https://orcid.org/0000-0001-7735-5131
Abstract
The rhizobiome is being increasingly acknowledged as a key player in plant health and breeding strategies. The pine pitch canker (PPC), caused by the fungus Fusarium circinatum, affects pine species with varying susceptibility degrees. Our aims were to explore the bacterial rhizobiome of a susceptible (Pinus radiata) and a resistant (Pinus pinea) species together with other physiological traits, and to analyze shifts upon F. circinatum inoculation. Pinus seedlings were stem inoculated with F. circinatum spores and needle gas exchange and antioxidant-related parameters were analyzed in non-inoculated and inoculated plants. Rhizobiome structure was evaluated through 16S rRNA gene massive parallel sequencing. Species (non-inoculated plants) harbored distinct rhizobiomes (<40% similarity), where P. pinea displayed a rhizobiome with increased abundance of taxa described in suppressive soils, displaying plant growth promoting (PGP) traits and/or anti-fungal activity. Plants of this species also displayed higher levels of phenolic compounds. F. circinatum induced slight changes in the rhizobiome of both species and a negative impact in photosynthetic-related parameters in P. radiata. We concluded that the rhizobiome of each pine species is distinct and higher abundance of bacterial taxa associated to disease protection was registered for the PPC-resistant species. Furthermore, differences in the rhizobiome are paralleled by a distinct content in phenolic compounds, which are also linked to plants’ resistance against PPC. This study unveils a species-specific rhizobiome and provides insights to exploit the rhizobiome for plant selection in nurseries and for rhizobiome-based plant-growth-promoting strategies, boosting environmentally friendly disease control strategies.
Introduction
Fusarium circinatum Nirenberg and O’Donnell (teleomorph = Gibberella circinata) is the causal agent of pine pitch canker (PPC), one of the most serious pine diseases (Wingfield et al. 2008). Sporadic outbreaks and epidemics caused by this fungus have been reported worldwide, threatening natural pine forests, commercial plantations, and particularly nurseries (Martín-García et al. 2019, Quesada et al. 2019, Wingfield et al. 2008). PPC is characterized by the obstruction of water and nutrients flow in the stem (Martín-Rodrigues et al. 2013), which leads to symptoms including damping off and wilting of seedlings; or branch die-back, stem cankers, pitch formation and mortality in adult trees (Wingfield et al. 2008).
It is well-known that Pinus susceptibility to F. circinatum is species dependent. Several reports indicate that Pinus radiata, Pinus patula and Pinus elliottii are highly susceptible to PPC, whereas Pinus tecunumanii, Pinus oocarpa, Pinus canariensis, Pinus pinea, and Pinus thunbergii are highly resistant (reviewed by Martín-García et al. 2019). This differential response may be dependent on a wide range of genetic or physiological differences among species. Recently, several reports (Amaral et al. 2019a, 2019b, Amaral et al. 2021, Zamora-Ballesteros et al. 2021) investigated pine species with different responses to PPC adding new insights into the physiological and molecular mechanisms behind PPC susceptible and resistant host phenotypes. The susceptible P. radiata suffered changes in plant water status and stomata closure, as well as photosynthetic impairment upon inoculation with F. circinatum (Amaral et al. 2019a, Amaral et al. 2021, Cerqueira et al. 2017). Sink metabolism induction, accumulation of amino acids and overexpression of pathogenesis-related genes were also observed (Amaral et al. 2019a). On the other hand, P. pinea, resistant to PPC, was able to maintain the stomatal opening and increased transpiration after inoculation with F. circinatum (Amaral et al. 2019a, 2019b, Amaral et al. 2021).
The chemical composition of woody plants, including pines, may be determinant in host defense mechanisms against pathogenic fungi (Witzell and Martín 2008). Phenolic compounds can have a fungicidal effect, antioxidant properties, modulate the activity of precursors of defense-related compounds and enhance mechanical barriers (Daayf et al. 2012, Witzell and Martín 2008). The variability of these compounds due to genetic intraspecific variation among hosts has been recognized in examples such as the Chrysomyxa rhododendri-Norway spruce pathosysthem (Ganthaler et al. 2017) and are now being explored as markers of resistance to pathogens in tree protection (Sherwood and Bonello 2013).
The outcome of plant genotype-based traits, plant physiological performance and the ability to respond to environmental variables may modulate root exudates composition driving changes in soil chemical properties and microbial communities (Hubbard et al. 2019, Lareen et al. 2016, Olanrewaju et al. 2019). These microbial communities interact with each other impacting hosts’ fitness, by enhancing nutrient assimilation, producing plant hormones and respective precursors, as well as boosting plant immunity (Backer et al. 2018). The microorganisms present in root-associated communities form the rhizobiome (Cúcio et al. 2016). These range from bacteria, to archaea, fungi, viruses and protozoa, and may be found inside the roots (endophytes), at its surface (epiphytes) or live in the nearby soil (Hinsinger et al. 2009). Recent findings suggest that host defense systems not only provide protection against pathogens but may also regulate the composition of its microbial community (Jones et al. 2019), including the rhizobiome in trees (Habiyaremye et al. 2020, Li et al. 2018). From a practical point of view, the fact that the rhizosphere may be controlled by a selective pressure driven by host plants is of prime importance for plant breeders, as well as developing new phenotypes (Wei and Jousset 2017).
