Structural and Dynamic Analysis of Leaf-Associated Fungal Community of Walnut Leaves Infected by Leaf Spot Disease Based Illumina High-Throughput Sequencing Technology
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Introduction
There are many unique microbial communities inside and outside plants, such as plant rhizosphere microorganisms, seed microorganisms, vascular microbes, and phyllosphere microorganisms. They are formed by the long-term co-living of plant microorganisms and specific parts of plants (Blakeman 1981; Bennett and Whipps 2008; Hartmann et al. 2008). The community composition and abundance of leaf-related microorganisms are highly complex and abundant, including bacteria, actinomycetes, and fungi (Sivakumar et al. 2020). These microbes play a vital role in helping the host against pathogens (Lacava et al. 2006; Mejía et al. 2008; Rajendran et al. 2011). According to the principle of plantable microecology to prevent and control plant diseases, beneficial microorganisms, and harmful microorganisms exist in the microbial community
on the surface of plants (Elad and Pertot 2014). When harmful microorganisms replace the dominant population in the community, the plant will enter the pathological process, that is, the susceptible state. The microbes on the surface of plants play essential roles in defense of plants against pathogens (Ritpitakphong et al. 2016). For a long time, the research on phyllosphere microbes has been far behind rhizosphere microbes and soil microbes studies. Most previous studies have focused on screening beneficial microorganisms with biocontrol effects through experiments such as confrontation assays to aid in managing plant diseases (Nam et al. 2016; Yuan et al. 2017). The composition of the foliar microbial community is not a random combination of various microorganisms but is composed after a rigorous selection. Microecological regulation prevents and controls plant diseases by regulating the balance between pathogens and the micro-environment.
Walnuts (Juglans regia L.) are one of the four most important worldwide nuts, considering that this has an impact on brain aging (Joseph et al. 2009) and cancer prevention (Soriano-Hernandez et al. 2015). Walnut anthracnose and brown spot are both critical diseases in the growth of walnuts, which can affect their leaves, fruits, and branches. Colletotrichum gloeosporioides, Colletotrichum fructicola, Colletotrichum siamense, Colletotrichum acutatum, Colletotrichum aenigma, and Marssonina juglandis all are pathogens that have been reported for walnut anthracnose, with C. gloeosporioides being the significant pathogen (Saremi et al. 2010; Huang et al. 2016; Wang et al. 2017; Da Lio et al. 2018; Wang et al. 2018; Wang et al. 2021). When the leaves are infected, irregular long or round spots appear, and the disease forms yellowing symptoms at the edge of the leaves in the later stages. The primary pathogens of the walnut brown spot are Fusarium spp., Alternaria spp., Phomopsis sp., Cladosporium sp., and Colletotrichum sp. (Belisario et al. 2001); when the leaves are infected, brown spots first appear and then gradually expand into nearly round or irregular spots. The middle part of the spot is brown, the edge is green, and a yellow halo around the periphery (Chen et al. 2021). Most previous studies of these two diseases have focused on the fruit of walnuts, but they also seriously endanger the health of walnut leaves. Walnuts are deciduous trees with limited leaf growth time, so if the health of their leaves is not taken seriously, the growth and development of subsequent flowers and fruits will also be significantly affected.
The proportion of culturable microorganisms to total microorganisms in the phyllosphere is hard to be established using culture-based or other methods (Rastogi et al. 2010; Yashiro et al. 2011; Rastogi et al. 2012). High-throughput sequencing technologies can characterize microbial communities’ composition in complex environmental ecosystems and are widely used in many fields to study microbial diversity and environmental diversity (Lentendu et al. 2013; Bork et al. 2015). Exploring the structure and succession of plant microbiota is an integral part of microecological regulation, and the underlying processes of fungal microbial population dynamics are not yet known. This study uses high-throughput sequencing technology to analyze the species diversity, abundance, dynamics, and composition of microbial communities associated with healthy and infected walnut leaves. Moreover, we discuss differences in fungal community diversity between healthy and infected walnut leaves and their relationships. In addition, a theoretical basis for investigating and controlling walnut leaf diseases is provided.
