The specialization of Esteya vermicola hyphae in infection to Bursaphelenchus xylophilus and its colonization of pine tree

Pine wilt disease (PWD) caused by the nematode Bursaphelenchus xylophilus is a serious problem on pines, and there is currently no effective control strategy for this disease. Although the endoparasitic fungus Esteya vermicola showed great effectiveness in controlling pine wilt disease, the colonization patterns of the host pine tree xylem by this fungus are unknown. To investigate the colonization patterns of pine xylem by this fungus, the species Pinus koraiensis grown in a greenhouse was used as an experimental host tree. The fungal colonization of healthy and wilting pine trees by E. vermicola was quantified using PCR with a TaqMan probe, and a green fluorescence protein (GFP) transformant was used for visualization. The results reported a specific infection approach used by E. vermicola to infect B. xylophilus and specialized fungal parasitic cells in PWN infection. In addition, the inoculated blastospores of E. vermicola germinated and grew inside of healthy pine xylem, although the growth rate was slow. Moreover, E. vermicola extended into the pine xylem following spray inoculation of wounded pine seedling stems, and a significant increase in fungal quantity was observed in response to B. xylophilus invasion. An accelerated extension of E. vermicola colonization was shown in PWN-infected wilting pine trees, due to the immigration of fungal-infected PWNs. Our results provide helpful knowledge about the extension rate of this fungus in healthy and wilting PWN-susceptible pine trees in the biological control of PWD and will contribute to the development of a management method for PWD control in the field. Author summary Pine wilt disease, caused by Bursaphelenchus xylophilus, has infected most pine forests in Asian and European forests and led to enormous losses of forest ecosystem and economy. Esteya vermicola is a bio-control fungus against pinewood nematode, showed excellent control efficient to pine wilt disease in both of greenhouse experiments and field tests. Although this bio-control agent was well known for the management of pine wilt disease, the infection mechanism of fungal infection and colonization of host pine tree are less understand. Here, we use GFP-tagged mutant to investigate the fungal infection to pinewood nematode; additionally, the temporal and spatial dynamics of E. vermicola colonize to pine tree were determined by the TaqMan real-time PCR quantification, as well as the response to pinewood nematode invasion. We found a specific infection approach used by E. vermicola to infect B. xylophilus and specialized fungal parasitic cells in PWN infection. In addition, the fungal germination and extension inside of pine tree xylem after inoculation were revealed. In addition, the quantity of E. vermicola increased as response to pinewood nematode invasion was reported. Our study provides two novel technologies for the visualization and detection of E. vermicola for the future investigations of fungal colonization and its parasitism against pinewood nematode, and the mechanisms of the bio-control process.


Introduction
The pinewood nematode (PWN), Bursaphelenchus xylophilus, infects pine trees as the causal agent of pine wilt disease (PWD) [1] and is transmitted through the wounds created by the sawyer beetle (Monochamus spp.) feeding [2]. The PWN feeds on the epithelial cells and resin ducts of its host, which leads to dysfunction of vascular organs.
Subsequently, the PWN is distributed throughout the sapwood of the branches, trunk, and roots of susceptible host plants [3]. Polymerase chain reaction (PCR) amplification has been applied for the molecular detection of various organisms. To detect E. vermicola after inoculation or spraying on pine trees, Wei et al. (2014) [16] developed a simple and easy method with FTA-DNA extraction and PCR amplification from environmental samples based on the presence of a specific 176 bp fragment. Subsequently, a novel pinewood sample preparation and DNA extraction method were established [17], in which another three primer pairs based on chitinase, β-tubulin, and large subunit ribosomal RNA genes were designed for PCR amplification. However, the above two techniques cannot provide quantification analyses to reveal the colonization patterns of E. vermicola. More recently, a precise and accurate quantitative technique for E. vermicola in environmental samples was developed with TaqMan PCR quantification [18].
The fungus was inoculated into pine tree xylem as a nonnative species due to its parasitic behavior against PWNs. Thus, the colonization and development of E. vermicola in host pine trees before or during PWN invasion are the most concerning.
However, few studies on the temporal and spatial dynamics of fungal colonization of susceptible tree xylem following artificial inoculation has been conducted. Thus, the above-developed approaches were used in the present study to reveal the temporal and spatial patterns of pine xylem colonization by E. vermicola, with or without the presence of PWN.