Studies on forest tree rhizobiomes and their roles in plant health and fitness lag far behind when compared to parallel research on crops (Maghnia et al. 2019, Terhonen et al. 2019). However, some advances regarding these communities have been made. Li et al. (2018) suggested that the plant genotype shapes its rhizobiome by driving the development of plant phenotypes in spruce. To our knowledge, in Pinus spp., studies underlying the role that the host may play in the structure of its rhizobiome, alongside the Pinus-Fusarium pathosystem, do not exist. The few reported studies were performed to assess microbiome diversity and its potential as bio-inoculant in Pinus roxburghii (Naz et al. 2018) and to describe the rhizobiome-soil interactions in Pinus contorta (Chow et al. 2002). Both studies identified the prevalence of bacterial families such as Sphingomonadaceae, Acetobacteraceae, Rhizobiaceae, Caulobactereaceae, Burkholderiaceae, Comamonadaceae, Pseudomonadaceae and Acidimicrobiaceae (Chow et al. 2002, Naz et al. 2018).
Another understudied question in forest trees is related with the dynamics of the rhizobiome assembly under pathogen attack. The bacterial community structure of P. thunbergii naturally infected with the nematode Bursaphelenchus xylophilus was recently reported (Ma et al. 2020). Results revealed a significant difference between healthy and diseased trees (roots and needles) in terms of community structure and functional profile. In healthy roots, the bacterial community was dominated by Acidobacteria, Planctomycetes and Rhizobiales, whereas in the diseased roots, Proteobacteria, Firmicutes and Burkholderiales were the dominant ones. Given the wide-scale planting and economic importance of pines and the need to control emerging diseases such as PPC there is an urgent need to understand in detail host defense associated players, such as the rhizobiome, and how these influence host–pathogen interactions in order to design novel forest management and health strategies. The inoculation of specific bacterial strains in disease suppressive soils is a proven effective method of counteracting fungal infections in various plants (Schlatter et al. 2017) such as in banana orchards against soil-borne Fusarium wilt disease (Shen et al. 2015) or against downy mildew, a foliar pathogen, in Arabidopsis thaliana (Berendsen et al. 2018). Its use in woody plants is still sporadic and often accompanied by other chemical methods (Kyselková and Moënne-Loccoz 2012, Mazzola et al. 2002). More information is needed to foster a broader application.
Bearing this in mind, the aim of this study was to explore the bacterial rhizobiome (hereinafter referred to as rhizobiome for simplification) of two pine species with contrasting responses to F. circinatum infection: P. radiata (susceptible) and P. pinea (resistant). The characterization of the rhizobiome dynamics induced by F. circinatum inoculation was carried out together with the analysis of their constitutive background. Both species were germinated and kept under standard greenhouse production conditions until the controlled laboratory experiment was carried out excluding the influence of environmental drivers. The physiological traits related with gas exchange parameters and antioxidant capacity were evaluated to assess plant performance and its possible influence in plants’ rhizobiome. To our knowledge, this is the first study analyzing the rhizobiome of these two pine species and the pathosystem Pinus spp.–F. circinatum.
Material and methods
Plant material
Eight months-old pine P. radiata D. Don (Turkish provenance, highly susceptible to PPC) and P. pinea L. (Portuguese provenance region 5, relatively resistant to PPC) seedlings with 15 ± 2 cm height, were obtained from Sociedade Agrícola Pecuária Melo a Cancela Lda (Anadia, Portugal) and transplanted to 200 mL pots containing a 3:2 (w/w) peat:perlite mixture. The plants were acclimatized during one month in a climate chamber (Fitoclima D1200, Aralab, Portugal) under controlled day/night conditions: 16/8 h photoperiod, 25/20°C mean temperatures; 60/65% relative humidity; 500 μmol m2 s−1 photosynthetic photon flux density (PPFD). The plants were watered daily and fertilized weekly (Frutifol, Nufarm, Portugal, N:P:K 5:8:10).
Fungal material and plant inoculation
The F. circinatum isolate (FcCa6; mating type 2; Martínez-Álvarez et al. 2012) was obtained from the collection of the Forest Entomology and Pathology Lab at the University of Valladolid. The isolate was grown in Potato Dextrose Agar (PDA; Merck, Darmstadt, Germany), at room temperature for 5 days. Afterward, small square incisions were made on the border of the isolate (1–2 cm) and subsequently, immersed in Potato Dextrose Broth (PDB; Merck, Darmstadt, Germany) and incubated at 20 ± 2°C for 1 day under continuous agitation. In order to obtain a solution with 1 × 106 spores mL−1, the liquid medium with the fungus was filtrated through a sterile gauze to remove most of the hyphae, and the concentration of spores was verified using a Neubauer chamber on an optical microscope. Fungal artificial inoculation was performed by wounding the plants’ stems (app. 4 cm aboveground) using a sterile scalpel, followed by the inoculation of 10 μL of PDB with F. circinatum spores (Amaral et al. 2019a). The wounded site was sealed using Parafilm®. For control groups, the same procedure was followed using 10 μL of sterile PDB.
