Soil cadmium stress affects the phyllosphere microbiome and associated pathogen resistance differently in male and female poplars

Abstract

Microorganisms associated with the phyllosphere play a crucial role in protecting plants from diseases, and their composition and diversity are strongly influenced by heavy metal contaminants. Dioecious plants are known to exhibit sexual dimorphism in metal accumulation and tolerance between male and female individuals. Hence, in this study we used male and female full-siblings of Populus deltoides to investigate whether the two sexes present differences in their phyllosphere microbiome structures and in their associated resistance to the leaf pathogenic fungus Pestalotiopsis microspora after exposure to excess soil cadmium (Cd). We found that Cd-treated male plants grew better and accumulated more leaf Cd than females. Cd stress reduced the lesion areas on leaves of both sexes after pathogen infection, but male plants exhibited better resistance than females. More importantly, Cd exposure differentially altered the structure and function of the phyllosphere microbiomes between the male and female plants, with more abundant ecologically beneficial microbes and decreased pathogenic fungal taxa harbored by male plants. In vitro toxicity tests suggested that the sexual difference in pathogen resistance could be attribute to both direct Cd toxicity and indirect shifts in the phyllosphere microbiome. This study provides new information relevant for understanding the underlying mechanisms of the effects of heavy metals involved in plant–pathogen interactions.

Introduction

As heavy metals are biologically toxic to herbivorous insects and pathogenic microorganisms, the ‘elemental defense hypothesis’ (EDH) suggests that plants can utilize metal accumulation as a novel defensive strategy to protect them from herbivory and pathogen attack (Boyd et al., 1994). To date, the EDH has mainly been tested with different types of herbivores whilst only a few studies have focused on pathogen resistance (Fones et al., 2010; Hörger et al., 2013), and the evidence collected so far has not been consistent enough to support the EDH with regard to pathogens. For example, some studies have reported that plants exposed to metal stress do indeed show increased resistance to pathogen infection (Boyd et al., 1994; Ghaderian and Baker, 2000; Hanson et al., 2003), whereas other studies have found an increased susceptibility to infection in metal-accumulator plants (Davis et al., 2001). Boyd (2012) has argued that defensive enhancement might be species-specific at both the plant and pathogen levels, but the underlying mechanism by which heavy metals might be involved in plant–pathogen interactions still remains unclear.

The phyllosphere, which refers to the aerial surfaces of the leaves of plants, harbors a diverse variety of microorganisms that are dominated by bacteria and fungi (Vacher et al., 2016). The microorganisms associated with the phyllosphere interact with their host plants, influencing plant fitness and contributing to several ecosystem functions (Stone et al., 2018; Gong and Xin, 2021). For example, the phyllosphere microbes can act as mutualists to increase plant growth and health (Liu et al., 2019, 2020b), to improve plant nutrient acquisition (Pankievicz et al., 2015; Chhabra and Dowling, 2017), to enhance plant tolerances to environmental stressors (Curá et al., 2017; Tiwari and Lata, 2018; Liu et al., 2020a), and to detoxify environmental pollutants (Weyens et al., 2015; Wei et al., 2017; Hussain et al., 2018). More importantly, microorganisms present in the phyllosphere are also capable of improving plant resistance to diseases (Ritpitakphong et al., 2016; Berg and Koskella, 2018; Vannier et al., 2019).

Due to the fact that the phyllosphere is an open environment, its associated microbiome is reported to be strongly affected by several environmental factors, such as ultraviolet radiation, water availability, and nutrient limitation (Vorholt, 2012; Gomes et al., 2018; Leveau, 2019). However, the diversity of phyllosphere microbial communities has rarely been examined under conditions of heavy metal stress, although there is limited evidence to show that the structure and function of the phyllosphere microbiome of some herbaceous species are both strongly influenced by heavy metal contaminants (Dobrone Toth et al., 2009; Jia et al., 2018). Hence, whether shifts that occur in the structure and function of the microbiome associated with the phyllosphere in response to heavy metal stress will contribute to plant resistance against pathogen infection still needs further investigation.

Dioecious plant species exhibit sex-specific differences in terms of plant growth, defense, and tolerance under various abiotic stresses due to the trade-offs in resource allocation between vegetative growth and reproduction (Stevens and Esser, 2009; Juvany and Munné-Bosch, 2015). In general, female individuals are considered to invest more energy into reproduction compared to their conspecific males, and this results in lower growth performance and stress tolerance (Cornelissen and Stiling, 2005). For example, male poplar plants are found to have better growth performance, and increased metal tolerance and herbivore resistance than their conspecific females when grown in soil contaminated with heavy metals (Chen et al., 2016; Lin et al., 2022). Therefore, combining the assumptions of the EDH with our knowledge of the sex-specific responses of dioecious plants to heavy metal stress, we hypothesize that the sexual dimorphism in heavy metal tolerance and accumulation of a dioecious plant will result in differences in the structure and function of the phyllosphere microbiome between male and female conspecifics, and this in turn will contribute to sexual differences in plant resistance against leaf pathogen attack.

