DOI: https://doi.org/10.21203/rs.3.rs-2267712/v1
To ensure sustainable agricultural productivity, the crop's dependency on chemical fertilizers and pesticides has to be reduced; the crop should maintain its yield potential in varied marginal soil types, and the crop should be resilient against biotic and abiotic stresses. All these goals of agricultural sustainability can be achieved through effective plant-microbe interactions. Several reviews and concept papers have already shown the potential of plant microbiome on agricultural sustainability (de Souza et al. 2020; de Souza et al. 2016; Pozo et al. 2021; Ray et al. 2020). The complexity of plant microbiome in terms of community and functioning often does not allow to proceed the resources for application in agriculture.
The host plant recruits its microbiome from the surrounding environment to improve its growth and resiliency (de Souza et al. 2020). The crop's wild types, landraces, and ecotypes smartly chose their microbial partners through co-evolution processes for their benefit. In contrast, modern high-yielding cultivars and hybrids lose their partner due to breeding, agronomical, or management strategies (Nerva et al. 2022). Hence, the consequences of the domestication of plants into crop cultivars and hybrids on plant microbiomes were addressed recently (Chang et al. 2022; Martínez-Romero et al. 2020; Pérez-Jaramillo et al. 2018; Pérez-Jaramillo et al. 2016). The microbiome profile of cultivars and their wild relatives were compared in several crops, including rice (Chang et al. 2021), wheat (Abdullaeva et al. 2021), barley (Bulgarelli et al. 2015), tomato (Smulders et al. 2021), and soybean (Chang et al. 2019). The rhizo-microbiome of wild rice and high-yielding cultivars were profiled through the 16S rRNA sequencing approach (Chang et al. 2021; Peng et al. 2021; Shenton et al. 2016; Shi et al. 2019; Sun et al. 2021a; Tian et al. 2022; Zhang et al. 2022). All these reports cautioned the loss of microbiome diversity in cultivated plants, and the wild species got more benefits from the associated microbiome than the cultivars developed through human-centered breeding (Nerva et al. 2022). Consequently, the high-yielding cultivars depend heavily on external resources/inputs (chemical fertilizers and pesticides), while the landraces manage their needs through their microbial partners. Likewise, the landraces and native cultivars have more resilience to biotic and abiotic stresses than cultivars and hybrids (Pérez-Jaramillo et al. 2016). Hence, domestication induced microbiome changes in cultivars is not beneficial to the plants.
During the microbiome recruitment by a crop, soil types followed by cultivation methods play a crucial role, while the genotype-induced changes are trivial in most of the cultivated crops like wheat (Simonin et al. 2020), Medicago (Brown et al. 2020), soybean (Liu et al. 2019), and common bean (Stopnisek and Shade 2021). In contrast, detectable rhizo-microbiome changes were also documented among the genotypes of Saccharum spp. (Ishida et al. 2022), rice (Dhondge et al. 2022; Kim et al. 2022; Wu et al. 2022; Xiong et al. 2021), Arabidopsis thaliana (Urbina et al. 2018), and olive (Malacrinò et al. 2022). In Sorghum, the genes responsible for genotype-specific rhizo-microbiome were identified in Chromosome 4 loci (Deng et al. 2021). The cultivar and wild species or landrace had wide variability in rhizo-microbiome composition, while the cultivar's genotypes are less different microbiome compositions.
Rice is susceptible to drought due to its small root architecture and semi-aquatic growth habit (Hirasawa 1999). The mean yield loss of rice due to drought is 25.4% in India, and the yield loss varies depending on the crop stage. For example, drought during the vegetative stage cause 21-50.5% yield loss, the flowering stage accounts for 42-83.7% and drought during the entire reproductive phase of rice lead to 51-90.6% yield loss (Zhang et al. 2018). Rice uses several molecular, physiological, and morphological responses for drought mitigation after receiving water for growth (Pandey and Shukla 2015). Exploring the plant microbiome can improve the resistance by modulating the plant responses (Compant et al. 2010).