Construction of calibration curves
In the nested-based TaqMan PCR, the 473-bp and 213-bp DNA products were generated using primers 28S-1F/28S-1R and 28S-2F/28S-2R, respectively. The application of nested TaqMan PCR allowed the detection of target DNA at quantities as low as 10 -5 ng. The standard curves for real-time TaqMan PCR quantification of both wt E. vermicola and CNU120806gfp obtained in the present study are shown in Supplementary Fig. S1. The curves showed good fits, with R 2 values of 0.9838 and 0.9969 for traditional and nested TaqMan PCR of wt E. vermicola, respectively (Supplementary Fig. S1a -b). The C t values of traditional TaqMan PCR ranged from 25.81 ± 0.82 to 41.24 ± 1.04, and those of nested TaqMan PCR ranged from 14.86 ± 0.16 to 32.93 ± 0.07. For CNU120806gfp, R 2 values of 0.9759 and 0.9927 for traditional and nested TaqMan PCR, respectively, were obtained ( Supplementary Fig. S1c -d).
The detectable concentration of the genomic DNA of pure cultured E. vermicola with conventional TaqMan PCR ranged from 10 -2 ng/µl to 10 2 ng/µl. However, the lowest detectable concentration of template using nested PCR, which significantly increased the detection sensitivity, was 10 -5 ng/µl.

Temporal infection dynamics of PWN by E. vermicola
The fungal infection pattern of PWN by the nematophagous fungus E. vermicola was studied by quantification of fungal hyphae. The single-PWNs that were classified by the infection stage were assayed, the infection status of each stage of these tested PWNs is shown in Supplementary Fig. S2 and the quantification results of fungal hyphae are illustrated in Fig. 1. The results showed that no genomic DNA of E. vermicola was detected in non-infected PWNs. A small amount of fungal DNA was detected in the infected-PWNs that were under the adhesion and infection stage, as well as the earlyfungal elongation stage. The fungal hyphal quantity, as expected, was significantly increased in the stages of mid-fungal elongation, late-fungal elongation, and conidium production.

Morphological observation of parasitic fungal hyphae
The morphological characteristics of fungal hyphae parasitizing PWN were observed with the help of green fluorescence emission of GFP-tagged mutant CNU120806gfp ( Fig. 2 and Fig. 3). According to our observations, fungal infection occurred 20 hours after the lunate conidia adhered to the cuticle of the PWN. An ovoid propagule was produced and implanted into the pseudocoelom of PWN by the fungal lunate conidia ( Fig. 2a-c). The propagule found in the PWN pseudocoelom showed morphological features similar to those of fungal cells germinated by the lunate conidia cultured on a PDA plate (Fig. 2d-e). In addition, a papillary bulge was found at the center of the concave side of the germinated lunate conidia (Fig. 2d). Additionally, a similar structure was found at the germinated lunate conidia that were cultured on the PDA plate (Fig.   2e). The dumbbell-like fungal cells produced by the ovoid propagule were found in the PWN (Fig. 3a-d), which was significantly different from the fungal hyphae grown either on PDA plates (Fig. 3e) or in PDB fermentation (Fig. 3f) in morphological features. The dumbbell-like fungal hyphae released from infected PWNs were cultured in situ for morphological comparison. The newly produced fungal cells, as expected, showed morphological features similar to those of the fungal hyphae grown on PDA plates (Fig. 3d).

Temporal and spatial dynamics of E. vermicola colonization in healthy pine xylem
To examine the colonization patterns of E. vermicola, a blastospore suspension of CNU120806gfp was used for the inoculation of pine tree xylem. Real-time TaqMan PCR assays of DNA extracts from wood with primers and probes were conducted to discover the temporal and spatial dynamics of E. vermicola colonization. After inoculation, the conidia of E. vermicola accumulated in the wood within 1 cm of the injection site (Fig. 4a). Fifteen days later, E. vermicola was detected in the wood 2 cm away from the injection site (Fig. 4b). Wood samples within 5 cm of the injection site harbored this nematophagous fungus 2 months after inoculation (Fig. 4c-e). Although a small number of fungal hyphae were found in the pine tree, the extension of E. vermicola in the pine tree xylem at lengths indicated that the fungus germinated and grew inside the host plants. Although the relative quantities were low, as expected, the fungal hyphae were found at 90 and 180 dai, suggesting that the fungus E. vermicola can stably survive in host pine trees (Fig. 4e-f). Additionally, a slight increase was found at 6 months after inoculation.
With the help of GFP fluorescence emission under a microscope, the visualization of fungal germination and colonization of E. vermicola is shown in Fig. 5a Fig. S3a-c), which led to a subsequent fungal hyphal colonization. Moreover, newly produced lunate conidia, which are mainly responsible for the effectiveness of fungal infection of PWNs, were found in the inoculated tree xylem (Fig. 5a).

