Fungi and Oomycetes in the Irrigation Water of Forest Nurseries

: The aim of the present study was to assess fungal and oomycete communities in the irrigation water of forest nurseries, focusing on plant pathogens in the hope of getting a better understanding of potential pathogenic microorganisms and spreading routes in forest nurseries. The study sites were at Anykšˇciai, Dubrava, Kretinga and Trakai state forest nurseries in Lithuania. For the collection of microbial samples, at each nursery ﬁve 100-L water samples were collected from the irrigation ponds and ﬁltered. Following DNA isolation from the irrigation water ﬁltrate samples, these were individually ampliﬁed using ITS rDNA as a marker and subjected to PacBio high-throughput sequencing. Clustering in the SCATA pipeline and the taxonomic classiﬁcation of 24,006 high-quality reads showed the presence of 1286 non-singleton taxa. Among those, 895 were representing fungi and oomycetes. The detected fungi were 57.3% Ascomycota, 38.1% Basidiomycota, 3.1% Chytridiomycota, 0.8% Mucoromycota and 0.7% Oomycota. The most common fungi were Malassezia restricta E. Gu é ho, J. Guillot & Midgley (20.1% of all high-quality fungal sequences), Pezizella discreta (P. Karst.) Dennis (10.8%) and Epicoccum nigrum Link (4.9%). The most common oomycetes were Phytopythium cf. citrinum (B. Paul) Abad, de Cock, Bala, Robideau, Lodhi & L é vesque (0.4%), Phytophthora gallica T. Jung & J. Nechwatal (0.05%) and Peronospora sp. 4248_322 (0.05%). The results demonstrated that the irrigation water used by forest nurseries was inhabited by a species-rich but largely site-speciﬁc communities of fungi. Plant pathogens were relatively rare, but, under suitable conditions, these can develop rapidly, spread e ﬃ ciently through the irrigation system and be a threat to the production of high-quality tree seedlings.


Introduction
The production of high-quality tree seedlings in forest nurseries is of key importance for forestry. Historically, artificial reforestation, i.e., the replanting or sowing of forest reproductive materials, has constantly increased [1,2], and today about 30% of the European Union (EU) forests are artificially reforested [3]. Besides, as ca. 30 million of forest tree seedlings are traded annually within the EU [2], the supply of healthy seedlings is a major challenge in order to prevent the spread and the introduction of fungal diseases to new areas [4].
Different abiotic and biotic factors may stress tree seedlings and predispose them to infections by pathogenic microorganisms [5]. Although a number of such pathogens are often air-or soil-borne, these can also spread through the irrigation water [6][7][8]. The irrigation water can either be from natural often associated with the surface water, careful assessment and management of the irrigation water used in forest nurseries is of key importance.
The aim of the present study was to assess fungal and oomycete communities in the irrigation water of forest nurseries, focusing on plant pathogens in the hope of getting a better understanding on potential pathogenic microorganisms and the spreading routes in forest nurseries in Lithuania, as, in the country, similar studies have not been done before, but is of considerable practical importance.

Study Sites and Sampling
The study sites were at Anykščiai, Dubrava, Kretinga and Trakai state forest nurseries in Lithuania. For each forest nursery, information on geographical position, the total land area, the number of seedlings produced annually and the area of the water pond, which is used for seedling irrigation, is in Table 1. All forest nurseries were situated within a radius of ca. 300 km. In all nurseries, seedlings are produced using a bare-root cultivation system. For the irrigation of seedlings, at Anykščiai, Dubrava and Trakai forest nurseries, the water is taken from ponds that are situated in the forest, while at Kretinga the water is taken from the dammed bog. Sampling was carried out during the dormancy period, i.e., between November 2017 and April 2018. At the time of sampling, the mean monthly temperature and precipitation were as presented in Table 1.  For the collection of microbial samples, at each nursery, five 100-L water samples, which were collected at a 50-m distance from each other along the coast of the pond, were separately passed through a 90-mm diameter cellulose filter paper (particles ≥10µm are retained) (Ahlstrom, Falun, Sweden) placed in a holder under the funnel. Between different samples, the filtering equipment was cleaned using 70% ethyl alcohol. Water samples were taken at a depth of ca. 0.5-1.0 m below the water surface without disturbing the bottom sediment. In total, 20 filter papers with water residuals were collected, labelled, placed individually into the plastic zip lock bags, transported to the laboratory and kept at −20 • C before further analyses.

