The 21st century is witnessing an unprecedented loss of biodiversity that affects all species groups, from birds to mammals, amphibians, and trees (Ripple et al., 2017). Among these organisms, tree declines have significant impacts. They play a major role in providing habitat for countless species and many ecosystem services, including maintaining biosphere dynamics, storing atmospheric carbon, and protecting soils and freshwater resources (Ripple et al., 2017). It is estimated that more than 40% of the forest area has been lost during the industrial era and continues to decline, while at least 25 countries have completely lost their forest cover (Estoque et al., 2022). These losses in tree biodiversity are caused by several anthropogenic factors, including deforestation, climate change, agricultural intensification, and invasive alien species (Hadziabdic et al., 2021; Rivers et al., 2023). In the temperate zones of Europe and North America, the main threat to plant diversity comes from invasive pests and diseases. At the same time, climate change can create an environment favorable to the establishment and dispersal of invasive alien species, influencing their distribution area and hastening their spread (Rivers et al., 2023; Sturrock et al., 2011).
It is generally accepted that international trade and the movement of nursery stock for planting in urban environments and for field cultivation are predominant factors in introducing invasive alien species (Liebhold et al., 2012; Paap et al., 2017; Santini et al., 2013). In addition, agricultural production provides conditions conducive, such as having a pool of susceptible species and irrigation or tillage practices, to the establishment of invasive alien species populations, their dispersal, and their emergence in new areas (Stukenbrock & McDonald, 2008). In other words, urban environments, nurseries, and agrosystems could be reservoirs or bridgeheads for these species; they establish and multiply in these artificial environments, facilitating their movement to natural ones (Dale et al., 2022; Paap et al., 2017).
Historically, members of the oomycetes were often responsible for severe damage in agricultural and forest landscapes (Hansen et al., 2012; Schroeder et al., 2013). Given the extended life span of resting structures and the potential for long-distance transport in contaminated soils, asymptomatic plants, and through airborne sporangia, the risk of disease outbreaks with irreversible consequences is high with oomycetes, especially when they become established in ecosystems where native hosts are not adapted to the introduced species or lineages (Fones et al., 2020; Fry et al., 1992; Rossmann et al., 2021). The genus Phytophthora, with over 212 described species so far (Abad, Burgess, Bourret, et al., 2023; Abad, Burgess, Redford, et al., 2023), is one of the most important groups of plant pathogens with major economic and cultural damages. In recent years, several cases of forest decline and dieback have been associated with Phytophthora species, and at least 41 species are associated with these declines in forests and natural ecosystems (Brasier et al., 2022).
The most infamous example worldwide is unquestionably P. cinnamomi. Infecting close to 5000 plant species, P. cinnamomi causes dramatic losses for agriculture and forestry but, most importantly, represents a threat to entire natural ecosystems, especially in temperate climates (Burgess et al., 2017; Cahill et al., 2008; Hardham & Blackman, 2018). In Western Australia, for example, it was estimated that more than 40% of the described plant species were susceptible to P. cinnamomi (Shearer et al., 2004). Severe impacts were also observed on chestnut stands in the United States and Europe and native oak species in Europe (Brasier et al., 1993; Cardillo et al., 2021; Diogo et al., 2022; Prigigallo et al., 2015). Another important example in forestry is the invasive species P. ramorum, which is responsible for sudden oak death in North America (LeBoldus et al., 2022; Rizzo et al., 2002) and sudden larch death in the UK (Brasier & Webber, 2010). The introduction and spread of P. ramorum on the West Coast of the United States and Canada stress the critical importance of surveillance and the need for monitoring tools (LeBoldus et al., 2022). Other species of Phytophthoras are important tree pathogens, for example, P. alni causing alder decline (Aguayo et al., 2012; Bjelke et al., 2016; Brasier et al., 2004), or P. plurivora associated with root rot, dieback or canker in various species (Beaulieu et al., 2017; Bily et al., 2021; Jankowiak et al., 2023; Matsiakh et al., 2021; Schoebel et al., 2014). The rise, since the mid-2000s, of real-time qPCR has been a primary driver for the improvement of detection methods and is now commonplace in surveillance programs (Bilodeau et al., 2007, 2009, 2014; Feau et al., 2019; Ioos et al., 2009; Miles et al., 2017; Schenck et al., 2016). However, broader screening is also essential to characterize the communities associated with declines, wilting, or dieback syndromes to understand the diversity of pathogenic species that may be involved in disease development.
In recent years, metabarcoding (massively parallel amplicon sequencing) has proven to help identify target species and characterize oomycete communities from environmental DNA (eDNA) (Burgess et al., 2022). Metabarcoding involves sequencing amplicons by high-throughput sequencing following PCR amplification of a barcode region, often the internal transcribed spacer 1 (ITS-1) for fungi and oomycetes (Burgess et al., 2022). The technique allows the detection of various organisms directly from complex matrix samples, such as soil and plant tissues, without relying on culture-based methods or enrichment. Hence, it enhances the detection of obligate biotrophic pathogens that cannot or are difficult to isolate on culture media. In a surveillance context, implementing a metabarcoding approach can significantly improve the discovery of new or introduced plant pathogens, developing specific diagnostic tools, contributing to developing adapted management strategies, and thus limiting the spread of potential threats.