Temporal dynamics of E. vermicola colonization of pine seedling xylem and its response to PWN invasion
The xylem of pine seedlings was colonized by the E. vermicola mutant CNU120806gfp via artificial wounds on the branches. After spraying to the surface of wounds, the fungus E. vermicola was detected in the xylem from 7 to 28 dai (Fig. 6). According to our quantifications, however, this biological control agent cannot be detected at 45 and 60 dai. In addition, the fungal hyphae of CNU120806gfp were observed in wood samples from the tested pine seedlings (Fig. 5c). Although the frequency of fungal hyphae that appeared in wood sections of tested pine seedlings was very low, the occurrence of fungal hyphae suggested successful colonization by the fungus E.

vermicola.
To reveal the response of E. vermicola in pine seedlings to PWN invasion, the seedlings were infected with PWNs at 60 days after fungal inoculation (Fig. 6, black arrowhead).
The results showed that the quantities of E. vermicola slightly increased from 4 days after PWN infection, although in extremely low quantities. This result suggested that E. vermicola inhabited the pine tree xylem as a low-richness species; however, E. vermicola quantities increased when PWN invasion occurred.
The PWNs were extracted from wood samples of these tested seedlings. B. xylophilus was first extracted from pine seedlings at 2 dai, including the fungus-infected and uninfected PWN ( Fig. 7a and 7b). In the pine seedlings inoculated with E. vermicola, the total number of extracted PWNs increased within 6 dai and then decreased from 8 dai onward (Fig. 7a); while in the control group, the number of extracted PWNs increased continuously after inoculation (P < 0.01) (Fig. 7b). The PWN infected by CNU120806gfp extracted from pine seedlings was visualized using white light and fluorescence microscopy (Fig. S4a). In most cases, the infected PWN was in the early stage of fungal infection by E. vermicola. The relative quantity of E. vermicola assayed by TaqMan probe quantification illustrated the infection status of extracted PWN from pine seedlings (Fig. 7c). The relative quantity of E. vermicola DNA in PWNs recovered from fungus-treated plants increased from 4 dai, and reached the top value at 10 dai.
Subsequently, the relative quantity of fungal genomic DNA was decreased followed by a decline in the PWN amount recovered from pine seedlings from 12 dai.  Table 1). The results indicated that E. vermicola was only detected in the wood tissue around the inoculation site within 7 dai. Two weeks later, however, the fungus E. vermicola was detected 10 cm away from the inoculation point. Subsequently, the genomic DNA of E. vermicola was found in the xylem 50 cm away from the inoculation point at 21 dai. However, the fungal colonization patterns showed jumping spread in the wilting pine tree. The extension patterns that appeared in wilting plants were thought to be due to the immigration of parasitized PWNs in the pine xylem. The hyphal morphology of E. vermicola inoculated in PWN-infected wilting pine xylem was observed (Fig. 5d). The fungal hyphae in the wilting pine tree showed higher abundance and more irregular development than those in healthy trees.

PWN density in the wilting pine tree
The PWNs inhabiting wilting host plants were measured in the present study, and the results are shown in Fig. 8. As shown in the figures, the PWN density in the xylem of both E. vermicola and sterile ddH 2 O-treated wilting pines was increased. However, the number of PWNs per 10 g of the wood sample around the E. vermicola inoculation point decreased from 7 dai to 21 dai ( Fig. 8a-d), and PWN could not be extracted at 28 and 42 dai ( Fig. 8e-f). Additionally, the PWN density in the wood 10 and 20 cm away from the inoculation site significantly declined after fungal inoculation ( Fig. 8c-f).