To test this hypothesis, in this study we employed full siblings of Populus deltoides as a study model. As a typical dioecious species, it has the advantage of fast growth ability and high adaptation to environmental stresses (Yang et al., 2011). Consequently, this species has been introduced into China from North America and used as phytoremediation material at sites polluted with heavy metals (Gao et al., 2009). This species has been recorded to exhibit typical sexual dimorphism in metal accumulation and tolerance in response to heavy metal stress (Chen et al., 2016; Xu et al., 2016; Hao et al., 2020). We first investigated the differences in structure and function of the phyllosphere fungal and bacterial communities between male and female conspecifics after exposure to Cd using a method based on an Illumina 16S rRNA/ITS1 MiSeq System. In addition, we compared the differences in leaf lesion area between the two sexes with or without Cd exposure after artificial inoculation with a leaf pathogenic fungus. We also further examined the inhibition effects of Cd and of the cultivable phyllosphere microorganisms on the growth of the fungus through in vitro toxicity tests. To our knowledge, this is the first study to disentangle the underlying mechanism of the involvement of heavy metals in a plant–microbe interaction. Understanding the ecological relationships between plant dioecy and pathogen resistance in the light of EDH provides insights for predicting the disease incidence of dioecious plant populations naturally distributed in heavy metal-contaminated ranges.

Materials and methods

Plant materials

In this study, five male and five female Populus deltoides parental plants were collected from the same population growing in mineralized soil near a mining field in Southwestern China (Meigu County, 28°04ʹN, 102°11ʹE, mean altitude 2450 m). The mean soil Cd²⁺ concentration at the collecting site was 60 mg kg^–1 dry soil. For each pair of parental plants, one male and one female F₁ full-sibling was derived via artificial pollination, making a total of five male and five female F1 plants that were used for treatment.

Leaf pathogen

Pestalotiopsis microspora was selected as a representative leaf pathogenic fungus in this study. Previous studies have shown that this species can successfully infect the leaves of poplar plants and other woody species, resulting in leaf spot disease (Keith et al., 2006; Jeon et al., 2010; Shen et al., 2014; Lin et al., 2020). The strain used in this experiment was purchased from the China Forestry Culture Collection Center with the ID number of IDN88154.

Plant cultivation

Annual shoot cuttings of 10 cm height from the F₁ females and males were planted individually into plastic pots (25 × 25 × 30 cm) filled with 7 kg clean soil, which was collected from a non-metal-polluted farmland and had the following characteristics: pH 6.95, organic matter 7.12%, total N 0.53 g kg^–1 dry soil, total phosphorus 0.87 g kg^–1 dry soil, and total Cd 0.15 mg kg^–1 dry soil. Based on the mean soil Cd concentration recorded at the collecting site where the parental plants were growing, the clean soil of each pot was mixed with CdCl₂·2.5H₂O to reach a concentration of 60 mg Cd²⁺ kg^–1 dry soil. There were four treatments, as follows: control/male (CKM), control/female (CKF), Cd/male (CDM), and Cd/female (CDF). For each treatment, there five individual plants cultured as biological replicates. The plants were grown in a greenhouse with natural light for 90 d at Chengdu city (Southwestern China, 30°42ʹN, 103°51ʹE). During the growing period, the mean daytime temperature was 24.8 °C and mean night-time temperature was 14.3 °C, while the mean relative humidity was 52.5%.

Measurement of host plant traits

After the 90 d of Cd treatment, the height and basal stem diameter of each plant were determined. A portable leaf chlorophyll meter (SPAD-502, Minlota) was used to measure the leaf chlorophyll content of the sixth and seventh fully expanded leaves of each plant. The fresh mass of each of the same leaves was then recorded and the dry mass was measured after 5 d of oven-drying at 60 °C. The leaf water contents were calculated and no significant difference were observed among the four treatments (one-way ANOVA, F_1,3=0.204, P>0.05). After weighing, the dried leaf samples were ground into fine powder and passed through a 100-mesh sieve for subsequent measurement of leaf Cd concentration using an inductively coupled plasma atomic emission spectrometer (Avio 550 Max ICP- AES, PerkinElmer).

DNA extraction of phyllosphere epiphytic and endophytic microorganisms

Epiphytic microorganisms were washed from the surface of leaf samples as described by Gomes et al. (2018) with a few modifications. The third to fifth fully expanded leaves were collected from each plant and put into a conical flask with 100 ml of 0.1 M potassium phosphate buffer, pH 8. The solution was first sonicated in a water bath for 15 min at 47 kHz and then shaken for 30 min at room temperature. This process was repeated two more times, and then the supernatant was transferred into a new tube and centrifuged at 10 000 g for 20 min. The supernatant was removed and the pellet was resuspended with 5 ml cetyltrimethylammonium bromide (CTAB) extraction buffer (pH 8.0) pre-heated to 65 °C, and then homogenized at 6.0 m s^–1 for 30 s with an EZNA plant DNA Kit (Omega Biotek, Norcross, GA, USA) for DNA extraction according to the manufacturer’s protocol. For the extraction of endophytic microorganisms, the surfaces of the leaf samples were sterilized by consecutive immersion in 75% ethanol for 1 min, in 3.25% sodium hypochlorite for 3 min, and in 75% ethanol for 30 s, followed by rinsing three times in sterile distilled water. After that, the leaves were freeze-dried in liquid nitrogen and homogenized using a sterilized mortar and pestle before the extraction of the DNA of the endophytic microorganisms. The samples were transferred to tubes with 5 ml CTAB extraction buffer pre-heated to 65 °C and then homogenized at 6.0 m s^–1 for 30 s with an EZNA plant DNA Kit. The quantity and quality of extracted DNA were determined using a NanoDrop ND-1000 UV-Vis Spectrophotometer (NanoDrop Technologies).