Canada grows approximately 23,800 ha of fir trees for the Christmas tree market. Of this number, 8255 ha are produced in Québec, making the province the country's largest producer (Anonymus, 2023). The production is concentrated in two central regions, Estrie and Chaudière-Appalaches (Beauce and Québec), which produce mainly Abies balsamea and A. fraseri. Several pathogens may threaten the production, but species of the genus Abies are often highly susceptible to Phytophthora root rot (PRR), which is considered to be one of the most critical problems in Christmas tree cultivation in the United States (Benson & Grand, 2000; Frampton & Benson, 2012; McKeever & Chastagner, 2016; Pettersson et al., 2019). In recent years, PRR incidence has increased up to 20% in Estrie (Dominique Choquette, MAPAQ, personal communication) and up to 25% in Chaudière-Appalaches (Christian Lacroix, MAPAQ, personal communication).
In the Eastern United States, at least 17 Phytophthora species have been recorded in association with infections of Abies fraseri (P. abietivora, P. cactorum, P. cambivora, P. capsici, P. cinnamomi, P. citricola, P. citrophthora, P. cryptogea, P. drechsleri, P. europaea, P. kelmanii, P. megasperma, P. nicotiniae, P. pini, P. plurivora, and P. sansomeana) (Hoover & Bates, 2013; Huang et al., 2004; Li et al., 2019; McKeever & Chastagner, 2016; Pettersson et al., 2017; Quesada-Ocampo et al., 2009). However, the situation in Eastern Canada, particularly in Québec, is not as well documented. Hence, a research program aimed at describing the diversity of oomycetes in cultivated fir trees was therefore implemented in 2019, during which sampling was carried out in plantations showing PRR-like symptoms in the central growing regions of the province. In the first phase of this research, 44 strains were isolated using a rhododendron leaf baiting assay from symptomatic and asymptomatic trees. They revealed the presence of at least six Phytophthora spp. (P. abietivora, P. chlamydospora, P. gonapodyides, P. gregata, P. megasperma, and P. kelmanii) in Southern Québec plantations, reported P. abietivora for the first time in Canada, and confirmed its pathogenicity on A. balsamea and A. fraseri (Charron et al., 2022).
In this study, using a metabarcoding approach, we build on the sampling conducted by Charron et al. (2022) to investigate oomycete diversity on a broader scale. The specific objectives were to (i) identify oomycetes associated with trees showing PRR-like symptoms in root samples and soil samples (before and after enrichment by leaf baiting for soils); (ii) identify patterns of co-occurrence between pairs of oomycete species; (iii) characterize oomycete communities from soil samples taken from beneath healthy- and diseased-looking fir trees in plantations and the adjacent natural forests; and (iv) test the hypothesis that land use is a major factor in shaping this diversity.
MATERIALS AND METHODS Sample collection and DNA isolationThe samples used in this study were initially collected to characterize Phytophthora spp. associated with PRR in Christmas tree plantations; the complete sampling methodology is presented in Charron et al. (2022). In brief, plantations reported with PRR-like symptoms were identified with the help of agronomists and crop specialists from the Québec Ministry of Agriculture, Fisheries and Food in Estrie and Chaudière-Appalaches (Québec, Canada). Sampling was conducted in the fall of 2019 and 2020 at 30 sites (Figure 1). When sampling, 12 trees were identified in each site (four diseased, four healthy from plantations, and four healthy from neighboring natural forests). For each tree, three 17 cm deep soil samples were taken at the crown edge, the first 5 cm of the organic layer was removed, and the three sub-samples were pooled together.
Root fragments were removed from the soil using a 4 mm mesh sieve, and the roots of samples from diseased-looking trees, including both symptomatic and asymptomatic root fragments, were stored at 4°C until DNA extraction. The soil was then divided into two parts, one for DNA extraction and the other for enrichment by baiting. DNA extraction was done on 250 mg of soil or diseased roots using the DNeasy PowerSoil Kit (Qiagen, Toronto, ON, Canada) according to the manufacturer's instructions. Enrichment was performed by baiting on rhododendron leaves according to the method described by Charron et al. (2022). After 4–7 days, the leaves were washed with tap water and dried for 10 minutes before taking 5 mm disks from the edges of areas of the rhododendron leaves displaying water-soaked lesions. The five-leaf disks were pooled and subjected to DNA extraction for each baited soil sample using the DNeasy Plant Mini Kit (Qiagen, Toronto, ON, Canada) and quantified using a Qubit 4 Fluorometer and the dsDNA HS Assay Kit (Thermo Fisher, Ottawa, ON, Canada). All DNA samples were standardized at a final concentration of 0.1 ng/μL for library preparation.
PCR amplification and metabarcodingSince the Ion Torrent technology was used, fusion primers were required to generate tagged amplicons. These fusion primers included the sequencing adapters (A and P1), a key, a unique barcode for each sample, and the primer itself. With this approach, there are two pairs of primers for each sample: one pair of forward sequencing primers and the second pair of reverse sequencing primers (Tremblay et al., 2022). Hence, for the PCR amplification of the ITS1 region from oomycetes, fusion primers OOM-LO5.8S47 (ATTACGTATCGCAGTTCGCAG) and OOM-UP18S67 (CTCGCCATTTAGAGGAAGGT) were used, as described in Tremblay et al. (2019). Each PCR (one reaction per direction) was performed in a volume of 25 μL with 1X PCR buffer, 1 mM MgCl2, 0.25 mM dNTP, 0.50 mM forward and reverse primers, 0.04 U Platinum Taq Polymerase (Life Technologies, Carlsbad, CA), and 0.1 ng purified DNA. PCR conditions were 3 min at 95°C for initial denaturation followed by 30 cycles at 95°C for 30 s, 52°C for 30 s, and 72°C for 60 s, and a final 10-min extension step at 72°C. Each PCR run included a negative control (1 μL H2O) and a positive control (1 ng P. ramorum), the latter consisting of a gDNA extract of P. ramorum. PCR products were run on a 1.5% agarose gel to confirm amplification and amplicon sizes. Only samples with an expected fragment length of 350–400 bp were taken to the next step. All fragments smaller than 150 bp were removed following purification with Agencourt AMPure XP magnetic beads (Fisher Scientific, Mississauga, ON, Canada) with a bead: amplicon ratio of 0.7:1.
