The Jinsha River is a critical component of the Yangtze River (Su et al. 2022). Wudongde dam was established on the Jinsha River for the Wudongde hydropower station, crossing the left bank of Huidong county (Sichuan Province) and the right bank of Luquan county (Yunan Province) (Zhao et al. 2018). The study area is located at the WLFZ of Wudongde reservoir, Jinsha River Basin, Xinmin Village, Jiyi Town, Wuding County, Chuxiong Yi Autonomous Prefecture, Yunnan Province (26.1667° N, 102.2167° E). Monthly average maximum temperature, minimum temperature and precipitation in this area during 1981-2010 were 16.2-26.3 °C, 0.3-17.5 °C and 12.1-198.2 mm, respectively, according to Chinese Central Meteorological Station. This area belongs to the low-latitude plateau subtropical monsoon climate with sufficient sunlight and strong evaporation (Niu et al. 2015). With the Wudongde hydropower fully functioning in 2021, the water level changes between 952 m and 975 m in the Wudongde reservoir. Due to long submergence time (about 270-330 days per year), the vegetation coverage is very low in this area, where it was found that A. mexicana L. is a dominant plant in the present study area.

The date for collecting the samples of rhizophere from A. mexicana was April 17, 2022. The rhizosphere soil samples were obtained from both wilted and unwilted A. mexicana L. individuals (each with more than 4 individuals) at different elevations and sites from the WLFZ by shaking the roots to make the adhering soil go into the sterile centrifuge tubes. Because the available amount of rhizosphere soil for a A. mexicana L. individual is very little each sample often requires sampling the rhizosphere soils from multiple A. mexicana L. individuals. All samples were obtained between elevations of 930-940 m (Fig. 1). Wilted and unwilted states can be observed for many plants. These states were assessed by plants’ growth performance, where the plant growth that is normal is called an unwilted state and the plant growth with the stem and leaf atrophying and bending toward the ground is called a wilted state. These soil samples from multiple wilted and unwilted A. mexicana L. individuals were mixed, respectively, with each having three replicates, and then stored at -80°C refrigerator before formal sequencing. For simplicity, the rhizosphere microbial samples from wilted and unwilted A. mexicana L. individuals were defined as JB and JA, respectively. Finally, genomic DNA of these samples was extracted by CTAB method.

Figure 1.  

Sampling site.

The amplification of 16S V3-V4 regions and ITS regions of each sample was carried out using primers of CCTAYGGGRBGCASCAG and GGACTACNNGGGTATCTAAT, and primers of CTTGGTCATTTAGAGGAAGTAA and GCTGCGTTCTTCATCGATGC, respectively. TruSeq® DNA PCR-Free Sample Preparation Kit (Illumina, USA) was adopted for the generation of sequence libraries. The Qubit@ 2.0 Fluorometer (Thermo Scientific) was used to analyze the library quality. The sequencing of rhizosphere microbes was performed using the Illumina NovaSeq platform with 250 bp paired-end reads produced. QIIME2 dada2 plugin was employed for the filtration, trim, denoising, and merging of sequences from each sample to get the feature table of amplicon sequence variant (ASV). The raw data of all samples were deposited in BioProject database (ID: PRJNA985046). Taxonomic information was obtained using the Greengenes (16S) and UNITE (ITS) databases based on sklearn algorithm. A wide number of bacteria and fungi were detected in the rhizosphere soil from A. mexicana L. under wilted and unwilted states, where unclassified microbes are removed for further exploration from the analyzed data. The bacterial and fungal compositions at phylum, class, order and genus levels were investigated. Chao1, Goods_coverage, Observed_features, Faith_pd, Simpson and Shannon indexes of bacterial and fungal communities were calculated to explore the alpha diversity. Finally, the functional profiles of rhizosphere microbiome were predicted using PICRUSt2 software with the MetaCyc metabolic pathways.

In this study, a total of 500375 reads (range: 75440-87185, mean: 83396) for rhizosphere bacteria and 482739 reads (range: 60818-87130, mean: 80457) for rhizosphere fungi were obtained after sequencing. The Venn diagrams of bacterial communities in JA and JB are shown in Fig. 2. It was noted that 837 common bacterial amplicon sequence variants (ASV) were presented in both JA and JB with 3108 unique to JA and 3569 unique to JB. The Venn diagram of fungal communities is shown in Fig. 2B with 212 unique to JA, 693 unique to JB, and 92 common between them. The microbial list is available in Suppl. material 1.