Illumina MiSeq sequencing

The sequencing of the phyllosphere fungal and bacterial communities of the P. deltoides plants was conducted using an Illumina MiSeq platform by Shanghai Majorbio Bio-pharm Technology Co., Ltd (Shanghai, China). The primers ITS1F (5´-CTTGGTCATTTAGAGGAAGTAA-3´) and ITS2R (5´-GCTGCGTTCTTCATCGATGC-3´) were used for the amplification of the fungal ITS1 region (Monard et al., 2013) while the primers 799F (5´-AACMGGATTAGATACCCKG-3´) and 1193R (5´-ACGTCATCCCCACCTTCC-3´) were used for the amplification of the bacterial 16S rRNA region (Thijs et al., 2017).

Bioinformatics analysis

Low-quality sequences including those with an average quality score <20, length <250 bp, containing ambiguous bases, or without a valid primer sequence or barcode sequence were eliminated from the raw sequences through the use of the DADA2 algorithm as a plugin for QIIME2 (v. 2017.12, https://qiime2.org/) (Callahan et al., 2016). The remaining sequences contained only high-quality reads. The taxonomic identification of amplicon sequence variants (with 99% of similarity) for the bacterial sequences was conducted using the VSEARCH consensus taxonomy classifier implemented in QIIME2 and the SILVA 16S rRNA database (https://www.arb-silva.de/) according to Liu et al. (2021), whilst the ITS1 region of fungal sequences was extracted using the fungal ITSx software package and identified by comparison with the UNITE database (v. 18.11.2018) in the MOTHUR software (v. 1.31.2) as described by Yao et al. (2019). The fungal and bacterial sequence data have been deposited in the NCBI database under the SRA accessions of PRJNA859240 and PRJNA859001, respectively.

Leaf pathogenic fungus infection and in vitro Cd toxicity tests

Artificial inoculation of the leaf pathogenic fungus P. microspora was performed as follows. After 90 d growth in the greenhouse, all the leaves of the five individual replicate plants from each treatment were sprayed with 200 ml spore suspension of P. microspora with a concentration of 1 × 10⁶ spores ml^–1. The plants were then grown in a separated greenhouse for 7 d before the total leaf lesion area of each plant was quantified by scanning and use of the Adobe Photoshop CS6 software.

Since poplar plants contain several defensive chemicals that might be involved in pathogen resistance (Yevtushenko and Misra, 2019), the growth of fungal mycelia of P. microspora on PDA plates amended with a series of Cd concentrations (0, 20, 40, 60, 80 and 100 mg l^–1 Cd²⁺) was determined in order to test the direct toxicity of Cd. An agar cutting (diameter 1 cm) with a P. microspora mycelium was collected from a Cd-free PDA plate and then placed in the center of each Cd-treated PDA plate (diameter 9 cm; three replicate plates). The diameter of the mycelium of each plate was then measured after 7 d of growth in an incubator at 25 °C with a relative humidity of 60%.

In vitro toxicity test of phyllosphere cultivable microorganisms on the growth of the leaf pathogenic fungus

To determine whether the phyllosphere microorganisms could inhibit the growth of the leaf pathogenic fungus P. microspora, an in vitro toxicity test was performed as described by Ritpitakphong et al. (2016), with modifications. The epiphytic and endophytic microorganisms were extracted from the plants in each treatment as described for the DNA extraction above. The extraction was diluted in ¼-strength PDB (13.5 μl) and mixed with 1.5 μl of P. microspore spore suspension to reach a final concentration of 1 × 10⁶ spores ml^–1. The mixture was then placed on a Petri dish, which was put in a humid box and incubated for 3 d in an incubator at 25 °C. The mean diameter of P. microspore on each dish was then measured (n=5 replicates for each treatment).

Statistical analysis

The differences in host-plant traits, leaf lesion areas, and Alpha diversity (Shannon Index and Chao1 Index) of the phyllosphere microorganisms among the four treatments were analysed using Generalized Linear Models followed by Wald pair-wise comparisons using the IBM SPSS 22.0 package (Venables and Ripley, 2002), with plant sex, Cd treatment, and their interaction as independent factors.

The structure and function of the phyllosphere fungal and bacterial communities were analysed using the free online platform of Majorbio (https://cloud.majorbio.com/; Ren et al., 2022). Blast results were sorted according to highest identity scores and lowest E-value, and the operational taxonomic units (OTUs) of both the phyllosphere fungi and bacteria were determined based on 97% sequence similarity. The Alpha diversity of the phyllosphere microorganisms was measured using the Shannon and Chao1 indices while the Beta diversity was analysed by principal co-ordinates analysis (PCoA) based on Bray–Curtis distances. Permutational multivariate analysis of variance (PERMANOVA) was carried out to measure the effect size and significance of each factor on the Beta diversity of the phyllosphere microbial communities. A one-way analysis of similarity (ANOSIM) was utilized to analyse the significant differences among groupings of microbial samples from each treatment obtained in the PCoA using Bray–Curtis distance matrices. Redundancy analysis was applied to determine the relationships between host-plant traits and the taxonomic distribution of the microbiome. A linear discriminant analysis coupled with effect size analysis was applied to analyse the relative abundance of fungal and bacterial taxa in the different treatments. The assignment of functional information to fungal and bacterial OTUs was performed using FUNGuild (http://www.funguild.org/) and PICRUSt2 (https://github.com/picrust/picrust2), respectively. Heatmap clustering was performed using MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/MetaboAnalyst/home.xhtml).