Experimental
Materials and Methods
Study site. The sampling location was in Ma Lie Township, Hanyuan County, Ya’an City, Sichuan Province, China (N29°20ʹ, E102°46ʹ). Ma Lie is located in the high mountain area, and the climate in this area is humid, with an average annual temperature and precipitation of 20°C and 800 mm, respectively. The study was conducted in a walnut orchard cv. ‘chuanzao’ with the landowner consent.
Sampling and processing. Five healthy and five diseased walnut trees, each of similar size and growth conditions, were selected for this study. All leaves showing abnormal spots were considered diseased, including irregular or round spots with yellow surroundings, and brown sub-circular spots. The canopy layer was divided into upper, middle, and lower layers. For healthy trees, five well-developed and fully mature leaves were collected from each canopy layer in each direction: east, west, south, and north. For trees showing symptoms of leaf-disease spots, five diseased leaves were collected similarly. A total of 60 leaves were collected from each tree. Samples of healthy or diseased leaves collected from different trees were placed in separate sterile Ziploc bags and returned to the laboratory for cryopreservation within 12 hours. Healthy/ infected leaves were mixed evenly and 50 of similar size were selected for high-throughput sequencing analysis. For diseased leaves, leaves with roughly the same degree of disease were selected for each sampling (the leaf area occupied by the disease spot area was approximately the same, Table SI).
The sampling timing depended on the phenological stage of walnuts to better reflect the microbial community’s dynamics on walnut leaves. The phenological walnut stages were determined based on the work of Ji et al. (2021). The annual cycle of walnut includes the dormancy stage, pre-budding, budding, flowering, leaf-expansion period, fruit swelling, flower-bud differentiation, core-hardening stage, kernel-filling, maturing, and defoliation period (Ji et al. 2021). Sampling periods included mid-May, mid-June (after the leaves were fully expanded), early July to late August, late August to early September, and early October (before leaves fell) in 2018. These sampling times corresponded to the leaf-expansion period, fruit swelling, core-hardening stage, maturing, and post-harvest before defoliation periods of walnuts, respectively. Healthy leaves were marked as HE groups, and infected leaves were marked as IN groups, as shown in Table SI.
Collection of microorganisms on the leaves and extraction of total DNA. The collection of microorganisms on the leaves referenced the approaches from Donegan et al. (1991). After mixing the 50 walnut leaves, 30 leaves were selected and cut into 5 × 5 mm fragments using scissors, sterilized at 121°C. Subsequently, the pieces were transferred to a sterilized triangular flask, and 200 ml of sterile 0.9% NaCl solution was added. The flask was sealed with a breathable sealing membrane and incubated on a shaker at 4°C, 120 rpm for 1 h. Afterward, the walnut leaf residue was filtered through a layer of sterile gauze, and the filtrate was subsequently centrifuged at 14,000 × g, 4°C for 30 min. The precipitate was resuspended in a small sterile 0.9% NaCl solution. The same processing procedure was adopted for each sample. Healthy and diseased leaves were five groups of samples, respectively. Total genomic DNA was extracted using the modified cetyltrimethylammonium bromide (CTAB) method (Stewart and Via 1993).
PCR amplification, mixing, and purification of PCR product. Using diluted genomic DNA as a template, PCR was performed using specific Barcoded primers, Phusion® High-Fidelity PCR Master Mix (New England Biolabs, USA) with GC Buffer. They contained a high-efficiency, highfidelity polymerase selected based on the sequenced region to ensure efficiency and accuracy. The ITS1 region primers were ITS5-1737F (5’-GGAAGTAAAAGTCGTAACAAGG-3’) and ITS2-2043R (5’-GCTGCGTTCTTCATCGATGC-3’) (Usyk et al. 2017). The PCR amplification system was 20 μl (4 μl 5 × Fastpfu Buffer, 2 μl dNTPs (2.5 mmol/l), 0.8 μl 1737F (5 μmol/l), 0.8 μl 2043R (5 μmol/l), 0.4 μl Fastpfu Polymerase (2.5 U/μl), 0.2 μl BSA, 10 ng DNA, Make up to 20 μl with ddH2O). The PCR conditions were 95°C 5 min, 94°C 1 min, 58°C 50 s, 68°C 1 min; 30 cycles; 68°C 10 min; stored at 4°C.