Figure 2.  

Venn diagrams of bacterial (A) and fungal (B) communities in the rhizosphere soils of both unwilted (JA) and wilted (JB) individuals of A. mexicana L. in the WLFZ.

Bacteria play an important role in soils, including organic matter decomposition, elemental cycles, plant symbiosis, etc. (Song et al. 2017). Clarifying the diversity in soil bacterial communities is becoming increasingly important in soil microecology. In previous studies, alpha diversity indexes of microbes were often used to reveal the richness and diversity of soil microbes (Khan et al. 2020, Man et al. 2020, Tong et al. 2021). In this study, the bacterial richness and diversity of JA and JB from A. mexicana L. are thus analyzed and shown in Fig. 3. The Chao1 and Observed_features indexes in JA are 1634.93±204.11 and 1614.67±199.08, respectively, being lower than JB where they are 1812.35±150.49 and 1791.33±148.85, respectively. This indicated that the wilted individuals have higher richness than the unwilted ones (Zhang et al. 2017). The same is true for Shannon and Simpson indexes, they are 8.82±0.58 and 0.98±0.02 in JA, being lower than JB with a Shannon index of 9.15±0.62 and a Simpson index of 0.99±0.01. Lower Shannon and Simpson indexes found in JA than in JB suggested less bacterial diversity in JA (Chen et al. 2020, Li et al. 2021). In terms of mean value of Faith_pd, there is JA<JB. Thus, it can be speculated that the phylogenetic diversity of bacterial species in JB is more complex than in JA (Faith 1992).

Figure 3.  

Bacterial alpha diversity indexes of the rhizosphere soils of both unwilted (JA) and wilted (JB) individuals of A. mexicana L. in the WLFZ. A, Chao1 index; B, goods_coverage index; C, Faith_pd index; D, Observed_features index; E, Shannon index; F, Simpson index.

Soil fungi can act as biological controllers, decomposers and ecosystem regulators (Frąc et al. 2018). Fungal diversity is affected by plant traits, and in return regulate plant growth by symbiosis, pathogenicity and transformation. In this study, the rhizosphere fungal diversity of A. mexicana L. was thus investigated. Fig. 4 shows the fungal richness and diversity of JA and JB from A. mexicana L. It was found that the Chao1 and Observed_features indexes were consistent in JA (130.67±32.83), which were lower than those in JB (323±140.54). The Shannon and Simpson indexes in JB are 1.54 and 0.13 more than those in JA, respectively. Noteworthy is thatthe Faith_pd in JB is 24.21 more than that in JA. This suggested that the phylogenetic diversity of fungal community for evolution is higher in JB than in JA.

Figure 4.  

Fungal alpha diversity indexes of the rhizosphere soils of both unwilted (JA) and wilted (JB) individuals of A. mexicana L. in the WLFZ. A Chao1 index; B goods_coverage index; C Faith_pd index; D Observed_features index; E Shannon index; F Simpson index.

In summary, the coverage of both fungal and bacterial community in JA and JB is near 100%, characterized by goods_coverage index as shown in Figs 3, 4. The rhizosphere of wilted A. mexicana L. individuals has a higher microbial richness and diversity than the unwilted ones, irrespective of the bacterial and fungal communities. These findings suggest that the diversity of both fungal and bacterial community from the rhizosphere of A. mexicana L. has been affected by its wilted state. This may be because the rhizosphere microenvironment of A. mexicana L. has transformed under the wilted state as compared to the unwilted state.