Results

Effects of plant sex and soil Cd treatment on leaf Cd concentration and plant growth

Both the soil Cd treatment and plant sex as well as their interaction had significant effects on the leaf Cd concentration of P. deltoides (Fig. 1A). More specifically, both males and females from the control treatment only contained trace amounts of Cd in their leaves, whereas male plants accumulated 1.4 times more Cd than their female conspecifics after soil Cd treatment. In addition, Cd treatment and plant sex alone but not their interaction significantly affected plant height and basal stem diameter (Fig. 1B, C): on average, male plants were 11.0% taller and had a 14.7% wider basal stem than the female plants without Cd stress. However, Cd exposure drastically reduced the growth performance of both the sexes, resulting in 16.2% and 23.6% reductions in plant height and 20.0% and 20.2% decreases in basal stem diameter in male and female plants, respectively, compared to the corresponding controls. In addition, Cd stress significantly reduced the leaf chlorophyll content of both sexes, with the leaves of the female plants containing 29.4% less chlorophyll than their male conspecifics (Fig. 1D).

Effects of plant sex and soil Cd treatment on leaf lesion areas caused by P. microspora infection

At 7 d after artificial infection of the leaves with the pathogenic fungus Pestalotiopsis microspora, the soil Cd treatment and plant sex both had significant effects on the leaf lesion area of P. deltoides, and no interaction was observed (Fig. 1E; Supplementary Fig. S1). More specifically, without Cd exposure the female plants exhibited 1.2 times higher percentage of leaf lesion area than male plants. In contrast, Cd treatment significantly decreased the percentage of leaf lesion area of both sexes, resulting in 41.6% and 35.0% reductions in lesion area of male and female plants, respectively, compared to the corresponding controls.

Effects of Cd-amended growth medium and cultivable microorganisms on the growth of the leaf pathogenic fungus

The growth of P. microspora on PDA plates amended with a series of Cd concentrations indicated that the mycelium growth was significantly negatively correlated with the Cd concentration (Fig. 1F). The Cd concentration required for a 50% reduction in mycelium diameter (EC₅₀) was ~51.0 mg Cd²⁺ l^–1. In addition, the soil Cd treatment and plant sex but not their interaction significantly affected the growth of P. microspora in PDB culture mixed with the phyllospheric microorganisms extracted from the P. deltoides plants in the different treatments (Fig. 1G, H). More specifically, without soil Cd stress the growth of mycelium of P. microspora in the culture mixed with epiphytic microorganisms washed from the leaf surface of male plants was 1.3 times larger than that from female plants (Fig. 1G). Soil Cd exposure decreased the growth of P. microspora in microbial culture from the phyllosphere of male and female plants by 2.6 and 2.0 times, respectively, resulting in better growth in cultures from female plants compared to males. For the cultivable endophytic microorganisms, the presence of soil Cd stress reduced the growth of P. microspora growth with microbes from the phyllosphere of male and female plants by 1.9 and 1.5 times, respectively, whereas no significant difference was observed between the two sexes under control conditions (Fig. 1H).

General taxonomic distribution of phyllosphere microbial communities

After quality filtering, denoising, merging, and chimera screening, a total of 1 746 373 and 642 255 high-quality fungal and bacterial sequence reads with average lengths of 244 bp and 374 bp, respectively, were obtained across all samples. A total of 1468 fungal OTUs were isolated from the phyllosphere samples of P. deltoides plants subjected to the four treatments, and epiphytes had a greater diversity (1881 OTUs) compared to endophytes (559 OTUs). Since the bacterial 16S rDNA primers could also target the chloroplast and mitochondrial DNA, we removed the sequences affiliated to them and this resulted in 1840 bacterial OTUs. In contrast to the phyllosphere fungi, endophytic bacteria had greater diversity (1822 OTUs) compared to epiphytic bacteria (249 OTUs). The taxonomically classified OTUs of the fungal samples were associated with eight phyla, 36 classes, 94 orders, 251 families, 538 genera, and 874 species, while those of the bacterial samples were associated with 34 phyla, 100 classes, 227 orders, 386 families, 797 genera, and 1275 species. Venn diagrams indicated that there were 29 OTUs common to all fungal samples while only eight OTUs were common to all bacterial samples (Supplementary Fig. S2). In addition, across the four treatments there were 48 and 306 OTUs common to the endophytic and epiphytic fungal communities, respectively, while 185 and 13 OTUs were common to endophytic and epiphytic bacterial communities, respectively.

Effects of plant sex and soil Cd treatment on the diversity and composition of phyllosphere microbial communities

The Shannon and Chao1 indices were used to analyse the Alpha diversity of the phyllosphere fungal and bacterial communities. The results showed that Cd treatment, plant sex, and their interaction all significantly affected the Shannon indices of the fungal and bacterial communities for epiphytes but not for endophytes (Fig. 2). More specifically, male control plants had a higher Shannon index of epiphytic fungi than the other three treatments, while the female plants under Cd stress showed the highest Shannon index of epiphytic bacteria (Fig. 2A, B). With regard to the Chao1 index, there were no significant effects of Cd treatment, plant sex, or their interaction for either the endophytic or epiphytic fungi (Fig. 2C). In contrast, Cd exposure slightly reduced the Chao1 index of endophytic bacteria in the phyllosphere of both the male and female plants compared to their corresponding controls; however, only the difference between male controls and Cd-stressed females was significant (Fig. 2D). For the epiphytic bacteria, female plants exposed to Cd stress showed the highest Chao1 index whilst the other three treatments had no significant differences.

