Soil physicochemical properties were affected by continuous monocropping of S. flavescens
The results of a comparative analysis of the soil physicochemical properties among the four sampling sites, SCC, TCC, FCC (second-, third- and fifth-year continuous cropping, respectively) and CK (control) sites, are shown in Table 1. It was found that all of the soils were alkaline (pH value, 7.58 ~ 8.20) and that the CK soil had the highest pH value, which was significantly different from those of the SCC and TCC rhizosphere soil. The available phosphorus (AP) content was lowest in CK soil and highest in the SCC rhizosphere soil. The contents of soil organic matter (OM) and total nitrogen (TN) were lowest in SCC soil and highest in TCC soil (TCC > FCC > SCC). The sucrase content was lowest in FCC soil and highest in CK soil. The urease content was lowest in TCC soil and highest in FCC soil. Among these parameters, the OM and sucrase contents did not significantly differ with different years of continuous cropping with S. flavescens, and the other four indexes showed significant differences in rhizosphere soil with different years of continuous cropping (Table 1).
Table 1
The rhizosphere and bulk soil physical and chemical properties of S. flavescens
sample | pH | AP (mg/kg) | OM (mg/kg) | TN (g/kg) | Sucrase (mg/g) | Urease (mg/g) |
SCC | 7.59 ± 0.20b | 12.27 ± 0.19a | 115.63 ± 10.52b | 0.03 ± 0.01b | 1.16 ± 0.17a | 0.43 ± 0.03b |
TCC | 7.58 ± 0.04b | 8.32 ± 0.85b | 141.27 ± 10.02a | 0.63 ± 0.11a | 1.37 ± 0.14a | 0.38 ± 0.07ab |
FCC | 8.13 ± 0.14a | 2.19 ± 0.34c | 127.22 ± 5.41a | 0.07 ± 0.01b | 1.12 ± 0.09a | 0.67 ± 0.05a |
CK | 8.20 ± 0.01a | 1.57 ± 0.48c | 136.35 ± 14.04a | 0.10 ± 0.05b | 1.42 ± 0.09a | 0.63 ± 0.07ab |
Different lowercase letters indicate significant differences between different samples (P < 0.05) |
Field microbial community structure varied with continuous monocropping time
For the bacterial communities, the operational taxonomic units (OTUs) from four soil sites were found to belong to 40 phyla, 101 classes, 129 orders, 233 families and 318 genera. The composition of the bacterial community at the phylum level and its phylum abundances are shown in Fig. 1A and S1A. The top 10 relatively abundant bacterial phyla over all samples included Acidobacteria, Proteobacteria, Actinobacteria, Chloroflexi, Gemmatimonadetes, Bacteroidetes, Planctomycetes, Nitrospirae, Firmicutes, and Verrucomicrobia. The sum of these phyla accounted for more than 93% of the bacteriome. Acidobacteria and Proteobacteria accounted for the largest proportion (more than 51%). The abundance of Acidobacteria in the communities of SCC, TCC and FCC soils presented a decreasing trend with increasing time (Fig. S1A, 34.9% − 24.3%). The abundance of Proteobacteria increased in SCC soil (21.0%) compared with that of CK coil and reached its highest level in the FCC community (31.9%). The abundance of Actinobacteria in the SCC community was lower than those of other samples but with only small differences. Finally, the abundance of Bacteroidetes in the community presented an increasing trend from SCC to FCC soil (Fig. S1A, 2.7% − 4.6%).
For the fungal communities, OTUs detected from the four soil samples belong to 13 phyla, 35 classes, 89 orders, 170 families and 280 genera. Nine phyla and one unidentified phylum were identified from the soil samples (Fig. 1B, S1B) that accounted for over 99% of the fungal sequences. The nine determined phyla were Ascomycota, Mortierellomycota, Basidiomycota, Glomeromycota, Chytridiomycota, Kickxellomycota, Rozellomycota, Cercozoa and Chlorophyta. The fungal community of CK soil diverged from those of rhizosphere soils in the abundances of both Ascomycota and the unidentified phylum; the former dominated in rhizosphere soil, especially in TCC soil (70.0%), and the latter dominated in CK soil (39.6%). The abundance of Basidiomycota was highest in the FCC community (18.5%). Chlorophyta was present in only CK soil, in low abundance (0.1%), but was not detected in other rhizosphere soils. Moreover, the abundance of Rozellomycota in the community increased from SCC to FCC soil (Fig. S1B, 0.04% − 0.34%).
Accordingly, the OTUs identified in all analyzed samples were regarded as the core OTUs, and those identified in at least one sample were defined as pan-OTUs [29]. Here, core and pan-OTUs were identified for all soil samples (Fig. 2). In total, 879 core OTUs (Fig. 2A, Table S1) and 4,364 pan-OTUs were identified for the bacterial community (Fig. 2B), and 110 core OTUs (Fig. 2C, Table S1) and 1,178 pan-OTUs were identified for the fungal community at all sites (Fig. 2D). There were 2,225 bacterial OTUs shared by all four samples and 2,419 OTUs shared by SCC, TCC and FCC soils. The number of unique bacterial OTUs for these four sites was as follows: 182, 83, 98 and 118 for CK, SCC, TCC and FCC soils, respectively (Fig. 2B). A total of 377 fungal OTUs were identified as common OTUs for all samples (Fig. 2D). Four hundred and fifty fungal OTUs were shared by SCC, TCC and FCC soils. There were 44, 79, 60 and 78 fungal OTUs unique to CK, SCC, TCC and FCC soils, respectively. Before and after continuous cropping, the bacterial OTU numbers for CK, SCC, TCC and FCC soils were 3,570, 3,422, 3,220 and 3,350, respectively (Fig. 2B), and the fungal OTU numbers for them were 706, 782, 773 and 827, respectively (Fig. 2D).
