Tree dominance shapes soil and tree phyllosphere microbial communities in coniferous and broadleaf deciduous boreal forests

doi:10.21203/rs.3.rs-2238260/v1

Download PDF

Research Article

Tree dominance shapes soil and tree phyllosphere microbial communities in coniferous and broadleaf deciduous boreal forests

https://doi.org/10.21203/rs.3.rs-2238260/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 02 Mar, 2023

Read the published version in Plant and Soil →

You are reading this latest preprint version

Purpose: Natural and anthropogenic causes have produced changes in tree dominance from coniferous to broadleaf deciduous forests, generating shifts in litter inputs and plant understory composition. The impact of changes in canopy-associated factors on belowground microbial communities remain poorly understood. The objective of this study was to better understand how abiotic and biotic factors in black spruce and trembling aspen forests shape soil microbial community structure.

Methods: With high throughput sequencing, we first analyzed differences in microbial communities between microhabitats (tree phyllosphere vs. soil microbiome) and forest types (black spruce vs. trembling aspen). Second, we analyzed how shifts in factors related to each forest type (litter deposition and understory vegetation) affected soil microbial community composition.

Results: We found a high microhabitat specificity of bacterial communities interacting with forest type. Shifts in litter deposition and understory vegetation between forest types did not influence microbial community composition, but the legacy effects of each forest type defined soil bacterial and fungal communities. Fungal community composition was more strongly influenced by forest type compared with bacterial communities, and both were correlated with several soil physicochemical properties that differed among forest types.

Conclusion: This study expands our knowledge of the microbial composition of tree phyllosphere and soil microbial communities in black spruce and trembling aspen forests and their correlation with abiotic and biotic factors in each forest type. Our study demonstrates the resistance of microorganisms to variation in canopy-related factors and the importance of legacy effects of forest type in defining soil microbial community composition.

Aboveground-belowground interactions

tree dominance

tree phyllosphere

soil microbiome

Picea mariana (black spruce)

Populus tremuloides (Trembling aspen)

Aboveground-belowground interactions and their feedback loops shape community composition and ecological functions in forest ecosystems (Augusto et al. 2015; Kardol et al. 2018; Nilsson and Wardle 2005). Several studies have demonstrated that tree composition defines local environmental conditions (Augusto et al. 2015; Cavard et al. 2011), understory plant communities (Barbier et al. 2008; Bartels and Chen 2013; Qian et al. 2003), microbial communities in litter (Prescott and Grayston 2013; Urbanová et al. 2015) and soil microbiome composition (Ghotsa Mekontchou et al. 2022; Hannam et al. 2006; Urbanová et al. 2015). The soil microbiome and plant understory communities both affect tree seedling regeneration, decomposition, and nutrient cycling, but in turn, their community composition is shaped by different factors related to tree dominance, such as different kinds of litter deposition, as well as changes in local abiotic conditions, such as light, temperature and soil moisture and nutrient composition (Laganière et al. 2010; Nilsson and Wardle 2005; Sercu et al. 2017).

Litter deposition is one of the most important factors influencing the soil microbial community, as it changes soil nutrient inputs and dynamics in the understory (Chen et al. 2017; Laganière et al. 2010). One example is the difference between coniferous needleleaf versus broadleaf deciduous tree litter that has different chemical compositions and decomposition rates (Laganière et al. 2010; Prescott et al. 2000). For example, aspen litter contains more N, P, K, Ca and Mg and decomposes faster than spruce needleleaf litter (Laganière et al. 2010; Prescott et al. 2000). A different soil microbial community structure and a higher microbial biomass and community composition in aspen soils compared to spruce soils has been associated to differences in litter quality, plant understory composition and forest floor physicochemical properties (Hannam et al. 2006; Lamarche et al. 2004; Nagati et al. 2018; Prescott and Grayston 2013). Also, phylogenetically different trees have distinct associated phyllosphere microbiomes (i.e., microorganisms living on leaves) (Kembel et al. 2014; Laforest-Lapointe et al. 2016), which can affect belowground nutrient dynamics through litter inputs. Although there have been studies comparing litter microbial and soil microbial community composition (Hannam et al. 2007; Prescott and Grayston 2013; Urbanová et al. 2015), we have less information on the possible influence of microbial communities from the phyllosphere of dominant trees to the soil microbiome, or how canopy-related factors affect soil microbial community structure, particularly in coniferous and broadleaf deciduous forests.

The boreal forest has been changing in tree dominance from coniferous to broadleaf deciduous forests due to natural and anthropogenic disturbances (Baltzer et al. 2021; Laquerre et al. 2009; Marchais et al. 2020). These changes in turn affect understory composition and dynamics (Barbier et al. 2008). In the natural boreal forests of eastern Canada, the dominant black spruce forests (Picea mariana (Miller) Britton, Sterns & Poggenburgh) have a homogeneous understory with a relatively thick organic layer dominated by feather mosses and ericaceous plants (Légaré et al. 2001; Nilsson and Wardle 2005). These forests are characterized by acid and nutrient-poor soils with coniferous leaf litter inputs that are difficult to decompose (Laganière et al. 2010), which affects the soil microbial community composition (Lamarche et al. 2004). In contrast, the increasingly abundant trembling aspen forests (Populus tremuloides Michx.) have nutrient-rich litter inputs that promote nutrient cycling and decomposition (Cavard et al. 2011; Laganière et al. 2010), which leads to a higher abundance of microbial communities than in coniferous forests (Hannam et al. 2006). These deciduous forests have a heterogeneous understory composed by a high diversity of shrubs, herbs and bryophyte species (Cavard et al. 2011; Laganière et al. 2011; Qian et al. 2003). Hence, these increasing changes in the boreal tree dominance are likely to have a strong effect on understory dynamics, which is essential to better understand in forest management.

Litter deposition and understory plants main factors associated with tree dominance of coniferous and broadleaf deciduous trees that shape the understory composition (Rodríguez-Rodríguez et al. 2022a, in revision). We analyzed factors shaping the soil microbial community structure in coniferous and broadleaf deciduous forests. First, we analysed differences in bacterial communities between microhabitats (soil microbiome vs. tree phyllosphere) and forest types (black spruce vs. trembling aspen forests). We hypothesized that microbial communities will be different between forest types, but within forest types, bacteria will co-occur between both microhabitats, because we expect a sharing of bacterial communities from tree phyllosphere to soil microbiome through litter fall. Secondly, we analyzed if bacterial and fungal communities were affected by changes in factors associated with tree dominance, including shifts in litter deposition and understory transplantations between forest types, and correlated them with abiotic and biotic factors. On the one hand, shifts in litter deposition consisted in adding needle leaves in trembling aspen forests and broad leaves in blacks spruce forests. On the other hand, we transplanted complete plots of understory vegetation from black spruce to trembling aspen forests, and vice-versa and analyzed their microbial composition after 5 years. We hypothesized that factors associated with tree dominance will produce a microbial community composition different than the natural conditions (control) but more similar to the opposite forest type, and will be correlated with abiotic and biotic factors in each forest type.

Study area and experimental design

The study area corresponds to the spruce-moss forest domain of the Clay Belt in western Quebec (Site A: 49°11'46'' N − 78°50'33'' W; Site B: 49°09'20''N − 78°47'56''W and Site C: 49°09'39''N − 78°47'55''W), located in the Boreal Shield of Canada. Three similar sites (Fig. 1), approximatively between 0.3 and 2.3 km apart, were selected with adjacent stands of around dominant compositions of black spruce (≥ 75 of Picea mariana canopy cover) and trembling aspen (≥ 75% of Populus tremuloides canopy cover) forests. Each stand was approximatively 1 ha in size. These sites are natural forests that originated from the same wildfire ca. 1916 (Bergeron et al. 2004; Légaré et al. 2005) and have comparable abiotic conditions, such as surface deposit, slope, drainage, soil type and topography, regardless of their canopy composition, as has been described in previous studies (Cavard et al. 2011; Laganière et al. 2010; Légaré et al. 2005). These subhygric soils are composed of a fine-textured parent material dominated by clay (more than 50%) (Cavard et al. 2011).