The soil Cd treatment and microbial location significantly affected the fungal communities in terms of their composition (Supplementary Table S1), while samples from different treatments could clearly be distinguished from each other by their most common classes (Fig. 3A). The most abundant classes in the endophytic fungal communities from the phyllosphere of control plants were Dothideomycetes (55.3%), Sordariomycetes (12.9%), Leotiomycetes (12.3%), Eurotiomycetes (4.5%), Agaricomycetes (4.5%), Tremellomycetes (3.5%), and Saccharomycetes (3%). Soil exposure to Cd reduced the relative abundance of the phylum Ascomycota by 3.5% and 14.5% in the leaves of male and female P. deltoides, respectively, while the abundance of the phylum Basidiomycota was increased by 5.2% and 15.0% in males and females, respectively. The most abundant classes of the epiphytic fungal communities from the phyllosphere of control plants were Dothideomycetes (85.0%), Leotiomycetes (5.0%), and Sordariomycetes (2.0%). Both plant sexes from the control and the Cd treatment all showed similar abundance of the Ascomycota and Basidiomycota phyla in their epiphytic fungal communities. In addition, the ANOSIM suggested that the composition of the bacterial communities was significantly affected by both Cd and microbe location (Supplementary Table S1), and the bacterial taxa from the different treatments differed from each other in terms of their most common classes (Fig. 3B). More specifically, the most abundant classes of the endophytic bacterial communities from control plants were Gammaproteobacteria (66.2%), Alphaproteobacteria (7.8%), Bacilli (7.7%), Actinobacteria (5.6%), and Bacteroidia (2.4%). Cd stress reduced the relative abundance of the Proteobacteria phylum in leaves of male and female P. deltoides by 22.0% and 33.34%, respectively, while the abundance of the Firmicutes phylum was increased by 18.6% and 35.5%, respectively. For the epiphytic bacterial communities, Cd exposure only had large effects in reducing the abundance of the Proteobacteria phylum (by 69.4%) and in increasing the abundance of the Firmicutes phylum (by 68.8%) in leaves of female P. deltoides, whereas both these two phyla showed similar abundances in male plants from the control and Cd-stress treatments.

Principal co-ordinates analysis (PCoA) revealed the Beta diversity of the phyllospheric microbial communities. With regard to the fungal communities, the endophytic and epiphytic samples exhibited a clear separation (Fig. 3C). More specifically, the endophytic samples from control and Cd-treated plants were separated from each other while the epiphytic samples were clustered together, except for the Cd-treated male plants. A PERMANOVA further indicated that the fungal location and Cd treatment explained 27% (P=0.001) and 8% (P=0.005), respectively, of the variation in the phyllospheric fungal communities (Supplementary Table S2). For the bacterial communities, PCoA indicated a clear separation in the endophytic samples between the control and Cd-treated plants, whilst the epiphytes from female Cd-treated plants were separated from the clustered epiphytic samples of the other treatments (Fig. 3D). A PERMANOVA further confirmed that Cd treatment and bacterial location explained 31% (P=0.001) and 39% (P=0.001), respectively, of the variation in the phyllospheric bacterial communities (Supplementary Table S2).

Effects of plant sex and soil Cd treatment on the relative abundances of key taxa in the phyllosphere

Heatmap clustering was performed to examine the influence of Cd treatment on the differences in abundance of fungal and bacterial taxa at the genus level. For the epiphytic fungi, male plants from controls exhibited a relatively higher abundance of the genera Periconia and Candida than corresponding female plants, whilst the soil Cd treatment decreased the relative abundance of Phyllactinia in males but not in females (Fig. 4A). The taxonomic abundances of the endophytic fungi differedamong the different treatments (Fig. 4B). For example, without Cd stress the genera Cercospora, Cyberlindnera, and Apiotrichum were more highly enriched in female plants than males while Sterigmatomyces, Penicillium, and Simplicillium were more abundant in male plants. Cd treatment considerably decreased the abundances of Penicillium, Phyllactinia, and Sterigmatomyces in male plants but increased those of Cercospora, Rhodotorula, Holtermanniella, and Cyberlindnera. On the other hand, in female plants Cd stress reduced the abundances of Cercospora, Coprinellus, Phyllactinia, and Sterigmatomyces but increased those of ­Holtermanniella and Schizophyllum. As a result, female plants exhibited higher abundances of Phyllactinia, Penicillium, Pestalotiopsis, and Gibberella, whereas male plants were enriched in Cercospora and Cyberlindnera.

With regard to the epiphytic bacteria, male control plants were enriched in the family Paenibacillaceae and the genera Pseudomonas and Acinetobacter relative to female controls (Fig. 4C). Cd treatment considerably increased the abundance of the Paenibacillaceae family and the Pseudomonas, Acinetobacter, and Bacillus genera in females but not in males. Hence, female plants subjected to Cd exposure showed a relatively higher abundance in the genus Bacillus while male Cd-treated plants were more enriched in the family Comamonadaceae. For the endophytic bacteria, male controls contained a higher abundance of the genus Lactobacillus but a lower abundance of the genus Streptococcus compared to their female conspecifics (Fig. 4D). Meanwhile, soil Cd treatment considerably increased the relative abundances of the genera Bacteroides, Faecalibacterium, Bifidobacterium, Prevotella, and Lachnospiraceae NK3A20 in both the male and female phyllospheres.

