The roles of microorganisms in enhancing crop production have been demonstrated for a range of cropping systems. Most studies to date, however, have been confined to a limited number of locations, making it difficult to identify general soil biotic and abiotic characteristics underpinning the yield-promotion across various locations. This knowledge gap limits our capacity to harness soil microbiome to improve crop production. Here we used high-throughput amplicon sequencing to investigate the common features of bacterial community composition, ecological networks and physicochemical properties in six yield-invigorating and adjacent yield-debilitating orchards. We found that yield-invigorating soils exhibited higher contents of organic matter than yield-debilitating soils and harbored unique bacterial communities. Greater alpha diversity and higher relative abundances of Planctomycetota and Chloroflexota were observed in yield-debilitating soils. Co-occurrence network analysis revealed that yield-invigorating soils displayed a greater number of functionally interrelated modules (meta-modules) and a higher proportion of negative links to positive links. Chloroflexota was recognized as a keystone taxon in manipulating the interaction of bacterial communities in yield-invigorating soils. Altogether, we provide evidence that yield-invigorating soils across a range of locations appear to share common features, including accumulation of soil organic matter, higher microbial diversity, enrichment of key taxa like Chloroflexota and maintaining a competitive network. These findings have implications for science-based guidance for sustainable food production.
Soils are essential to human well-being due to their great contributions to the production of food, fiber, feed and medicine (Raaijmakers and Mazzola, 2016). Soil organisms play critical roles in maintaining these ecosystem services, such as driving nutrient cycling, maintaining soil fertility, improving plant productivity and suppressing plant diseases (Bender et al., 2016; Barrios, 2007). Microorganisms participate in nearly all soil biological processes, and the microbial abundance, community composition and activity primarily determine the sustainable productivity of agricultural lands (Philippot et al., 2013). Fungi participate in decomposition of organic matter and deliver nutrients for plant growth (Frąc et al., 2018); however, considering that bacteria are the most diverse and abundant group of microorganisms in soil, bacterial communities and their functions can be pivotal indicators for crop production in agroecosystems (van der Heijden et al., 2008).
In general, an increase in microbial diversity is linked to a high-yielding crop production mainly through improving the host resilience to physical or chemical disturbances, modifying plant competition and facilitating plant access to nutrients (Chaer et al., 2009; Kennedy and Smith, 1995). Since individual organisms do not live in isolation but rather form a complex system of inter-species interactions in soil, interactions among community members were found to be related to crop production in the potato monoculture system (Lu et al., 2013). Enrichments of key functional microbes in soil were deemed to serve specific soil system functions, such as suppressing soil-borne pathogens and maintaining sustainable crop production (Banerjee et al., 2018). However, the relative contributions of microbial diversity, interactions among community members or enrichment of key taxa to crop production remain largely unknown. Therefore, it is highly desirable to identify pivotal indicators of bacterial community composition in response to high-yielding crop production.
Changes in composition of soil bacterial communities across space are often strongly correlated with soil pH (Fierer and Jackson, 2006). Soil pH has been recognized as a key driver in determining the assembly of bacterial communities in arable soils by field experiments (Rousk et al., 2010). However, recent studies have demonstrated that compositions of soil bacterial communities were driven by a myriad of soil abiotic traits, such as organic matter contents, forms and contents of soil nutrients (Tian et al., 2018; Wang et al., 2018). For example, soil bacterial community composition, which determines the ability of soil to suppress soil-borne pathogens, was found to be strongly correlated with soil organic matter (Shen et al., 2018). An imbalanced ratio of soil nutrients, i.e., the ratio of nitrogen to phosphorus or potassium, could be a driving force altering the bacterial community composition in long-term fertilized soils (Eo and Park, 2016). Key soil chemical properties identified in controlling the distribution and abundance of bacterial communities largely depends on the geographical distributions of soils. As a consequence, a better understanding of the relationship between soil edaphic properties and bacterial community composition is critical to develop targeted manipulation options to increase soil service provisions.
“Sucui No. 1” pear is an early-maturing variety bred by the Jiangsu Academy of Agricultural Sciences, China, and has been popularly cultivated in eastern and central China, due to certain advantages including being easy to produce, adaptable to the environment and having good quality and high economic benefits (Lin et al., 2013). With the increasing demand in China, sustainable production of high-quality pear is becoming increasingly important. Manipulation of soil microbiomes has shown to be an effective way to increase soil productivity (Chaparro et al., 2012). Considering that large-scale surveys could exhibit the diversity of soil microbial communities exceeds what is found in host-associated communities (Toju et al., 2018), it is necessary to explore the general microbial characteristics of multiple yield-invigorating soils and identify key environmental drivers in assembling bacterial communities.