For a complete random block design, six blocks were placed in each site, three in black spruce and three in trembling aspen stands, each block separated by at least three meters. Therefore, our study corresponds to a nested block experimental design (Fig. 1), in which we collected a total of 54 leaf samples (3 sites x 2 forest types x 3 blocks x 3 leaf samples), and 72 plots (3 sites x 2 forest types x 3 blocks x 4 treatments) in which we collected soil samples and determined understory vegetation composition.

Within each block, four 1 m² treatment-plots were randomly installed. Treatments corresponded to the 1) Control (C), as unmanipulated plots. ; 2) Litter addition (F), in which we exchanged litter inputs between forest types (i.e. adding needleleaf litter in plots in trembling aspen stands and broadleaf litter in plots in black spruce stands); 3) Transplants-out (To), corresponding to transplantations of understory vegetation from the opposite forest type (i.e. black spruce understory in trembling aspen stands and vice-versa) and 4) Transplants-in (Ti), corresponding to transplantations of understory vegetation from the same forest type, to distinguish the effect that the transplantation itself can have on the understory vegetation and soil microbial communities (i.e. black spruce understory in another black spruce stand and similarly for trembling aspen understory). These treatments were stablished in 2013 and were maintained over 5 years until 2018, as a part of the previous study in which the details of the complete in situ experiment have been fully described (Rodríguez-Rodríguez et al. 2022a, in revision). We chose these four treatments among others, as they were the two most important factors associated with tree dominance and had the stronger effect on plant understory composition (Rodríguez-Rodríguez et al. 2022a, in revision). With these shifts in litter deposition and in understory vegetation we could analyze the cumulative 5 years-effect of treatments on soil microbial communities among forest types.

To collect broadleaf litter for the litter treatment, we used 2 m² sheets placed on the ground, every year at the beginning of autumn, and the collected leaves were added to the plots. To collect needleleaf litter, 50 cm x 100 cm litter traps were placed in black spruce forests to retain the litter during winter and the collected needles were added to the plots in spring. We added the equivalent of the amount of litter that fell that year in an approximately 1 m² area in a single layer of needleleaf litter or broadleaf litter on each plot from the opposite forest type (broadleaf litter on black spruce stands and needleleaf litter on trembling aspen stands). Regarding the transplantations treatments (transplants-in and transplants-out), each transplant consisted of a 1m² surface of understory vegetation and approximately 30 cm of soil depth that was rolled on itself in a sheet and transferred to the corresponding plot in November 2013. The transplants-in (Ti) was a treatment control to ensure changes in microbial communities were not due to the shock the transplantation itself.

Sampling and DNA extraction

Leaf sampling and DNA extraction

For leaf sampling, three trees within each block in the corresponding forest type (black spruce or trembling aspen) were randomly selected at least 2 meters apart and leaves (needle leaves or broad leaves) were collected by shooting the very high branches of the bottom third of the crown with a 12 mm caliber rifle and the leaves were caught before they reached the soil to avoid possible contamination of the leaf phyllosphere with the soil microbiome. Between 50–100 g of leaf mass per sample was collected in a single day in September 2018, and the samples were stored in sterile roll bags, in a cooler on ice at approximately 4°C and then transferred to a -20°C freezer before phyllosphere DNA extraction. Epiphytic microorganisms in the tree phyllosphere were extracted from the 54 samples, following a modified version of existing protocols (Kembel et al. 2014). We added 100 ml of washing solution (1:50 diluted Redford Buffer: 1 M Tris·HCl, 0.5 M Na EDTA, and 1.2% CTAB) (Kadivar and Stapleton 2003) into the sterile bags of each sample which were shaken by hand for 5 minutes to wash epiphytic microbes from leaves. The fluid was transferred to 50ml Falcon tubes that were centrifuged (4.000 × g, 20 min, 4°C) and the supernatant was discarded by pipetting. The pellet with microbial cells was resuspended with 500 µL PowerSoil bead solution and up to 1ml was returned to the Powerbead tube and the DNA extraction was performed according to the manufacturer’s protocol of the PowerSoil DNA extraction Kit (QIAGEN). One negative extraction control per kit was processed and sequenced following the same extraction conditions. Finally, all samples were stored at -80°C prior to sequencing.

Soil sampling and DNA extraction

From each treatment plot, four soil samples were taken leaving a buffer area (as shown in Fig. 1) and were combined in a sterile bag and kept in a cooler on ice at approximately 4°C and then transferred to a -20°C freezer. Soil samples corresponded to the interface between the organic and the mineral soil horizon, taking the last organic layer just before the clay layer, with a consistent texture that sticks together. The organic soil layer was thicker in black spruce (10 cm in average) than in trembling aspen forests (5 cm in average) and is underlain by a clay layer (representative of the clay belt of Québec and Ontario). Soils were collected in the same week in August 2018 and were sieved to 2mm. We then added 0.25 g of soil into Powerbead tubes and which were then stored at -20°C before microbial DNA extraction. The rest of the soil material was dried at room temperature and stored in plastic bags for physicochemical analysis. We used the PowerSoil DNA Kit (QIAGEN) for microbial DNA extraction following the manufacturer’s protocol and a negative control of each extraction kit was processed and sequenced following the same conditions as the rest of the samples. Samples were stored at -80°C prior to sequencing.

Vegetation sampling

Considering that this project corresponded to a long-term study started in 2013 (Rodríguez-Rodríguez et al. 2022a, in revision), we used the data of understory vegetation abundance from control plots in 2018 from each forest type. These percentage abundances were calculated from counts of all species (vascular and non-vascular plants) present in a 50 x 50 cm grid placed in the center of each 1 m²-plot, leaving around 25 cm of buffer area to avoid edge effects (Fig. 1). Species taxonomy was based on the Canadensys database (http://www.canadensys.net/) and abbreviations were assigned based on a guide for our study region (Blouin and Berger 2002).

Microbial DNA sequencing

Soil DNA samples

To identify soil bacterial communities, we used a one-step PCR protocol to produce amplicon libraries containing the specific sequence for amplicon, dual indexes and all the motifs (Nextera adapters) needed for Miseq sequencing. The PCRs were performed in a 25 µL mixture containing 1 µL of 1/10 diluted DNA extract, 0.2 µM of each forward and reverse primers, 0.5 U of Phusion Hot Start II High-Fidelity DNA Polymerase (ThermoFisher), 1X of Phusion HF Buffer, 0.2mM of dNTPs, 3% DMSO following the thermal cycling conditions at 98°C (30 s and 15 s), 64°C (30 s x 33), 72°C (30 s and 10 min) and 8°C and hold. Soil DNA was PCR-amplified with the chloroplast-excluding 16S primers 799F/1115R (799F: AACMGGATTAGATACCCKG − 1115R: AGGGTTGCGCTCGTTG) from the region V5-V6 of the bacterial 16S rRNA gene (Parada et al., 2016). In order to identify potential contaminants and verify sequencing run quality we used negative and positive controls for both PCR amplification and sequencing. As negative controls we used sterile water instead of DNA, and as positive controls we used the ZymoBIOMICS™ Microbial Community DNA Standard (Zymoresearch). A total of 77 samples were sequenced (72 samples of soil microbiome, two negative extraction controls from each extraction kit, one negative PCR control, one negative extraction-PCR control and one positive PCR control), to identify soil bacteria and obtain the relative abundance of each taxon with respect to other identified bacterial taxa. Likewise, soil fungal communities were identified using the same PCR conditions explained above, with the exception that the initial mixture contained 1 µL of DNA extract (not 1/10 diluted) and we used the nuclear fungal ribosomal Internal Transcribed Spacer (ITS) region with the primers ITS1F/ITS2 (ITS1F: CTTGGTCATTTAGAGGAAGTAA - ITS2: GCTGCGTTCTTCATCGATGC).

Leaf DNA samples

To identify the epiphytic bacterial communities, we used the 16S primers 799F/1115R and the same PCR conditions as for soil DNA samples, except that the initial mixture contained 1 µL of DNA extract (not 1/10 diluted). We used sterile water instead of DNA as negative control and there was no positive control. Therefore, a total of 57 samples were sequenced (54 samples of soil microbiome, two negative extraction controls from each extraction kit and one negative PCR control).

Sequencing

After amplification, purification and normalization, samples to identify bacterial communities from both soils and leaves were sequenced in a single run, whereas samples to identify fungal communities were sequenced in a different run, with the same conditions. DNA libraries were sequenced using MiSeq paired-end 2 x 300 base pair, v3 Kit (600-cycles, Illumina reference MS-102-3003) at the UQAM CERMO-FC Genomics Platform.

