Introduction

The boreal forests, which are the second most extensive terrestrial biome on Earth after tropical forests, occupy an area of 18.8 million km2. Picea mariana forest ecosystems in Canada are a dominant part of this biome located north of 50°N latitude. Close to one million hectares of forest were harvested in Canada in 1990 [35], and the harvested area has increased by more than 3% per year ever since [7]. Because of the huge economical pressure on black spruce forests and the impact of climate change, fire, insects, and diseases, an extensive reforestation program was established in Canada. The efficiency of this program relies, in good part, on the safe production of high-quality pathogen-free plants. Although beneficial soil inhabitants and the undetected pathogenic fungi that can occur in the rhizosphere of healthy (asymptomatic) plants may be important factors influencing their survival, our knowledge of these microbial communities is very limited. Despite large replant efforts in the last decade, most Canadian provinces have a regeneration gap between the area harvested and the area regenerated, either naturally or through planting programs. Canada produces approximately one billion tree seedlings per year to reforest over 400,000 hectares [7]. The province of Quebec, in eastern Canada, contributes approximately 15% to this effort, mostly as black spruce [P. mariana (Mill.) B.S.P.]. The St-Modeste Forest Nursery, owned by the Government of Quebec, has developed a black spruce production approach based on controlled hybridization, propagation by cuttings, and growth in containers [51]. This technology provides genetically selected and more uniform plants, thus ensuring a more successful reforestation that depends strongly on a continuous supply of high-quality seedlings [10]. Unfortunately, as in other production approaches, the plants may suffer from diverse fungal pathogenic infections causing root-rot and damping-off outbreaks [25, 27, 32]. Indeed, the young conifers are threatened by these pathogens in the early stages of regeneration, significantly affecting the success of forest regeneration programs throughout North America [17, 28, 42].

In the 1990s, more than four million nursery seedlings showing root-rot or damping-off were destroyed in the province of Québec [25]. Despite large losses, little is known about the diversity of the fungal pathogens involved. Identification of the causal agents is often contradictory, even for the main pathogenic fungal species that belong to the Ascomycota genus Cylindrocladium. Previously, Cylindrocladium floridanum Sobers and Seymour has been considered to be a prevalent pathogenic species of black spruce in eastern Canada [15], although some genetic differences were found between Canadian and American isolates [16, 27]. However, based on cultural, phenotypic, and rDNA internal transcribed spacer (ITS) sequence analyses, Vujanovic and St-Arnaud [53] have recently shown that Cylindrocladium isolates from the St-Modeste nursery belong to Cylindrocladium canadense Kang, Crous, and Schoch [29] rather than C. floridanum. If C. canadense is prevalent in at least some eastern Canadian nurseries, the potential differences with C. floridanum ecology and a lack of knowledge about the ecological preferences of C. canadense may reduce the effectiveness of management strategies to control root diseases.

With the increasing concern about the environmental consequences of fungicide applications, microbial diversity associated with various coniferous species has recently received attention to explore the potential for alternative strategies of disease management [18, 31]. Among the various microbial ecological niches associated with plants, the rhizosphere is the most challenging. The rhizosphere is the soil zone directly influenced by the plant roots [58]; it is inhabited by complex microbial communities, which depend on the nutrients released by the roots [5]. Fungal mycelia constitute a significant part of the microbial biomass in close proximity to roots and some taxa may be important competitors of pathogens, acting as biocontrol agents [8, 14, 15, 30]. A shift in the fungal community structure related to plant development during the various nursery production stages may consequently be an important factor affecting disease outbreaks. Change in the plant health status and photosynthetic activity may also influence root exudation and modify microbial community structure or activity and, therefore, rhizosphere suppression of root pathogens [44]. Many fungal communities or functional groups are poorly known, particularly those classified as Ascomycota, which could be important competitors of fungal pathogens. The available data concerning spruce rhizosphere microbiota are very scarce, with the exception of mycorrhizal root symbionts [14, 31, 36].