In this study, orchards with higher pear yield production compared with local average yield was recognized as yield-invigorating (YI) orchards, while orchards having lower pear yield production in comparison with local average yield was regarded as yield-debilitating (YD) orchards. After field surveys accomplished in 2019, six separately located YI pear orchards and adjacent YD pear orchards were selected for further analysis of soil chemical properties and microbiome. We hypothesized that high input of organic fertilizer could improve soil structure and modify soil chemical properties, which leads to YI soils harboring unique bacterial communities associated with high-yielding pear production. To address this, soil bacterial communities and edaphic properties of the study sites were compared to (1) decipher the differences of taxonomic diversity and composition of the bacterial communities, and (2) determine the contributions of environmental variables to the changes in the structure of bacterial communities.
From July–August 2019, a field production survey of orchards cultivated with “Suci No. 1” pear was performed after pear fruits harvest to compare the differences of soil nutrients and microbiota between yield-invigorating with yield-debilitating orchards. The locations, planting density, cropping years, soil type and total yield were recorded. To minimize the effects of microclimate at each site, only pair-located pear orchards with invigorating and debilitating yield and at similar growth stage were selected for this research. In total, six pair-located yield-invigorating and yield-debilitating pear orchards distributed in four cities of Jiangsu province, China, were selected in the main pear production areas (Fig. 1; Table S1 in the Supplement). The yield per tree was obtained by dividing the total yield per hectare by plant density.
Distribution of studied field sites. Map showing the sites of six pair-located orchards sampled in this study.
Paired yield-invigorating and yield-debilitating orchards from Fengxian
(FX), Suining (SN) and Tongshan (TS) were maintained in the Xuzhou city
under the warm temperate subhumid monsoon climate. This site has a mean
annual temperature (MAT) of 14.5
Along with the field survey, soil sampling campaigns were performed from
July to August 2019 after pear fruits harvest. For each yield-invigorating or
yield-debilitating orchard, four subplots with three pear trees in each subplot
were randomly selected for soil sampling. Subsequently three soil cores
(0–20 cm) under the trunk base for each tree were collected using a 25 mm
soil auger. In total, nine soil cores for each subplot were pooled as a
composite sample and finally four composite soil samples for each orchard
were collected and promptly transported on ice to the laboratory. After
sifting through a 2 mm sieve and thoroughly mixing, one portion of each soil
sample was air-dried for chemical property analyses while the remainder was
stored at
Soil chemical properties, including soil pH, content of organic matter (OM),
total nitrogen (TN), available phosphorus (AP), available potassium (AK),
alkali-hydrolyzable nitrogen (N), exchangeable calcium (Ca), effective
magnesium (Mg), effective iron (Fe), effective manganese (Mn), effective
copper (Cu) and effective zinc (Zn), were measured according to methods
described by Shen et al. (2018) and Huang et al. (2019). Briefly, soil pH
was determined using a glass electrode meter in a suspension with a 1 : 5
soil
Genomic DNA from 0.25 g soil for each sample was extracted by using the
DNeasy® PowerSoil® Kit
(QIAGEN GmbH, Germany) according to the manufacturer's instructions. The
abundances of soil bacteria were determined with the Eub338F/Eub518R primer
using a 7500 Real Time PCR System (Applied Biosystems, USA). Standard curves
were generated by using 10-fold serial dilutions of a plasmid containing a
full-length copy of the 16S rRNA gene from
The gene-specific primers 515F/806R with 12 bp barcode were used to amplify the V4 region of bacterial 16S rRNA gene on the BioRad S1000 (Bio-Rad Laboratory, CA) according to the protocols described by Caporaso et al. (2011). All constructed libraries were sequenced using the Illumina NovaSeq 6000 at the Guangdong Magigene Biotechnology Co., Ltd. (Guangzhou, China).
Quality filtering of the paired-end raw reads was performed to obtain the
high-quality clean reads according to the Trimmomatic V0.33 (Bolger et al.,
2014) quality control process. Sequences were assigned to each sample based
on their unique barcode, after which the barcodes and primers were removed.