Molecular Analysis
Prior to isolation of DNA, individual filters with water filtrate samples were freeze-dried at −60 • C for 24 h. From each lyophilised filter, 1/4 part was taken (remaining was kept as a backup), cut into smaller pieces and placed in three 2-mL screw-cap centrifugation tubes (60 in total; 20 samples × 3 DNA extraction replicates) together with sterile glass beads and homogenised using a Precellys 24 tissue homogenizer (Bertin Technologies, Montigny-le-Bretonneux, France). The total DNA was isolated from each sample by adding 1000 µL of CTAB buffer (0.5 M EDTA pH 8.0, 1 M Tris-HCL pH 8.0, 5 M NaCl, 3% CTAB) and incubating at 65 • C for 1 h. Following centrifugation at 8000 rpm for 5 min, the supernatant was transferred to a new 1.5-mL centrifugation tube, mixed by pipetting with an equal volume of chloroform, centrifuged for 8 min at 13,000 rpm and the upper phase was transferred to a new 1.5-mL centrifugation tube. Then, an equal volume of 2-propanol was added to precipitate the DNA that was pelleted by centrifugation at 13,000 rpm for 20 min. The DNA pellet was washed in 500 µL 70% ethanol by centrifugation at 13,000 rpm for 5 min, dried and dissolved in 50 µL of sterile milli-Q water. The DNA concentration of each sample was measured using a NanoDrop™ One spectrophotometer (Thermo Scientific, Rodchester, NY, USA) and if needed diluted to 1-10 ng/µL.
The amplification by PCR of the ITS rDNA region was done using forward ITS6 primer [43] and reverse ITS4 primer with barcodes [44]. The primer pair used was shown to amplify both fungi and oomycetes [43]. PCR reactions 50 µL in volume were performed using the following final concentrations: 200 µM of dNTPs; 750 µM of MgCl 2 ; 200 nM of each primer; 0.025 µM DreamTaq Green polymerase (5 U/µL) (Thermo Scientific, Waltham, MA, USA); and 0.02 ng/µL of template DNA. Sterile milli-Q water was added to make the final volume (50.0 µL) of the reaction. Amplifications were done using the Applied Biosystems 2720 thermal cycler (Applied Biosystems, Foster City, CA, USA). The PCR program started with initial denaturation at 95 • C for 2 min, followed by 35 cycles of 95 • C for 30 s, annealing at 55 • C for 30 s and 72 • C for 1 min, followed by a final extension step at 72 • C for 7 min. The PCR products were assessed using gel electrophoresis on 1% agarose gels stained with GelRed (Biotium, Fremont, CA, USA). Following successful amplification, all three DNA extraction replicates of the same sample were pooled together, resulting in a total of 20 PCR samples. PCR products were purified using 3 M sodium acetate (pH 5.2) (AppliChem Gmbh, Darmstadt, Germany) and 96% ethanol mixture (1:25). After the quantification of PCR products using a Qubit fluorometer 4.0 (Life Technologies, Stockholm, Sweden), these were pooled in an equimolar mix and sequenced using a PacBio platform and one SMRT cell (SciLifeLab, Uppsala, Sweden).