In a previous work [19], profiling the rhizosphere microbial communities of fully grown (3-year-old) black spruce seedlings, we found 84 bacterial and 31 fungal sequences that belonged to a wide range of microorganisms. Based on cloned 18S rDNA gene sequence analysis, rhizosphere-associated fungal communities varied considerably between healthy and diseased plants. While accurate species identification was not possible in most cases, the results raised the possibility that some fungal taxa frequently or exclusively found in healthy rhizospheres may have competed with the pathogen and reduced its ability to colonize the rhizosphere and invade roots. Still, we lack knowledge of the fungal communities that characterize the P. mariana plants at the species level, and also of the population dynamics and dominant taxa throughout the production chronosequences. Moreover, propagation of genetically improved P. mariana plants by cloning of selected lines yield more uniform plants with superior phenotypes, but may also reduce the diversity of rhizosphere microbial communities and potentially lower the overall substrate antagonism toward root pathogens.

Recently, Mazzola [34] stated that effective implementation of strategies to manage root-rot and damping-off pathogens requires (1) initial identification of the biological components involved in the disease’s etiology and (2) monitoring of the environmental impact on the abundance of the target microbial population’s growth stages. Furthermore, additional data on the fungal diversity associated with economically important forest trees is urgently required to prevent and manage the involuntary introduction of genetically unique foreign pathogens associated with world trade [12]. The aim of this study was to compare the fungal community bioprofile and to identify dominant species and assemblages in the rhizosphere of healthy and diseased containerized P. mariana plants, in relation to the production chronosequences of a forest nursery. Fungal assessment and identification at the species level were based on a combination of phenotypic, cultural, and ITS rRNA gene sequence analyses. Special attention was given to the Ascomycota fungal communities, as they are known sources of both black spruce pathogens and potential biocontrol species.

Materials and Methods

Experimental Design and Sampling

Black spruce cuttings and rhizosphere soil samples were collected in September 2001 and 2002 from a governmental tree nursery located in the eastern region of Canada (St-Modeste, Québec: 47°49′55″N, 69°24′11″W) where a high incidence of root rot diseases were previously reported (Innes, pers. comm.). In this nursery, black spruce propagation is routinely conducted by cuttings from 2–3-year-old mother plants (MP) produced from black spruce seeds harvested from controlled pollination orchards [10]. Plants are rooted in compartmentalized production trays in a mixture of peat, vermiculite, and perlite (2:1:1) and grown for 1 year under greenhouse conditions, followed by 2 years of growth outdoors before being shipped for reforestation [51]. Sampling was conducted over the regular P. mariana production chronosequences and included the following four key production stages: (1) MP, (2) 8-week-old cuttings rooted in compartmentalized production trays in a greenhouse (B + 0), (3) second-year cuttings where the production trays were arranged on outdoor plots for an additional 6-month growth period after 6 months in a greenhouse (B + 1), and (4) third-year cuttings having grown for a second year in outdoor plots (B + 2). Each year, five healthy-looking plants and five plants showing symptoms associated with root rot were randomly chosen within each of the four growth stages. Plants were carefully examined under a dissecting microscope to confirm the absence or presence of root necrosis. All plants were maintained in compartmentalized containers. For each plant, a 100-g soil sample was collected from the rhizosphere by vigorously shaking the P. mariana plant and collecting the soil portion in close contact with the root system. Roots and soil samples were placed in individual bags and transported back to the laboratory, where they were rapidly stored at 4°C before the isolation of fungi and determination of root biomass.

Data Collection and Identification of Fungal Taxa

Roots from asymptomatic and symptomatic MP, B + 0, B + 1, and B + 2 plants from 2001 and 2002 samplings were separately washed. Fresh and dry weights were determined after drying each sample for 24 h at 70°C. A 3-g subsample was taken from each 100-g soil sample after thorough mixing in a sterile bag. Dilution series were prepared in sterile water and inoculated onto plates of 2% potato dextrose agar (PDA, Difco) supplemented with antibiotics (100 mg l−1 streptomycin sulfate and 12 mg l−1 neomycin sulfate, Sigma, St. Louis, MO, USA). The dilution series were repeated five times for each plant. Assay plates were incubated at 22°C in the dark and observations were made after 7 and 14 days. All fungi were isolated generally from the 10−3 dilution, separated in morphotypes and quantified as colony-forming units (CFUs). Pure cultures were established using standard procedures. Fungi were morphologically identified at least to the genus level using standard diagnostic characteristics in fungal taxonomy. Representative isolates of Ascomycota occurring at a frequency of >1% in all samples, as well as ambiguous rare isolates, were further analyzed by polymerase chain reaction (PCR) amplification and sequencing of the ITS rRNA gene region. Most isolates belonging to Basidiomycota and Zygomycota were identified by morphology only. Reference DNA samples were deposited in the Institut de recherche en biologie végétale Micro fungus collection (MTF) under numbers MTF A01-F05 (see Table 2).

