GFP-expressing plant. In these infected cells, the green fluorescence of the PM marker disappeared compared to adjacent epidermal cells where no pathogen was visible and the PMA4-GFP signal was retained (Figure 6b). These observations suggest that either the PM has been specifically depleted of the PMA4-GFP proteins or that it generally disintegrated when intracellular primary plasmodia were formed within the infected cells. In either case, whether this is actively triggered by the pathogen during pathogenesis or constitutes a defense response of the host cell remains to be elucidated.

| D ISCUSS I ON
The aim of this study was twofold: (a) to establish an axenic culture system for the production of pure inoculum of P. brassicae; and (b) to develop an in planta staining protocol for observation and detection of P. brassicae infection structures at the early stages of pathogenesis on medium plates using pure inoculum from the axenic culture. Arabidopsis and canola roots (Figure 4). These observations are consistent with previous reports that several root hair deficient mutants of Arabidopsis, including the hairless mutants rhd4-1 and axr1-12, are as susceptible to P. brassicae infection as wild-type Arabidopsis plants (Siemens, Nagel, Ludwig-Müller, & Sacristán, 2002). Rapid expansion of cells at the root elongation zone is coordinated by increased hydraulic pressure exerted from an osmolyte-filled vacuole together with the controlled relaxation of cell walls, as a result of polysaccharide interlinks being broken (Verbelen, Cnodder, Le, Vissenberg, & Baluška, 2006). This process likely contributes to ease of initial were conducted in soil conditions. We also observed that the level of the primary and secondary infection under the axenic culture system was much lower than that observed on soil-grown plants at the same stages after inoculation with the same inoculum density. This discrepancy could be due to the difference of the microenvironment between culture media and soil, which may involve pH, exposure to light, nutrient availability, presence of other microorganisms, and plant growth. For example, at the same growth stages, Arabidopsis plants grown on medium plates are generally much smaller than plants grown in soil. Since P. brassicae is an obligate intracellular parasite, and its nutrients are being supplied almost exclusively from the infected host cells. Restricted growth of the host plant and its cells on the medium plates may provide the nutrient limitation that will affect the rapid, massive growth, and reproduction of P. brassicae in host tissues, thus reducing gall enlargement.
One of the key challenges in clubroot research is the lack of a in secondary plasmodia and resting spores of P. brassicae (Deora et al., 2013;Williams & McNabola, 1967;Williams & Yukawa, 1967).
This key finding has led us to identify new histological probes for the observation of P. brassicae structures during the infection process.
We have employed Nile red, which is cell permeable and binds to polar and nonpolar lipids in live cells, for labeling and visualization of intracellular lipid droplets of P. brassicae by confocal microscopy using excitation/emission wavelengths of 488 nm/585-615 nm. Our results are indeed consistent with past reports, demonstrating that neutral, nonpolar lipid droplets are abundant in secondary plasmodia and resting spores as well as in primary plasmodia of P. brassicae (Deora et al., 2013;Williams & McNabola, 1967;Williams & Yukawa, 1967 whereas CFW binds to the β1-3 and β1-4 polysaccharides of cellulose and chitin, the latter is found in fungal cell walls and the cell walls of P. brassicae mature resting spores (Moxham & Buczacki, 1983;Schwelm et al., 2015). By using double-staining with Nile red and DAPI, multiple small spherical primary plasmodia with abundant Nile red staining and its compatibility with live cell imaging has been thoroughly studied in fungal, microalgae, animal and plant cells (Greenspan, Mayer, & Fowler, 1985;Guzmán et al., 2010;Kimura et al., 2004;Kopischke et al., 2013;Kucherak et al., 2010;Miquel et al., 2014;Rumin et al., 2015). In this study, simultaneous visualization of

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

E TH I C S S TATEM ENT
This study did not involve animal participants. Therefore, no ethics approval is required.

DATA ACCE SS I B I LIT Y
All data are included in the manuscript.