The PCR products were detected by electrophoresis using a 2% agarose gel. Subsequently, the target PCR product was purified and recovered from the gel using a gel recovery kit provided by Qiagen (Germany).
Library construction and sequencing. The library was constructed using a TruSeq® DNA PCR-Free Sample Preparation kit. The completed library was quantified by Qubit and Q-PCR and sequenced using a HiSeq 2500 PE250 platform.
Analysis of the data. The high-throughput sequencing data were processed with the pair-end process of preliminary original sequences. We used FLASH software to splice the original data and filter the spliced sequence to obtain a high-quality Tags sequence. Finally, the operational taxonomic unit (OTU) analysis sequence was acquired. Using QIIME software, the sequence with a more than 97% similarity is classified into an operating classification unit, OTU.
After the OTUs were obtained, a rarefaction curve was drawn to assess whether the sequencing depth of each sample was sufficient to reflect the microbial diversity contained in the community sample. Species annotation was to compare the OTUs sequence with UNITE database, and species annotation analysis (set threshold of 0.8–1) was performed using the Mothur method (Schloss et al. 2009) and the SSUrRNA database of SILVA132 (http://www.arb-silva.de) to obtain taxonomic information. Then, at each classification level: kingdom, phylum, class, order, family, genus, species, the community composition of each sample is counted. We calculated many indexes to investigate fungal species richness, diversity, and community composition differences. QIIME was used to analyze the samples’ PD_whole_tree, Shannon, Simpson, Chao1, and ACE diversity index to analyze the alpha diversity of the samples (Caporaso et al. 2010).
We used Ven and petal diagrams to understand the uniqueness and overlap of the different grouped samples. The number of OTUs detected in samples collected from healthy and diseased leaves at five different sampling time points. We look at the similarities and differences in fungal populations in different samples by employing a heat map. The top 35 most abundant genera in each sample were analyzed by cluster analysis at the taxon and sample levels to generate a heat map. Non-parametric Wilcox test was used for intergroup difference analysis. NMDS analysis is performed based on OTU level, UPGMA clustering tree analysis is performed based on weighted UniFrac distance, and beta diversity is analyzed. Rarefaction curve, NMDS analysis, Veen diagram, and heat map were used vegan package, VeenDiagram package, and ggplot2 package in R software (Version 2.15.3).
Results
In this study, 50 leaf samples were collected and sequenced for each group. The number of sequences by high-throughput sequencing obtained after quality control was 83,312, and the quality control efficiency was 94.52%. We got all Q30 values above 98%, indicating an absence of contamination and that the dataset’s quality was satisfactory, meeting the requirements for subsequent analysis. Raw sequence reads have been deposited in NCBI; the accession number is PRJNA600291 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA600291) The statistics obtained in each step of the data processing are shown in Table SII. After data analysis, high-quality sequences were clustered into 2,155 ITS operational taxonomic units (OTUs) with 97% identity. The rarefaction curve indicates that all samples’ sequence numbers reached the sequencing depth (Fig. S1).
Alpha diversity analysis. Alpha diversity indices include Shannon, Chao1, ACE, PD_whole_tree, and Simpson (Table SIII and SIV). The Chao1 index fluctuated between increasing and decreasing, with an overall trend toward increasing. The Shannon index of diseased leaves was consistently higher than that of healthy leaves throughout the study cycle. It increased and decreased with an increasing trend (Fig. 1). By late diseased onset, observed_species, population richness index (Chao1 index, ACE index, PD_whole_tree index), and population evenness index (Simpson index) were all higher than those of healthy leaves.
Fig. 1
Line chart of alpha diversity index for different sampling times; a) dynamic changes of Chao1 index, b) dynamic changes of Shannon index. The different letters indicate the significant difference at the 0.05 level (p < 0.05, n = 5).