The Ion Universal Quantification qPCR Kit (Life Technologies, Carlsbad, CA) was used to determine the concentration of each bi-directional barcoded library before pooling at an equimolar concentration of 100 pM. The Ion Chef and Ion 510 & 520 & 530 Kit—Chef reagents were used for template preparation and loading the libraries into nine Ion 530 chips in total (Life Technologies, Carlsbad, CA). The chips were processed using the Ion S5 sequencing system (Life Technologies, Carlsbad, CA), which provided a single FASTQ file per sample at the end of each run, while demultiplexing was performed using the S5 built-in software.
Sequence processingData quality was first examined using the FastQC software (v0.12.1) (
To assign taxonomy to the obtained representative sequences (QIIME2 rep-seq artifact), they were submitted to a nucleotide BLAST using the QIIME2 feature-classifier with the classify-consensus-blast command against a manually curated database. A “winner-takes-all” strategy was applied, in which the top hit was used to assign the taxonomic classification to the searched sequence (Hu et al., 2019, 2022). The database was built from the accession list used in Redekar et al. (2019), which themselves came from the Phytophthora-ID database (Grünwald et al., 2011) and (Robideau et al., 2011). Ex-types or selected specimens of 212 culturable Phytophthora species obtained from IDphy online resource (Abad, Burgess, Redford, et al., 2023), sequences of Pythium spp. from Van der Heyden et al. (2019), and the species of the P. europea complex from Charron et al. (2022) were added to the local database. In addition, because ITS1 does not resolve species identification for several Pythium and Phytophthora spp., representative sequences were imported in Geneious Prime (v2022.2.2) for local alignments and confirmation. Finally, species unresolved with the ITS1 region were grouped and labeled with the cluster or complex designation (Redekar et al., 2019, 2020). The description of each species cluster or complex used in this study is presented in Table 1.
TABLE 1 List of oomycete species grouped in cluster or complex based on Redekar et al. (2019) and Robideau et al. (2011).
Genus | Cluster or complex | Clade | Component species | Genus | Cluster or complex | Clade | Component species |
Phytophthora | P. cactorum cluster | 1a | P. cactorum | Phytophthora | P. megasperma cluster | 6b | P. megasperma |
P. alpina | P. crassamura | ||||||
P. idaei | P. chlamydospora cluster | 6b | P. chlamydospora | ||||
P. aleatoria | P. borealis | ||||||
P. pseudotsugae | P. gonapodyides | ||||||
P. citricola complex | 2c | P. citricola | P. mississippiae | ||||
P. acerina | P. europaea cluster | 7a | P. europaea | ||||
P. plurivora | P. abietivora | ||||||
P. nemorosa cluster | 3 | P. nemorosa | P. uliginosa | ||||
P. ilicis | P. flexuosa | ||||||
P. pseudosyringae | P. sp. Cadmea | ||||||
P. tyrrhenica | |||||||
P. pluvialis | P. pisi cluster | 7b | P. pisi | ||||
P. cooljarloo cluster | 6a | P. cooljarloo | P. asiatica | ||||
P. rosacearum | P. cryptogea complex | 8a | P. cryptogea | ||||
P. kwongonina | P. pseudocryptogea | ||||||
P. pseudorosacearum | P. erythroseptica | ||||||
P. inundata cluster | 6a | P. inundata | P. kelmanii | ||||
P. humicola | |||||||
P. condilina | |||||||
Elongisporangium | E. undulatum complex | H | E. undulatum | Pythium | P. myriotylum complex | B1 | P. myriotylum |
E. dimorphum | P. zingiberis | ||||||
Globisporangium | G. attrantheridium complex | F | G. attrantheridium | P. salpingophorum complex | B1 | P. salpingophorum | |
P. balticum | P. conidiophorum | ||||||
G. irregulare complex | F | G. irregulare | P. dissotocum complex | B2 | P. dissotocum | ||
G. cryptoirregulare | P. aff. dictyosporum | ||||||
G. cylindrosporum | P. coloratum | ||||||
G. heterothallicum complex | I | G. heterothallicum | P. diclinum | ||||
G. glomeratum | P. lutarium | ||||||
G. megalacanthum complex | J | G. megalacanthum | P. pachycaule complex | B2 | P. pachycaule | ||
G. aff. polymastum | P. oopapillum | ||||||
G. glomeratum | P. aquatile complex | B2 | P. aquatile | ||||
G. sylvaticum complex | F | G. sylvaticum | P. sukuiense | ||||
G. terrestres | P. minus complex | E2 | P. minus | ||||
P. aff. pleroticum | |||||||
P. arrhenomanes cluster | B1 | P. arrhenomanes | |||||
P. aristoporum | |||||||
P. phragmitis |
QIIME2 artifacts (table.qza, rooted-tree.qza, and taxonomy.qza), along with the metadata file, were imported into R (R Core Team, 2022) and converted to a phyloseq object using the qiime2R package version 0.99.6 (Bisanz, 2018). Next, read distribution was visualized using the microbiomeutilities R package (Shetty & Leo, 2020); samples with less than 400 reads were removed from the dataset before data analysis. The number of reads in each sample was standardized using the median sequencing depth. Phylogenies were assembled and investigated for each sample type (soil, baiting, and diseased roots) using the R packages phyloseq v1.40.0 (McMurdie & Holmes, 2013) and MicrobiotaProcess v1.10.3 (Xu et al., 2023). To highlight the effect of sample type on species composition, a Venn diagram was used to depict the distribution of Phytophthora spp. across soil, baiting, and root fragment samples. The “plot tree” function implemented in phyloseq was used to describe the species distribution of all ASVs using the tax_glom function for all sample types and three groups (i.e., diseased plantations, healthy plantations, and healthy forests).