Bioinformatics
The sequences obtained were analysed using the Sequence Clustering and Analysis of Tagged Amplicons (SCATA) next-generation sequencing (NGS) pipeline available at http://scata.mykopat.slu.se. Quality filtering included the removal of sequences shorter than 200 bp, sequences of low quality, homopolymers and primer dimers. Sequences lacking a tag or primer were also excluded. Following quality filtering, the primer and sample tags were removed from the sequence, but this information was stored as meta-data. The high-quality sequences were clustered into different OTUs using single-linkage clustering based on 98% sequence similarity. For each cluster, the most common genotype (a real sequence read) was used to represent each OTU. A consensus sequence was produced for clusters that were composed of only two sequences. Fungal OTUs were assigned taxonomic names using an Ribosomal Database Project (RDP) pipeline classifier at https://pyro.cme.msu.edu/index.jsp (centre for Microbial Ecology, Michigan State University, Michigan, MI, USA). Sequences of 80% of higher similarity to the phylum level were considered to be of fungal or oomycete origin and were retained, while the remaining sequences were considered to be of non-fungal or non-oomycete origin and excluded from further analyses. The 30 most common fungal taxa and all oomycete taxa from water filtrate samples were identified using GenBank (NCBI) database and the BLASTn algorithm. The criteria used for taxonomic identification were: sequence coverage > 80%; similarity to species level 98%-100%; and similarity to genus level 94%-97%. Sequences deviating from these criteria were considered unidentified to species or genus level and were given unique names as shown in Tables 3 and 4 and Table S1. Representative sequences of fungal and oomycete non-singletons are available from GenBank under accession numbers MT236332-MT237171.

Statistical Analyses
Rarefaction analysis was carried out using Analytical Rarefaction v.1.3 (http://www.uga.edu/strata/ software/index.html). Differences in the richness of fungal taxa (for simplicity, here and onwards, oomycetes are referred to as fungi) in water filtrate samples among different forest nurseries were compared by nonparametric chi-square test [45]. As each of the datasets was subjected to multiple comparisons, confidence limits for the p-values of chi-square tests were reduced the corresponding number of times as required by the Bonferroni correction [46]. The Shannon diversity index, qualitative Sørensen similarity index and correspondence analysis (CA) in SAS v. 9.4 (SAS Institute, Cary, NC, USA) [45,47] were used to characterise the diversity and composition of fungal communities. The nonparametric Mann-Whitney test in SAS v. 9.4 was used to test if the Shannon diversity index differed among different forest nurseries.

Results
High-throughput sequencing of the 20 amplicon samples generated 35,179 reads. Following quality filtering, 24,006 high-quality reads (668 bp on average) were retained. Clustering analysis showed the presence of 1286 non-singleton taxa ( Table 2 and Figure 1), while singletons were removed.    Among the non-singletons, 895 (69.6%) were representing fungi and oomycetes, and the remaining 391 (30.4%) were non-fungal, which were excluded. The number of high-quality sequences and fungal taxa from each study site are in Table 2. A plot of fungal taxa from four forest nurseries vs. the number of fungal sequences resulted in species accumulation curves that did not reach the species saturation ( Figure 1), indicating that a potentially higher diversity of taxa could be detected by deeper sequencing. In this study, the detected fungi were 57.3% Ascomycota, 38.1% Basidiomycota, 3.1% Chytridiomycota, 0.8% Mucoromycota, and 0.7% Oomycota.
Twenty-three oomycete taxa were found in the irrigation water in all forest nurseries (Table 4). Among these, five taxa were from the genus Pythium, four from Phytophythium, three from Phytophthora, five from Saprolegnia Nees, and one was from each Achlya Nees and Peronospora Corda. The four taxa representing 0.03% of all high-quality sequences could not be identified to taxon or genus level. The most common oomycete taxon was P. cf. citrinum, which was detected in Anykščiai, Dubrava and Trakai, but not in the Kretinga forest nursery. Peronospora sp. 4248_322 occurred in the Kretinga forest nursery only (Table 4).
Correspondence analysis of fungal communities explained 13.6% variation on Axis 1 and 10.7% on Axis 2. The CA showed that different samples of the same forest nursery more or less well clustered together and were largely separated from other forest nurseries (Figure 3). An exception was the K4 sample from the Kretinga nursery, which clustered more closely with Anykščiai samples (Figure 3). The Sørensen similarity index of fungal communities ranged between 0.12 and 0.  Table 2). The Shannon diversity index was significantly higher in the Kretinga forest nursery than in the Trakai forest nursery (p < 0.05), while no significant differences were found between other forest nurseries (p > 0.05).  Table 2). The Shannon diversity index was significantly higher in the Kretinga forest nursery than in the Trakai forest nursery (p < 0.05), while no significant differences were found between other forest nurseries (p > 0.05).