The bacterial phylum, class, order and genus at top relative abundances in rhizosphere of A. mexicana L. in the WLFZ are shown in Fig. 5. In the rhizosphere soils of A. mexicana L., Proteobacteria are the most abundant phylum in both JA (58%) and JB (37.48%). Similar trend was also observed in a previous study where Proteobacteria phylum was the most dominant in the rhizosphere soils of cucumber (Li et al. 2021). The second most abundant phylum is Actinobacteria (17.42%) for JA and Firmicutes (22.80%) for JB. Actinobacteria is reported to be more abundant in soils than in other media (Barka et al. 2016, Goodfellow and Williams 1983). For per gram of soil, there are 10⁶-10⁹ Actinobacteria cells. Some Actinobacteria are pathogen, but have a relatively minor role in plant pathogenicity compared to other bacteria. Instead, many Actinobacteria are able to provide beneficial effects to plant growth by competing with plant pathogen, production of antibiotics and nutrient supply. Like Actinobacteria, Firmicutes is also a frequently examined bacterial phylum in the rhizosphere soils (Lee et al. 2021, Zhang et al. 2019). The species from this phylum are often used as biofertilizers for plant growth promotion, biocontrol agents for the inhibition of plant pathogens, and bioremediation agents for pollutant removal (Hashmi et al. 2020). The order of other dominant phyla in JA is Acidobacteria (5.71%) > Firmicutes (5.55%) > Bacteroidetes (4.52%) > Gemmatimonadetes (2.98%) > TM7 (2.04%) > Chloroflexi (1.60%) > Nitrospirae (0.74%) > Verrucomicrobia (0.54%), while that is Bacteroidetes (14.28%) > Acidobacteria (9.99%) > Actinobacteria (5.91%) > Chloroflexi (4.48%) > Gemmatimonadetes (1.49%) > Verrucomicrobia (1.15%) > Nitrospirae (0.90%) > TM7 (0.19%) in JB.

Figure 5.  

Rhizosphere bacterial community composition of both unwilted (JA) and wilted (JB) individuals of A. mexicana L. at the phylum (A), class (B), order (C) and genus (D) levels in the WLFZ, characterized by sequence number (relative abundance). Different colors indicate the top 10 phyla, classes and orders, and the top 20 genera; the rest of the bacteria are shown as “other”.

The rhizosphere bacterial class was dominated by Alphaproteobacteria (18.59%), Gammaproteobacteria (18.36%), Betaproteobacteria (17.50%), Actinobacteria (11.46%), Clostridia (4.53%), Deltaproteobacteria (3.67%), Acidobacteria_6 (2.76%), Gemmatimonadetes (1.91%), Bacteroidia (1.76%) and Chloracidobacteria (0.98%). Alphaproteobacteria exhibited several kinds of morphologies (Stalked, stellate and spiral), involving in various metabolic strategies such as nitrogen fixation, ammonia oxidation and photosynthesis (Williams Kelly et al. 2007). Unuofin et al. (2019) found that some Gammaproteobacteria such as Stenotrophomonas maltophila BIJ16, Pseudomonas aeruginosa DEJ16, and Pseudomonas mendocina AEN16 could produce laccase and degrade phenolic and non-phenolic contaminants. Betaproteobacteria presents the interactions with fungi by lichens and endosymbionts (Degli Esposti et al. 2018). Differing from Alphaproteobacteria with symbiosis genes located in plamids, chromosomes or chromids, Betaproteobacteria has its symbiosis genes in plasmids. However, the bacterial abundance pattern in JB is totally different from that in JA. The top ten bacterial Class in JB are Clostridia (19.74%), Gammaproteobacteria (13.94%), Alphaproteobacteria (12.43%), Bacteroidia (12.42%), Betaproteobacteria (6.44%), Deltaproteobacteria (4.62%), Chloracidobacteria (4.38%), Actinobacteria (3.17%), Acidobacteria_6 (2.25%) and Gemmatimonadetes (0.86%). It was reported that a number of compounds can be degraded by Clostridia, and it prefers strictly anaerobic or moderately aerotolerant environment (Dürre 2007).