OTU-based Venn and petal diagrams. Venn and petal diagrams are shown in Fig. S2. Analysis of the shared OTUs in each sampling time’s diseased and healthy groups was 58.17%, 54.27%, 30.96%, 49.33%, and 46.37%, respectively. It showed that healthy and diseased leaf fungi composition was more similar between the first two sampling times. There was a significant difference between healthy and diseased samples in the third sampling time and had the highest number of unique OTUs. The similarity of the communities among the five groups of infected leaves (IN1 ~ 5) was higher than that among the five groups of healthy leaves (HE1 ~ 5). The non-parametric Wilcox test results of Chao1 and Shannon index between groups are shown in Table I. The significant difference was only observed in the group HE5-IN5 (p‑value = 0.0005) in the former. While in the latter, the HE2-IN2, HE3-IN3, and HE4-IN4 groups have significant differences.
The non-parametric Wilcox test results.
The non-parametric Wilcox test results of Chao1 index
Difference
p-value
sig.
LCL
UCL
HE1-IN1
–4.6
0.3909
–15.3189
6.118942
HE2-IN2
1.2
0.8221
–9.51894
11.91894
HE3-IN3
–8.6
0.1128
–19.3189
2.118942
HE4-IN4
–9.8
0.072
–20.5189
0.918942
HE5-IN5
–20
0.0005
***
–30.7189
–9.28106
The non-parametric Wilcox test results of Shannon index
Difference
p-value
sig.
LCL
UCL
HE1-IN1
–6.8
0.1353
–15.8168
2.216805
HE2-IN2
–17.8
0.0003
***
–26.8168
–8.78319
HE3-IN3
–26.2
0
***
–35.2168
–17.1832
HE4-IN4
–12.4
0.0083
**
–21.4168
–3.38319
HE5-IN5
–8.2
0.0735
–17.2168
0.816805
sig. – indicates whether it is significant or not, * – p-value < 0.05, ** – p‑value < 0.01,
Fungal taxonomic identification of healthy vs. infected leaves. The number of sequenced tags from each sample was annotated to the number distributed at each taxonomic level (Fig. S3). At the phylum level, the detected fungi were similar in the healthy leaf group (HE) and the infected leaf group (IN) (Fig. 2); the main phyla are Ascomycota, Basidiomycota, and Glomeromycota. Representative sequences of each of the top 100 genera were obtained by multiple sequence alignment, and the phylogenetic tree at the genus-level analysis is shown in Fig. 3. The relative abundance of the top 10 and top 30 genera at the genus level is shown in Fig. 4 (top 10 and top 30). The top 35 most abundant genera in each sample were analyzed by cluster analysis at the taxon and sample levels to generate a heat map (Fig. 5). The composition of fungi in the samples at different sampling times was significantly different; the relative abundance of fungi of the same genus in the healthy leaf group (HE) and the infected leaf group (IN) differed significantly.
Fig. 2
Column map of species relative abundance at phylum level (Group Analysis).
Fig. 3
The group’s phylogenetic relationship of genus species at the level of analysis. A phylogenetic tree constructed from representative sequences of genus levels. The color of the branches and sectors represents the corresponding phylum. The stacked column diagram outside the fan ring illustrates the abundance distribution information of the genus in different samples.
Fig. 4
Species relative abundance of species at the genus level (by group analysis); a) relative abundance of the top 10 genera, b) relative abundance of the top 30 genera.
Fig. 5
Heat map of the Genus-level species abundance clustering (by group analysis). The longitudinal direction is the sample information, and the horizontal direction is the species annotation information. The cluster tree on the left side of figure is the species clustering tree. Different colors represent different relative abundances, red represents the high relative abundance, and blue represents the low relative abundance.
At the genus level, the dominant taxa were Vishniacozyma, Cercospora, and Ramularia, and the dominant genus varied with sampling time (Fig. 3). Among them, Colletotrichum spp., Fusarium spp., and Alternaria spp., pathogens of walnut leaf diseases, were consistently present in both healthy and infected groups (Fig. 6). The Alternaria spp. had the highest relative abundance at the first sampling (leaf-expansion stage). At the fourth sampling (fruit maturing stage), Colletotrichum spp. was dramatically increased in the infected group and much larger than in the healthy group. The relative abundance of Fusarium spp. in the infected group also increased substantially, while Alternaria spp. was decreased. However, the relative abundance of the genera Fusarium spp. and Alternaria spp. was much smaller than that of Colletotrichum spp.