To investigate the distribution of amplicon sequence variant (ASV) abundance among the three groups, a partitioning analysis was performed and visualized using the R package ggtern v3.3.5 (Hamilton & Ferry, 2018). In addition, a linear discriminant effect size analysis (LEfSe) was used to determine which species might explain the differences between the three groups. This analysis used the R package MicrobiotaProcess v1.10.3 (Xu et al., 2023). This approach uses a non-parametric Kruskal–Wallis test to detect species with significant differential abundance relative to the groups. Then, biological significance was tested using pairwise Wilcoxon's rank-sum tests, with the Benjamini–Hochberg correction to control the false discovery rate (FDR). In the final step, a linear discriminant analysis (LDA) was used to estimate the effect size of each differentially abundant feature. Finally, pairwise co-occurrence patterns were investigated using the R package cooccur v1.3 (Griffith et al., 2016).
Diversity analysesDifferences in the α-diversity metrics (richness [observed ASVs], Shannon's diversity, and Pielou's evenness) between diseased plantations, healthy plantations, and healthy forests, tree species, and site location as a proxy of microenvironment were tested for significance using a pairwise Wilcoxon's test with the Benjamini–Hochberg correction to adjust p-values for multiple comparisons (Benjamini & Hochberg, 1995). Beta diversity analysis was performed on the weighted UniFrac distance matrix and visualized using principal coordinate analysis (PCoA). To test the null hypothesis of no differences between disease status, permutational analysis of variance (PERMANOVA) with 9999 permutations was performed using the Adonis function of the R package vegan v2.6-2 (Oksanen et al., 2023).
RESULTS Data and sequencing summaryA total of 864 samples were processed by PCR, and of these, 319 soil samples, 125 baited samples, and 128 diseased root samples were positive for oomycetes, for a total of 572 positive samples. For soils, 107, 108, and 104 samples were positive for diseased plantation, healthy forest, and healthy plantation categories, respectively. For baiting, 58, 26, and 41 samples were positive for the diseased plantation, healthy forest, and healthy plantation categories, respectively. For root samples, all 128 samples tested were positive for oomycetes. The nine Ion Torrent reactions yielded a total of 181.03 million reads, with a median of 171,610 reads per sample (mean = 316,484 reads per sample) (Figure S1A–C). After quality control, ITS extraction, and the first pruning step, 97.786 million reads were kept (median = 91,418 reads per sample and mean = 170,954 reads per sample) (Figure S2A). After the last pruning steps, 97.785 million reads were kept (median = 92,247 reads per sample and mean = 172,158 reads per sample) (Figure S2B). Six samples with low read counts were removed, and the final dataset consisted of 566 samples.
The reads could be assembled into 7225 ASVs, of which 4663 (92.14% of all reads) were assigned to known oomycetes. The remainder were considered to be either from fungi, novel species or variants, or of known species missing in the database. These 4663 ASVs were assembled into 142 oomycete species, clusters, or complexes. The most abundant belonged to the Pythiales (70.70% of the reads), followed by the Saprolegniales (12.60% of the reads), the Peronosporales (5.09% of the reads), the Lagenidiales (0.16% of the reads), the Leptomitales (0.06% of the reads), and the Albuginales (0.003% of the reads), while only 11.40% of the reads were unassigned. In the soil, the Pythiales (78.80%), the Saprolegniales (10.60%), and the Peronosporales (1.44%) were the most abundant orders; the others being below 0.01% of the reads obtained for soil samples (Figure 2). The baiting approach significantly enriched the species belonging to the Peronosporales (13.90%) and Saprolegniales (28.70%) (Figure 2). The Pythiales (69.10%) and the Peronosporales (5.54%) in the diseased roots were predominant, while an important share of the reads (23.60%) were not assigned to known oomycetes (Figure 2).
When examining the distribution of Phytophthora spp. across the different sample types, it can be observed for the samples taken beneath healthy-looking trees (soil and baiting only) that the two sample types shared eight species. In contrast, 2 and 11 species were private to baiting and soil samples, respectively (Figure 3a). Similarly, for samples taken under trees showing above-ground PRR-like symptoms (soil, baiting, and roots), six species were shared between all sample types, while four, one, and eight species were private to baiting, root, and soil samples, respectively (Figure 3b). One species (P. nemorosa cluster) was shared between soils and baiting samples, and three were shared between root and soil samples (Figure 3b).