Fungal quantity of E. vermicola in extracted PWNs
Additionally, the amount of E. vermicola in the body of extracted PWNs was quantified using TaqMan PCR quantification ( Table 2). The results showed that E. vermicola

Discussion
Fungal colonization and in vivo infection of B. xylophilus by E. vermicola inside susceptible pine trees is crucial for the biocontrol of PWD. Moreover, fungal colonization of the susceptible pine xylem before potential PWN invasion can significantly improve the control effects [8,9]. For a biological agent that infects and controls PWN in pine trees, the colonization and extension of E. vermicola inside of its host tree are essential. However, the temporal and spatial dynamics of E. vermicola colonization of the host pine xylem are unknown in both healthy and wilting pine trees.
Thus, a time-course investigation on the presence of E. vermicola in healthy and PWNinfected pine trees is important to understand fungal colonization and is helpful to improve forest management.
By visualizing the parasitic hyphae in infected PWNs, the specific infection approach used by E. vermicola to infect B. xylophilus was explored. Although the infection of B.
xylophilus by E. vermicola has been studied [11,19], the specific implantation method and specialized fungal hyphae in PWN infection were first reported in the present study.
The infection process of the plant-parasitic nematode by endoparasitic fungi has been investigated using the GFP-labeled fungal transformant [13,20,21]. Spatial and temporal analyses of E. vermicola within healthy pine tree xylem using TaqMan qPCR techniques revealed the patterns and dynamics of fungal colonization and the extension rate inside pine trees. After inoculation, the fungal conidia accumulated within 1 cm of the inoculation sites. This result validates the hypothesis that the fungal conidia of E. vermicola cannot pass through the pits on the tracheid wall of a healthy pine tree [11]. Subsequently, the conidia of E. vermicola were germinated and grew within a few days after inoculation and extended upward and downward. A small quantity of fungus, however, extended to a sampled site more than 1 cm away from inoculation sites. Although the extension rate is slow, fungal hyphae of E.
vermicola were indeed growing through the wall of the tracheid. Regarding the low quantity of fungal hyphae detected in the farther samples from inoculation sites, the inhibitory effects of plant defense compounds and segregation of the tracheid wall [22,23] are supposed to be the main limiting factors. Moreover, the fungal hyphae were detectable and observable at 180 dai after fungal inoculation, suggesting that the fungal hyphae could stablely colonize pine tree xylem.
The fungal colonization of pine seedlings by E. vermicola was investigated, and the pine seedling stem can be colonized via artificial wounds [9]. After spraying inoculation of the wounds, as expected, the quantity of E. vermicola inhabiting the pine seedling xylem showed a moderate increase. However, the target DNA could not be detected at 45 and 60 dai. The colonization behavior detected for E. vermicola was probably responsible for the inhibition of the plant defense systems, as described in previous studies [24,25]. Then, the inoculated pine seedlings were infected with PWN suspension. The quantity of E. vermicola increased in response to PWN invasion in our study, indicating that the fungal hyphae in pine tree xylem parasitized the invading PWNs and led to a moderate increase in fungal quantity from the undetectable value.
Previous reports suggested that invasion of PWNs would enhance the defense system of the host plant due to the destruction of plant tissues [22,26]. However, the fungal quantities were moderately increased after PWN invasion in our study, despite increasing defense compound levels in the pine seedlings. The increase in fungal quantity could have occurred due to the elicitation of PWN infection cycles by E.
vermicola once the PWNs invaded and were exposed to lunate conidia, providing nutrients for fungal propagation. These results are supported by the observation of fungal hyphae and successful extraction of fungal-infected PWNs.
Regarding the inoculation of wilting pine trees, the fungal hyphae of E. vermicola extended rapidly and illustrated a jumping spread in the pine xylem. Similar to the fungal inoculation of healthy pine trees, the conidia inoculated with wilting pine xylem also accumulated around the inoculation sites, followed by rapid germination and colonization. However, the PWNs in the wilting pine trees supply plenty of nutrients for the growth of the nematophagous fungus E. vermicola. Thus, the fungal hyphae were extended rapidly. Moreover, PWNs parasitized by E. vermicola do not die within 3 days [27]. The PWNs attached by lunate conidia or infected by E. vermicola can move in the first 2 days after infection. In this case, the jumping spread of E. vermicola is suggested to be triggered by the immigration of infected PWNs, which was confirmed by the distribution pattern of fungi detected in wood samples, as well in extracted PWNs.
Furthermore, the significant decline in PWN density in wilting pine xylem after inoculation with E. vermicola indicates the great potential of this biological control agent in the management of PWD.
It is well known that resin is the main chemical defense substrate in conifers and that it includes the volatile compounds α-pinene, β-pinene, myrcene, limonene, and 3-carene [28,29], which injure or kill beetles, nematodes and other invaders. According to a previous study, these compounds are the main volatile monoterpenes obtained from