It was found that the most dominant bacterial order is Clostridiales with a relative abundance of 20.45% in JB, but it only accounts for 4.93% in JA. In a previous study performed by Li et al. (2019), Clostridiales is also examined to be the most abundant order in pentachlorphenol-contaminated soils, and was critical to pentachlorophenol dichlorination. The addition of both citrate and lactate could modify the soil bacterial community with an increase of Clostridiales. Other dominant bacterial orders in JB are Bacteroidales (12.87%), Sphingomonadales (4.39%), RB41 (4.37%), Rhizobiales (3.98%), Xanthomonadales (3.55%), Actinomycetales (3.19%), Burkholderiales (2.89%), iii1_15 (1.94%) and Pseudomonadales (1.13%). In particular, the Bacteroidales dominating in JB only covers 1.95% in JA, and Bacteroidales has been considered a potential alternative to traditional fecal indicator bacteria for water quality monitoring in a previous study (Weidhaas et al. 2015). Burkholderiales (13.79%), Pseudomonadales (12.89%) and Actinomycetales (12.24%) occupied preponderance in the competition with many other bacteria in JA. It was also shown that Burkholderiales were enriched in maize-root soil samples, and harbored several beneficial effects on their linked plants, such as pathogen inhibition and nitrogen fixation (Aguirre-von-Wobeser et al. 2018). Both Rhizobiales (7.69%) and Sphingomonadales (7.28%) were also dominant in JA.

For the bacterial genus level, the abundance pattern between JA and JB is totally inconsistent. Compared to phylum, class, order, “other” groups in genus account for a very high percent in both JA and JB: >50% for JA and >60% for JB, if only top 10 bacteria were shown. Thus, we showed 20 top genera in this point. The most dominant genus Pseudomonas (21.57%) in JA only covers 2.66% in JB. Pseudomonas has been found to be able to increase the drought resistance of willows by enhancing nitrogen uptake (Kong et al. 2022).

The Linear discriminant analysis Effect Size (LEfSe) analyses were performed to explore the key bacterial taxa (genus level or higher) contribution mostly to the detected differences between JA and JB (Fig. 6A). It should be noted that LEfSe analysis has been believed to be efficient for identifying soil microbes that significantly responded to karst rocky desertification progression (Qi et al. 2018), chilli pepper-banana rotation (Hong et al. 2020) and soil-plant compartments in grapevine (Martínez-Diz et al. 2019). The major bacterial groups examined in JA were Proteobacteria, Massilia, Devosia, Actinobacteria, Micrococcaceae, Rhizobiaceae, Sphingomonas, Oxalobacteraceae, Agrobacterium, Actinobacteria, Arthrobacter, Burkholderiales, Betaproteobacteria, Actinomycetales. As compared to JA, Bacteroidales, Turicibacteraceae, Turicibacterales, Bacteroidetes, Verrucomicrobiales, Verrucomicrobiaceae, Bacteroidia, Firmicutes, Turicibacter, Verrucomicrobiae, Bacteroidaceae, Bacteroides, Akkermansia, Verrucomicrobia and Roseburia are highly represented in JB. Among these significantly different taxonomical groups in JA, 9 belong to Proteobacteria phylum, and 5 belong to Actinobacteria phylum. None of the other taxonomical groups were detected in this study. These findings suggested significant variations in bacterial community composition in the unwilted state of A. mexicana L. only occurred in both Proteobacteria and Actinobacteria phyla as compared to the wilted state of A. mexicana L. However, this is not the situation for JB, where 5 for Bacteroidetes, 5 for Firmicutes and 5 for Verrucomicrobia among these significantly different taxonomical groups in JB.

Figure 6.  

Linear discriminant analysis effect Size (LEfSe) analyses identify the taxa (phylum, class, order and genus) that have the most contribution to observed differences between JA and JB. A bacteria; B fungi. Relative abundance of bacteria or fungi is significant at P < 0.05 with a logarithmic LDA score threshold of 4.0.