Experimental design
To evaluate the impact of plant inoculation with F. circinatum and of host genotype on the pine rhizobiome composition, two groups of plants were set up per pine species: (i) the control group (Control, CR for P. radiata and CP for P. pinea) (n = 8 per species); and (ii) the fungus-inoculated group (Inoculated, FR for P. radiata and FP for P. pinea), in which plants were inoculated with F. circinatum (n = 16 per species). Plants were randomly placed in the climate chamber and were periodically and randomly moved to the neighboring position during the entire experiment so to minimize the effect of possible environmental heterogeneity. They were verified daily for disease symptoms, such as apical dieback and needle wilting and chlorosis. The trial was conducted until an inoculated group displayed at least 50% symptomatic plants (as in Cerqueira et al. 2017, Amaral et al. 2019a, 2019b). The experiment was carried out under the same conditions described for the acclimatization period. Only asymptomatic plants were sampled. Pine needles were collected, immediately frozen in liquid nitrogen, and kept at −80°C for antioxidant capacity analysis. Three small stem pieces of each inoculated and non-inoculated plant were collected (at the inoculation point and 1 cm above and below this point) and plated in PDA to re-isolate and confirm the presence or the absence of the fungus in each group. When present (inoculated plants), the pathogen was identified by micromorphological analysis.
Needle gas exchange-related parameters
Needle gas exchange-related parameters were measured at the time of the sampling as described by Amaral et al. (2019a) before plants were processed for further analysis. Net CO2 assimilation rate (A, μmol CO2 m−2 s−1 g DW−1), stomatal conductance (gs, mmol H2O m−2 s−1 g DW−1), transpiration rate (E, mmol H2O m−2 s−1 g DW−1), and intercellular CO2 concentration content (Ci, vpm) were measured with a gas exchange system (LCpro-SD, ADC BioScientific Limited, Hertfordshire, UK) coupled to a conifer-type chamber. Inside the chamber, the following conditions were maintained during all the measurements: ambient CO2 concentration: 406.66 ± 6.72 μmol m2 s −1; air flux: 201 μmol s−1; block temperature: 29.2 ± 0.56°C. To find out the saturation light intensity, light response curves of CO2 assimilation curves (A/PPFD) were performed with the following PPFD: 2500, 2000, 1500, 1000, 750, 500, 250, 100, 50 and 0 μmol m2 s−1. Measurements at saturation light intensity were performed at 1000 μmol m2 s−1. Data were recorded when the measured parameters were stable (2–6 min). At least four biological replicates of each group were analyzed.
Antioxidant-related parameters
At least four asymptomatic individuals of each treatment were selected for this analysis. Samples were processed as described by Dinis et al. (2012) with small modifications. Briefly, forty milligrams of frozen needles were ground to powder with liquid nitrogen using a pestle and mortar, which was followed by the addition of 1.5 mL of 70% (v/v) methanol. Samples were successively shaken in an orbital shaker at 700 rpm, 25°C for 1 h, followed by centrifugation at 10,000 g, 4°C for 15 min. This process was repeated four times until obtaining a final volume of ~6 mL, which was then frozen at −80°C until quantification.
Total phenolic compounds and ortho-phenols
Total phenolic content was estimated by the Folin Ciocalteu’s method (Singleton and Rossi 1965). Briefly, 20 μL of gallic acid standards/needle extracts were loaded in each microplate well. Afterward, 90 μL of distilled water and 10 μL of Folin-Ciocalteau reagent solution were added to each well and the microplate was kept in the dark for 6 min. Subsequently, 80 μL of 7% (w/v) sodium carbonate was added and the microplate was kept in the dark for 2 h. After this period, absorbance was measured at 750 nm using a microplate reader (Synergy HT, BioTek Instruments, Winooski, VT).
For ortho-phenols quantification (Singleton et al. 1999), 160 μL of gallic acid standards/needle extracts were loaded in each microplate well. Afterward, 40 μL of 5% (w/v) sodium molybdate solution were added to each microplate well and the microplate was kept in the dark for 15 min at room temperature. After this period, absorbance was measured at 370 nm using a microplate reader (Synergy HT, BioTek Instruments, Winooski, VT).
For both total phenolic and ortho-phenols content, calibration curves were performed using the following concentrations of gallic acid in 70% (v/v) methanol: 1, 0.500, 0.250, 0.125, 0.063, 0.031, 0.016, 0.008, 0.004 and 0 mg mL−1. Blanks were performed using 70% (v/v) methanol. All measurements were performed in triplicate.
Data of total phenolics and ortho-phenols contents were expressed as mg−1 of gallic acid equivalents g−1 of fresh weight (mg −1 gFW−1).
Total flavonoid content
Total flavonoid content was measured with the aluminium chloride colorimetric assay (Chang et al. 2002). Catechin standards/needle extracts (60 μL) were loaded in triplicate into microplate wells. Then, 28 μL of 5% (w/v) sodium nitrite solution were added to each microplate well and kept in the dark for 6 minutes at room temperature, followed by the addition of 28 μL of 10% (w/v) aluminium chloride solution to each microplate well, being also kept in the dark for 6 min. Lastly, 120 μL of 4% (w/v) sodium hydroxide solution was added to the microplate and the absorbance was read at 510 nm (Synergy HT, BioTek Instruments, Winooski, VT). A calibration curve was built using catechin diluted in 70% (v/v) methanol as standard, with the following concentrations: 0.500; 0.250; 0.125; 0.063; 0.031; 0.016; 0.008; 0.004 and 0 mg/mL Blanks were performed using 70% (v/v) methanol. The data for total flavonoids contents were expressed as mg−1 of catechin equivalent by g−1 of fresh weight (mg −1 gFW−1).