Environmental data and soil physicochemical analysis

Abiotic factors were measured in each treatment-plot in 2018. A spherical convex densiometer (No. 43887/8, Forestry Suppliers, Inc.) was used to calculate the percentage of overstory density in each plot, to obtain an average of four measures by plot. Light inputs obtained in µmol m²s^− 1 units were measured in August 2018 using the device Sunfleck Ceptometer (Decagon, USA), by making an average of four measures by plot. Soil moisture was measured during one day in summer, after a 48 h period without rain to avoid a saturated relative humidity. We placed a sensor (FIELDSCOUT, TDR 300 soil moisture meter, Spectrum Technologies, Inc.), in four spots in the buffer zone of each plot to calculate an average of soil moisture per plot. To measure the soil temperature, i-Buttons® (DS1922L) were individually packed in 2 Ziploc bags for protection, configured to register the temperature every six hours in summer (May to August 2018), and placed in the buffer zone of every plot at a depth of 5–10 cm under the soil surface.

From the collected soil samples, physicochemical analysis were performed at the “Station de recherche agroalimentaire de l’Abitibi-Témiscamingue”. We obtained total carbon and nitrogen by total combustion and gas detection with a Thermal conductivity detector using the vario MAX cube analyzer. The pH was calculated in two different ways: 1) pH-H₂O with a dilution 1:1 V/V with water, and 2) pH-buffer with the Shoemaker-McLean-Pratt (SMP) method. The Mehlich III extraction method was used to measure soil minerals, that were analyzed with the Inductively coupled plasma - optical emission spectrometry (ICP-OES); and finally, we calculated the cation exchange capacity (CEC = meq / 100 g).

Bioinformatic analysis of microbial communities

Bacterial communities in soils and leaves

We analyzed the complete bacterial composition of the soil microbiome and the tree phyllosphere that were sequenced in the same run using the 16S bacterial primers to quantify variation in bacterial taxa in each forest type. A total of 5,555,443 reads were produced by high-throughput Illumina sequencing. We used the amplicon sequence variant (ASV) approach in the DADA2 package version 1.6 (Callahan et al. 2016b; Callahan et al. 2016a) in R software, version 3.6.0 (R Core Development Team 2019) for the bioinformatic analysis, following the DADA2 sequence processing workflow (Callahan et al. 2016b) with default parameters except as noted. We trimmed the first 20 nucleotides from the beginning of both forward and reverse reads to eliminate primers. We truncated the reads where quality decreased sharply at positions 270 (forward reads) and 205 (reverse reads) with maxEE setting of 2 and we used the pseudo-pooling approach to infer amplicon sequence variants (ASVs). Then, we merged forward and reverse reads, obtaining a total of 20,326 ASVs. The chimeric ASVs were removed using the consensus method, obtaining 11274 non-chimeric ASVs that were used for subsequent analyses. We used the RDP Naive Bayesian Classifier algorithm method (Wang et al. 2007) with the SILVA version 132 database (www.arb-silva.de) for taxonomic assignment from phylum to genus, and the RefSeq + RDP database (NCBI RefSeq 16S rRNA database supplemented by RDP) (Sousa 2019) to assign species-level taxonomy, to obtain more accurate species assignments. The phyloseq R package (McMurdie and Holmes 2013) was used to analyze the taxonomically annotated ASV data, containing 11,274 ASVs from 132 samples (52 from tree phyllosphere, 72 from soil microbiome and a total of 6 controls) with a total of 2,919,227 sequences.

We filtered out ASVs annotated as originating from plant chloroplasts, leaving only sequences corresponding to the kingdom Bacteria for a total of 8782 taxa and 2,622,199 sequences. The compositions of both positive and negative control samples were completely different from the evaluated samples according to a PCA ordination and all negative controls contained few sequences (between 0 and 192 sequences). Hence, we excluded all positive and negative controls prior rarefaction. The rarest ASVs were excluded, keeping ASVs with more than one sequence per sample in and with at least 10 sequences per ASV. All ASVs that occurred in at least two samples and with a minimum total abundance of 10 sequences were selected. This led to the exclusion of 4,626 ASVs. A threshold of 7300 sequences per sample was chosen to randomly rarefy the data for subsequent analysis, as this rarefaction cutoff was sufficient to capture the vast majority of ASVs in samples (Fig. S1), leaving a total of 912,500 sequences from 4156 ASVs in 125 samples after rarefaction.

Fungal communities in soils

The complete fungal composition of the soil microbiome was sequenced using the ITS primers with regions ITS1F/ITS2 (ITS1F: CTTGGTCATTTAGAGGAAGTAA - ITS2: GCTGCGTTCTTCATCGATG) (Gardes and Bruns 1993; White et al. 1990), to quantify variation in fungal taxa in each forest type. A total of 2,491,414 reads were produced by high-throughput Illumina sequencing. As for bacterial sequences, we used the ASV approach with DADA2 package version 1.6 (Callahan et al. 2016b; Callahan et al. 2016a) and sequence processing workflow in R software for the bioinformatic analysis (Callahan et al. 2016b), with default parameters except as noted. We trimmed the first 20 nucleotides from the beginning of both forward and reverse reads to eliminate primers. We truncated the reads where quality decreased sharply at positions 270 (forward reads) and 240 (reverse reads) with maxEE setting of 2 and we used the pseudo-pooling approach to infer amplicon sequence variants (ASVs). Then, we merged forward and reverse reads, obtaining a total of 3,679 ASVs. Chimeric ASVs were removed using the consensus method, obtaining 2,757 non-chimeric ASVs that were used for subsequent analyses. We determined the taxonomic identity of each ASV using the UNITE QIIME database for fungi, version 8.3 (Abarenkov et al. 2021). The phyloseq R package (McMurdie and Holmes 2013) was used to analyze the taxonomically annotated ASV data, containing 2,757 ASVs from 74 samples (72 from soil microbiome and 2 controls) with a total of 1,580,209 sequences.

Both negative controls of PCR and kit DNA extraction contained few sequences (5 and 9 sequences, respectively) and their composition were completely different from the evaluated samples according to a PCA ordination. Hence, we filtered out both controls prior rarefaction. The rarest ASVs were excluded, keeping ASVs with more than one sequence per sample in and with at least 10 sequences per ASV. This led to the exclusion of 1,655 ASVs. A threshold of 2902 sequences per sample was chosen to randomly rarefy the data for subsequent analysis, as this rarefaction cutoff was sufficient to capture the vast majority of ASVs in samples (Fig. S2), leaving a total of 206,042 sequences from 1076 ASVs in 71 samples after rarefaction.

Statistical analysis

This study analyses the bacterial and fungal communities in soils of black spruce and trembling aspen forests, and we further analyzed the possible link between the bacterial phyllosphere of each tree species to their soil bacterial microbiome. We analyzed differences in the bacterial diversity, composition, and relative abundance between microhabitats (soil microbiome and tree phyllosphere) and forest types (black spruce and trembling aspen forests). All statistical analysis were performed using R software, version 3.6.0 (R Core Development Team 2019) and figures were made using the ggplot2 package, version 3.3.5 (Wickham 2016), unless otherwise specified, and then modified in Adobe Illustrator (version 26.3.1) for a better data visualization. Differences in bacterial communities between microhabitats (phyllosphere vs. soil microbiome) and forest types (black spruce - BS vs. trembling aspen - TA) were analyzed first by randomly taking 20 subsamples per microhabitat in each forest type to use an equally amount of samples to compare and to represent the co-occurrence of bacterial ASVs in a Venn diagram, using ps_venn function (Russel 2021). The bacterial ASVs Shannon diversity (sample-level α-diversity) between tree phyllosphere and soil microbiome in black spruce and trembling aspen forests was calculated with the vegan package, version 2.5-7 (Oksanen et al. 2020) based on the rarefied relative abundances. Differences in Shannon diversity among microhabitats and forest types were tested with an ANOVA of a mixed effect model in which we included the interaction of microhabitat and forest type, using Sites and Blocks as random factors, using the stats package, version 4.1 (R Core Development Team 2019) and the lme4 package, version 1.1–27.1 (Bates et al. 2015). From the model, we also performed post hoc pairwise Tukey comparisons between microhabitat and forest type with the emmeans package, version 1.6.2-1 (Lenth et al. 2021). Finally, the Shannon diversity among microhabitats and forest types is presented as boxplots, including both the median and mean.