Discussion
The available knowledge indicates that water from open sources, such as rivers, canals, ponds, and reservoirs, may constitute a significant risk for disease spread to forest nurseries [11]. Some soil-inhabiting pathogenic fungi can enter the open water source together with the rainwater and, if a fine filtering of this water is lacking [48], these can be disseminated through the irrigation system. Runoff from the nursery can also carry plant pathogens back to the pond, thereby increasing the risk of new infections following irrigation [20,31]. Therefore, the irrigation water used in different cultivation systems can be one of the most efficient vehicles for the spread of plant pathogens [49,50] Our results show that the richness of fungal taxa and the fungal community composition differed in water ponds of different forest nurseries. It appears that the characteristics of fungal communities in water filtrate samples can depend on specific parameters of each water body. In Kretinga, where water samples were from the damped bog, we found the highest diversity and the highest relative abundance of fungal taxa (Figure 2, Table 2). In Anykščiai, the pond was heavily overgrown by water plants, which could have also affected the diversity and composition of fungal communities in water [51]. In Dubrava, the pond is regularly refilled from a large lake, while, in Trakai, the pond is refilled using rainwater, which likely contributed to the observed lowest richness of fungal taxa as compared to other forest nurseries (Figure 2). The latter demonstrates that the heterogeneity of environmental conditions has likely contributed to the observed differences in richness and composition of fungal communities in different forest nurseries (Figures 1 and 2; Tables 2 and 3). This is in agreement with similar studies on aquatic fungi for which the diversity and composition may change depending on the source, location, and time of the year [52][53][54][55]. Indeed, as the sampling at different sites was carried out both in the autumn (Anykščiai, Kretinga and Dubrava) and in the spring (Trakai) ( Table 1), the possibility should not be excluded that this has also contributed to the observed variation in richness and the composition of fungal communities among different forest nurseries. Besides, the fungal community structure may also vary significantly depending on both physical and chemical properties of water and physical properties of the habitat, e.g., the size and depth of the water body [56,57].
The detected principal taxonomic groups of fungi coincided with those in similar studies on freshwater habitats [12,58,59]. Among the 30 most commonly encountered fungi, several seedling pathogens have been identified, including Fusarium avenaceum (Fr.) Sacc., which is known as a cosmopolitan fungus occurring in most soils of the temperate climate zone. It is often considered as a soil-borne pathogen that causes numerous diseases, including seedling blights. Apart from soil, it can be transmitted with seeds and/or plant debris. Fusarium avenaceum has also been shown to be a problem in forest nurseries [60][61][62]. Nowadays, in Lithuanian forest nurseries, the use of chemical fungicides for seedling protection is limited due to the forest certification by the Forest Stewardship Council (FSC) [63]. Therefore, only five fungicides are registered by the State Plant Service [64] for the use in forest nurseries against a very narrow range of seedling diseases, such as needle cast, leaf mildew or rust. Therefore, seedlings can potentially be infected by various diseases, which are not controlled by these fungicides.
The second most common pathogen was Botrytis cinerea Pers. (Table 2), which is one of the main foliar pathogens in forest nurseries. Infections often appear after abiotic damages, such as caused by frost, fertilizers or herbicides [65], or environmental stress (e.g., high temperature or drought). This pathogen can also infect plants simultaneously with other fungi, or to colonize necroses caused by other diseases, such as pine twisting rust [66,67]. Petäistö [68] has found B. cinerea to be an important fungal pathogen showing frequent infection spread from diseased seedlings to healthy ones. The conidia of B. cinerea can be spread by insects (e.g., Bradysia spp.: [69]) or wind [65]. Our results provide additional information on the possible spread of this economically important fungal pathogen with the irrigation water. Generally, an infection by B. cinerea can take place just after three hours at 15-20 • C and 98% relative humidity in the presence of water on plant surfaces, while for Norway spruce it can be already at 6 • C and 80-90% relative humidity [68].