Paired-end clean reads were merged using FLASH V1.2.11 (Magoč and
Salzberg, 2011). Raw tags were processed to generate the final Amplicon
Sequence Variant (ASV) table file at 97 % pairwise identity according to
the QIIME2 pipeline (Bolyen et al., 2019). The nonbacterial and
mitochondrial ASVs and extremely low frequency ASVs (relative abundance
Statistical analyses were performed using the software SPSS 20.0 (SPSS
Technologies, Armonk, NY, USA) and R (
Principal Coordinate Analysis (PCoA) based on the Bray–Curtis distance was
performed in MOTHUR V1.38.1 (Schloss et al., 2009) and visualized by the
“ggplot2” package (Wickham and Chang, 2015) in R to explore the differences
in microbial community composition. Permutational multivariate analysis of
variance (PERMANOVA) was performed to evaluate the significant differences
of microbial community composition according to sample locations and orchard
yield using the “vegan” package in R. Microbial alpha diversity indexes
(Chao, Shannon) were calculated based on randomly resampled ASV abundance
matrices at the same depth (23 800 sequences) in MOTHUR. A Venn diagram was
generated based on the final ASVs to compare microbial community composition
between yield-invigorating and yield-debilitating orchard soils. The affiliations
of unique and shared ASVs in yield-invigorating and yield-debilitating soils were
compared to evaluate the differences in the bacterial community composition
and plotted using the “pheatmap” package
(
The phylogenetic molecular ecological networks (pMEN) were constructed using
the random matrix theory-based approach to explore the organization of
bacterial communities in yield-invigorating (YI) or yield-debilitating (YD)
soil samples. Potential ecological interactions among bacteria were
determined by modeling the microbial community using Molecular Ecological
Network Analysis (
In total, 1 622 858 16S rRNA sequences were retained after quality control and a total of 9394 ASVs were obtained for the 16S rRNA gene sequences based on 97 % similarity. Among the total 16S rRNA gene sequences, 159 ASVs with 74 372 sequences were classified as archaea while 9235 ASVs with 1 548 486 sequences were identified as bacteria. Among Bacteria, Acidobacteriota, Pseudomonadota, Chloroflexota, Planctomycetota and Actinomycetota were the most abundant phyla (Fig. S1 in the Supplement).
Soil chemical properties differed among the locations and orchard yield
types (Table S3 in the Supplement). On average, yield-invigorating orchards showed obviously
higher contents of OM, AP, Mg and Fe, and a lower content of Mn, in
comparison with those in yield-debilitating orchards. However, when taking
all sites together, only a higher relative abundance of OM, on average, was
observed in yield-invigorating orchards compared with that in
yield-debilitating orchards based on the Wilcoxon test (
Yield-invigorating orchards together displayed significantly higher
abundances of total bacteria than those in co-located yield-debilitating
orchards based on real-time PCR results (Fig. 2a). Meanwhile, bacterial
community compositions at the ASV level were significantly correlated to
pear yield (
Quantitation of the abundance of bacteria population, and
linkage of microbial composition to pear yield.
PCoA based on Bray–Curtis distance matrices clearly revealed location-based
differences in bacterial community compositions (Fig. 3a). Six distinct
groups representing samples from different locations (FX, GC, SN, TS, TX and
ZJ) were obviously separated and confirmed by the PERMANOVA test (
Overview of bacterial composition and alpha diversity.
The Venn diagram showed that 4540 ASVs occupying over 90 % of total sequences were shared between yield-invigorating and yield-debilitating orchards (Fig. 3c). Among these shared ASVs, the fold changes larger than 2 of ASVs in yield-invigorating compared with yield-debilitating orchards were recognized as potential responders linking to yield improvement. Surprisingly, none of these ASVs potentially linked to yield improvement were shared among six separated co-located orchards (Fig. S3 in the Supplement). A total of 2546 unique ASVs with 53 222 sequences were found in all yield-invigorating orchards while 2308 unique ASVs with 44 389 sequences were observed in all yield-invigorating orchards. Among these unique ASVs, almost 70 % of ASVs were shared between yield-invigorating orchards and yield-debilitating orchards; however, no shared unique ASVs were found among six separately located orchards. The affiliation of unique and shared ASVs at the phylum level exhibited that the Pseudomonadota, Planctomycetota, Chloroflexota, Acidobacteriota and Actinomycetota were the top five phyla (Fig. 3d).