The bacterial community composition among microhabitats (tree phyllosphere and soil microbiome) and forest types (black spruce and trembling aspen) was analysed using a Non-Metric Multidimensional Scaling (NMDS), based on the ASVs relative abundance data previously rarefied, Hellinger transformed and with Bray-Curtis distance, using the vegan package (version 2.5-7), with bacterial phyla abundances added a posteriori to the ordination. In order to test for differences in bacterial community composition explained by microhabitat and forest type, a PERMANOVA analysis carried out using the adonis2 function of vegan package (Oksanen et al. 2020), including Site and Block as random factors.

After separating the analysis of tree phyllosphere and soil microbiome, we analyzed the relative abundance of ASVs assigned to bacterial phyla in the phyllosphere of black spruce needles and trembling aspen leaves in a stacked barplot plot using a linear model of log₁₀-transformed bacterial relative abundances (function lm of stats4 package, version 4.1.0), with a cut-off of adjusted P < 0.05 using the method of Benjamini and Hochberg (1995) to correct for multiple hypothesis testing.

In order to analyze the effect of treatments on soil bacterial and fungal communities, we explored in detail the Shannon diversity, relative abundance and community composition between treatments. Similarly as before, the Shannon diversity (sample-level α-diversity) of soil bacterial and fungal communities were calculated for the different treatments and among forest types with the vegan package (Oksanen et al. 2020) based on the rarefied relative abundances. Differences in Shannon diversity were tested with an ANOVA of a mixed effect model in which we included the interaction of forest type and treatments, using Sites and Blocks as random factors, using the stats package, version 4.1 (R Core Development Team 2019) and the lme4 package, version 1.1–27.1 (Bates et al. 2015). From the model, post hoc pairwise Tukey comparisons between microhabitat and forest type were performed with the emmeans package, version 1.6.2-1 (Lenth et al. 2021). Finally, the Shannon diversity among microhabitats and forest types is presented as boxplots, including both the median and mean.

The relative ASVs abundance of the assigned bacterial and fungal phyla in soils were compared among treatments and forest types using a linear model of log₁₀-transformed ASV relative abundances (stats4 package), with a cut-off of adjusted P < 0.05 using the method of Benjamini and Hochberg (1995) to correct for multiple hypothesis testing. The soil microbial community composition among treatments and forest types were independently presented for bacteria and fungi in a NMDS, based on the rarefied ASVs relative abundance data Hellinger transformed and with Bray-Curtis distance, using the vegan package (version 2.5-7). The ordination was presented three times, adding a posteriori with the envfit function (P < 0.05, vegan package), the microbial phyla, the abiotic factors and understory vegetation. The goodness of fit of environmental variables and soil physicochemical properties was tested from each NMDS ordination of bacterial and fungal communities, based on 999 random permutations. Also, in order to test for differences in their microbial community composition, we performed again for these ordinations a PERMANOVA analysis, including Site and Block as random factors.

Finally, we evaluated the relative importance of forest type, treatments, abiotic and biotic factors (understory vegetation) affecting soil bacterial and fungal community structure. We tested for differences in the abiotic factors including both environmental properties (soil temperature, light, overstory density and soil moisture) and soil physicochemical properties between treatments, forest types and their interaction, with a linear mixed-effect ANOVA with the nlme package (Pinheiro et al. 2021) and the Stats package (R Core Development Team 2019), respectively, using Site as a random factor. iButtons from certain plots did not work so we did data imputation for the missing values, using the averaged temperature from each Site in each forest type.

Differences in bacterial communities between microhabitats and forest types

The identified bacterial ASVs varied depending on the microhabitat (tree phyllosphere vs. soil microbiome) and forest type (black spruce vs. trembling aspen). The bacterial Shannon diversity (sample-level α-diversity, Fig. 2), based on all identified ASVs, was significantly influenced by the interaction of both microhabitat (phyllosphere and soil microbiome) and forest type (black spruce and trembling aspen) (ANOVA, F_1,121 = 45.744, P < 0.0001, Table S1). All post hoc comparisons of Shannon diversity among categories were significantly different (Tukey tests, All P > 0.001 – Table S2), with a higher Shannon diversity in soils than in tree phyllosphere, a higher Shannon diversity in the phyllosphere of black spruce needles than in trembling aspen leaves, and the opposite pattern in the soil bacterial community with a Shannon diversity higher in trembling aspen than in black spruce soils.

The Venn diagram, based on 20 samples randomly taken for each microhabitat and forest type, show the cooccurrence of bacterial ASVs between in the leaves phyllosphere and soil microbiome of black spruce and trembling aspen (Fig. 2b-h). A 65% of bacterial ASVs (2666 ASVs) was shared between black spruce and trembling aspen for both microhabitats (Fig. 2c), whereas 12% (476 ASVs) cooccurred between the phyllosphere and the soil microbiome for both forest types (Fig. 2f). In both black spruce (Fig. 2-d) and trembling aspen forests (Fig. 2e), the soil microbiome had a higher percentage of about ~ 70% of bacterial ASVs (2355 ASVs), compared to its phyllosphere (~ 22%, ≥ 748 ASVs). Therefore, a small proportion of ASVs (~ 7–9%) was shared between the tree leaves and the soil in both forest types. Comparing between forest types for each microhabitat, the needle leaves phyllosphere shared 28% of bacterial ASVs (452 ASVs) with broad leaves (Fig. 2g). In contrast, the soil microbiome shared 77% of bacterial ASVs (2300 ASVs) between black spruce and trembling aspen dominated forests (Fig. 2h). Thus, a higher proportion of soils ASVs were shared between forest types than phyllosphere ASVs between leaf types.

The bacterial community composition was different depending on microhabitat, differing between soil microbiome and tree phyllosphere (Fig. 3, first axis), and between forest types (Fig. 3, second axis), with a significant interaction between microhabitat and forest type (PERMANOVA, R² = 0.125, F_1,124 = 35.573, P < 0.0001, Table S3). The differences in bacterial phyllosphere communities between forest types were more distinct than those in soils, and the relative abundance of ASVs assigned to the different bacterial phyla in tree phyllosphere varied between forest types (Fig. 4).

The relative abundance of ASVs from the tree phyllosphere assigned to bacterial phyla between forest types (Fig. 4 and Table S4) show that the most abundant bacteria in black spruce needles were Proteobacteria, Acidobacteria and Actinobacteria, whereas in trembling aspen leaves the most abundant phyla were Actinobacteria, Proteobacteria and Bacteroidetes. Also, Planctomycetes and Verrucomicrobia were only present on black spruce leaves, whereas Deinococcus-Thermus, Fusobacteria and Gemmatimonadetes were only present in trembling aspen leaves. All bacteria were significantly different between forest types (ANOVA, all Benjamini–Hochberg-adjusted P < 0.05), except Armatimonadetes, Verrucomicrobia and Fusobacteria. However, the more abundant phyla differentiating the most in relative abundance between forest types were Proteobacteria and Acidobacteria being more abundant in black spruce leaves, whereas Actinobacteria and Bacteroidetes were more abundant in trembling aspen leaves.

Effects of shifts in factors associated with tree dominance on soil bacteria and fungi

We evaluated the effect of shifts in factors associated with tree dominance, including treatments of litter inputs and understory transplantations, on bacterial and fungal Shannon diversity, relative abundance and community composition in black spruce and trembling aspen forests and how they are correlated with abiotic and biotic factors in each forest type.

Soil bacterial Shannon diversity (Fig. 5a) was significantly different between forest types (ANOVA, F_1,62 = 18.139, P < 0.0001; Table S5), being higher in trembling aspen (emmean of 6.05 ± 0.067 SD) than in black spruce forests (emmean of 5.73 ± 0.067 SD), but did not differ significantly between treatments. In contrast, fungal Shannon diversity (Fig. 5b) was not significantly different between forest types, nor treatments (ANOVA, All P > 0.05; Table S5). Furthermore, bacterial communities were more diverse than fungal communities in soils of both forest types, according to the Shannon index (Fig. 5a,b).