Gremmeniella abietina (Lagerb.) M. Morelet was the other pathogenic fungus, which is known as a causal agent of Scleroderris canker in conifer tree seedlings [70]. Alternaria alternata (Fr.) Keissl. was also detected and it is a destructive pathogen, which can affect sprouting seeds, first-year coniferous and deciduous tree seedlings [71]. The presence of Verticillium sp. has also indicated the potential threat to seedling production as these diseases are among the most devastating and can affect numerous plant species worldwide, ranging from herbaceous annuals to woody perennials [72]. Although the above-mentioned plant pathogens can cause a significant damage in forest nurseries worldwide, in our samples their abundance was relatively low.
The detected oomycete community included primarily plant pathogens (Pythopythium, Pythium, Phytophthora and Peronospora) and pathogens of fish and other aquatic organisms (Saprolengia) [6,73,74]. Interestingly, a similar oomycete community was reported by Redekar et al. [32] from the recycled irrigation water in a forest nursery using a containerized cultivation system. Nevertheless, during the last century, plant pathogenic oomycetes, such as Phytophthora and Phythium were detected in the irrigation water in different countries [10,30,32]. The latter repeatedly demonstrates the risk of pathogens spread with the irrigation water and highlights the importance of appropriate water management in plant production systems.
The detected diversity and abundance of oomycete taxa was relatively low, i.e., 23 taxa (Table 4), as compared to Redekar et al. [32] who, among other taxa in the irrigation water, detected 48 species of Phytophthora and 36 of Pythium. The lower diversity and relative abundance of oomycetes in our study could be due to climatic conditions as samples were taken during the dormancy period and oomycetes are known to be temperature dependant [10]. As sampling was carried out at the 0.5-1.0 m depth, this may have also affected the detected diversity of oomycetes as these are more common to the water surface. Besides, as structures of oomycetes are rather small, some of these could have passed through the filter, thereby affecting species richness and abundance estimates. Alternatively, this can partly be due to PCR biases that selected for shorter fragments of fungal DNA. Nevertheless, the oomycete diversity in the irrigation water may change depending on environmental conditions and the time of the year, as these factors may affect the survival and activity of individual taxa, and thus, may affect the degree of plant infections [6,16,[52][53][54][55]75,76].
The oomycete genus Phytopythium was represented by four taxa, among which P. cf. citrinum was most common (Table 4). In Poland, P. cf. citrinum has been shown to be commonly isolated from the rhizosphere soil of oak stands with different health status [40]. As in our study, P. cf. citrinum predominated the oomycete community in Anykščiai, Dubrava and Trakai forest nurseries, it can be a threat to the seedlings of pedunculate oak, which comprises about 10% of the annual production in all of these forest nurseries [77].
The detected aggressive tree pathogens included Phytophthora cactorum (found in Kretinga and Trakai) and Phytophthora plurivora T. Jung & T.I. Burgess (found in Trakai). Phytophthora cactorum was previously reported from Finland, where it was isolated from the necrotic stem lesions of the container-grown seedlings of Betula pendula Roth. It is an omnivorous pathogen that can infect over 200 plant species in 160 genera, including Betula spp., Salix spp. and many other woody plants [30]. Phytophthora plurivora has been extensively isolated in Europe from natural forests and a variety of hosts in other parts of the world. It has been recovered from numerous hosts with symptoms of crown dieback, small-sized and often yellowish foliage, fine root dieback, root lesions, collar rots, cankers and shoot dieback [78]. Among the remaining oomycetes, Phytopythium litorale (Nechw.) Abad, de Cock, Bala, Robideau, Lodhi & Lévesque, Pythium mamillatum Meurs and Pythium dissotocum Drechsler have been reported as pathogens causing severe damping-off and root-rot diseases to different agricultural crops [54,79,80].

Conclusions
The irrigation water used by forest nurseries was inhabited by a species-rich but largely site-specific community of fungi. Although plant pathogens were relatively rare, under suitable conditions these can develop rapidly, spread efficiently through the irrigation system and be a threat to the production of high-quality tree seedlings. To prevent seedling infections and losses, the appropriate management of the irrigation water should be among the practices of integrated disease management in forest nurseries.

Conflicts of Interest:
The authors declare no conflict of interest.