Total antioxidant capacity
Total antioxidant capacity based on the DPPH (2,2-diphenyl-1-picrylhydrazyl)-free radical scavenging capacity of needle extracts was accessed (Xu and Chang 2007). Briefly, 22 μL of needle extracts were loaded in a microplate where 200 μL of 120 μM DPPH solution were added to each microplate well, which was kept in the dark for 30 min at room temperature. After this period, the absorbance for the sample (A1) and negative control (A0) was measured at 517 nm against a 70% (v/v) methanol blank. All measurements were performed in triplicate using a microplate reader (Synergy HT, BioTek Instruments, Winooski, VT).
The percentage of DPPH radical reduction was calculated as follows: |$\frac{\Big(A0-A1\Big)}{A0}\times 100$| and represented as a percentage (%) of DPPH inhibition.
Rhizobiome analysis
For rhizobiome analysis, samples were collected and processed following the protocol described by Simmons et al. (2018), with minor changes. Briefly, the roots collected were shaken vigorously in order to retain only the soil closely adherent to the roots and were then placed in 50 mL falcon tubes filled with 35 mL of phosphate-buffered saline (PBS). The tubes were agitated at 35 rpm for 25 minutes. Between each one of the three washing cycles, tubes were centrifuged (4,000 g, 10 min, 4°C) and the supernatants were discarded. The resulting rhizospheric soil was frozen and kept at −80°C. About 0.25 g of rhizospheric soil was used for DNA extraction and purification following the manufacturer’s instructions of the ‘DNeasy Powersoil Kit’ (MoBio, Qiagen).
Twelve samples were selected to be sequenced: six samples from control plants (3 per pine species) and six samples from plants inoculated with F. circinatum (3 per pine species). Rhizobiome profiling with Illumina MiSeq was performed by Eurofins Genomics (Ebersberg, Germany) on the V3-V4 region of the 16S rRNA gene. Data processing and taxonomical classification consisted of removing all reads with errors, as well as removing all ambiguous bases (‘N’). The remaining reads were processed using minimum entropy decomposition, which partitioned the marker gene dataset into operational taxonomic units (OTU). Each OTU consisted of a cluster with significant sequence divergence to another cluster. Taxonomic data were assigned to each OTU by DC-MEGABLAST alignments of representative cluster sequences to the National Center for Biotechnology Information (NCBI) database (https://www.ncbi.nlm.nih.gov/nucleotide/, assessed 2019-01-05). Only sequences with 70% sequence identity across at least 80% of the members of the cluster were considered for reference purposes. Sequences were not assigned if they were considered as noise (which includes potential chimeric sequences and singletons). Normalization of the number of reads was achieved by taking into account the estimated number 16S gene copy number per phylogenetic marker (Angly et al. 2014).
The sequencing data were deposited in the NCBI database under accession number PRJNA750408.
Statistical analysis
Statistical analysis was conducted in R version 3.6.1 (R Core Team 2014), using the ‘vegan’ (vegdist function) (Oksanen et al. 2008), ‘phyloseq’ (phyloseq, plot_heatmap, plot_bar, plot_richnesss functions) (McMurdie and Holmes 2013) and ‘DESeq2’ (phyloseq_to_deseq2 and DESeq functions) (Love et al. 2016) software packages.
All treatments were tested for their normality (Shapiro–Wilk test) and homogeneity of variances (Levene test), and, consequently, compared accordingly with parametric [one-way analysis of variance (ANOVA), analysis of variance, followed by Tukey post hoc test] or nonparametric tests (Kruskall–Wallis test) that suited the analysis. Both α, i.e., Shannon’s index and richness (OTUs abundance), and β diversity metrics were calculated based on OTUs abundance tables resulting from massive parallel sequencing analysis, transformed by log(x + 1).
A Bray–Curtis dissimilarity measure was applied to infer a distance matrix among the rhizobiome structure of the samples, resulting in a similarity dendrogram and in a principal coordinate analysis (PCoA) plot. Furthermore, the data concerning taxonomic groups was verified through PERMANOVA, followed by a Monte-Carlo test. The P value deemed significant for all statistical tests was considered to be below 0.05. All figures were constructed using the R software package ‘phyloseq’ using the functions specified above, with the exception of the differential abundance figures (based on the log base 2), which were used to assess significantly different OTUs and families using the Wald test (DEseq function) under the ‘DESeq2’ software package.
Results
Plant symptomatology and physiology
Plant symptomatology
P. radiata inoculated plants displayed the first disease symptoms (needle wilting and apical dieback) 7 days post-inoculation with F. circinatum, with 85% of inoculated plants being asymptomatic at this point (Figure 1). This value decreased to 30% just 10 days post-inoculation, reaching the criteria selected to perform the sampling. P. pinea inoculated group did not show symptoms of disease throughout the entire experiment (Figure 1). Koch’s postulates validated the inoculation with F. circinatum in inoculated plants.
Percentage of inoculated asymptomatic plants of P. radiata and P. pinea following fungal inoculation with F. circinatum.