The community composition for each sample type (soil, baiting, and diseased roots) and disease status (diseased plantation, healthy plantation, and healthy forest) is presented in Figures 4 and 5. Overall, the most abundant species were those belonging to the Globisporangium attrantheridium complex, followed by Saprolegnia megasperma, G. sylvaticum complex, Pythium monospermum, Phytopythium. sp. CAL-2011b G. heterothallicum complex, P. dissotocum complex, Elongisporangium undulatum complex, G. macrosporum, Thraustotheca terrestris, and Phytophthora europaea complex, which represent 47.46%, 3.25%, 2.48%, 2.46%, 2.27%, 1.88%, 1.84%, 1.83%, 1.76%, 1.61%, and 1.60% of all reads, respectively. In soils, the G. attrantheridium complex was by far the most abundant, followed by S. megasperma, P. monospermum, species of the G. heterothallicum complex, and S. unispora, which represent 66.42%, 5.15%, 2.20%, and 2.09%, of all reads associated with soil samples, respectively. On the other hand, baiting significantly decreased the proportion of G. attrantheridium complex (3.95% of reads associated with baiting). For this sample type, the most abundant were Pp. sp. CAL-2011b, E. undulatum complex, T. terrestris, P. dissotocum complex, and P. europaea complex, which represent 8.10%, 7.64%, 7.17%, 6.20%, and 5.35% of reads associated with baiting, respectively. From the diseased root end, the G. attrantheridium complex was the most abundant, with 42.77% of the root-associated reads, followed by the G. sylvaticum complex (9.19%), the P. salpingophorum complex (3.05%), and the G. heterothallicum complex (2.65%).
To further investigate the distribution of taxonomic groups across soils from forest samples, healthy plantations, and diseased plantations, the relative abundance of each species was partitioned for these three soil types (healthy forest, healthy plantation, and diseased plantation) (Figure 6). Pythium spp. and Globisporangium spp. are mainly distributed along the diseased plantation–healthy plantation axis, while species of the genus Saprolegnia were almost exclusively present in healthy forest soils. The species belonging to the Phytophthora genus were mainly grouped in the soils of diseased plantations, apart from a few species present in the soils of healthy plantations and healthy forests.
Linear discriminant analysis effect size, performed to determine which species might be more abundant in the soils for each disease status, revealed that 12 species were significantly more abundant in soils collected from diseased trees and seven species were significantly more abundant in soils collected from healthy forest trees (Figure 7). In particular, it was shown that several species of the genus Saprolegnia (S. megasperma, S. turfosa, S. eccentrica, S. sp. CAL-2011b), P. cactorum complex, Achlya papillosa, and Brevilegnia macrospora were enriched in soils collected from under healthy trees in a forest environment (LDA score >4, p < 0.05). On the other hand, four Phytophthora spp. (P. cryptogea complex, P. europaea cluster, P. sansomeana, and P. chlamydospora cluster) were significantly more abundant in soils collected from plantations under diseased trees (LDA score >4, p < 0.05). We also found that G. minor, S. anisospora, P. chamaehyphon, L. giganteum, P. ornamentum, P. torulosum, G. megalacanthum complex, and P. inflatum were more abundant in soils raised under diseased trees compared with soils sampled under healthy trees in plantations and in forest areas. However, no species were found to be representative of soils sampled in plantations under healthy trees.
Patterns of co-occurrence of oomycetes in soil samples were investigated for 3903 species pairs, while 6393 pairs were removed from the analysis because the expected co-occurrence was smaller than one. For most pairwise comparisons (62%), expected and observed co-occurrence rates were not significantly different, indicating a random distribution among samples. However, 38.4% of the species pairs showed a positive association, while 38.11% of the analyzed pairs showed a negative co-occurrence pattern (Figure 8, Figure S4). Several species of the genus Pythium (P. vanterpoolii, P. inflatum, P. longandrum, P. pectinolyticum, P. sulcatum, P. volutum, P. salpingophorum complex, P. sp. CAL-2011e, and P. dissotocum complex), Globisporangium (G. megalacanthum complex, G. perplexum, and G. rostratifingens), and Phytophthora (P. transitoria, P. sansomeana, and P. europaea cluster) showed the highest positive associations. The correlation between positive association and frequency of detection was weak (τ = 0.49) but significant (p < 0.0001).
Alpha diversity, richness (observed ASV counts), diversity (Shannon), and evenness (Pielou) were compared with regard to growing region, fir species, and disease status (Table 2). The three factors had a significant effect on ASV richness. For the growing region, Québec (66.89 ± 1.10) was found to be significantly higher than Beauce (43.74 ± 0.50) and Estrie (37.8 ± 0.67) (padj <0.001 for both combinations). Fir species also significantly influenced ASV richness, with observed ASVs being higher in A. balsamea (56.95 ± 0.58) than A. fraseri (42.67 ± 1.05) and A. balsamea collected in forest environments (30.33 ± 0.53) (padj <0.005 for all combinations) (Table 2). Similarly, for disease status, ASV richness was significantly higher in the soil collected in plantations underneath diseased trees (59.36 ± 0.74) compared to soils collected underneath healthy trees in plantations (48.96 ± 0.69) or forests (30.33 ± 0.53). None of the factors under study influenced Shannon's diversity or Pielou's evenness, except for a significant difference in Pielou's evenness between A. balsamea and A. fraseri.
TABLE 2 Diversity estimates for the oomycete community inferred from soil samples.
