Microbial community abundance and diversity
Fungal and bacterial diversity, richness and observed-species were higher in the rhizosphere soil than those in the unplanted soil. Shannon index of fungi of chemical fertilizer only was more than that of furfural residue treatments with planting history longer than 20 years. While Shannon index of bacteria in all the fields was more than furfural residue treatments. Chao 1 and observed-species in furfural residue-treated soils were more than those of chemical fertilizer treatments. In addition, microbes dominated more in unplanted than those in the rhizosphere soil (Table 1).
Insert Table 1 here.
Venn diagram showed that among different continuous cropping year, 159 fungal OTUs, while 172, 530 were common among unplanted soil (Fig. 1A), chemical fertilizer-treated soils (Fig. 1B), and furfural residue treatments(Fig. 1C), respectively. With increase of continuous cropping years, unique fungal OTUs of the rhizosphere soil was more than that of unplanted (Fig. 1A). Similar difference was observed in OTUs of both common and unique bacterial of the rhizosphere soil (Supplementary Fig. 1A–1C). Microbial community diversity was significant increased, and more OTUs were induced in the rhizosphere soil and the species richness increased after planting and fertilization.
Insert Fig.1 A-1C here.
Insert supplementary Fig.1A-1C here.
Composition of the fungal microbiota of rhizosphere soil
Phylum composition of the fungal community in rhizosphere soil
A total of101,307, 211,144, and 206,009 raw tags were analyzed and approximately, 99.21% of the reads were utilized for metagenomic contig construction. Significant differences in fungal phyla were observed between the soil treated with and without furfural residue. Ascomycota andMortierellomycota were dominant in soils treated with chemical fertilizer only, while AscomycotaandBasidiomycota were dominant in soil treated with furfural residue( Fig. 2A). The relative abundance of Basidiomycota increased while that of Ascomycota decreased with the increase of continuous cropping year of furfural residue treatments (Fig. 2A).
Soil rhizosphere fungi of furfural residue treatments were significantly different with that of unplanted soil, and the richness of dominant phyla differed significantly among the seasonal soil batches. Ascomycota and Zygomycota were dominated in phyla unplanted soil, Chytridiomycota andGlomeromycota were least abundant (Supplementary Fig. 2A). In the rhizosphere of furfural residue-treatments, Ascomycota, Basidiomycota and Chytridiomycota were dominant, while Zygomycota was the least abundant (Supplementary Fig. 2B–C). A new inhabitant was Neocallimastigomycota, which increased with the increase of cropping years in the furfural residue treatments. Streptophycophyta was occasionally detected in the seasonal soil batches (Supplementary Fig. 2B), Fungi_ phy_ Incertae _ sedis was occasionally presented at the bell mouth stage (Supplementary Fig. 2C).
Insert Fig.2A-2B here.
Insert supplementary Fig.2A-2C here.
Genus composition of fungal community in rhizosphere soil
The community composition and diversity of fungal genera differed significantly between chemical fertilizer treatments and furfural residue treatments. Chaetomium, Humicola, Monographella and Cortinarius were more in furfural residue treatments, while Mortierella, Chrysosporium and Podospora were more in that without furfural residue(Fig. 2B). The top 30 fungal genera found in soils treated with or without furfural residue were totally different with exception of Chaetomium, Fusarium, Podospora and Acrimonium. The community of fungal genera shifted distinctly, and the microbial structure of rhizosphere soil was reestablished. The potential pathogen of Alternaria, Trichocladium, Bipolaris, Solicoccozyma, Gibberella and Cladosporium were not detected, while the beneficial Penicillium, Schizothecium and Rhizophlyctis were observed in the furfural residue treatments (Fig. 2B).
Similarly, the composition and diversity of fungal genera differed significantly among the treatments over different seasonal soil batches. During seedling stage, Alternaria, Trichocladium, Gibberella, Ciliophora, Candida, Solicoccozyma, Ilyonectria, Pseudogymnoascus, Metarhizium and Clonostachys were specific inhabitants in the unplanted soil (Fig. 3A), while Mycosphaerella, Camarops, Cercophora and Cryptococcus were prevalent. Fusarium increasedwith the increase of continuous cropping year at seedling stage (Fig. 3B). At bell mouth stage, Mortierella, Spizellomyces, Trichocladium and Verticillium were dominated in the unplanted soil, Typhula, Staphylotrichum, Coprinopsis and Ustilago were new inhabitants in the rhizosphere soil. Penicillium in rhizosphere soil increased within 5-20 years ; however, it gradually depleted after 20-30 years of continuous cropping (Fig. 3C). Spizellomyces was detected in soil of March and July; Myrothecium and Ophiosphaerella were dominant in March and May, while Humicola, Rhizophlyctis, Cortinarius, Acremonium, Cyathus, Madurella, Phaeoacremonium, Geomyces, and Sarocladium were dominant in May and July (Fig. 3). The increase of fungal genera during the seedling stage was notable. Pathogenic Alternaria, Trichocladium, Gibberella and Solicoccozyma disappeared in soil of May and July after the furfural residue application. And there was a significant decrease of Fusarium, Humicola, Chaetomium, Microdochium and Monographella during the bell mouth stage.The relative abundances of Aspergillus and Verticillium decreased with the plant growth. The relative abundance of Mortierella, Conocybe and Schizothecium in rhizosphere soil decreased at the seedling stage, and then increased at the bell mouth stage (Fig. 3A-C).