At the phylum level, the relative abundances of bacterial dominant phyla varied across the location and orchard yield condition. Pseudomonadota, Acidobacteriota, Actinomycetota, Chloroflexota and Planctomycetota were the top five abundant phyla (Fig. 4). The mean abundance of Chloroflexota and Planctomycetota was significantly higher, while Firmicutes was significantly lower, in yield-invigorating orchards compared with yield-debilitating orchards based on the Wilcoxon test (Fig. S4 in the Supplement).
Key taxonomic groups in distinguishing yield-invigorating
(YI) and yield-debilitating (YD) orchards.
At a finer resolution, 967 genera were observed for all soil samples, among
which 299 genera appeared in more than half of soil samples in
yield-invigorating or yield-debilitating orchards; however, only 34 genera
displayed significant differences between yield-invigorating or
yield-debilitating orchard soils based on STAMP analysis (Fig. S5 in the Supplement).
Interestingly,
The yield-invigorating network contained 302 nodes, 448 edges and 11 larger
modules (
Co-occurrence networks of bacterial communities and identified
keystone taxa in distinguishing yield-invigorating (YI) and
yield-debilitating (YD) orchards.
Analysis using the threshold values of
Soil chemical properties were significantly correlated to the bacterial
community compositions (Mantel:
Relationships among bacterial communities, soil edaphic factors
and pear yield.
After forward stepwise selection, the module including soil OM, TN, alkaline N, AP and AK, available calcium (Ca), copper (Cu) and manganese (Mn) explained the majority of the variation in bacterial community composition (Fig. 6b). As evidenced by the RDA vectors, OM within the module was identified as the most important soil property that determines the composition of bacterial communities. Random forest analysis showed that contents of soil Mn, OM and Ca were the top parameters for predicting the orchard yield (Fig. 6c). Furthermore, soil OM was also significantly correlated with bacterial communities as revealed by the Mantel test (Fig. 6d; Table S6 in the Supplement).
Although pear is among the most important fruits worldwide, soil microbial communities in pear orchards have been largely under-investigated (Huang et al., 2019). The present study attempts to decipher the bacterial communities linked to high-yield production of pear. Our results based on Mantel analysis suggested significant correlations among bacterial communities, soil chemical properties and pear yield. Microbial characteristics responding to yield promotion have repeatedly been observed on several crops depending on single experimental site (Zhong et al., 2020; Qiao et al., 2019; Shen et al., 2013). It remained unclear, however, whether these distinctions are ubiquitous at a large-scale. By comparing multiple co-located yield-invigorating and yield-debilitating orchards, we demonstrate that high-yielding pear production soils exhibited high organic matter contents and harbored bacterial communities with high diversity, significantly enriched indigenous microbes and more interactive network, which was triggered by high inputs of soil organic fertilizer. Here we have discussed these main results and potential mechanisms in detail.
Microbial diversity is critical to soil ecosystems in maintaining the integrity, function and long-term sustainability (Kennedy and Smith, 1995). Higher soil biodiversity is considered to be linked to a more stable system and enhance the combination of vital microbial functions and processes (Cardinale et al., 2006; Bell et al., 2005). In line with a previous report that crop yield was correlated to the soil bacterial diversity (Zhao et al., 2014), greater diversity of bacterial communities in yield-invigorating soils was observed in the present study. Hence, we infer that higher microbial diversity may result in a more productive agroecosystem, contributing to sustainable pear production.
In this study, we found that Pseudomonadota, Acidobacteriota, Actinomycetota, Planctomycetota and Chloroflexota were the top abundant phyla. This result roughly agreed with previous studies showing that Pseudomonadota, Acidobacteriota and Actinomycetota are usually dominant bacterial taxa in agricultural soils (Xun et al., 2019; Dai et al., 2018), while Planctomycetota and Chloroflexota exhibit an unexpectedly high relative abundance in rice cropped soil (Edwards et al., 2015) and sandy loam soil (Pathan et al., 2021). The highest relative abundance of Pseudomonadota was probably explained by the fact that Pseudomonadota are considered as copiotrophic bacteria and flourish in soils with large amounts of available nutrients (Fierer et al., 2007).
Moreover, a significantly higher abundance of Planctomycetota and Chloroflexota was observed in yield-invigorating orchards, indicating that Planctomycetota and Chloroflexota may be associated with pear yield improvement. There is no direct evidence showing that Planctomycetota could improve plant growth. However, Planctomycetota has been reported to be involved in many soil biological processes such as ammoxidation as well as carbohydrate and polysaccharide metabolism (Fuerst, 2017). This implies that Planctomycetota may promote plant production through improving soil fertility. Chloroflexota is a facultative anaerobic phylum including autotrophic, heterotrophic and mixotrophic taxa (Speirs et al., 2019). Considering that soil amended with organic fertilizer may enhance the soil water holding capacity, the yield-invigorating soils with more organic material input have a higher soil moisture content, especially after irrigation, probably leading to the enrichment of Chloroflexota in soil.