Needle gas exchange related parameters
Significant differences between inoculated P. radiata and its control were found regarding needle gas exchange related parameters (Figure 2). Inoculated P. radiata plants showed lower transpiration rate, stomatal conductance, and net CO2 assimilation rate; yet exhibited higher levels of internal CO2 concentration when compared to the respective control. For P. pinea no statistical differences were found between control groups and inoculated ones for internal CO2 concentration and net CO2 assimilation rate (Figure 2). Nonetheless, inoculated samples of P. pinea showed higher significant transpiration rate and stomatal conductance than the controls of the same species.
Needle gas exchange related parameters: (a) transpiration rate (E), (b) stomatal conductance (gs), (c) intercellular CO2 concentration (Ci) and (d) net CO2 assimilation rate of control and inoculated P. radiata and P. pinea plants after F. circinatum inoculation. Data are presented as mean ± SD. Statistically similar data are grouped by similar lower-case letters (post-hoc Tukey test, P value < 0.05).
Antioxidant-related parameters
No significant differences in the levels of total phenolic compounds, ortho-phenols, total antioxidant capacity and flavonoid content were registered between non-inoculated and inoculated plants for both pine species (Figure 3). However, P. pinea samples exhibited a significantly higher basal level of phenols, ortho-phenols and flavonoids in comparison with P. radiata (Figure 3).
Antioxidant-related parameters plots: total phenolic compounds (a), DPPH scavenging activity (b), ortho-phenols (c) and flavonoids content (d) of control and inoculated P. radiata and P. pinea plants after F. circinatum inoculation. Data are presented as mean ± SD. Statistically similar data are grouped by similar lower-case letters (post-hoc Tukey test, P value < 0.05).
Rhizobiome analysis
In total, 513,269 raw reads were obtained, with an average of 42,772 ± 5,753 (mean ± standard error) reads per sample, which were assigned to a total of 725 OTUs (234 ± 18 per sample) (see Figure S1 available as Supplementary Data at Tree Physiology Online). The rarefaction curve for each sample reached a point of full saturation (see Figure S2 available as Supplementary Data at Tree Physiology Online).
In terms of β-diversity, the rhizobiomes of P. pinea and P. radiata shared <40% of similarity (Figure 4a). The two main axis in the PCoA plot explained 69% of the data variation (Figure 4b). The two pine species are clearly separated along principal coordinate 1 (PCoA1), which explains 61.8% of the variation. The pairwise PERMANOVA confirmed significant differences between the control groups of the two pine species (Figure S3). No significant differences were found between inoculated and control groups for both pine species.
Beta diversity analysis. (a) Similarity dendrogram using Bray–Curtis distance; (b) PCoA mapping each sample by using Bray–Curtis distance; CP: P. pinea control group; FP: P. pinea inoculated group; CR: P. radiata control group; FR: P. radiata inoculated group.
At the OTU level, both control (84 OTUs) and inoculated (91 OTUs) plants of P. radiata displayed a higher quantity of exclusive OTUs when compared to the same treatments in P. pinea (45 OTUs and 72 OTUs, respectively), as seen in Figure 5. In P. radiata groups (control and inoculated) 155 OTUs were shared, whereas between P. pinea groups there were 107 common OTUs. At the core, 101 OTUs were shared between all treatments.
Venn diagram of the OTUs common and exclusive to each of the four groups: P. pinea control group; P. pinea inoculated group; P. radiata control group; P. radiata inoculated group.
The rhizobiome of P. radiata control group displayed a non-significant higher richness and diversity (Shannon index) than the rhizobiome of control P. pinea plants (Figure 6). On the other hand, the control samples of both pine species showed non-significant higher richness and diversity than the inoculated samples (Figure 6).
Boxplot of alpha diversity metrics: Richness (observed OTUs) and Shannon index for both control and inoculated groups of P. pinea and P. radiata.
The three most abundant phyla were similar for all samples, regardless of the pine species and treatment (Figure 7): Proteobacteria (CR: 56.4 ± 2.4%; FR: 48.9 ± 4.9%; CP: 42.4 ± 3.6%; FP: 39.6 ± 2.4%); Acidobacteria (CR: 16.1 ± 0.7% vs. FR: 18.9 ± 3.6%; CP: 30.4 ± 2.8% vs. FP: 27.0 ± 1.6%); and Actinobacteria (CR: 15.9 ± 4.1%; FR: 19.7 ± 3.7%; CP: 20.3 ± 1.6%; FP: 26.8 ± 3.0%).
The relative abundance of phyla among the samples and respective treatment. Because of a large portion of reads belonging to non-dominant phyla, a category designated ‘Others’ (containing OTUs with <2% relative abundance) was included for clarity. CP: P. pinea control group; FP: P. pinea inoculated group; CR: P. radiata control group; FR: P. radiata inoculated group.
The dominant bacterial families in all samples were Acidobacteriaceae (CP: 13.4 ± 3.0%; FP: 10.8 ± 1.4%; CR: 6.4 ± 1.1%; FR: 8.33 ± 1.7%) and Acidimicrobiaceae (CP: 6.2 ± 1.1%; FP: 7.9 ± 1.9%; CR: 3.0 ± 0.3%; FR: 5.3 ± 1.7%) (Figure 8). Bacterial families for which abundance was significantly different between P. radiata and P. pinea controls are displayed in Figure 9 according to their logarithm (base 2) based differences. Nocardioidaceae, Burkholderiaceae, Roseiarcaceae, Acetobacteraceae, Xanthomonadaceae, Sphingobacteriaceae, Methylocystaceae and Acidobacteriaceae were statistically more abundant in P. pinea. On the other hand, in the P. radiata rhizobiome, Bradyrhizobiaceae, Micropepsaceae, Xanthobacteraceae, Hyphomicrobiaceae, Rhodobacteraceae, Gemmatimonadaceae and Phaselicystidaceae were significantly more abundant. No significant differences were found between control and inoculated groups of both pine species concerning the abundance of dominant bacterial families.