The relative abundance of ASVs assigned to soil bacterial and fungal phyla was similar between treatments and some phyla differed between forest types (Fig. 6). The more relative abundant ASVs assigned to bacterial phyla (Fig. 6a) were Actinobacteria, Proteobacteria and Acidobacteria, followed by Chloroflexi and Bacteroidetes. At the phyla level, all bacteria were similar in relative abundance among treatments, whereas 12 phyla were significantly different between forest types (ANOVA, all Benjamini–Hochberg-adjusted P < 0.05, Fig. 6a). Other phyla, including Fusobacteria, Deinococcus-Thermus and Spirochaetes, had a very low relative abundance to find significant differences between treatments or forest types. Furthermore, comparing the tree phyllosphere (Fig. 4) to the soil microbiome (Fig. 6a), most of the identified bacterial ASVs co-occurred in both, except for Dependentiae, Nitrospirae, Fibrobacteres, and Elusimicrobia, which where only present in soils.

The most relative abundant ASVs assigned to fungal phyla (Fig. 6b) were Basidiomycota, Ascomycota and Mortierellomycota, with a relative abundance similar between treatments. However, in black spruce soils, Chytridiomycota and Zoopagomycota were only present in treatment To, whereas in trembling aspen soils, Deinococcus-Thermus was only present in treatment F, and Spirochaetes was absent in the control. We found significant differences in ASVs relative abundance between forest types for the assigned phyla Mortierellomycota and Murocomycota (ANOVA, all Benjamini–Hochberg-adjusted P < 0.05, Fig. 6b and Table S3). The other phyla (Chytridiomycota, Zoopagomycota and Deinococcus-Thermus) had a very low relative abundance to find significant differences between treatments or forest types.

The soil microbial community composition was analyzed among treatments in black spruce and trembling aspen forests. We analyzed independently the soil bacterial (Fig. 7) and fungal (Fig. 8) community composition and their correlation with abiotic factors and understory plant communities in each forest type. We found significant differences between forest types of different abiotic factors, including environmental conditions (i.e. light, canopy cover, soil moisture, soil temperature) and soil physicochemical properties between forest types (All ANOVA, P < 0.05 – Table S6), but any significant differences among treatments (All comparisons ANOVA, P > 0.05). The different factors were correlated with the soil microbial community composition (Table S7).

Soil bacterial community composition (Fig. 7a) was grouped by forest type in the NMDS ordination but not among treatments, which is statistically supported by significant differences between forest types (PERMANOVA, R² = 0.131, F_1,71 = 10.195, P < 0.0001), but no effect of treatment or their interaction (Table S8). The first axis of the NMDS was correlated with a gradient in cations, including H, Ca and soil pH, but most of elements were significantly correlated with each forest type (Fig. 7b and Table S7). Black spruce soils were significantly correlated with higher concentration N, C, Al, Fe, H, S, CEC and C : N and N : P ratios, as well as higher concentration of soil moisture and light inputs than trembling aspen soils (Envfit, P < 0.05 - Table S7). In contrast, trembling aspen soils were significantly correlated with soil pH, Ca, Mn and P : Ca ratio than black spruce soils (Envfit, P < 0.05 - Table S7). Finally, the composition of understory vegetation (Fig. 7c) was also significantly correlated with each forest type (Envfit, P < 0.05 - Table S7). Black spruce forests were correlated with moss species such as Pleurozium schreberi (PLS), Ptilium crista-castrensis (PTC), Dicranum polysetum (DIP) and Sphagnum spp. (SPS), as well as small plants including Gaultheria hispidula (GAH) and Geocaulon lividum (COL). In contrast, trembling aspen stands were associated with several herbs and shrubs in the understory, including Aralia nudicaulis (ARN), Viola spp. (VIS), Viburnum edule (VIE), Lysimachia borealis (TRB), Spinulum annotinum (LYA), Rubus idaeus (RUI), Petasites frigidus var. palmatus (PES), Rubus pubescens (RUP), Cornus canadensis (CON) and Mitella nuda (MAC). Therefore, forest type shapes soil bacterial communities, soil physicochemical properties and plant understory vegetation.

Soil fungal community composition (Fig. 8a) was significantly different between forest types in the first axis of the NMDS ordination (PERMANOVA, R² = 0.082, F_1,70 = 6.105, P < 0.0001), but not among treatments or their interaction (Table S8). The second axis of the NMDS ordination was correlated with soil moisture and driven by the presence of Mucoromycota and Ascomycota, which were only present in the transplants from black spruce to trembling aspen forests. Abiotic factors were significantly correlated with forest type in the first axis (Fig. 8b), with higher concentration of N, C, Na, H, Al, Fe and C : N ratio in black spruce soils (Envfit, P < 0.05 - Table S7), whereas a higher concentration of Ca, Mn and soil pH in trembling aspen soils (Envfit, P < 0.05 - Table S7). Finally, understory vegetation (Fig. 8c) was also significantly correlated with each forest type (Envfit, P < 0.05 - Table S7). Black spruce forests were correlated with moss species, such as Pleurozium schreberi (PLS), Ptilium crista-castrensis (PTC), Dicranum polysetum (DIP) and Sphagnum spp. (SPS), as well as small plants including Gaultheria hispidula (GAH) and Kalmia angustifolia (KAA). In contrast, trembling aspen stands were associated with several herbs and shrubs in the understory, including Rubus pubescens (RUP), Petasites frigidus var. palmatus (PES), Mitella nuda (MAC), Galium asprellum (GAA), Aralia nudicaulis (ARN), Rubus idaeus (RUI), Lysimachia borealis (TRB), Cornus canadensis (CON), Poa spp. (POA), Linnaea borealis (LIB), Viburnum edule (VIE), Viola spp. (VIS), Spinulum annotinum (LYA) and Clintonia borealis (CLB). Therefore, forest type was the principal factor shaping soil fungal communities, soil physicochemical properties and plant understory vegetation.

Soil bacterial communities differed between microhabitats and forest types

We analyzed the variation in soil bacterial communities between microhabitats of tree phyllosphere vs. soil microbiome, and between forest types of black spruce vs. trembling aspen forests. We found that the diversity and community composition of bacterial communities were influenced by both microhabitat and forest type. Comparing between microhabitats, we found a higher sample-level diversity (Shannon α-diversity) of bacteria in soil microbiome than in tree phyllosphere for both forest types. More than the half of identified bacterial ASVs were exclusively identified in soil microbiomes (~ 61% of ASVs) from both forest types, whereas only ~ 28% of ASVs were exclusively found in in the phyllosphere of black spruce and trembling aspen trees. Comparing between forest types, black spruce needle leaves had a higher Shannon diversity in their phyllosphere of trembling aspen broad leaves, whereas in soils the bacterial Shannon diversity was higher in trembling aspen than in black spruce forests.

We were expecting a bacterial sharing (co-occurrence) from tree phyllosphere to the soil microbiome in each forest type. However, our results indicate a high microhabitat specificity within forest types, as only a small percentage (6–7%) of ASVs co-occurred between tree phyllosphere and soil microbiome in both forest types. Although we found major differences in ASVs composition between microhabitats, at the phylum level we observed that the most relatively abundant groups present in soil microbiomes (Actinobacteria, Proteobacteria, Acidobacteria, Cloroflexi and Bacteroidetes) co-occurred with tree phyllosphere of black spruce and trembling aspen trees, indicating a prevalence of these abundant bacterial phyla in the evaluated boreal forests. Moreover, Proteobacteria and Acidobacteria were dominant phyla in black spruce needle-leaves, whereas Actinobacteria, Proteobacteria and Bacteroidetes were dominant in trembling aspen broad leaves, groups that have also been previously associated with the phyllosphere of different dominant trees (Khlifa et al. 2021; Laforest-Lapointe et al. 2016; Mason et al. 2015).