The fungal phylum, class, order and genus at top ten relative abundance are shown in Fig. 7. Ascomycota constituted the majority of fungal communities in the JA (95.24%). Chytridiomycota and Basidiomycota accounted for 2.65% and 2.04% of total fungal sequences, respectively. Other phyla indicated a relative abundance of < 0.04% in JA. Within JB samples, Ascomycota (75.61%) and Basidiomycota (19.63%) were the most abundant, followed by Chytridiomycota (1.54%) and Aphelidiomycota (1.01%), while the rest of the phyla were found to have a relative abundance less than 1%. Similar phenomena were also found in the rhizosphere soils of wheats (Gqozo et al. 2020), Stipa purpurea (Lu et al. 2016), potato (Hou et al. 2020), Adenium obesum, and Aloe dhufarensis (Khan et al. 2020), where Ascomycota and Basidiomycota were dominant. Ascomycota fungi have wide ecological niches, being correlated with carbon and nitrogen cycles, plant biomass decomposition and pathogenesis (Challacombe et al. 2019), while Basidiomycota fungi contain major pathogen lineages and mushroom-forming species (Coelho et al. 2017). For class level, the maximum variation occurred in Sordariomycetes between JA and JB, with 62.81% in JA changed to 39.77% in JB. It has been reported that Sordariomycetes is well known as one of the largest classes of Ascomycota (Maharachchikumbura et al. 2016), being extensively presented in marine, freshwater and terrestrial ecosystem. Several species such as Trichoderma viride, T. harzianum and Beauveria bassiana belonging to Sordariomycetes are important biocontrol agents (Kaewchai 2009, Maharachchikumbura et al. 2016). The second obvious variation appeared in Agaricomycetes between JA and JB, with an increase of 0.85% in JA to 17.61% in JB. Many Agaricomycetes are composed of wood-decaying fungi that play an important role in the carbon cycle. White rot and brown rot fungi of Agaricomycetes can decompose each part of plants’ cell walls and depolymerize cellulose, respectively (Kersten and Cullen 2014). Dothideomycetes is comparable between JA (33.73%) and JB (32.83%).

Figure 7.  

Rhizosphere fungal community composition of both unwilted (JA) and wilted (JB) individuals of A. mexicana L at the phylum (A), class (B), order (C) and genus (D) levels in the WLFZ, characterized by sequence number (relative abundance). Different colors indicate the top 10 phyla, classes and orders, and the top 20 genera; the rest of the bacteria are shown as “other”.

The diversity of fungal orders was lower in JA than in JB. Both Hypocreales (68.04%) and Pleosporales (21.49%) constituted almost 90% of all fungal sequences in JA. The order Hypocreales was reported to be the best-selected biocontrol fungi source for suppressing the deleterious plant pests, and many of them in the rhizosphere environment are able to outcompete plant pathogens and produce the promoters for plant growth (Contreras-Cornejo et al. 2016, Kepler et al. 2017). In addition, Capnodiales (5.87%) is a unique fungal order with a relative abundance > 1%. Different from JA, there are four fungal orders that have a relative abundance > 10%: Pleosporales (31.27%), Agaricales (15.17%), Microascales (15.07%) and Sordariales (10.65%). It should be noted that many other fungal orders were relatively abundant in JB, such as Hypocreales (6.53%), Saccharomycetales (2.33%) and Filobasidiales (2.09%).

The same is true for fungal genera, where the high abundance is only concentrated on several genera in JA, including Neocosmospora (58.90%), Alternaria (14.22%), Cladosporium (6.02%) and Epicoccum (4.43%). Neocosmospora and Alternaria covers a wide range of species that belong to endophytes, pathogens and saprobes (Lou et al. 2013, Sandoval-Denis et al. 2019). They are ubiquitous in soils, water and air. It has been demonstrated that more than 268 metabolites could be produced from Alternaria fungi, of which some compounds have the properties of anti-microbes (Lou et al. 2013). Cladosporium spreads in a variety of terrestrial and marine ecosystems, and often develops symbiotic relationships with many plants (Bensch et al. 2015, Salvatore et al. 2021). Epicoccum can be found in soils, air and decaying vegetation, and secretes various secondary metabolites (antioxidant, antimicrobial, and anticancer compounds) for phytopathogens and biotechnological applications (Braga et al. 2018). Other genera were in a very low abundance in JA. The most dominant genera were Acrocalymma (30.47%), Canariomyces (18.25%) and Coprinellus (17.11%) in JB, while other many genera were in a relatively even distribution, e.g., Wickerhamomyces, Neocosmospora, Fusarium, Microdochium, Gibberella and Naganishia, with a relative abundance of 1.5%-3%.