Effects of plant sex and soil Cd treatment on the taxonomic composition of the phyllosphere microbial communities

A linear discriminant analysis coupled with effect size analysis was utilized to compare the microbial composition from phyla to genera between the Cd treatment and the controls as well as between the two plant sexes. Both the epiphytic and endophytic fungal communities showed significant differences in composition between the Cd treatments and between the sexes (Fig. 5A–D). For the epiphytic communities, the phylum Ascomycota, the classes Dothideomycetes and Eurotiomycetes, and the order Capnodlales were predominant in the Cd-treated plants, while the phylum Basidiomycota, the classes Agaricomycetes, Saccharomycetes, and Leotiomycetes, and the order Pleosporales were enriched in control plants (Fig. 5A). In addition, the phylum Basidiomycota, the class Saccharomycetes, and the family Didymellacaeaeae were dominant in phyllosphere of male plants, while the phylum Ascomycota was enriched in that of females (Fig. 5B). For the endophytic fungal communities, the class Saccharomycetes, the orders Hypocreales and Trichosporonales, and the family Meruliaceae were more abundant in the control plants, whilst the phylum Rozellomycota, the classes Agaricostilbomycetes and Leotiomycetes, the orders Capnodiales and Filobasidiales, the families Ramularia, Xylariaceae, Microdochiaceae, and the genus Letendraea were more abundant in the Cd-treated plants (Fig. 5C). In addition, the endophytic fungal communities of male plants were relatively abundant in the order Tremellales while the families Ophiocordycipitaceae and Trichomeriaceae were enriched in the communities of the female plants (Fig. 5D).

The compositions of both the epiphytic and endophytic bacterial communities were also significantly influenced by Cd treatment and plant sex (Fig. 5E–H). With regard to the epiphytes, the phyla Firmicutes and Myxococcota, the class ­Thermoleophilla, the orders Rhizobiales, Oceanospirillales, Propionibacteriales, Streptosporangiales, Pseudomonadales, Caulobacterales, Pseudonocardiales, Corynebacteriales, and Sphingomonadales, and the genera Kineococcus and Terrabacter were all predominant in Cd-treated plants, while the order Burkholderiales was enriched in control plants (Fig. 5E). The order Bacillales was more prevalent in female plants while the family Comamonadaceae was enriched more in male plants (Fig. 5F). For the endophytic bacterial communities, soil Cd stress enriched the abundance of the phyla Firmicutes and Bacteroidota, the orders Corynebacteriales and Bifidobacteriales, whilst in control plants the phyla Proteobacteria and Chloroflexi, the class Thermoleophilia, and the order Frankiales were more prevalent (Fig. 5G). Lastly, the phylum Planctomycetota, the class Acidimicrobiia, the order Chitinophagales, the family Alcaligenaceae, and the genus Erythrobacter were more abundant in the endophytic bacteria of male plants, while the family Propionibacteriaceae was enriched in the female plants (Fig. 5H).

Correlations between host-plant traits and phyllosphere microbial taxa

A redundancy analysis was performed to investigate the effects of host-plant traits on the distribution of microbial taxa associated with the phyllosphere. The results showed that leaf lesion area was the best predictor of the shifts in community structure of the phyllosphere fungi (Fig. 6A; Supplementary Table S3; R²=0.3083, P=0.001), while leaf Cd concentration also contributed significantly to the composition of the fungal community (R²=0.2198, P=0.012). Heatmaps of Spearman correlations further indicated that leaf Cd concentration was positively correlated with the relative abundance of the genera Cyberlindnera and Cercospora and negatively correlated with the genera Phyllactinia and Coprinellus in the phyllosphere of males (Fig. 6B). Meanwhile, leaf lesion area was positively correlated with the genera Phyllactinia and Coprinellus and negatively correlated with the genus Aspergillus. In the phyllosphere of female plants, leaf Cd concentration was positively correlated with the genus Aspergillus but negatively correlated with the genera Phyllactinia and Cercospora, whilst leaf lesion area was positively correlated with the genera Cercospora and Purpureocillium (Fig. 6C).

Changes in the relative abundances of phyllosphere bacterial taxa were significantly associated with both the leaf Cd concentration (Fig. 6D; Supplementary Table S3; R²=0.4032, P=0.001) and the leaf lesion area (R²=0.3188, P=0.002). More specifically, heatmaps of Spearman correlations revealed that leaf Cd concentration in the male phyllosphere was positively correlated with the genera Acinetobacter, Paenibacillus, unclassifiedf. Paenibacillaceae, Pseudomonas, and Bacillus while negatively correlated with the genus Comamonas (Fig. 6E). In addition, leaf lesion area was positively correlated with the genus Comamonas whilst negatively correlated with the genus Acinetobacter. In the phyllosphere of females, the genera Comamonas and Acinetobacter were both positively correlated with leaf lesion area (Fig. 6F).

Plant sex and soil Cd treatment affect predicted microbial functions

Differences in predicted functions of the fungal and bacterial communities from the four treatments were explored using FUNGuild and PICRUSt2, respectively. For the epiphytic fungi, the relative abundance of fungal taxa related to the functional group ‘Animal Pathogen-Endophyte-Lichen Parasite-Plant Pathogen-Wood Saprotroph’ for control male plants was the lowest compared to the other three treatments, whilst the male plants from the Cd treatment contained the lowest abundance of fungi related to the functional group ‘Plant Pathogen’ (Fig. 7A). Exposure to soil Cd decreased the relative abundance of the endophytic fungi related to the functional groups ‘Animal Pathogen-Endophyte-Lichen Parasite-Plant Pathogen-Wood Saprotroph’ and ‘Plant Pathogen’ while it increased that of the endophytic fungi related to the functional groups ‘Undefined Saprotroph’ and ‘Soil Saprotroph’ in both the male and female plants (Fig. 7B). As a result, the relative abundances of the functional group related to ‘Plant Pathogen’ in male and female Cd-treated plants were 1.7% and 5.5%, respectively.