DNA Extraction, PCR Amplification and ITS Sequencing

Fungal isolates were grown for 14 days on 2% PDA plates, and mycelial mats were collected using a sterile scalpel blade. Mycelia were freeze-dried and ground with a mortar and pestle in liquid nitrogen [57]. Genomic DNA was extracted using the DNAeasy plant mini kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions. Internal transcribed spacer region 1, the 5.8S rRNA gene, and ITS region 2 were amplified using the primer pair ITS1F and ITS4 [21, 56]. PCR amplicons were purified on agarose gels using the QIAquick gel extraction kit and sequenced at the Montreal Genomics Centre (Montreal, Canada). Sequences were identified by similarity searches in GenBank using the BLASTn search algorithm. These identifications were confirmed in comparison with phenotypical characterization. All fungal ITS sequences generated in this study were deposited in GenBank under accession numbers DQ132810–DQ132848.

Data Analyses

Sequences were aligned using Clustal X software (version 1,82) [50]. Distance trees were produced with the PAUP 4.0b10 software [48] using a neighbor-joining approach, and support for groups in the tree was assessed using a bootstrap analysis with 1000 repetitions. A fungal distance tree was prepared with sequences that matched at 91% and higher.

Fungal diversity indices were computed for healthy and diseased health classes using the Shannon–Wiener index (H′) calculated from \( H\prime = - \sum p_{{\text{i}}} \ln p_{{\text{i}}} \) [40]. A Taxon contribution index was defined as \( {\text{Tc}} = {{\left( { - p_{{\text{i}}} \ln p_{{\text{i}}} } \right)}} \mathord{\left/ {\vphantom {{{\left( { - p_{{\text{i}}} \ln p_{{\text{i}}} } \right)}} {H\prime }}} \right. \kern-\nulldelimiterspace} {H\prime } \) and is derived from the Shannon–Wiener index. Tc is a fraction and describes the contribution of each taxon to the normalized value of diversity. Furthermore, the persistence of a species through the production stages (MP, B + 0, B + 1, and B + 2) was analyzed using a persistence index defined as \( {\text{Pi}} = { - {\left[ {{\sum q_{{\text{i}}} } \mathord{\left/ {\vphantom {{\sum q_{{\text{i}}} } {Q\ln {\left( {{q_{{\text{i}}} } \mathord{\left/ {\vphantom {{q_{{\text{i}}} } Q}} \right. \kern-\nulldelimiterspace} Q} \right)}}}} \right. \kern-\nulldelimiterspace} {Q\ln {\left( {{q_{{\text{i}}} } \mathord{\left/ {\vphantom {{q_{{\text{i}}} } Q}} \right. \kern-\nulldelimiterspace} Q} \right)}}} \right]}} \mathord{\left/ {\vphantom {{ - {\left[ {{\sum q_{{\text{i}}} } \mathord{\left/ {\vphantom {{\sum q_{{\text{i}}} } {Q\ln {\left( {{q_{{\text{i}}} } \mathord{\left/ {\vphantom {{q_{{\text{i}}} } Q}} \right. \kern-\nulldelimiterspace} Q} \right)}}}} \right. \kern-\nulldelimiterspace} {Q\ln {\left( {{q_{{\text{i}}} } \mathord{\left/ {\vphantom {{q_{{\text{i}}} } Q}} \right. \kern-\nulldelimiterspace} Q} \right)}}} \right]}} {{\left[ {2\ln 2} \right]}}}} \right. \kern-\nulldelimiterspace} {{\left[ {2\ln 2} \right]}} \), where q i is the CFU value for one production stage, Q is the sum of all the CFU values of a particular taxon, and 2 ln 2 is a normalizing constant. A value of Pi equal to 1 means that the fungal CFU values were constant through all production stages, whereas a value of Pi equal to 0 means that the fungal taxon was isolated in only one stage.