The relative abundance of bacterial families among samples. CP: P. pinea control group; FP: P. pinea inoculated group; CR: P. radiata control group; FR: P. radiata inoculated group.
Differential abundance analysis of bacterial families between P. radiata and P. pinea for control treatments. Samples farther to the right are prominently abundant in P. pinea, whereas samples farther to the left are prominently abundant in P. radiata. For a clearer analysis of the abundances, only taxa seen more than once in each treatment were filtered.
For an overview of the top 40 most abundant OTUs considering all samples, a heatmap was built including the bacterial genera designation (Figure 10). Concerning OTUs showing statistically different abundances, their logarithm (base 2) based differences are displayed in Figure 11 for each pine species (control samples). Among these, OTU_34 (affiliated with genus Rhizomicrobium), OTU_18 (Devosia), OTU_20 (Pseudolabrys), OTU_30 (family: Rhodobacteraceae), OTU_108 (Phaselicystis) and OTU_59 (Steroidobacter) were significantly more abundant in P. radiata, with OTU_63 (family: Oxalobacter), OTU_17 (Paucibacter) and OTU_104 (order: Enterobacterales) being exclusive to this pine species. On the other hand, OTU_4 (Paraburkholderia), OTU_5 (Acidocella), OTU_25 (Roseiarcus), OTU_3 (Mucigilinibacter) and OTU_11 (Acidipila) were significantly more abundant in P. pinea, with OTU_110 (Nocardioides) being exclusive to this pine species. No significant differences in terms of dominant OTUs abundance were found between control and inoculated samples of each pine species.
Heatmap of the top 40 most abundant OTU in all samples. Darker blue indicates higher abundance, whereas white indicates absence. CP: P. pinea control group; FP: P. pinea inoculated group; CR: P. radiata control group; FR: P. radiata inoculated group. Significant differences between CP and CR are highlighted with asterisk (Kruskall–Wallis test, P value < 0.05). Genus affiliation (or previous available taxa is shown) is presented in the figure.
Differential abundance analysis of OTUs (genus-level) between P. radiata and P. pinea for control treatments. Samples farther to the right are prominently abundant in P. pinea, whereas samples farther to the left are prominently abundant in P. radiata. Samples denoted with an asterisk are exclusive to the respective treatment. When assignment to the genus was not possible, previous taxonomic level was used. Only significantly different OTUs are presented.
Discussion
Recently, pine species with different responses to PPC have been deeply investigated, adding new insights into the physiological and molecular mechanisms behind PPC susceptible and resistant host phenotypes (e.g., Amaral et al. 2021, Zamora-Ballesteros et al. 2021). Yet, although plant health is known to be intimately influenced by the rhizobiome structure (Bonito et al. 2014), this community has been understudied in the context of this emergent disease. We provide here the first study analyzing the rhizobiome of Pinus—F. circinatum pathosystem by exploring two pine species with contrasting responses to PPC. Figure 4b showed a clear separation between P. pinea and P. radiata rhizobiomes, regardless F. circinatum inoculation. These rhizobiomes showed a low similarity (<40%) and their spatial distribution on the PCoA is mainly according to the PCoA1, which explains a large proportion of the data variation (61.8%), being therefore likely related to pine species constitutive differences, rather than to the effect of F. circinatum inoculation (Figure 4). Moreover, high levels of similarity (>60%), as seen on the dendrogram (Figure 4a), were found among samples from each pine species, regardless fungus inoculation.
Based on these results, this section discusses three main points: the impact of pathogen inoculation in plant performance and in each rhizobiome, the species-specific physiological traits that may shape each rhizobiome assemblage, and the constitutive nature of the rhizobiome and its potential relation with PPC.
The slight changes in the rhizobiome of the susceptible host upon F. circinatum inoculation are unlikely to be due to photosynthesis impairment
Photosynthesis suffered stomatal and non-stomatal limitations in P. radiata plants upon inoculation with F. circinatum, which is aligned with previous studies (Amaral et al. 2019a, 2019b, Amaral et al. 2021, Cerqueira et al. 2017). Changes in host metabolism, particularly at the photosynthetic level, were associated to a source-to-sink transition (Amaral et al. 2019a) and an increased susceptibility to photoinhibition (Amaral et al. 2021) in P. radiata plants inoculated with F. circinatum. This may affect rhizobiome–host interactions as plants provide up to 40% of complex carbons produced by photosynthesis via roots to nourish the rhizobiome (Whipps et al. 1990). In fact, photosynthetic products deposited belowground can shape the plants’ rhizobiome by changing physicochemical characteristics of soil-root microenvironment (Hartmann et al. 2009, Yarwood et al. 2009).