The PICRUSt2 analysis indicated that Cd stress reduced the relative abundance of epiphytic bacterial taxa that related to the KEGG pathway functions of ‘ABC transporters’, ‘Two-component system’, ‘Carbon metabolism’, ‘Quorum sensing’, ‘Valine, leucine, and isoleucine degradation’, ‘Butanoate metabolism’, ‘Propanoate metabolism’ and ‘Carbon fixation pathways in prokaryotes’ only in female plants and not in males (Fig. 7C). Exposure to soil Cd decreased the relative abundance of all the top 25 abundant functions in KEGG pathways of endophytic bacterial taxa in the leaves of both male and female plant, with male plants exhibiting relatively lower abundance of bacterial taxa related to the function ‘Biosynthesis of amino acids’ compared to female plants (Fig. 7D).

Discussion

In this study, we found that male P. deltoides plants showed a better growth performance than female co-specifics in the absence of soil Cd contamination. Numerous studies have observed similar male-biased performance in vegetative traits in several dioecious plant species under either natural or environmentally stressful conditions (Munné-Bosch, 2015). However, in contrast to male-biased herbivory as documented in several dioecious species under either natural or laboratory conditions (Boecklen and Hoffman, 1993; Lin et al., 2022), pathogens are the only guild that showed a female-biased tendency when comparing the preference of natural enemies such as herbivores, pathogens, and seed predators between the two plant sexes of dioecious plant species in a meta-analysis (Cornelissen and Stiling, 2005). Indeed, in the present study we found that female P. deltoides plants had 22% greater leaf lesion area than male plants after artificial infection with the leaf pathogenic fungus, Pestalotiopsis microspora. A few studies have also documented consistent female-biased infection rates in terms of plant susceptibility to pathogens among several dioecious plant species (Williams et al., 2011). For example, Ahman (1997) found that Melampsora rust disease was more severe on female individuals than on male conspecifics of Salix viminalis, and Lee (1981) showed that female individuals of Silene dioica suffered more severe infection by Microbotryum violaceum than males.

The plant-associated microbiome has been suggested to be able to protect host plants against pathogen invasion either through direct inhibition of pathogen proliferation via production of antimicrobial compounds or competition for nutrients and space (Ritpitakphong et al., 2016; Harish and Bhargavi, 2021), or through indirect stimulation of plant innate immunity, which confers resistance against various pathogens (Teixeira et al., 2019). Interestingly, in this study we found that leaves of male P. deltoides plants harbored more abundant ecologically beneficial fungal taxa than their female conspecifics, such as Candida, Simplicillium, and Penicillium as well as bacterial taxa such as Paenibacillus, Pseudomonas, and Acinetobacter (Fig. 4). All these taxa have been recorded as antagonists or biological controls that are able to suppress the growth or infection of some pathogenic microorganisms (El-Ghaouth et al., 1998; De Cal et al., 2008; Ward et al., 2012; Grady et al., 2016). For example, gray mold disease caused by Botrytis cinerea in apple trees has been shown to be strongly inhibited by Candida saitoana (El-Ghaouth et al., 1998), while Simplicillium lanosoniveum can be used as biocontrol agent to control the soybean rust pathogen Phakopsora pachyrhizi (Ward et al., 2012). Interestingly, the bacterial genera Paenibacillus and Acinetobacter have also been found to exert plant growth-promoting effects besides their antimicrobial abilities (Sachdev et al., 2010; Grady et al., 2016). In contrast, the common leaf pathogenic fungal genus Cercospora, which is reported to cause leaf spot in several plant species (Alderman et al., 1987; Skaracis et al., 2010), is more abundant in the phyllosphere of female sugar beet than in males (Steddom et al., 2005). Therefore, the microbiome associated with the phyllosphere of male P. deltoides might potentially act as a protective shield against leaf pathogen attack.

In accordance with our expectations, soil Cd exposure decreased the disease incidence for both plant sexes after P. microspora infection, and the reduction in the leaf lesion area in male P. deltoides plants was greater than that for the female ­conspecifics (Fig. 1E; 42% versus 35%). This difference in pathogen resistance might firstly be attributable to differences in leaf Cd accumulation between the two plant sexes. The mean Cd concentrations determined in P. deltoides after exposure to soil Cd stress were 116.7 mg Cd²⁺ kg^–1 DW in leaves of male plants and 84.5 mg Cd²⁺ kg^–1 DW in leaves of female plants. Given that the leaves had a mean water content of 64%, these values were respectively equivalent to concentrations of 42.0 mg Cd²⁺ l^–1 and 30.4 mg Cd²⁺ l^–1 in the PDA medium that we used for mycelium growth experiments. These concentrations in the PDA medium could be calculated as being able to reduce the mycelium growth by 42.3% and 32.4%, respectively (Fig. 1F). Correspondingly, we found that soil Cd treatment decreased the percentage of leaf lesion area in male and female plants by 41.6% and 35.0%, respectively (Fig. 1e). There is similar compelling evidence from previous studies that the accumulation of heavy metals in tissues can protect plants from pathogen infection (Boyd et al., 1994; Hanson et al., 2003). We have previously found that Cd stress has no significant influence on the main defensive chemicals in P. deltoides leaves that might be involved in pathogen resistance, such as tannins, total phenolics, and total flavonoids (Lin et al., 2022), and this strongly suggests that the enhanced pathogen resistance of the Cd-stressed plants in this current study might be attributable to a direct Cd toxicity rather than to the changes in leaf defensive chemicals.