To explore the fungal taxa that has the greatest contribution to the observed differences between JA and JB, the LEfSe analyses were carried out (Fig. 6B). The major taxa in fungal group detected in JA samples were Leotiomycetes, Hypocreales, Ascomycota, Neocosmospora and Nectriaceae. It is very interesting that the Ascomycota fungal composition at all analyzed four taxonomical levels (phylum, class, order, family and genus) in JA were different from JB. Whereas in JB, As shown in Fig. 6B, represented fungi were 1 phylum (Basidiomycota), 1 class (Agaricomycetes), 2 orders (Microascales and Agaricales), 5 families (Aspergillaceae, Psathyrellaceae, Morosphaeriaceae, Phaffomycetaceae and Microascaceae) and 5 genera (Coprinellus, Canariomyces, Acrocalymma, Aspergillus and Wickerhamomyces). These taxa belong to Basidiomycota or Ascomycota phylum.

The functional profiles of bacterial and fungal communities were analyzed by predicting metabolic pathways using PICRUSt2 software (Douglas et al. 2020), i.e., the metabolic pathway abundances were explored by structured mappings of EC gene families to MetaCyc pathways. The top 20 MetaCyc pathways in JA or JB are shown in Tables 1, 2. MetaCyc is a frequently-adopted database of metabolic pathways from all domains of life in the functional analyses of rhizosphere soils (Caravaca et al. 2022, Jiménez et al. 2020), covering 2749 pathways in the recent version (Caspi et al. 2020).

A total of 18 metabolic pathways were observed to significantly differ between the bacterial communities of both JA and JB (ANOVA, p<0.01), where the degradation/reduction pathways were highly represented with 13 among them directly related to these processes involved in the nitrate reduction and the degradation of anaerobic aromatic compounds, catechol, creatinine, toluene, adenosine nucleotides, salicylate, 1,5-anhydrofructose, pyrimidine ribonucleosides, L-arabinose, adenosine nucleotides, and L-valine. For the rhizosphere fungi, 14 metabolic pathways were found to present significant difference between JA and JB (p<0.01), of which 7 pathways were involved in the degradation/oxidation/metabolism/reduction, 4 pathways were involved in biosynthesis, 2 pathways were correlated with interconversion, and 1 pathway was phospholipid remodeling (phosphatidylethanolamine, yeast). Also, a large scale of rhizosphere bacteriome structure and function analysis based on 557 pairs of published sequencing data for the rhizosphere and bulk soils showed that organic compound conversion pathway was highly enriched in the rhizosphere (Ling et al. 2022). This can be expected, since plants’ rhizosphere is the pool of root litter that requires the participation of degradation-related microbes.

Notably, each of the top 20 dominant bacterial metabolic pathways in JA have very approximate relative abundance with JB (Table 1), where biosynthesis pathways (12/20) are highly dominant. The biosynthesis pathways were also reported to be enriched in the rhizosphere microbiome of Barley (Ye et al. 2022) and Carpobrotus edulis (Caravaca et al. 2022). Interestingly, almost each pathway of these pathways exhibited an even relative abundance distribution (0.6-1.5%) in the rhizosphere bacterial communities. However, this situation related to relative abundance for fungi is totally different from that in bacterial communities (Table 2). The dominant fungal metabolic pathways in JA were Aerobic respiration I (cytochrome c) (7.10%) and Aerobic respiration II (cytochrome c) (yeast) (7.10%), followed by adenosine ribonucleotides de novo biosynthesis (3.74%), D-myo-inositol (1,4,5)-trisphosphate biosynthesis pathway (3.34%), Pentose phosphate pathway (non-oxidative branch) I (3.29%) and glyoxylate cycle (3.14%), and Superpathway of adenosine nucleotides de novo biosynthesis I (2.44%), pyruvate fermentation to isobutanol (engineered) (2.42%), TCA cycle II (plants and fungi) (2.41%), tRNA charging (2.32%) and L-valine biosynthesis (2.27%), Superpathway of adenosine nucleotides de novo biosynthesis II (2.22%) and Superpathway of L-serine and glycine biosynthesis I (2.16%) are also abundant. The palmitate biosynthesis I (type I fatty acid synthase) pathway is the most abundant (6.02%) in JB, but it disappeared in JA. The same is true for stearate biosynthesis III (fungi), mitochondrial NADPH production (yeast), Sucrose degradation III (sucrose invertase), Fatty acid β-oxidation VII (yeast peroxisome) and they are enriched in JB with abundances of 2.85%, 2.44%, 2.06% and 2.00%, respectively, but they disappeared in JA.