α-diversity index | Observed ASVs (richness) | Shannon (diversity) | Pielou (evenness) |
Growing region | |||
Beauce (n = 176) | 43.74 ± 0.50 | 1.14 ± 0.08 | 0.32 ± 0.04 |
Estrie (n = 85) | 37.8 ± 0.67 | 1.25 ± 0.12 | 0.35 ± 0.06 |
Quebec (n = 55) | 66.89 ± 1.10 | 1.34 ± 0.15 | 0.35 ± 0.08 |
padj (Beauce–Estrie) | 0.055 | 0.200 | 0.120 |
padj (Beauce–Québec) | <0.001 | 0.200 | 0.560 |
padj (Estrie–Québec) | <0.001 | 0.720 | 0.560 |
Fir species | |||
A. balsamea (n = 170) | 56.95 ± 0.58 | 1.19 ± 1.43 | 0.31 ± 0.46 |
A. balsamea (forest) (n = 107) | 30.33 ± 0.53 | 1.18 ± 0.11 | 0.36 ± 0.06 |
A. fraseri (n = 39) | 42.67 ± 1.05 | 1.35 ± 0.19 | 0.49 ± 0.11 |
padj (balsamea–balsamea (forest)) | <0.001 | 0.110 | 0.062 |
padj (balsamea–fraseri) | 0.004 | 0.500 | 0.032 |
padj (balsamea (forest)–fraseri) | <0.001 | 0.110 | 0.568 |
Disease status | |||
Diseased_plantation (n = 107) | 59.36 ± 0.74 | 1.24 ± 0.11 | 0.31 ± 0.05 |
Healthy_forest (n = 107) | 30.33 ± 0.53 | 1.18 ± 0.11 | 0.36 ± 0.06 |
Healthy_plantation (n = 102) | 48.96 ± 0.69 | 1.20 ± 0.11 | 0.32 ± 0.06 |
padj (DP–HF) | <0.001 | 0.760 | 0.370 |
padj (DP–HP) | 0.017 | 0.780 | 0.730 |
padj (HP–HF) | <0.001 | 0.780 | 0.380 |
Note: Values in bold are significant at the 0.05 level.
Beta diversity of oomycete communities in soilThe effect of the disease status, fir species, and the growing region was subjected to a PERMANOVA to understand which factors affected the structure of oomycete communities (beta diversity). Among these factors, disease status explained the largest share of the variability (R2 = 0.168, p < 0.001), followed by the interaction between disease status and growing region (R2 = 0.033, p < 0.001), the region alone (R2 = 0.031, p < 0.001), the year (R2 = 0.019, p < 0.001), and the status by year interaction (R2 = 0.011, p = 0.020) (Table 3). The interactions between disease status and fir species and between disease status, growing region, and year were not significant. The principal coordinate analysis supported the PERMANOVA results and suggested clustering based on disease status (Figure 9). According to the PCoA, soil oomycete communities were separated into two main clusters representing healthy forest soils and plantation soils, while there was little clustering between diseased and healthy plantations (Figure 8). The PCoA did not show any distinct clustering in relation to the three regions covered by the study (Figure S4).
TABLE 3 Results of the permutational multivariate analysis of variance (PERMANOVA) of weighted UniFrac distances between samples.
Factor | DF | Sum of squares | R 2 | F | p |
Status | 2 | 6.857 | 0.168 | 34.728 | <0.001 |
Region | 2 | 1.248 | 0.031 | 6.319 | <0.001 |
Year | 1 | 0.770 | 0.019 | 7.804 | <0.001 |
Species | 1 | 0.235 | 0.006 | 2.381 | 0.041 |
Status:Region | 4 | 1.333 | 0.033 | 3.377 | <0.001 |
Status:Year | 2 | 0.437 | 0.011 | 2.214 | 0.020 |
Status:Species | 1 | 0.060 | 0.001 | 0.606 | 0.721 |
Region:Year | 1 | 0.232 | 0.006 | 2.352 | 0.040 |
Status:Region:Year | 2 | 0.199 | 0.005 | 1.006 | 0.375 |
Residual | 299 | 29.517 | 0.722 |
Forests provide numerous ecosystem services (ecological, environmental, social, and economic) (Hadziabdic et al., 2021), harbor a substantial portion of terrestrial biodiversity, are at the root of terrestrial geochemical processes (Brockerhoff et al., 2017), and play a significant role in climate change mitigation through carbon sequestration (Di Sacco et al., 2021; Palmer, 2021). Still, forests are threatened by many natural stressors exacerbated by climate change (higher temperatures and lower water availability) and anthropogenic activities (land use and trades) (Pautasso et al., 2015; Sturrock et al., 2011), which tend to increase their susceptibility to pests and pathogens (Dale et al., 2022; Hadziabdic et al., 2021). In addition, the conversion of forests into large-scale plantations, often in the vicinity of natural ecosystems, can become a reservoir of pathogenic organisms and exacerbate the risk of disease outbreaks in natural forests and plantations (Dale et al., 2022). Hence, to anticipate the risk for natural environments, it is crucial to understand the risk and estimate the pathogen's diversity in plantations and planted environments. From that perspective, this study provides an overview of the diversity, richness, and composition of the oomycete community in fir plantations compared to surrounding natural forests. Hence, the approach presented in this study becomes helpful as part of an early warning strategy to guide the development of targeted molecular tools.
Although the use of the ITS1 region may be limiting for taxa identification, primarily because certain species cannot be identified at the species level and have to be grouped in species complex or cluster (Redekar et al., 2019; Robideau et al., 2011), the approach used in this study can be considered successful in that a total of 142 oomycete species, cluster, or complex belonging to 21 genera were detected among the different sampling sites and sample types. Among the genus of particular interest, 26 clusters or complexes of Phytophthora species were found, confirming the genus's remarkable diversity and ability to adapt to various natural and anthropogenically disturbed environments. However, with 40 species or complexes identified, the genus Pythium was the most diverse, as it is often the case in several natural, urban, and agricultural environments (Cerri et al., 2017; Dale et al., 2022; Fiore-Donno & Bonkowski, 2021; Navarro et al., 2021; Sapp et al., 2019). In addition, 26 species of Globisporangium, a genus recently outgrouped from Pythium (Nguyen et al., 2022; Uzuhashi et al., 2010) that includes species potentially pathogenic to coniferous species (Weiland et al., 2013), were also detected.