Real-time PCR quantification and standard curves
To quantify the fungal hyphae of E. vermicola in the tested samples, the TaqMan Supplementary Table S1.
In the present study, two calibration curves targeting different concentration ranges of target DNA were constructed to quantify both wt E. vermicola and CNU120806gfp.
Traditional TaqMan PCR was performed for the target DNA at concentrations ranging from 10 -2 to 100 ng of total template DNA with the specific primers 28S-2F/28S-2R.
In addition, nested TaqMan PCR-based quantification was performed for target DNA ranging from 10 -5 ng to 10 -1 ng with the primer pairs 28S-1F/28S-1R and 28S-2F/28S-2R to detect the target DNA at an extremely low concentration. The C t values obtained from the reactions were correlated with the amount of DNA and were used to construct calibration curves. Three replicates were conducted for each treatment.

Visualization of E. vermicola in infected PWNs
The

Temporal and spatial dynamics of fungal colonization in healthy pine trees
To investigate the temporal and spatial dynamics of E. vermicola colonization in

Temporal dynamics of fungal colonization inside healthy pine seedlings
To study the temporal dynamics of E. vermicola colonization in pine tree xylem and its response to PWN invasion, a series of experiments were performed. Twenty-seven 4- year-old P. koraiensis pine seedlings grown in a greenhouse were used to study the temporal dynamics of E. vermicola colonization in pine seedlings. For fungal inoculation, the bark of pine seedling stems was randomly cut to create artificial wounds, and then a blastospore suspension (3 × 10 8 ml -1 ) of E. vermicola mutant CNU120806gfp was sprayed onto the wounds of the seedlings. Subsequently, the wounded branch of seedlings was individually covered by cling film for 7 days to maintain moisture. After that, the pine seedlings were sampled at 1, 2, 4, 7, 14, 21, 28, 48, and 60 dai. The wood samples were used to determine the population dynamics of E. vermicola inside pine seedlings after inoculation, as well as to reveal the fungal location in the pine seedlings.

Response of the fungal quantity inside pine seedlings to PWN invasion
In this experiment, eighteen 4-year-old P. koraiensis seedlings were treated with a blastospore suspension (3 × 10 8 ml -1 ) of the E. vermicola mutant CNU120806gfp following the method described in the previous section and cultured in a greenhouse for two months. Subsequently, the pine seedlings were artificially infected by PWNs.

Total DNA extraction of wood and PWN samples
All of the sampled wood specimens were removed from the bark and surface-sterilized with 5% NaOCl for 1 min to avoid the influence of hyphae attached to the surface of the materials [38,39]. Then, the wood samples were ground into coarse powder and

Quantification of E. vermicola in wood and PWN samples
The total DNA samples were extracted from wood and PWNs using the method described by Wang et al. (2020a) [18]. Then, the DNA extracts were used for quantification of E. vermicola using the TaqMan probe PCR technique to discover the fungal colonization pattern in the host pine tree and infection of PWN. In each reaction, 100 ng of total DNA extract was included as a DNA template to quantify the amount of target DNA. The C t values obtained from TaqMan PCRs containing total DNA extracts of wood samples were interpolated using calibration curves to calculate the quantity of target DNA with respect to the total DNA.

Observation of fungal colonization in pinewood
To observe the fungal germination and colonization of E. vermicola in the host pine tree, wood inoculated with mutant CNU120806gfp was used for visualization by sectioning and fluorescence microscopy. The pine xylem of wood samples was freehand sectioned with a slicing knife to obtain thin wood slices. Subsequently, the fragments of the sectioned wood slices were randomly picked and mounted on glass slides with a cover glass. Then, the wood slices were observed and recorded with white and fluorescence microscopy.

Statistical analyses
Data obtained from experiments were compared using ANOVA tests for differences             The relative quantity of fungal genomic DNA of E. vermicola in PWNs extracted from pine seedlings. Black column, the relative fungal quantity of E. vermicola in PWN extracted from fungal inoculated pine seedlings; gray column, the relative fungal quantity of E. vermicola in PWN extracted from control pine seedlings. *, P value < 0.05; **, P value < 0.01; ***, P value < 0.001.