Despite the high similarity observed in the PCoA between P. radiata control and inoculated samples, the 91 exclusive OTUs found in inoculated plants, combined with lower richness and diversity (although non-significant), suggest that F. circinatum inoculation induces slight changes in the P. radiata rhizobiome. Regarding P. pinea, 72 exclusive OTUs and non-significant lower levels of richness and diversity were also observed in inoculated plants, also suggesting a slight effect in the P. pinea rhizobiome yet any photosynthetic changes were observed. Our findings may indicate that the infection progressed too quickly to induce significant changes on the rizhobiome community after inoculation and that the photosynthetic responses were not responsible for the changes in P. radiata rhizobiome upon F. circinatum inoculation. The changes observed are probably due to other pre-existing constitutive differences between both Pinus spp. However, rhizobiome alterations in response to leaf or stem infectious diseases have been previously reported. For instance, Berendsen et al. (2018) reported that A. thaliana alters the bacterial composition of the rhizosphere in plants challenged by downy mildew, increasing protection against this pathogen in a second population of plants growing in the same soil. The slight changes observed in our study upon fungus inoculation may suggest a similar effect that should be further explored in future studies, which may consider other pine species and other isolates of P. circinatum. Given that these changes are subtle and not easily detected, it is also important to include a larger number of replicates to account for inra-specific variability.
Constitutive needle phenolic composition may confer PPC resistance and potentially shape pine rhizobiome
Several studies point out that plants usually accumulate antioxidant compounds such as phenols in response to pathogenic attack (Beimen et al. 1992, Gayoso et al. 2004, Petkovšek et al. 2009). However, no significant differences were found between inoculated and non-inoculated plants for both pine species regarding the antioxidant parameters tested. This might be related to the type of plant tissue analyzed, given that phenols accumulation after fungal infection is mainly reported nearby the fungus entrance point (Gayoso et al. 2004), which is corroborated by the increased antioxidant activity reported in stems of Pinus pinaster plants inoculated with F. circinatum (Vivas et al. 2014). Moreover, Erbilgin et al. (2017) found constitutive differences (concentration and profile) between the phloem and the needles of Pinus banksiana and P. contorta, as well as a decrease of P. contorta phenolics in the phloem but not in the needles after fungal inoculations. Nonetheless, phenolics analysis after nematode inoculation in P. pinea and P. radiata stems, revealed no change in total phenolic compounds (Nunes da Silva et al. 2015).
Regarding non-inoculated plants, higher levels of constitutive phenolic compounds (total phenolics, ortho-phenols, and flavonoids) were determined for P. pinea in comparison with P. radiata. Hypocotyls of Pinus sylvestris genotypes resistant to Fusarium also exhibited higher phenolic compounds levels than susceptible ones (Shein et al. 2003a). Also, Ganthaler et al. (2017) reported that a combination of constitutive and inducible accumulation of foliar phenolic compounds was associated with lower susceptibility of individual spruce trees to C. rhododendri. Phenolic compounds are known to have antioxidant properties and improve mechanical barriers (Witzell and Martín 2008). These are involved in lignin production which can serve as a physical deterrent against fungal progression (Shein et al. 2003b). Furthermore, the potential fungicide effect of phenolic compounds to limit hyphal growth and prevent the development of infection symptoms has been highlighted in conifers, including pine trees (Evensen et al. 2000, Mohareb et al. 2017). In P. radiata, a gene implied in flavonoids biosynthesis and in flavonoid phytoalexins production (important plant defense metabolites) was found to be overexpressed after F. circinatum infection (Donoso et al. 2015); with more resistant genotypes tending to activate this response faster at the infection site. Altogether, we hypothesize that constitutive phenolics levels may be determining in conferring PPC resistance.
Moreover, it has been proven that phenolic compounds in leaves can affect the rhizosphere and its bacterial community (Hunter et al. 2010), including different conifers (Ushio et al. 2008). Li et al. (2018) studied the correlations between six spruce species’ microbiome and plant phenotypic traits, suggesting that the plant’s genotype shapes its microbiome by driving the development of plant phenotypes. Root exudates are one of the main phenotypic drivers for rhizobiome diversity, affecting the rhizosphere chemical composition and nutrient cycling (Olanrewaju et al. 2019) and are related with several forest tree constitutive characteristics (Lambais et al. 2014, Prescott and Grayston 2013). The different concentration of phenolic compounds between P. radiata and P. pinea, also reported by (Nunes da Silva et al. 2015), represents a constitutive difference that may affect their rhizobiomes, a research topic that deserves more attention.
Constitutive differences on P. pinea and P. radiata rhizobiome may contribute to its defense vs. growth strategies
The rhizobiome of each plant varies according to the age and developmental stage (Chaparro et al. 2014) due to the production of phytohormones characteristic of each developmental stage, which are exuded to the rhizosphere. In younger seedling stages, Acidobacteria is a very predominant phylum, which is in accordance with our results. Furthermore, Acidobacteria and Actinobacteria are generally predominant in the rhizobiome of pine species (Lottmann et al. 2010, Naz et al. 2018), which was also observed in our study.