Seven of the ten most abundant species in this study belonged to the Pythiaceae. In this family, most species are saprophytic organisms, often facultative plant pathogens. However, several of them are important pathogens capable of causing significant damage to stressed and vulnerable plants and young seedlings (Lévesque, 2011). Despite Pythiaceae's pervasiveness in different environments, their pathogenicity on various plant species, particularly coniferous, is poorly documented. This is the case for species of the G. attrantheridium complex, which were by far the most abundant in this study, in fir plantations or natural forests. G. attrantheridium is a ubiquitous species that predominates in various agricultural (Allain-Boulé et al., 2004; Rojas et al., 2019), urban, and natural environments (Dale et al., 2022; Reinhart et al., 2010). It has been suggested that G. attrantheridium may play a role in regulating the diversity of plant communities by affecting the growth and establishment of young seedlings, especially at the periphery of parent trees, an effect potentially exacerbated in human-managed environments (Dale et al., 2022; Packer & Clay, 2004). Among the Pythiaceae, we also found species previously identified as potentially pathogenic to coniferous species (i.e., P. dissotocum complex, P. irregulare complex, G. macrosporum, G. rostratifingens (Weiland et al., 2013), and E. undulatum (Weber et al., 2004)), which may play a similar regulatory role.
Several Phytophthora spp. were also identified through this metabarcoding survey. Many had already been identified in Canada, such as P. drechsleri and P. syringae, while others, such as those of the P. megasperma cluster (which may include P. megasperma, and P. crassamura) or those of the P. europaea cluster (which may include P. abietivora, P. uliginosa, P. flexuosa, P. tyrrhenica, and P. sp. cadmea), were not yet reported or are still regulated species (Anonymous, 2016; Charron et al., 2022; Dale et al., 2022; Feau et al., 2019). However, although the resolution of metabarcoding data is sufficient for many species, care must be taken when inferring the distribution of species belonging to specific clusters or complexes for which the resolution does not allow discrimination between species in the said group or cluster. For example, the P. cactorum cluster contains species already reported in the country and not regulated (P. cactorum) and other species not reported or regulated, such as P. alpina, P. idaei, and P. pseudotsugae. This is also the case for the P. cryptogea or P. citricola complexes. Hence, as reported by others, the metabarcoding approach is very interesting from a pre-surveillance point of view (Burgess et al., 2022; Tremblay & Bilodeau, 2022). Still, in some instances, it remains necessary to confirm the actual presence of certain species using specific molecular methods and isolation.
Soil as a proxy for oomycete diversity assessmentAs expected, the relative abundance of the species detected through leaf baiting was considerably higher when compared with direct extractions from soil or roots. Conversely, more species were detected from soil samples, and only a subset of the total oomycetes was found with rhododendron leaf baiting. In this study, of the Globisporangium, Pythium, and Phytophthora species detected, 30.8%, 62.5%, and 53.8% were caught on leaf bait, compared to 100%, 97.5%, and 84.5% for the soil samples, respectively. These results are analogous to those obtained for irrigation water (Redekar et al., 2019) or rhizospheric soils from trade plants (Rossmann et al., 2021). However, in this study root samples came from diseased-looking trees only, and oomycete diversity might have been biased toward fir pathogens while missing other species. Oomycete diversity could have been greater with root samples if the sampling strategy included field roots from diseased- and healthy-looking trees. For example, Khaliq et al. (2018) suggested that eDNA obtained from field roots was better than eDNA obtained from soil to describe Phytophthora diversity. Nevertheless, in samples taken under PRR-like symptomatic trees, six Phytophthora species, complexes, or clusters (P. sansomeana, P. personensis cluster, P. cryptogea cluster, P. europea cluster, P. megasperma cluster, and P. chlamydospora cluster) were common to soil, root, and baiting samples. They could be considered core species for this system.
Analysis of indicator species showed that 12 species were more abundant in soils sampled from beneath trees showing PRR-like symptoms compared with those from healthy-looking trees in plantations or natural forests. Among these, four Phytophthora species or clusters (Phytophthora sansomeana, P. cryptogea cluster, P. megasperma, and P. europaea cluster) were already reported to be associated with PRR in Abies spp. (Charron et al., 2022; McKeever & Chastagner, 2016; Molnar et al., 2020; Pettersson et al., 2017, 2019; Talgø et al., 2007). Here, P. sansomeana, P. cryptogea cluster, and P. megasperma cluster were present in 11.2%, 12.1%, and 15% of the soil samples taken under PRR-like symptomatic trees but almost absent (below 5%) in those from healthy-looking trees, while the species most frequently associated with symptomatic trees were those of the P. europaea cluster (29.9% of samples). These results are consistent with those presented by Charron et al. (2022), who found, by isolation directly from root tissues and after a baiting step, that P. megasperma and P. gonapodyides (included in the P. megasperma cluster in this study) were associated with PRR-like diseased trees. However, it was also shown that P. abietivora (included in the P. europaea cluster) was prevalently associated with trees exhibiting above-ground PRR-like symptoms. We also found that P. europaea cluster was present in almost 15% of soil samples taken from healthy-looking trees in plantations and natural forests, suggesting that the problem is either underestimated or, more importantly, that there may already be a migration from diseased plantations to natural forests.