The isolation frequency of a single fungal taxon was calculated (N i/N t × 100; where N i is the CFU value of the isolated fungus and N t the total CFU values for all combinations of health class and production stage). Data were then submitted to canonical and noncanonical correspondence analyses (CCA and CA) in CANOCO (Windows version 4.5) [49] using χ 2 (P < 0.05) distance to test the similarity of the isolated fungal assemblages between health class and production stages combinations. Outliers were eliminated by following the empirical method described by Borcard (unpublished; available at http://biol10.biol.umontreal.ca/BIO6077/outliers.html). Clustering analyses were carried out with the R package [9] using asymmetrical quantitative coefficient to relate production stage with shift in associated fungal assemblages in the rhizosphere and Fager’s coefficient for fungal taxa ordination based on detected isolate variations through all chronosequences. The similarity matrix was then exported and the final figure was redrawn in Statistica 6.0 (StatSoft, Tulsa, OK, USA) using the unweighted pair-group arithmetic average (UPGMA) clustering method [43].

Results

Fungal Taxa Prevalence and Phylogenetic Affiliation

Most isolated fungal taxa belonged to phylum Ascomycota, whereas relatively few Basidiomycota and Zygomycota were recovered from the rhizosphere of either healthy or diseased black spruces (Table 1). Dominant taxa obtained from the rhizosphere of healthy plants were all classified as Ascomycota species of the genera Cylindrocladium, Hypocrea, Penicillium, and Pestalotiopsis (Tc > 0.2), followed by Acremonium, Alternaria, Cladosporium, Paecilomyces, Penicillium, and Sporothrix taxa (Tc = 0.1–0.2). Subdominating taxa were Aureobasidium, Chaetomium, Fusarium, Microsphaeropsis, Oidiodendron and Verticillium species (Tc = 0.05–0.1). The main rhizosphere inhabitants of diseased plants were Cylindrocladium and Fusarium species (Tc > 0.2), followed by Alternaria, Aureobasidium, an unidentified anamorphic Basidiomycete, Coniothyrium, Fusarium, Hypocrea, Mycelium radicis atrovirens I, Pestalotiopsis, and Xenochalara (Tc = 0.1–0.2), and then by Geniculosporium, Microsphaeropsis, Paecilomyces, Papulaspora, Penicillium, Ramularia, Rhizoctonia, Septonema, and Trichoderma (Tc = 0.05–0.1).

Table 1 Diversity, abundance and persistence of fungal taxa associated with the rhizosphere soil of healthy and diseased P. mariana plants throughout forest nursery production chronosequences (MP, B + 0, B + 1 and B + 2)
Full size table

Among the Basidiomycota members, Agaricus was exclusively isolated from diseased plant rhizospheres (Tc = 0.021) and Thelephora from healthy plant rhizospheres (Tc = 0.045), whereas Laccaria was more prevalent in healthy (Tc = 0.090) than diseased (Tc = 0.021) plant rhizospheres. One unidentified anamorphic Basidiomycete dominated in diseased plant rhizospheres (Tc = 0.146), occurring rarely in association with healthy plants (Tc = 0.045). Other minor Ascomycete isolates and a single isolate of Actinomucor/Umbelopsis sp. (Zygomycota) from a diseased B + 2 plant (Tc = 0.021) are listed in Table 1.