Fig. 6
Line chart of relative abundance for different samples; a) Relative abundance of Colletotrichum spp. at different sampling times, b) relative abundance of Fusarium spp. at different sampling times, c) relative abundance of Alternaria spp. at different sampling times, d) relative abundance of Colletotrichum spp., Fusarium spp., and Alternaria spp. The different letters indicate the significant difference at the 0.05 level (p < 0.05, n = 5).
NMDS analysis and UPGMA clustering tree. Fig. 7 is based on the NMDS analysis results at the OTU level, where the stress value = 0.138 shows that NMDS results could accurately reflect the degree of difference between samples. Samples from the same sampling period clustered together. In contrast, samples of healthy and infected leaves from the first two sampling periods showed more overlap, indicating that the difference between the healthy and infected leaves was more significant in the late stage of the disease. Fig. S4 shows the results of the UPGMA clustering tree based on the
Fig. 7
Phyllosphere fungal assemblage dissimilarity among healthy (HE1 ~ 5) and infected (IN1 ~ 5) leaves, represented by nonmetric multidimensional scaling (NMDS).
weighted UniFrac distance, which was generally consistent with the NMDS analysis results. The taxa with significant differences between the healthy and infected leaf groups were different at different taxonomic levels.
The taxa with substantial differences between the healthy and infected leaf groups were also different at different sampling times (Table SV). At the genus level, genera with significant differences included Cercospora, Cladosporium, Phoma, and Symmetrospora between the HE1 and IN1 (leaf-expansion stage) groups (Fig. S5a); Vishniacozyma, Cladosporium, Hannaella, Boeremia, Symmetrospora, Strelitziana, Colletotrichum, Aureobasidium, Botrytis, and Aspergillus between the HE2 and IN2 (fruit swelling stage) groups (Fig. S5b); Vishniacozyma, Cercospora, Ramularia, Cladosporium, Epicoccum, Phoma, Boeremia, Dioszegia, Symmetrospora, Stagonosporopsis, Strelitziana, Chrysozyma, Alternaria, Fusarium and Ceriporia between the HE3 and IN3 (core-hardening stage) groups (Fig. S5c). Ramularia, Phylactinia, Boeremia, Symmetrospora, Serendipita,
Colletotrichum, Chrysozyma, Septoriella, Exobasidium, Acremonium, Golubevia, Hirsutella, Plectosphaerella, and Gibberella between the HE4 and IN4 (fruit maturing stage) groups (Fig. S5d); and Vishniacozyma, Cercospora, Cladosporium, Microsphaera, Hannaella, Epicoccum, Phoma, Boeremia, Dioszegia, Symmetrospora, Strelitziana, Zasmidium, Serendipita, Chrysozyma, Septoriella, Acremonium, Golubevia, Cryptococcus, Ampelomyces and Ceriporia between the HE5 and IN5 (post-harvest before defoliation stage) groups (Fig. S5e).
Discussion
Many studies have investigated changes in microbial community diversity following plant infection, but the results are not entirely consistent. De Assis Costa et al. (2018) studied the fungal diversity of oil palm leaves infected with Fatal Yellow disease (Phytophthora palmivora) and healthy leaves, observing that the fungal diversity of healthy leaves was higher than that of infected leaves. Douanla-Meli et al. (2013) showed that the overall frequency of infection of yellowing citrus leaves was significantly higher, but species diversity was low. Abdelfattah et al. (2016) observed no significant difference in fungal diversity between anthracnose-infected and healthy citrus leaves. These indicate that different plants and diseases affect the leaf microbial community differently. However, it can be determined that pathogenic microorganisms can significantly impact microbial community composition and structure, and the results of this study confirm this theory for pathogenic fungal microorganisms. The sampling time referred to the phenological period of walnut
(the experiment was designed to sample over time). Late sampling time, i.e., leaf-diseased onset progresses with the phenological period to late, the difference between the microbial community richness of diseased leaves and healthy leaves becomes larger. The result is consistent with previous studies showing that pathogens cause an increase in microbial community abundance when disease stress is high (Luo et al. 2019; Karasov et al. 2020). It has been suggested that a healthy leaf environment usually colonizes unique OTUs (Zhang et al. 2018). This study showed that the number of unique OTUs was significantly different at different stages of disease development. The third susceptible group had the unique OTUs, which differs from the findings of Zhang et al. (2018). The current experimental design is not yet able to explain the cause and tentatively considers the interleaf of the plant affected by temperature and humidity, and UV radiation to appear different conclusions.