Bacterial community composition (β-diversity) was also structured by both microhabitat and forest type. Bacterial communities in soils were less variable between forest types than the tree phyllosphere, as needleleaf bacterial composition was very different compared to broad leaves. Our results agree with some European studies that found differences in microbial composition among a variety of microhabitats and among deciduous and broadleaf forests. For example, there has been reported differences among spruce needle leaves, roots and soil microbiome (Haas et al. 2018), among litter, roots, rhizosphere and soil of spruce forests (Starke et al. 2021), and among litter and soil microbial communities differing also between coniferous and broadleaf deciduous forests (Urbanová et al. 2015). Soil microbial communities are stable and resistant to local aboveground changes (Hannam et al. 2007), which explain the relatively similar soil microbial community composition within forest types. In contrast to soil microbiomes, bacterial epiphytic communities are exposed to harsh environmental conditions in the tree phyllosphere, such as ultraviolet radiation, low nutrient availability and constant fluctuations in temperature and moisture during a single day (Leveau 2006; Schlechter et al. 2019). Even the bacterial phyllosphere composition of small moss leaves in the understory is affected by forest type (Rodríguez-Rodríguez et al. 2022b). Furthermore, coniferous trees and broadleaf deciduous tress have a different physiological structure, chemical composition, and other distinct plant traits, that shape their bacterial community composition (Laforest-Lapointe et al. 2016; Lajoie and Kembel 2021; Leveau 2006). All these factors could explain the higher differences in community composition between phyllosphere of black spruce and trembling aspen, than between soil microbiomes in each forest type.

Factors associated with tree dominance affecting soil microbial communities

We analyzed if soil bacterial and fungal communities were affected by shifts in factors associated with black spruce and trembling aspen forests, including shifts in litter deposition and understory transplantations, and how the microbial composition was correlated with abiotic and biotic conditions of each forest type. We were expecting that treatments of shifts in litter deposition and understory vegetation exchanged between forest types would affect soil bacterial and fungal communities in both forest types. Despite that these same treatments were quite important field manipulations that altered understory vegetation after 5 years (Rodríguez-Rodríguez et al. 2022a, in revision), neither treatment had a significant impact on the soil bacterial and fungal Shannon diversity, relative abundance or community composition, but they were resilient to the aboveground changes made by treatments and only shaped by underlying effects of forest type. These results were surprising because litter deposition is likely one of the most important canopy-related factors affecting soil microbial communities, as it is the raw material to be decomposed and contribute with different kinds of nutrients, affecting soil pH and microbial activities (Haas et al. 2018; Laganière et al. 2010). The evaluated treatments were permanent plots, which have been submitted to leaves inputs Also, in a previous study, the transplanted understory vegetation from trembling aspen to black spruce forests remained similar to the original transplanted vegetation after 5 years, instead of becoming similar to the host stand (Rodríguez-Rodríguez et al. 2022a, in revision). Consequently, we were expecting a similar pattern for the soil microbial composition, being different as the host stand microbial composition, due to the important effect of understory plants on soil conditions and microbial communities. However, we did not detect effects on soil microbial communities, but they seem to be resistant to shifts in both litter deposition and understory vegetation, which show the importance of legacy effects of each forest type in defining microbial communities. Perhaps we could have observed a stronger effect in soil horizon above the evaluated soil layer, which corresponded to the last organic layer at the transition to the mineral soil. Also, since decomposition could be a long process, perhaps the litter inputs each year could need more time to achieve significant changes in the evaluated soil horizon. However, since these were 5-years permanent plots, we were expecting to observe an effect as it was observed for the plant understory composition (Rodríguez-Rodríguez et al. 2022a, in revision). However, our results agree with previous studies comparing spruce and aspen forests, in which shifts in litter inputs and in environmental conditions (fine root biomass, soil moisture, pH, nitrate concentrations and mesofauna abundance) did not significantly affected soil microbial communities but instead, there was a strong legacy effect of the forest type (Hannam et al. 2007; Norris et al. 2016). However, a substantial disruption in the forest floor, such as forest harvesting, can have a strong negative effect on soil microbial communities, specially for fungal more than bacterial communities, even after 15 years (Hartmann et al. 2012). Therefore, microbial communities seem to be resistant to aboveground changes, such as shifts in litter type and understory vegetation, but stronger disturbances in the forest floor (e.g. forest harvesting) or changes in tree composition can affect microbial communities.

The strong influence of forest type on soil microbial community composition found in our study agree with previous studies between coniferous and broadleaf deciduous forests (Hannam et al. 2006; Nagati et al. 2018; Norris et al. 2016; Urbanová et al. 2015). We observed that forest type had a stronger effect on fungal than bacterial communities in soil, which seems to be also a trend in other studies (Chen et al. 2019; Urbanová et al. 2015). Tree dominance of these forest types influence local conditions in the understory by affecting soil nutrient dynamics and decomposition processes (Augusto et al. 2015). Despite that the study sites had adjacent stands dominated by each forest type, with comparable soil abiotic conditions and topography, we found differences in soil physicochemical properties between forest types that were correlated with a different microbial community structure. Our results of the soil pH and chemical concentrations for both forest types agree with results of Nagati et al. (2020) from the same study sites. Forest type of broadleaf and coniferous trees, soil pH and base cation content have been considered to be main factors associated with differences in soil microbial communities (Prescott and Grayston 2013). Therefore, the influence of different factors associated with tree dominance, such as soil physicochemical properties, affects the composition of both bacterial and fungal communities.

We found differences of our results among microbial groups. Soil fungi had a lower sample-level diversity (Shannon α-diversity) than soil bacteria, a similar pattern reported in Urbanová et al. (2015) for European soils associated with different coniferous and broadleaf trees. Only three fungal phyla significantly influenced the community composition (i.e. Ascomycota, Mortierellomycota and Mucoromycota), whereas several groups of the soil bacterial community composition were significantly associated with each forest type, which included the relative abundant Actinobacteria, Proteobacteria and Acidobacteria. The differences in bacterial and fungal communities can be an indicator of forest conditions, as each microbial group carries out different ecological functions and affects the physicochemical conditions in the forest floor (Augusto et al. 2015; Van Der Heijden et al. 2008). However, the presence of the identified microbial taxa could be further analyzed in other studies to know how these microbial groups vary across seasons and which ecological functions they carry out in these boreal ecosystems, for example through metaproteomics, as it has been recently evaluated for temperate coniferous forests (Starke et al. 2021).

The relative importance of the evaluated factors associated with tree dominance, including forest type, treatments (shifts in litter deposition and plant understory), abiotic factors (local environmental conditions and physicochemical properties) and biotic factors (plant understory composition) was different for bacteria and fungal communities. Forest type was important in defining the bacterial community composition and soil physicochemical properties were also explaining part of the variance of soil bacterial communities. Several soil characteristics, including soil pH, organic carbon, soil (O₂) and redox status are considered important factors in structuring soil bacterial communities (Fierer 2017; Rousk et al. 2010). Our results indicate that soil pH, Ca and H seem to be associated with the main variation in the bacterial composition. The higher relative abundance of Acidobacteria in the more acid black spruce soils compared to trembling aspen soils are consistent with other studies highlighting the presence of this phyla in low-pH soils (Rousk et al. 2010; Urbanová et al. 2015). Several elements were also strongly associated with the bacterial community composition and differed in concentration between forest types. In black spruce soils, we found higher levels of N, C, Al, Fe, H, S, CEC and C : N and N : P ratios, as well as higher levels of soil moisture and light inputs. In contrast, trembling aspen soils had higher levels of pH, Ca, Mn and P : Ca ratio. Finally, the understory vegetation was also correlated with the bacterial community composition in each forest type, with selveral moss species and ericaceous plants in black spruce forests, whereas a high diversity of herbs and shrubs in trembling aspen forests, as previously reported (Rodríguez-Rodríguez et al. 2022a, in revision). However, it is the combined effect of all factors that explain 22% of the variation of soil bacterial communities, and the half of the variation was explained by the evaluated factors.

Regarding the soil fungal community composition, forest type had the strongest effect, followed by abiotic factors and understory vegetation. This agree with previous studies in which the effect of tree composition was more important than soil properties, explaining a large proportion of the fungal community structure (Urbanová et al. 2015). Also, in previous articles from the same study sites, the fungal composition varied between black spruce and trembling aspen forests (Ghotsa Mekontchou et al. 2022; Nagati et al. 2018). While the variability in the fungal community composition was mainly explained by forest type, soil moisture and the composition of either Mucoromycota or Ascomycota seem influence the variation of fungal communities. Both soil chemical properties and understory vegetation were associated with the fungal community composition in each forest type. In black spruce forests, we found higher levels of N, C, Na, H, Al, Fe and C : N ratio with a plant composition of mosses and ericaceous plants, whereas in trembling aspen forests, the levels of pH, Ca and Mn were higher and the understory composition was composed by diverse herbs and shrubs. However, 72% of the variation of fungal communities was not explained by the evaluated factors but remained to be elucidated.