Microbial diversity was influenced by the continuous monocropping time
The alpha diversity represents the measurement of within-community microbial diversity, which can be used to compare the diversities of S. flavescens rhizosphere soil among sites under different continuous monocropping times (Fig. 3). For the bacterial community, the Chao1 index values of CK, SCC, TCC and FCC rhizosphere soils were 3,384.07, 3,128.26, 3,006.66, and 2,988.44, respectively (Fig. 3, upper panel). According to a Shannon index analysis, the bacterial species richness from the rhizosphere soil of CK soil was the highest (9.82), followed by those of SCC (9.54), TCC (9.41) and FCC (9.47) soils (Fig. 3A). The diversity index results showed that the highest bacterial diversity of S. flavescens rhizosphere was found in the CK soil and that the lowest was found in TCC soil. For the fungi, the trends in rhizosphere richness and diversity index were basically the opposite for the three continuous cropping with S. flavescens periods. The Chao1 diversity and Shannon index values of FCC rhizosphere fungi were 654.75 and 6.07, respectively, followed by those of SCC rhizosphere fungi, which were 646.66 and 6.08, respectively, while the CK soil showed the lowest values, 637.93 and 5.09. The results showed that the number of years of continuous monocropping of S. flavescens and the presence of S. flavescens itself were the main factors affecting the diversity of rhizosphere fungi.
To better display the distance relationship between multiple samples, microbial β diversity was further assessed based on the unweighted UniFrac distance matrix [29]. The biological replicates clustered together, samples from the CK soil clustered with those from SCC soil, and samples from the other two sites (TCC and FCC) clustered together (Fig. 4A, Fig. S2). Importantly, great divergences in β diversity were identified between short-cropped and long-cropped sites (Fig. 4A). For the fungal community, a similar pattern was observed (Fig. 4B, Fig. S2). Both the bacterial and fungal community compositions varied among different samples, and this was also graphically illustrated in nonmetric multidimensional scaling (NMDS) [30] ordination and principal components analysis (PCA) (Fig. 4C, D and S2). Moreover, the community composition was apparently affected by the continuous cropping time and marginally influenced by the sampling sites.
Correlating physicochemical properties with microbial diversity
Soil physicochemical properties (the pH and AP, OM, TN, sucrase and urease contents) were significant explanatory factors that determined the observed clustering pattern of soil microbial communities with different years of continuous monocropping (Fig. 5A). To better understand the clustering and separation of samples caused by environmental factors, redundancy analysis (RDA) was conducted for both soil bacterial and fungal communities (Fig. 5). The bacterial community in the rhizosphere of SCC soil was related to the AP, that of TCC soil was related to soil TN and that of FCC soil was related to sucrase and OM. The soil pH and urease determined the pattern of the CK microbial communities (Fig. 5A). As expected, these six soil indexes also contributed to the composition of the fungal community (Fig. 5B). AP was identified as a primary explanatory factor responsible for the observed clustering pattern in the SCC rhizosphere fungal community. The TCC rhizosphere fungal community was strongly related to soil TN, urease and OM. The FCC rhizosphere fungal community was related to soil sucrase and pH. However, the driving factor for the formation of the CK fungal community was not identified. These results showed that for bacteria and fungi, microbial diversity in SCC soil was positively correlated with AP, and the diversities of the TCC and FCC rhizosphere communities were positively correlated with TN, OM and sucrase. Moreover, the continuous monocropping time had less correlation with soil pH and urease in fungal communities than in bacterial communities. AP is closely related to SCC soil and serves as a main explanatory factor for the diversity of rhizosphere bacterial and fungal communities in SCC soil.
Discovery of biomarkers of fields under different continuous cropping times
For the bacterial community, a linear discriminate analysis (LDA) effect size (LEfSe) [61] analysis identified 21, 16, 9 and 17 biomarkers for CK, SCC, TCC and FCC fields, respectively (Figs. 6A, S3A). The most abundant bacteria from CK soil belonged to Rhodobacterales. SCC fields were abundant in the bacteria Chloroflexi and Phycisphaerales. Bacilli were significantly enriched in TCC soil. Biomarkers in samples from the FCC field mainly comprised members of Betaproteobacteria, Myxococcales and Solirubrobacterales (Fig. 6A). For the fungal community, the LEfSe analysis identified 29, 27, 23 and 32 biomarkers for CK, SCC, TCC and FCC fields, respectively (Figs. 6B, S3B). For the CK sample, fungi that were relatively abundant included members of Kickxellomycota and Hymenochaetales. Biomarkers in samples from the SCC field included members of Tremellomycetes and Pezizomycetes. For the TCC field, fungi that were differentially abundant included members of Nectriaceae, Didymellaceae and Herpotrichiellaceae. The most differentially abundant fungi from the FCC field mainly included members of Basidiomycota and Glomeromycota (Fig. 6B).