In the present work, some bacterial families found prominently in the P. pinea rhizobiome included Nocardioidaceae, Burkholderiaceae and Xanthomonadaceae. The role of Nocardioidaceae and Xanthomonadaceae as important players in disease suppressive soils has been addressed in several studies (Rosenzweig et al. 2012). For instance, Nocardioides (OTU_110) have been found in disease suppressive soils used against the fungal root pathogen Rhizoctonia solani (Cordovez et al. 2015). Disease suppressive soils are designed to support the production of antimicrobial compounds that selectively inhibit pathogen growth, particularly soil-borne pathogens (Mendes et al. 2011). However, its potential to enhance host’s immune system by activating induced systemic resistance, also contributes to disease suppressiveness and potentialize its application to control pathogens with other origins such as those causing foliar diseases (Berendsen et al. 2018). This is a field of research in its infancy for forest species.
Furthermore, some genera of Xanthomonadaceae found within our control P. pinea samples (e.g., Lysobacter) can express anti-fungal compounds (de Bruijn et al. 2015). Regarding Burkholderiaceae, this family is commonly found in soils (Spain et al. 2009), having a widespread capability of producing ACC (1-aminocyclopropane-1-carboxylate) deaminase (Onofre-Lemus et al. 2009), among other plant growth promoting (PGP) traits. In addition to these bacterial families, the Paraburkholderia genus (OTU_4), significantly more abundant in P. pinea than in P. radiata (control samples), has been found endophytically in P. contorta (Puri et al. 2018) being diazotrophic (PGP trait). Furthermore, some members of this genus have prominent antifungal activity either directly (Ravi et al. 2020) or by boosting plants’ phenolic metabolism (Miotto-Vilanova et al. 2019). This is in accordance with our data showing greater Paraburkholderia abundance and higher phenolics concentrations in P. pinea than in P. radiata.
On the other hand, the bacterial families Micropepsaceae and Rhizobiaceae were more abundant in the rhizobiome of P. radiata than in P. pinea. Rhizobiaceae (OTU_59) is known for its PGP characteristics (Kannaiyan et al. 2004), specifically related with nitrogen fixation (Grobelak et al. 2015) and there are many studies indicating immediate plant benefits, such as the increase of roots length and fruit yield (in non-legumes) (García-Fraile et al. 2012). However, these Rhizobiaceae’s PGP traits have not been directly related with fungal interactions. Members of the family Micropepsaceae, e.g., Rhizomicrobium (OTU_34), are also associated to nitrogen fixation (Ueki et al. 2010). Other OTUs significantly more abundant in the P. radiata rhizobiome in comparison with P. pinea were: Oxalobacteraceae (OUT_63), a family commonly found in soils (Baldani et al. 2014) and associated with arbuscular mycorrhizal fungi, which have been reported in some Pinus species (Scheublin et al. 2010); Enterobacteriales (OTU_104), which also includes bacteria with PGP traits linked to abiotic stress relief (Acuña et al. 2019, Barra et al. 2016); and Paucibacter (OTU_17), which was mostly found in aquatic environments and for which data concerning plant interactions is lacking (Pheng et al. 2017).
The rhizobiomes’ composition of each pine species seems to be in line with its differences in terms of resistance to PPC (Martín-García et al. 2019) and growth performance, given that P. radiata is a fast-growing coniferous tree (Mead 2013). Nonetheless, additional studies are needed to clarify the role of the rhizobiome in forest species under pathogen attack, such as analysis focusing on other parts of the plant (e.g., the stem), which may provide complementary information regarding factors that may contribute to the mechanisms behind resistant phenotypes that could be further explored in plant protection strategies.
Conclusion
The full extent to which the species of Pinus influences its rhizobiome can be a strong point of interest for breeders, as it represents a promising tool for selection and enhancement of new plant traits. In the present work, we have assessed the rhizobiome of seedlings of two pine species with contrasting levels of resistance to F. circinatum. We concluded that pine species differentially shaped the rhizobiome either toward a higher abundance of bacterial taxa that can potentially provide complementary protection against fungal infections (PPC resistant P. pinea) or a higher abundance of bacteria that enhance plant vigor and growth (PPC susceptible and fast-growing P. radiata). This work contributes to discover new aspects of the rhizobiome of plants produced under nursery conditions, which is still understudied yet essential to stimulate new approaches to enhance plant resistance against forest diseases, including PPC which is known to have a great impact at the nursery phase.
- DPPH
2,2-diphenyl-1-picrylhydrazyl
- PPC
Pine pitch canker
- PGP
Plant growth promoting
- PPFD
Photosynthetic photon flux density
- OTU
Operational taxonomic unit
Data availability statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Conflict of interest
None declared.
Author contributions
G.P. and I.H. conceptualized and designed the experiment. F.L. and J.A. performed the plant upkeep and plant related assays/trials. P.M. and F.L. performed the photosynthetic parameters and antioxidant assays. F.L. performed the rhizobiome analysis. F.L., G.P. and I.H. analyzed the data. F.L. drafted the manuscript. G.P., I.H., J.A., and P.M. revised the manuscript.
Funding
Fundação para a Ciência e Tecnologia (FCT) provided funding through project ‘URGENTpine: UnRaveling hostpathoGEn iNteracTions in pine pitch canker disease’ (PTDC/AGRFOR/2768/2014; POCI-01-0145-FEDER016785), and granted PhD fellowships to J.A. (SFRH/BD/120967/2016), P.M. (SFRH/BD/143879/2019) and F.L. (2021.06400.BD). FCT also funded CESAM (UIDP/50017/2020 + UIDB/50017/2020), through national funds.