As a conclusion, this study expands our knowledge of the microbial composition of tree phyllosphere and soil microbial communities in black spruce and trembling aspen forests and the correlation with abiotic and biotic factors in each forest type. Our study demonstrates a legacy effect of forest type in shaping soil microbial communities that were resistant to aboveground changes of litter inputs and understory vegetation. Considering the strong effect of forest type in defining the soil microbial community composition, it is to be further analyzed how increasing changes in tree dominance of different regions in the boreal system (Danneyrolles et al. 2019; Mack et al. 2021; Marchais et al. 2020) may affect soil microbial communities and their associated ecological functions.

Acknowledgements

We thank the Laboratory of Bryology of the UQAT for their support in field work, sample processing and inputs on the manuscript. We thank Steven Kembel’s Lab at UQAM for their essential support for laboratory work, bioinformatic and statistical analysis. We thank the UQAM CERMO-FC Genomics Platform and Geneviève Bourret for the high-throughput sequencing and constant help in laboratory work. We thank Danielle Charron and Julie Arsenault for her support on this project in all logistics of field and laboratory work. We thank Armand Séguin from Natural Resources Canada and Sonia Légaré from the Ministère des Forêts, de la Faune et des Parcs (MFFP), for their valuable advice at the beginning of the project. We thank Frédéric Coulombe, Tatiana Corredor and Marion Barbé for their field assistance in tree-leaves hunting and Roberto Sepulveda for his help in soil sampling and environmental measurements.

Author Contribution

The field and laboratory work, the statistical and bioinformatics analysis and the draft of the manuscript was led by JCRR with guidance from NJF and SWK (co-senior authors). Funding was held by NJF and YB and laboratory work materials by SWK lab. All authors contributed to conceive the ideas and study approach, edited the manuscript and gave final approval for publication. We have no conflicts of interest to disclose.

Data Availability Statement

The raw sequence data generated for this study will be deposited in the Sequence Read Archive (SRA) of NCBI – BioProject database. All additional data as well as R scripts for statistical analysis that support the findings of this study will be available in the Github repository (https://github.com/juanitarodriguez). Datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

This project was funded by Mitacs Acceleration program (IT06831-Drobyshev) in partnership with the Norbord Inc. and Ryam Gestion Forestière. It is a contribution of the UQAT-UQAM Chair in Sustainable Forest Management. This project was also funded by the Fonds de recherche du Québec – Nature et Technologies (FRQNT) for PhD (B2X - FRQ: 0000285039), 2020-2021 program.

Competing interests

The authors declare no competing financial interests.