Ascomycota taxa were the most diversified and abundant fungal isolates in the P. mariana rhizosphere, with 39 taxa isolated (Table 1). Among those, 37 taxa had ITS rRNA sequence similarities ≥91% with sequences from GenBank (Table 2). An unrooted neighbor-joining tree based on ITS rDNA sequences shows the phylogenetic affiliation of the dominant Ascomycota isolates into seven well-defined families and three incertae sedis (Fig. 1), whereas seven other families are represented with at least one taxon (Table 1). Fungal community structure was similar between 2001 and 2002, so results are presented as the mean of the two sampling years. Most rhizosphere isolates belonged to order Hypocreales, families Hypocreaceae or Nectriaceae, or incertae sedis. Many Hypocreales isolates belong to species having strains known to be pathogenic to plants such as Acremonium strictum, C. canadense (strains F01, F02, F04, F05, and B05), Fusarium species complex (Fusarium avenaceum, Fusarium oxysporum B09, Fusarium sambucinum C12, and Fusarium tricinctum D03), Neonectria radicicola F03 [anamorph: Cylindrocarpon destructans (Zinssm.) Scholten] and Nectria mauritiicola [anamorph: Rhizostilbella hibisci (Pat.) Seifert]. On the other hand, other isolates of Hypocreales are from fungal genera containing biocontrol strains, such as Gliocladium sp., Hypocrea sp. C06, anamorphic Trichoderma isolates (Trichoderma atroviridae C08, Trichoderma viride A10, and Trichoderma sp. D05), and one Verticillium sp. A01 isolate. The second largest clade was formed by members of the family Trichocomaceae and included Aspergillus niger, Geniculosporium sp., Paecilomyces (A02), and eight Penicillium taxa (Penicillium brevicompactum A09, Penicillium daleae C10, Penicillium commune A05 and E02, Penicillium griseoroseum D04, Penicillium spinulosum C02, Penicillium thomii A06 and B11, Penicillium sp.1 A08, and Penicillium sp.2 D06). Eleven less represented families were Amphisphaeriaceae (two taxa), Chaetomiaceae (one taxon), Dothioraceae (one taxon), Leptosphaeriaceae (two taxa), Microascaceae (one taxon), Montagnulaceae (two taxa), Mycosphaerellaceae (two taxa), Myxotrichaceae (one taxon), Ophiostomataceae (two taxa), Pleosporaceae (two taxa), Xylariaceae (one taxon). Six other taxa were of uncertain position.

Table 2 ITS gene sequence analysis of Ascomycota associated with the rhizosphere soil of P. mariana plants grown in a nursery
Full size table
Figure 1
figure 1

Unrooted neighbor-joining tree showing similarities between ITS gene sequences from 39 Ascomycota isolates from rhizosphere soil of P. mariana plants. Bootstrapping values greater than 60% calculated from 1000 replicates are given above the branches. Scale bar indicates the number of substitutions per site. Identification numbers of DNA samples are given after species names (see Table 2).

Full size image

Relation Between Fungal Diversity, Production Stages and Plant Health

Both production chronosequence and plant health influenced fungal taxa assemblages considerably, as shown by an unweighted pair-group analysis using an asymmetrical quantitative coefficient (Fig. 2). Isolates from MP, B + 0, and B + 1 plant rhizospheres clustered together within healthy (linkage distance = 0.546) or diseased (linkage distance = 0.661) health classes. Isolates from healthy and diseased B + 2 plants clustered together but with considerably less linkage distance (linkage distance = 0.695). Root biomass was significantly higher (P < 0.05) in healthy plants compared to diseased plants for all production stages (Fig. 3). The highest proportion of fungal diversity was observed in diseased plant rhizospheres. From the 45 identified taxa (Table 1), 35 (78%) were associated with diseased (H′ = 3.077) and 28 (62%) with healthy (H′ = 2.846) plants. Fungal abundance was also generally higher in the rhizosphere of diseased plants. The range in the average number of viable fungal particles isolated from all four production stages was 160 CFU g−1 dry soil in the rhizosphere of healthy plants (450–610 CFU g−1 dry soil) and was considerably lower than the 790 CFU g−1 dry soil range measured from the diseased plants (310–1100 CFU g−1 dry soil). The total CFU number for all stages was also higher in samples from diseased (2700 CFU g−1 dry soil) than from healthy (2070 CFU g−1 dry soil) plant rhizospheres.

Figure 2
figure 2

Unweighted arithmetic average clustering (UPGMA) between black spruce plant production stage and associated fungi isolated in the rhizosphere. The distance used is the asymmetrical quantitative coefficient. H healthy seedling, D diseased seedling.

Full size image
Figure 3
figure 3

Relationship between root dry weight of healthy (H) and diseased (D) P. mariana containerized plants. Bars are the standard error of the mean.