We also found that soil Cd exposure altered the structure and function of the microbiomes associated with the phyllosphere between the two plant sexes. Interestingly, male control plants exhibited higher Alpha diversity than female plants in epiphytic fungal and endophytic bacterial communities, while Cd stress largely enhanced the Alpha diversity of epiphytic bacterial taxa in the females (Fig. 2). Other studies have also observed an increased microbial diversity in the phyllosphere promoted by plant heavy metal stress. For instance, Dobrone Toth et al. (2009) found a positive correlation between phyllospheric microbial population densities and Cd, Ni, and Zn contents in leaves of Ambrosia elatior, while Jia et al. (2018) have shown that the Alpha diversity of phyllosphere bacteria is positively correlated with Cd, Cr, Zn, and Pb contents in shoots of Bothriochloa ischaemum. However, although we observed that Cd-stressed female plants harbored a higher diversity of epiphytic bacterial communities in their phyllosphere, they were still more susceptible to pathogen attack than their male conspecifics. We argue that the protective effect of phyllospheric microbes for plant pathogen resistance might only be attributable to a few specific microbial species rather than to the whole collective microbial community. For example, Ritpitakphong et al. (2016) isolated and identified one Pseudomonas species from the phyllospheric microbiome of Arabidopsis as being able to increase plant resistance to the fungal pathogen Botrytis cinerea.

We have shown that exposure of male P. deltoides plants to soil Cd led to a higher abundance in the phyllosphere of several ecologically beneficial microorganisms such as the genera Cyberlindnera, Candida, Paenibacillus, Pseudomonas, and Acinetobacter, whilst the relative abundance of some plant pathogenic fungal genera such as Phyllactinia, Pestalotiopsis, and Fusarium were decreased compared to that of female plants (Fig. 4). For instance, Cyberlindnera sargentensis presents strong antagonistic activity against several plant pathogenic fungi (Rueda-Mejia et al., 2022), while Pseudomonas strains act as efficient biological control agents through the production of antifungal metabolites (Walsh et al., 2001). On the other hand, Phyllactinia species have been documented as causing powdery mildew disease in several woody plant species such as oak, poplar, and pear trees (Surma et al., 2022), while Fusarium species are the most common soil-borne fungal pathogens on crops and can lead to enormous agricultural losses (Ma et al., 2013). In addition, our functional prediction analysis further demonstrated that whilst soil Cd stress could greatly reduce the abundance of fungal taxa related to the group ‘Plant Pathogens’ in both plant sexes, the decrease was greater in the phyllosphere of males than females (Fig.7). Lastly, our in vitro toxicity tests of cultivable phyllosphere microbial extractions on the growth of P. microspora strongly suggested that both the epiphytic and endophytic microorganisms extracted from the phyllosphere of Cd-treated P. deltoides plants could suppress pathogen growth, with a better inhibition from the male phyllosphere than that of the female. Thus, taking our results together, we suggest that Cd accumulation in the leaves of male and female P. deltoides plants could differentially alter the structure and function of the microbiome associated with the phyllosphere, which in turn will alter its function as an additional layer of defense against infection by pathogenic microorganisms.

Conclusions

In summary, our study clearly showed that soil Cd exposure significantly reduced leaf lesion area in both sexes of P. deltoides after inoculation with a leaf pathogenic fungus, with male plants accumulating higher Cd concentrations in the leaves and exhibiting better resistance than conspecific females. Moreover, Cd exposure differentially altered the structure and function of the microbiome associated with the phyllospheres of the male and female plants, resulting in a greater abundance of ecologically beneficial microorganisms and a decreased abundance of plant pathogenic microbes being harbored in the phyllosphere of male plants. In vitro toxicity tests further suggested that these sexual differences in pathogen resistance could be attributed to direct Cd toxicity and to the changes in the structure and function of the phyllosphere-associated microbiome. However, it is worth noting that the P. deltoides plants used in this study were collected from a natural population growing in a metal-contaminated site, and hence they might have developed an enhanced Cd accumulation and tolerance ability in response to the stress. Therefore, more plant species grown on non-contaminated soil need to be further investigated in order to draw more general conclusions.

Supplementary data

The following supplementary data are available at JXB online.

Fig. S1. Representative images of leaf lesions on P. deltoids plants from the different treatments.

Fig. S2. Venn diagrams of the epiphytic and endophytic microbial operational taxonomic units.

Table S1. One-way analysis of similarity of the compositions of the phyllospheric microbial communities.

Table S2. Permutational multivariate analysis of variance for the Beta diversity of the microbial communities.

Table S3. Redundancy analysis for the effects of leaf Cd concentration and leaf lesion area on the divergence of microbial taxa.

Author contributions

TL and SH planned and designed the research; QL, ZZ, SyL, and SjL performed the experiments and analysed the data; TL, YL, TZ, LC, and SH interpreted the data; TL, CY, and SH wrote the original manuscript; all authors contributed to the subsequent revisions.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Funding

TL thanks the National Natural Science Foundation of China (Grant no. 32001204) and the Natural Science Foundation of Sichuan Province (Grant no. 2022NSFSC1633) for financial support.

Data availability

The data that support the findings of this study are openly available in the Dryad Digital Repository at https://doi.org/10.5061/dryad.79cnp5hzr (Lin et al., 2023). The fungal and bacterial sequence data have been deposited in the NCBI database (https://www.ncbi.nlm.nih.gov) under the SRA accessions of PRJNA859240 and PRJNA859001, respectively.

References

Author notes