Land use is a major driver of oomycete diversityOur results showed no significant differences between measures of diversity and evenness between treatments. However, oomycete species richness was highest in soils collected from plantations under PRR-like diseased trees and tended to be significantly lower in natural forest soils than in plantations, whether collected from diseased or healthy trees (Table 2). Several factors can influence species richness, but it is clear that land use is a determining factor, especially for oomycetes, which include primarily pathogenic and opportunistic species (Makiola et al., 2019). Trees in anthropogenically perturbated landscapes are more exposed to environmental stressors (e.g., chemical inputs, pollution, compaction, wounds, or low plant diversity), which increases their vulnerability to pests and diseases (Paap et al., 2017). Hence, disturbed environments and environments in the early stages of succession, as in fir plantations, have been shown to favor greater species richness of plant pathogens (Dickie et al., 2019; Makiola et al., 2019; Mangelsdorff et al., 2012). Interestingly, the data provided in this study showed significant differences in oomycete species richness when tested against the fir species, suggesting a difference in sensitivity to some oomycetes, as shown in Charron et al. (2022), which revealed differences in sensitivity of A. fraseri and A. balsamea to P. abietivora.
Several environmental factors can influence the composition of the oomycete community. For example, precipitation and temperature were shown to be the most important factors in partitioning oomycete diversity in soybean seedlings in North America (Rojas et al., 2017), while environment (described as a combination of cultivation history, precipitation, and temperature) had the greatest influence on oomycete community composition in soybeans in Ohio (Navarro et al., 2021). However, climate could be less important in anthropogenically disturbed environments compared to natural ones (Burgess et al., 2019; Delavaux et al., 2021; Sapp et al., 2019). Besides, communities of Phytophthora species, whose dispersal is now known to be human-accelerated, are more diverse in urban environments such as urban forests and botanical gardens and have often been identified as sentinels for surveillance and early detection (Dale et al., 2022; Hulbert et al., 2019; Paap et al., 2017). Similarly, the likelihood of finding Phytophthora spp. in recreational forests is higher than in natural forests (Riddell et al., 2019). In this study, disease status and grower's location were the factors that influenced the community composition the most, according to the PCoA and the PERMANOVA results (Table 3). The effect of disease status is likely due to the origin of the soil samples, whether they come from forest soils or plantations (Figure 3). In contrast, the location of the plantations is associated with microenvironmental factors (soil type, cultivation practices, grass cover, etc.). Together, these findings are consistent with the hypothesis that land use is a major factor in shaping oomycete communities.
CONCLUSIONThis study confirmed once again the value of metabarcoding in providing insightful information on oomycete species diversity present in the soil as a surveillance tool to monitor the introduction of alien invasive species. We also highlighted that the baiting method could underestimate species diversity. In addition to being part of an early warning system, metabarcoding can also be used to prioritize the development of specific molecular methods and to plan surveys. Our results reinforce the hypothesis that the P. europaea cluster represents a severe threat to fir production and report that land use (anthropogenic activities) shapes oomycete diversity and plantations can act as a gateway for invading natural forests. In fact, the results presented here suggest that the P. europaea cluster might already have crossed this boundary and that other species will likely follow in the near future, advocating the importance of improved surveillance of oomycete diversity in various environments. A better understanding of the composition of oomycete communities from healthy and unhealthy trees in natural and disturbed environments can help us to understand how anthropogenic activities influence oomycete communities, learn more about their role in ecosystems, and infer about the migration and introduction patterns of non-native species. However, the presence and abundance of these species still need to be confirmed either with direct isolation, species-specific molecular tools, or new metabarcodes, such as the rps10 gene (Foster et al., 2022), enabling the resolution of these clusters at the species level.
AUTHOR CONTRIBUTIONSHVH, PT, and GJB conceptualized the project and acquired the necessary funding. GC and PT carried out sampling, baiting experiments, and DNA extractions. GJB supervised library preparation and IonTorrent sequencing. With the help of M-OD, HVH carried out bioinformatics work and analyzed the data. HVH and GJB interpreted the data and planned the structure of the manuscript. HVH wrote the original version of the manuscript. All authors reviewed and approved the final manuscript.
ACKNOWLEDGMENTSThe authors would like to thank Jean-François Légaré, Éric Dussault, Christian Lacroix, Dominique Choquette, and Jacinthe Drouin for their contributions to fieldwork and sampling, Debbie Shearlaw, Miranda Newton, Andréanne Sauvageau, and Anne Piuze-Paquet for their technical assistance, and all the growers for giving us access to their plantations.
FUNDING INFORMATIONThis study was supported in part by a grant in the Cellule d'innovation des methodologies de diagnostic des ennemis de cultures from the Prime-Vert funding program of the Québec Ministry of Agriculture, Fisheries and Food (19-011-2.2-C-PHYTO).
CONFLICT OF INTEREST STATEMENTThe authors declare that they have no conflict of interest.
DATA AVAILABILITY STATEMENTThe pipeline used for quality filtering, ITS1 extraction, and generation of the representative sequence data is available via the GitHub page:
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2024. This work is published under http://creativecommons.org/licenses/by-nc/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
Forests are threatened by many natural stressors intensified by climate change and anthropogenic activities, which tend to increase their susceptibility to pests and pathogens. Consequently, oomycete-related forest decline or dieback cases are increasing in natural, urban, and agricultural landscapes. It is in this context that Christmas tree growers from Southern Québec, Canada, are experiencing root rot problems, with reported incidences up to 25%. In a previous study, seven
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Details


1 Agriculture and Agrifood Canada, Saint-Jean-sur-Richelieu Research and Development Centre, Saint-Jean-sur-Richelieu, Québec, Canada
2 Canadian Food Inspection Agency, Nepean, Ontario, Canada
3 Laurentian Forestry Centre, Canada Natural Resources, Québec, Québec, Canada