Full size image

Relation Between Fungal Communities

CCA revealed substantial alteration (P < 0.05) in fungal community assemblages with regard to plant health and yielded a principal axis direction that regrouped fungal taxa associated with diseased plant rhizosphere soil, whereas the opposite axis direction was associated with rhizospheres of healthy plants. The corresponding CA ordination is presented in Fig. 4. Four main fungal communities can be recognized in relation to: (1) healthy MP, B + 0, and B + 1 plant rhizospheres, (2) diseased MP, B + 0 and B + 1 plant rhizospheres, (3) healthy B + 2 plant rhizospheres, and (4) diseased B + 2 plant rhizospheres. The MP, B + 0, and B + 1 healthy plant rhizosphere fungal communities were dominated by isolates of Cladosporium cladosporoides, Chaetomium globosum, Hypocrea sp., Microphoma sp., Microsphaeropsis sp., N. mauritiicola, Oidiodendron tenuissimum, Paecilomyces sp., P. thomii, Scopulariopsis sp., and Verticillium bulbillosum. The MP, B + 0, and B + 1 diseased plant rhizospheres were associated with Alternaria alternata, Aureobasidium pullulans, Coniothyrium sp., Fusarium spp. (F. oxysporum, F. avenaceum, F. sambucinum, F. tricinctum), Papulospora, Paraphaeosphaeria sp., Phialocephala sp. and T. viride. The healthy B + 2 rhizosphere soil was associated with A. strictum, Actinomucor/Umbeliopsis sp., Gliocladium sp., Laccaria bicolor, Penicillium spp. (P. brevicompactum and P. commune), Sporothrix schenkii, Thelephora terrestris, and T. viride. Finally, the diseased B + 2 plant rhizospheres were mainly associated with Agaricus sp., A. niger, an unidentified anamorphic Basidiomycete, C. canadense, Exserohilum novae-zelandiae, Geniculosporium sp., M. radicis atrovirens (MRA-I and MRA-2), Ophiostoma nigrocarpum, P. daleae, Pestalotiopsis spp. (Pestalotiopsis uvicola and Pestalotiopsis aquatica), Ramularia sp., Rhizoctonia sp., Septonema sp., and Xenochalara juniperi.

Figure 4
figure 4

Correspondance analysis of the relationship between black spruce production chronosequences, health status, and fungal communities in rhizosphere soil. H healthy, D diseased plants. Circles nursery production stage, triangles fungal taxon. Fungal taxa numbers refer to identifications given in Table 1 and to ITS sequences closest matches in Table 2.

Full size image

The persistence of fungal taxa forming these communities varied widely. Indeed, some of the fungal taxa were isolated from a single production stage, whereas others were found in all production stages. In healthy plant rhizospheres, A. alternata, Cladosporium cladosporioides, C. canadense, Paecilomyces sp., P. thomii, and Hypocrea sp. taxa showed the highest persistence index (Pi > 0.7) (Table 1), whereas in diseased plant rhizospheres, the most persistent taxon was C. canadense (Pi > 0.9), followed by F. sambucinum, Fusarium trincinctum, and Hypocrea sp. (Pi > 0.7).

UPGMA clustering demonstrated relationships in the occurrence of the 45 fungal taxa isolated from the rhizosphere soil (Fig. 5). Cylindrocladium canadense clustered with Agaricus sp. and MRA-I (linkage distance = 0.546), whereas F. oxysporum clustered with A. strictum, E. novae-zelandiae, Penicillium spp. (P. brevicompactum, P. commune, and P. daleae) and S. schenkii (linkage distance = 0.578). Pestalotiopsis species formed a cluster with Geniculosporium sp., Gliocladium sp., C. globosum, F. tricinctum, and Parasphaeosphaeria sp. and with an unidentified anamorphic Basidiomycete (linkage distance = 0.816). Others, mainly numerous rare isolates, were distributed into three separate clusters having linkage distances >0.9.

Figure 5
figure 5

Unweighted arithmetic average clustering (UPGMA) between individual fungal taxa detected in rhizosphere soil of P. mariana plants. The distance used is Fager’s community coefficient. Numbers after fungal species correspond to isolate numbers in Table 1, and fungal taxa closest matches are given in Table 2.

Full size image

Discussion

This study provides the first comprehensive analysis of temporal changes in fungal communities in the rhizosphere of P. mariana plants over all nursery production stages. It was found that both plant health and production stages had a strong influence on fungal community structure. This may suggest that some fungal isolates found strongly associated with healthy plants have the potential to compete with root-rot pathogens in the spruce rhizosphere and are potential biocontrol candidates. However, it is also possible that these organisms were simply less competitive in diseased plant rhizospheres.