The composition of the microbial community changes during leaf development and in the presence or absence of the disease symptoms. In our study, the main ones annotated at the phylum level in healthy and diseased groups of leaves were Ascomycota and Basidiomycota. The result is similar to the effects of studies on Olea europaea, Fagus sylvatica, and cucumber (Cordier et al. 2012; Abdelfattah et al. 2015; Luo et al. 2019). Our experimental design considers the phenological stage of walnuts to sample, the relative abundance of ascomycetes in both healthy and disease-susceptible leaves increased with walnut growth (sampling time over time). Zhang et al. (2018) suggested that the abundance of Cysticercus phylum increases with disease severity. We obtained results similar to the previous study, indicating that Cysticercus is highly resistant to the external environment after colonizing the leaves. Vishniacozyma is the dominant species at the genus level, and the plant is an important reservoir for this species (Félix et al. 2020); it has biological control on blue mold and gray mold, affecting pears (Lutz et al. 2012; Lutz et al. 2013). The biological control potential of this Vishniacozyma genus to control fungal pathogens of walnut foliage is well worth studying.
The Colletotrichum spp., among the top 30 genera in relative abundance in our results, has become one of the top 10 internationally recognized phytopathogenic groups (Dean et al. 2012). The genus consists of more than 100 species, such as C. gloeosporioides, C. hanaui, and C. fioriniae, all of these species can cause foliar fungal diseases on walnut trees (Qiu et al. 2010; An and Yang 2014; Zhu et al. 2015; Wang et al. 2017; Varjas et al. 2021). Walnut brown spot is a complex pathogenic disease caused mainly by Fusarium spp. and Alternaria spp. (Belisario et al. 2010). Yang et al. (2017) isolated and identified the causal agent of walnut brown spot disease from diseased leaves and fruits as A. alternate. Fusarium spp. and Alternaria spp. were detected in both healthy and diseased leaves in this study. In the results of the fourth sampling, the relative abundance of Fusarium spp. and Colletotrichum spp. increased much more than the previous one; the last diseased leaf sample showed that the relative abundance of the Colletotrichum, Fusarium, and Alternaria tended to be close.
Existing research proves that walnut brown spot results from the interaction between Fusarium, Alternaria, and the environment. The maximum temperature is the environmental factor that significantly impacts the disease’s severity (Scotton et al. 2015). The fourth sampling period was during the ripening period of walnuts. The late August to early September was the high-temperature and precipitation season in the area where the sampling site was located (Table SVII). Therefore, the significant increase in the relative abundance of pathogenic bacteria Fusarium spp. and Colletotrichum spp. is consistent with the disease development pattern. In our results, the relative abundance of Alternaria spp. was the highest at the first sampling, which may be related to the fact that Alternaria overwintered as conidia on dead leaves and germinated to infest walnut leaves when the temperature was suitable (Yang et al. 2017).
Conclusion
The results herein show the differences in leaf-associated fungi on healthy and diseased walnut leaves. The main fungal phyla inhabiting walnut leaves were Ascomycota, Basidiomycota, and Glomeromycota. Fungal species differed and changed significantly between the healthy (HE) and infected (IN) leaves at different sampling times. The populations of foliar disease pathogens (Colletotrichum spp., Fusarium spp., and Alternaria spp.) showed dynamic changes with the development of the leaf at different walnut phenological stages. They increased dramatically at the fourth phenological stage (late August to early September). Thus, the selection of reasonable measures to control walnut foliar diseases before August can avoid the significant development of the disease. Understanding the microbial composition associated with leaves from which to explore changes in pathogen populations can provide the basis for developing more effective control measures. This study enabled us to infer the intricate link between leaf-associated fungi and walnut leaf diseases, pointing to the possibility of controlling leaf diseases through microecological regulation in the future.
| 24 set 2022