Abarenkov K, Zirk A, Piirmann T, Pöhönen R, Ivanov F, Nilsson RH, Kõljalg U (2021) UNITE QIIME release for Fungi. UNITE Community.
Augusto L, De Schrijver A, Vesterdal L, Smolander A, Prescott C, Ranger J (2015) Influences of evergreen gymnosperm and deciduous angiosperm tree species on the functioning of temperate and boreal forests. Biological Reviews 90: 444-466.
Baltzer JL, Day NJ, Walker XJ, Greene D, Mack MC, Alexander HD, Arseneault D, Barnes J, Bergeron Y, Boucher Y (2021) Increasing fire and the decline of fire adapted black spruce in the boreal forest. Proceedings of the National Academy of Sciences 118.
Barbier S, Gosselin F, Balandier P (2008) Influence of tree species on understory vegetation diversity and mechanisms involved—a critical review for temperate and boreal forests. Forest ecology and management 254: 1-15.
Bartels SF, Chen HY (2013) Interactions between overstorey and understorey vegetation along an overstorey compositional gradient. Journal of Vegetation Science 24: 543-552.
Bates D, Mächler M, Bolker B, Walker S (2015) Fitting linear mixed-effects models using lme4. Journal of Statistical Software 67: 1-48. doi: doi:10.18637/jss.v067.i01.
Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal statistical society: series B (Methodological) 57: 289-300.
Bergeron Y, Gauthier S, Flannigan M, Kafka V (2004) Fire regimes at the transition between mixedwood and coniferous boreal forest in northwestern Quebec. Ecology 85: 1916-1932.
Blouin J, Berger J-P (2002) Guide de reconnaissance des types écologiques de la région écologique 5a – Plaine de l’Abitibi. Ministère des Ressources naturelles du Québec, Forêt Québec, Direction des inventaires forestiers, Division de la classification écologique et productivité des stations: 180.
Callahan B, Sankaran K, Fukuyama J, McMurdie P, Holmes S (2016b) Bioconductor workflow for microbiome data analysis: from raw reads to community analyses [version 1; peer review: 3 approved]. F1000Research 5. doi: 10.12688/f1000research.8986.1.
Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP (2016a) DADA2: high-resolution sample inference from Illumina amplicon data. Nature methods 13: 581.
Cavard X, Bergeron Y, Chen HY, Paré D (2011) Effect of forest canopy composition on soil nutrients and dynamics of the understorey: mixed canopies serve neither vascular nor bryophyte strata. Journal of Vegetation Science 22: 1105-1119.
Chen HY, Brant AN, Seedre M, Brassard BW, Taylor AR (2017) The contribution of litterfall to net primary production during secondary succession in the boreal forest. Ecosystems 20: 830-844.
Chen L, Xiang W, Wu H, Ouyang S, Zhou B, Zeng Y, Chen Y, Kuzyakov Y (2019) Tree species identity surpasses richness in affecting soil microbial richness and community composition in subtropical forests. Soil Biology and Biochemistry 130: 113-121.
Danneyrolles V, Dupuis S, Fortin G, Leroyer M, de Römer A, Terrail R, Vellend M, Boucher Y, Laflamme J, Bergeron Y (2019) Stronger influence of anthropogenic disturbance than climate change on century-scale compositional changes in northern forests. Nature communications 10: 1-7.
Fierer N (2017) Embracing the unknown: disentangling the complexities of the soil microbiome. Nature Reviews Microbiology 15: 579-590.
Gardes M, Bruns TD (1993) ITS primers with enhanced specificity for basidiomycetes‐application to the identification of mycorrhizae and rusts. Molecular ecology 2: 113-118.
Ghotsa Mekontchou C, Houle D, Bergeron Y, Roy M, Gardes M, Séguin A, Drobyshev I (2022) Contrasting structure of root mycorrhizal communities of black spruce and trembling aspen in different layers of the soil profile in the boreal mixedwoods of eastern Canada. Plant and Soil. doi: 10.1007/s11104-022-05410-8.
Haas JC, Street NR, Sjödin A, Lee NM, Högberg MN, Näsholm T, Hurry V (2018) Microbial community response to growing season and plant nutrient optimisation in a boreal Norway spruce forest. Soil Biology and Biochemistry 125: 197-209.
Hannam K, Quideau S, Kishchuk B (2006) Forest floor microbial communities in relation to stand composition and timber harvesting in northern Alberta. Soil Biology and Biochemistry 38: 2565-2575.
Hannam K, Quideau S, Kishchuk B (2007) The microbial communities of aspen and spruce forest floors are resistant to changes in litter inputs and microclimate. Applied Soil Ecology 35: 635-647.
Hartmann M, Howes CG, VanInsberghe D, Yu H, Bachar D, Christen R, Henrik Nilsson R, Hallam SJ, Mohn WW (2012) Significant and persistent impact of timber harvesting on soil microbial communities in Northern coniferous forests. The ISME journal 6: 2199-2218.
Kadivar H, Stapleton AE (2003) Ultraviolet radiation alters maize phyllosphere bacterial diversity. Microbial Ecology: 353-361.
Kardol P, De Long JR, Mariotte P (2018) Soil biota as drivers of plant community assembly. Aboveground–Belowground Community Ecology. Springer.
Kembel SW, O’Connor TK, Arnold HK, Hubbell SP, Wright SJ, Green JL (2014) Relationships between phyllosphere bacterial communities and plant functional traits in a neotropical forest. Proceedings of the National Academy of Sciences 111: 13715-13720.
Khlifa R, Houle D, Morin H, Kembel SW (2021) Inconsistent effects of nitrogen canopy enrichment and soil warming on black spruce epiphytic phyllosphere bacterial communities, taxa, and functions. Canadian Journal of Forest Research 51: 1199-1207.
Laforest-Lapointe I, Messier C, Kembel SW (2016) Tree phyllosphere bacterial communities: exploring the magnitude of intra-and inter-individual variation among host species. PeerJ 4: e2367.
Laganière J, Angers DA, Paré D, Bergeron Y, Chen HY (2011) Black spruce soils accumulate more uncomplexed organic matter than aspen soils. Soil Science Society of America Journal 75: 1125-1132.
Laganière J, Pare D, Bradley RL (2010) How does a tree species influence litter decomposition? Separating the relative contribution of litter quality, litter mixing, and forest floor conditions. Canadian Journal of Forest Research 40: 465-475.
Lajoie G, Kembel SW (2021) Plant‐bacteria associations are phylogenetically structured in the phyllosphere. Molecular Ecology 30: 5572-5587.
Lamarche J, Bradley RL, Paré D, Légaré S, Bergeron Y (2004) Soil parent material may control forest floor properties more than stand type or stand age in mixedwood boreal forests. Ecoscience 11: 228-237.
Laquerre S, Leduc A, Harvey BD (2009) Augmentation du couvert en peuplier faux-tremble dans les pessières noires du nord-ouest du Québec après coupe totale. Ecoscience 16: 483-491.
Légaré S, Bergeron Y, Leduc A, Paré D (2001) Comparison of the understory vegetation in boreal forest types of southwest Quebec. Canadian Journal of Botany 79: 1019-1027.
Légaré S, Paré D, Bergeron Y (2005) Influence of aspen on forest floor properties in black spruce-dominated stands. Plant and Soil 275: 207-220.
Lenth RV, Singmann H, Love J, Buerkner P, Herve M (2021) emmeans: Estimated marginal means, aka least-squares means. R package version 162-1.
Leveau J (2006) Microbial communities in the phyllosphere. In: M Riederer, C Muller (eds) Biology of the Plant Cuticle. Annual Plant Reviews, Blackwell Pub., Oxford, UK.
Mack MC, Walker XJ, Johnstone JF, Alexander HD, Melvin AM, Jean M, Miller SN (2021) Carbon loss from boreal forest wildfires offset by increased dominance of deciduous trees. Science 372: 280-283.
Marchais M, Arseneault D, Bergeron Y (2020) Composition changes in the boreal mixedwood forest of western Quebec since Euro-Canadian settlement. Frontiers in Ecology and Evolution 8: 126.
Mason CJ, Pfammatter JA, Holeski LM, Raffa KF (2015) Foliar bacterial communities of trembling aspen in a common garden. Canadian journal of microbiology 61: 143-149.
McMurdie PJ, Holmes S (2013) phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PloS one 8: e61217.
Nagati M, Roy M, DesRochers A, Bergeron Y, Gardes M (2020) Importance of Soil, Stand, and Mycorrhizal Fungi in Abies balsamea Establishment in the Boreal Forest. Forests 11: 815.
Nagati M, Roy M, Manzi S, Richard F, Desrochers A, Gardes M, Bergeron Y (2018) Impact of local forest composition on soil fungal communities in a mixed boreal forest. Plant and Soil 432: 345-357.
Nilsson M-C, Wardle DA (2005) Understory vegetation as a forest ecosystem driver: evidence from the northern Swedish boreal forest. Frontiers in Ecology and the Environment 3: 421-428.
Norris CE, Quideau SA, Oh S-W (2016) Microbial utilization of double-labeled aspen litter in boreal aspen and spruce soils. Soil Biology and Biochemistry 100: 9-20.
Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin PR, O’Hara R, Simpson GL, Solymos P, Stevens MHH, Wagner H (2020) vegan: Community Ecology Package. R package version 2.5-7.
Prescott C, Zabek L, Staley C, Kabzems R (2000) Decomposition of broadleaf and needle litter in forests of British Columbia: influences of litter type, forest type, and litter mixtures. Canadian Journal of Forest Research 30: 1742-1750.
Prescott CE, Grayston SJ (2013) Tree species influence on microbial communities in litter and soil: current knowledge and research needs. Forest Ecology and Management 309: 19-27.
Qian H, Klinka K, Økland RH, Krestov P, Kayahara GJ (2003) Understorey vegetation in boreal Picea mariana and Populus tremuloides stands in British Columbia. Journal of Vegetation Science 14: 173-184.
R Core Development Team (2019) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
Rodríguez-Rodríguez JC, Bergeron Y, Fenton NJ, Kembel SW (2022a) Drivers of shifts in boreal understory vegetation between coniferous and broadleaf deciduous alternative states. Submitted to Ecological Monographs. In revision.
Rodríguez-Rodríguez JC, Bergeron Y, Kembel SW, Fenton NJ (2022b) Dominance of coniferous and broadleaved trees drives bacterial associations with boreal feather mosses. Environmental Microbiology n/a. doi: https://doi.org/10.1111/1462-2920.16013.
Rousk J, Bååth E, Brookes PC, Lauber CL, Lozupone C, Caporaso JG, Knight R, Fierer N (2010) Soil bacterial and fungal communities across a pH gradient in an arable soil. The ISME journal 4: 1340-1351.
Russel J (2021) Russel88/MicEco: v0.9.15. Zenodo.
Schlechter RO, Miebach M, Remus-Emsermann MN (2019) Driving factors of epiphytic bacterial communities: a review. Journal of advanced research 19: 57-65.
Sercu BK, Baeten L, van Coillie F, Martel A, Lens L, Verheyen K, Bonte D (2017) How tree species identity and diversity affect light transmittance to the understory in mature temperate forests. Ecology and Evolution 7: 10861-10870.
Sousa AG (2019) GTDB and RefSeq-RDP databases parsed for species assignment [Data set]. Zenodo. doi: http://doi.org/10.5281/zenodo.2658728.
Starke R, Mondéjar RL, Human ZR, Navrátilová D, Štursová M, Větrovský T, Olson HM, Orton DJ, Callister SJ, Lipton MS (2021) Niche differentiation of bacteria and fungi in carbon and nitrogen cycling of different habitats in a temperate coniferous forest: A metaproteomic approach. Soil Biology and Biochemistry 155: 108170.
Urbanová M, Šnajdr J, Baldrian P (2015) Composition of fungal and bacterial communities in forest litter and soil is largely determined by dominant trees. Soil Biology and Biochemistry 84: 53-64.
Van Der Heijden MG, Bardgett RD, Van Straalen NM (2008) The unseen majority: soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems. Ecology letters 11: 296-310.
Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Applied and environmental microbiology 73: 5261-5267.
White TJ, Bruns T, Lee S, Taylor J (1990) Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. PCR protocols: a guide to methods and applications 18: 315-322.
Wickham H, W. Chang, L. Henry, T. Pedersen, K. Takahashi, C. Wilke, K. Woo, H. Yutani, and D. Dunnington (2016) ggplot2: Elegant Graphics for Data Analysis. Springer Verlag, New York.

Supplementarymaterial.docx

Download PDF

Journal Publication

published 02 Mar, 2023

Read the published version in Plant and Soil →

Reviewers agreed at journal
19 Nov, 2022
Reviewers invited by journal
15 Nov, 2022
Editor invited by journal
08 Nov, 2022
Editor assigned by journal
08 Nov, 2022
First submitted to journal
08 Nov, 2022

You are reading this latest preprint version

Tree dominance shapes soil and tree phyllosphere microbial communities in coniferous and broadleaf deciduous boreal forests

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Materials And Methods

Study area and experimental design

Sampling and DNA extraction

Leaf sampling and DNA extraction

Soil sampling and DNA extraction

Vegetation sampling

Microbial DNA sequencing

Soil DNA samples

Leaf DNA samples

Sequencing

Environmental data and soil physicochemical analysis

Bioinformatic analysis of microbial communities

Bacterial communities in soils and leaves

Fungal communities in soils

Statistical analysis

Results

Differences in bacterial communities between microhabitats and forest types

Effects of shifts in factors associated with tree dominance on soil bacteria and fungi

Discussion

Soil bacterial communities differed between microhabitats and forest types

Factors associated with tree dominance affecting soil microbial communities

Declarations

Acknowledgements

Author Contribution

Data Availability Statement

Declarations

Competing interests

References

Supplementary Files

Status:

Journal Publication

Version 1