One important finding shown by our data is the dominance of the root pathogen C. canadense in the rhizosphere of diseased P. mariana plants. Cylindrocladium canadense is a mitosporic fungus belonging to the Nectriaceae family (Fig. 1) with a possible teleomorphic stage in Calonectria [13, 29]. To the best of our knowledge, the present work is the first ecological report of C. canadense and its occurrence throughout nursery production chronosequences. Differences in biology are found when compared to C. floridanum, Cylindrocladium scoparium, and Cylindrocladium crotalariae data from conifer nurseries in the USA [11, 22]. Although C. canadense occurred predominantly in the rhizosphere of diseased B + 2 plants (the last production stage before outplanting), the fungus was found in all production stages and in both healthy and diseased (H/D ratio = 1:1.85) plants (Table 1). Indeed, it is obvious from our results that this fungus could jeopardize a safe nursery production. This raises the question of the possible impact of symptom-free plant infections on reforestation success, considering that preinfected, asymptomatic black spruce seedlings can develop root-rot symptoms after transplantation, contributing to an increase in mortality of up to 25% [23].

In addition to C. canadense, several important sister phytopathogenic Hypocreales fungi were also frequently isolated in association with diseased plants. These are N. radicicola [anamorph: C. destructans (Zinssm.) Scholten], N. mauritiicola [anamorph: R. hibisci (Pat.) Seifert], and Giberella spp. (anamorph: Fusarium spp.) (Fig. 1, Table 1). Our results also showed the occurrence of a complex Fusarium clade represented by F. avenaceum, F. oxysporum, F. sambucinum and F. tricinctum. The Fusarium consortium as a whole was more abundant than C. canadense in the rhizosphere of diseased MP, as well as in the young B + 0 and B + 1 plants (Fig. 4). This suggests that the rhizosphere of younger plants may represent an excellent niche for colonization by fast-growing Fusarium species. Therefore, Fusarium taxa could very likely be a significant factor promoting the development of root-rot and damping-off symptoms preceding Cylindrocladium infection in these growth stages. Fusarium-associated root-rot has been previously documented in conifer seedlings grown in forest nurseries [2, 4, 6, 26] but has never been related to a specific P. mariana production stage in containers, nor to an associated fungal community. The third largest group of potentially deleterious species in our study belongs to the genus Pestalotiopsis (P. aquatica and P. uvicola) (Fig. 1). This group changes in abundance from common in young healthy MP, B + 0, and B + 1 plants to exclusively present in diseased B + 2 plants (Fig. 4). In this study, the ITS sequence similarity searches in Pestalotiopsis spp. (P. aquatica and P. uvicola, 95 and 97%, respectively) did not allow identification to the species level without any doubt. Previously, Pestalotiopsis funerea has been reported as a seed-born species frequently found on various pine hosts in eastern Canada [54]. Further taxonomical and ecological studies will therefore be needed to interpret the significance of this result.

In this study, the main fungal inhabitants of the rhizosphere of diseased plants (Table 1) were Cylindrocladium and Fusarium (Taxon contribution index, Tc > 0.2), followed by Alternaria, Aureobasidium, an unidentified Basidiomycota taxa, Coniothyrium, Hypocrea, MRA I, Pestalotiopsis, and Xenochalara (Tc = 0.1 to 0.2). Several of the fungal species isolated here are known to cause losses of conifer seedlings during germination, early emergence, and growth [35]. In our study, the diseased MP, B + 0, B + 1, and B + 2 plants had 27, 32, 40, and 80% lower root biomass, respectively, compared with the equivalent healthy plants (Fig. 3). The results indicated that a decreased root biomass in diseased plants was associated with an increased fungal taxa richness and abundance in several treatments, as shown in Table 1. This relation may be linked to increased root decomposition, leading to higher carbon availability in the soil to support fungal growth [1, 21]. Our results also support findings that the major pathogenic community (C. canadense, Fusarium spp., and Pestalotiopsis spp.), as well as the less abundant community (A. alternata, Paraphaeosphaeria sp., and Phoma herbarum), may have a seed-born origin. In fact, many of these taxa were isolated from the rhizosphere of black spruce MPs produced from seed under greenhouse conditions, and still persisted in later production growth stages. The postgermination spread of the pathogens throughout the successive production stages may occur from disease loci originating from a seed-born inoculum [46], as reported on Norway spruce in Europe [33].