Network analysis is a systems-level method to explore interactions within an ecosystem that cannot be directly observed through co-occurrence analysis (Fath et al., 2007). Similar to the food web network analyses in macro-ecosystems, microorganisms also form complex interactions with other species (Faust and Raes, 2012) and have been widely investigated to explore the linkage of microbial networks with soil function, such as nutrient supply (Fan et al., 2021) and disease suppression (Lu et al., 2013). Overall, in line with previous findings (Hu et al., 2020), the topological properties of the constructed networks, including connectivity, average clustering coefficients, average degree distance and modularity, indicate that these networks are scale-free, modular and small world. In short, a scale-free network represents a network whose connectivity follows a power law, and most of the nodes have only a few connections with other nodes, whereas a small-world network is a network in which most nodes are not neighbors but can be reached by a few paths. Modularity is a fundamental characteristic of biological network as a module in the network is a group of nodes that are highly connected within the group, but with very few connections outside the group (Deng et al., 2012). Our comparative network analysis indicated that microbial co-occurrence patterns in soils links to different pear production. As a meta-module is usually considered as a group of modules functionally interrelated (Langfelder and Horvath, 2007), a greater number of meta-modules were identified in the network constructed from yield-invigorating soils, suggesting that a greater number of network nodes in the yield-invigorating soils were functionally interrelated than those in the yield-debilitating soils. A majority of nodes in the meta-modules were not shared between yield-invigorating and yield-debilitating networks, indicating basal shifts in network architecture during pear production with contrasting yield performance.
Furthermore, a higher proportion of negative interactions to positive
interactions were identified in the network constructed from
the yield-invigorating network than the yield-debilitating network. Our results
indicated stronger resource competitions in yield-invigorating soils, which
means that the soil co-occurrence network was more stable to maintain soil
ecosystem function (Coyte et al., 2015). In our study, three module
connectors and three module hubs were identified as potentially key taxa in
the yield-invigorating network. Interestingly, among those key species,
ASV357 affiliated to
In this study, a significantly higher content of soil organic matter was observed in yield-invigorating orchards, demonstrating that soil organic matter could drive the assembly of bacterial communities. Consensus is emerging that microbial residues are an important constituent of soil organic matter (Kallenbach et al., 2016), which participate in almost all soil biological processes (Fierer, 2017). Despite that the quality of soil organic matter was not evaluated in this study, the quality of soil organic matter was associated with the diversity of microbial community (Ding et al., 2015), which implies that in our future work more attention should be paid to illustrate the relationship between the quality of soil organic matter and microbial community.
In conclusion, yield-invigorating soils displayed a higher content of organic matter and harbored unique bacterial communities with greater diversity than yield-debilitating soils. Further Chloroflexota was significantly enriched and identified as a potential keystone taxon in manipulating the interaction of bacterial communities in yield-invigorating soils. These findings indicated that soil organic matter triggered the assembly of soil microbiome, which both participated in maintaining crop production. Such knowledge is a first step toward harnessing soil microbiome in support of sustainable agroecosystems.
Raw amplicon sequencing data for each sample used in this study was deposited at the National Center for Biotechnology Information (NCBI) in the FASTQ format and is available under the accession number PRJNA749397. Other data that support the findings of this study are available on request from the corresponding author (Xiaomei Ye).
The supplement related to this article is available online at:
LW: performed all experiments; LW, XY and ZS: designed the study and wrote the majority of the paper; LW and ZS: analyzed the data; HH, JD, YX, JL and DC: participated in the design of the study, provided comments and edited the paper. All authors read and approved the final paper.
The contact author has declared that neither they nor their co-authors have any competing interests.
Publisher’s note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We sincerely thank all those who assisted with any part of this paper and all pear orchard owners for providing access to the soil sampling.
This research has been supported by the Jiangsu Agricultural Science and Technology Innovation Fund (grant no. CX(19)3094) and the National Natural Science Foundation of China (grant nos. 31801842 and 42090065).
This paper was edited by Axel Don and reviewed by two anonymous referees.