Insert Fig.3A-3C here.
Composition of the bacteria communication in rhizosphere soil
Phylum composition of the bacteria community
A total of91,973, 92,190, and 242,628 raw tags were analyzed and grouped into 68,626, 69,756, and 162,042 16S sequences. Unplanted soil had smaller species richness and greater evenness compared with furfural residue treatments (Table 1).
The 16S rRNA gene sequences from the continuous cropping year were affiliated with 30 bacteria phyla and corresponded to 1, 269 species-level OTUs. Phylum Actinobacteria and Thaumarchaeota decreased in furfural residue treatments and increased in chemical fertilizer only treatments (Fig. 4A). Compared with genus of rhizosphere soil, Acidobacteria and Bacteroidetes were dominant with furfural residue application, while Firmicutes and Chloroflexi were dominant in chemical fertilizer treatment. Rokubacteria and Gemmatimonadetes were less abundant in chemical fertilizer only treatments. Proteobacteria, Bacteroidetes and Acidobacteria were dominated under residue treatment over different continuous cropping years (Fig. 4A).
In different batch soil, phylum WPS-2 was prevalent in March and May; FCPU426 was specific in May and July; while Epsilonbacteraeota and Spirochaetes were specific in March and July. Proteobacteria and Bacteroidetes were abundant in May and sparse in July, while Acidobacteria was opposite. The bacterial genera Cyanobacteria, Firmicutes, Actinobacteria and Euryarchaeota depleted in unplanted soil and seasonal soil batches (Supplementary Fig. 3A–C).
Insert Fig.4A-3B here.
Insert supplementary Fig.3A-3C here.
Genus composition of the bacterial community of rhizosphere soil
The relative abundance of MND1, RB41, UTCFX1, Nitrospira, Ellin6067, Adhaeribacter, Flavisolibacter, Polycyclovorans and Terrimonas was greater than that of furfural residue treatments, while the abundances of Clostridium, Pseudarthrobacter and Roseiflexus was lesser(Fig. 4B).
In the seasonal soils, UTCFX1, RB41 and Terrimonas were the dominant genera in March (Fig. 5A), while MND1, RB41 and UTCFX1 were dominant in May and July (Fig. 5B–5C). Arcticibacter, ADurb.Bin063-1, Stenotrophomonas, UTBCD1 and Dongia were present only in March(Fig. 5A), while Caenimonas, Cellvibrio, Polycyclovorans, Gemmatimonas, and Arenimonas were restricted to the rhizosphere soil in May and July (Fig. 5B–C). RB41, Terrimonas, Nitrospira, Pir4_lineage, AKYG587 decreased at first, then increased with time, while Lysobacter, Adhaeribacter, Massilia and Pseudarthrobacter showed an opposite trend. The relative abundance of Haliangium and MND1 increased, while those of Pirellula, Ellin6067 and Luteimonas decreased in different season (Fig. 5).
Insert Fig.5A-5C here.
PCoA analysis
We compared the fungal and bacterial profiles obtained from the unplanted and planted soil samples with or without furfural residue application(Fig. 6). Taken together, they explained 97.45% of the total variability in fungal community at the phylum level and 37.25% in bacterial community in the rhizosphere soil. Apparent clustering was due to the variability with fertilization and growth stage. They together explained 96.36% variability in fungal communities. The first two axes of the PCA plot explained 94.09% of the total variability in data (Fig. 6A). Community diversity metrics clearly demonstrate that communities from the same treatments were similar and those from other treatments were distinct. The soil microbial community diversity after different continuous cropping years under the same treatment differed from each other; however, occasional overlap was observed between the samples. The variation of microbes in soils treated with furfural residue at different growth stages were similar and situated at the same quadrant. The dissimilarity in community diversity profiles among different treatments is notable and apparent from the PCA ordination, especially communities from unplanted soils and chemical fertilizer only treatments. The contribution of selected properties followed the trend of fertilizer > furfural residue > planting > seasonal soil batches. These results indicate that the microbial community structure is closely associated with fertilization (Fig. 6A). These variables together explained the variance in bacterial communities (Fig. 6B), of which the first two axes of the RDA plot explained 20.51% and 14.55% (35.06% of total) of the total variability. A significant correlation was observed between the microbial community diversity and the variations in bacterial community. Fertilization was the main factor that shaped the microbial community structure (Fig. 6B).
Insert Fig.6A-6B here.