Overall, ascomycetous fungi showed the highest biodiversity with 39 different isolated taxa (Table 1). In those fungi, CA ordination revealed that many taxa that were mainly associated with healthy P. mariana plants were also related to fungal species known to have biocontrol properties (Fig. 4). The most important species known for their biocontrol properties belong to genera such as Chaetomium, Gliocladium, Penicillium, Paecilomyces, Trichoderma (Hypocrea), and Sporothrix [3, 38, 39]. Our results showed a clear ecological relationship between pathogenic species and potential biocontrol taxa (Fig. 5): Cylindrocladium canadense and F. oxysporum were closely associated with Penicillium spp. (P. brevicompactum, P. commune, and P. daleae), whereas Pestalotiopsis and Paraphaeospheria (pathogens) formed a cluster with Gliocladium sp. and C. globosum. Moreover, we observed that other taxa, such as Hypocrea sp., anamorphic Trichoderma isolates (T. atroviridae and T. viride), Verticillium sp., Paecilomyces sp., and Penicillium (P. daleae, P. griseoroseum, and P. spinulosum), formed the largest clade among known biocontrol generalists [15, 30, 55]. However, it should be noted that, under favorable environmental conditions, some of these taxa could also act as weak, secondary seed-borne pathogens [47].

As previously observed [19], members of the Basidiomycota were abundant colonizers of the mature B + 2 plants grown outdoors. Indeed, L. bicolor and T. terrestris dominated in B + 2, although the latter was also found in B + 1. The two aforementioned species are known to form ectomycorrhizal associations and were mostly associated with healthy seedlings grown outdoors. This symbiotic association has been shown to protect host roots against root pathogens, including containerized Picea glauca and Picea engelmanii seedlings in western Canada [31]. Some isolates of M. radicis atrovirens are commonly found as inhabitants of the rhizosphere of P. mariana and were shown to be associated with the ectomycorrhizal mantle [45]. Here, M. radicis atrovirens isolates were exclusively found in diseased plants. Previously, they were shown to occur in up to 70% of nonmycorrhizal roots and were reported to have necrotrophic abilities in the presence of C. destructans in the rhizosphere of Picea abies [24].

This study showed the occurrence of marked contrast between specific fungal taxa assemblages associated with both production chronosequences and plant health that could affect (a) the effectiveness of the P. mariana clonal production system and, plausibly, (b) the survival of black spruce transplants in situ. Distance analyses based on ITS gene sequences showed that the ascomycetous taxa belong to seven large phylogenetic groups (Fig. 1). In addition, Basidiomycota were represented with four distinct groups and Zygomycota with one distinct group. Each of these groups showed differences in ecological preference to the health-production stage combination and might be seen as “ecological units” containing communities of species having similar interrelationships and mutually influencing each other. For instance, a recently described C. canadense and Fusarium spp. showed, among pathogenic associates, the highest persistence (Pi > 0.7) and a considerable environmental range of adaptability colonizing all production stages. Results also suggest that the best biocontrol candidates in this environment would be Penicillium taxa due to their uniform level of persistence, similar to the level measured for the main pathogens. In addition to being known antibiotic producers, Penicillium spp. occupied the same microniches as pathogens, showing competitiveness as a persistent population. Toxic metabolite production is involved in fungal competitive relationships in the soil biocenosis [52]. Penicillium damascenum was shown to protect Picea glehnii seeds and reduce infections caused by Pythium vexans [30]. Antagonism against Rhizoctonia solani by various Penicillium spp., including P. brevicompactum, was also reported [37]. Biocontrol of fusarium wilt of tomatoes by Penicillium oxalicum was recently described [41]. We suggest that a high persistence index of potential antagonistic taxa, such as the Penicillium group, may indicate a beneficial mycobiota effect on plant health. Plant and soil types and soil management regime were previously shown to affect the soil microbial diversity and soil’s disease suppressiveness [20]. Our results also support the hypothesis that taxon fluctuations may be driven by fungal interactions that may also affect the plant’s health and susceptibility to disease. Further research is, however, needed to fully evaluate the potential of such fungal associations to control root rot pathogens under conditions prevailing in modern conifer nursery production.

This study provides new data on fungal species associated with nursery trees in current reforestation efforts in boreal forests. It also supports the importance of monitoring, an important asset against exotic pathogens within an international trade of coniferous plants.