In this work, we compared the rhizosphere functioning and rhizo-microbiome recruitment of rice landrace, Norungan, and a high-yielding cultivar, Co51, under normal moisture and drought-induced conditions. Norungan is a potential landrace in Tamil Nadu state of India, known for its drought resistance and profuse growth in nutritionally limited soil conditions (Chandra Babu et al. 2001; Suji et al. 2012). It is the principal donor of drought-resistant genes for most established modern rice varieties (Divya 2020). Norungan has high zinc and iron accumulation (Anuradha et al. 2012) and possesses deep penetrating roots associated with plant production traits (Rajurkar et al. 2021). On the other hand, the Co51 is a high-yielding cultivar of Tamil Nadu, a derivative of a cross between ADT 43 and RR 272 -1745 characterized by short-duration and fertilizer-responsive (Robin et al. 2019). The questions we addressed in the present investigation are: 1) When the landrace and cultivar grown under same soil and environmental conditions, how do the biochemical attributes and microbiome community of rhizosphere vary between these two contrasting genotypes? 2) Does the moisture stress affect both the genotypes' rhizosphere uniformly? 3) Which bacterial taxa are common and different between these two genotypes under normal and moisture stress conditions? The answers to these questions through the soil biochemical and metagenomic analyses will provide the basis to understand the rice's genotype-mediated microbiome recruitment and functioning during stress conditions. The results could pave way for developing strategies to improve the microbiome-mediated drought mitigation of rice.
Rice genotypes
The drought-tolerant landrace Norungan and high-yielding cultivar Co51 were used in the present investigation. The characteristics of Norungan and Co51 are presented in Table 1 and Fig. 1.
Rice growth and soil sampling
The rice was grown under wetland conditions in cement tanks (1 m x 45 cm x 45 cm). The soils collected from the Tamil Nadu Agricultural University wetland farm, where no chemical fertilizers have been applied, were used for this experiment. The soil was characterized as clay loam textured with pH 7.97 (slightly alkaline), an electrical conductivity of 0.71 dS/m (non-saline), with available nitrogen 213 kg/ha; available phosphorus 31 kg/ha; available potassium 617 kg/ha, and soil organic carbon – 4.8 mg/g. The soil was filled in a cement tank, puddled with water, and paddy seeds were dibbled directly. After 15 days, two seedlings per hill were maintained. No exogenous chemicals, organic amendments, or chemical fertilizers were applied throughout the experimentation. The moisture stress was initiated on the 30th day by maintaining 50% field capacity for the next 25 days. The depletion of soil moisture on imposition of drought stress was determined by computing the moisture percentage of the soil randomly to maintain 50% field capacity. Hydration and dehydration were followed to maintain 50% soil moisture throughout the experiment. After 55 days of culturing, rice plants were uprooted randomly, the adhered soil was considered rhizosphere soil, and soil apart from roots was treated as bulk soil. The rhizosphere and bulk soil samples with three biological replications were collected on the 56th day. The samples were divided into two parts. One part was shade dried, powdered, and sieved using 2 mm for further soil analysis. The remaining part was packed in airtight sampling vials (HiMedia, India) and stored at 4°C and -20°C for biological attributes and metagenomic analyses, respectively.
Analyses of soil biological attributes
All the samples were analyzed for biological attributes using standard procedures. The potassium dichromate wet digestion method was adopted to estimate the soil organic carbon (Snyder and Trofymow 1984). The microbial biomass carbon (MBC) was measured by the fumigation-incubation technique (Jenkinson and Powlson 1976) and expressed as µg per g of soil. Soil labile carbon (SLC) was quantified by the permanganate method (Blair and Crocker 2000) and expressed as µg per g of soil. The substrate-induced respiration (SIR) was measured by the alkali-trap method as respiration rate using glucose as an amendment and expressed as µg of CO2 released per g of soil per h (Enwall et al. 2007). The basal soil respiration (without amendment) was calculated as µg of CO2 released per gram of soil per h (Stenberg et al. 1998), and metabolic quotient (qCO2) was calculated as the amount of basal respiration rate (µg C-CO2 per h) per unit of MBC of the corresponding sample (Dilly and Munch 1998). Dehydrogenase of samples was measured by the procedure described by Casida Jr et al. (1964) and expressed as µg of triphenyl formazan released per g of soil per h. Soil alkaline phosphatase was determined according to (Tabatabai and Bremner 1969) and reported as µg p-nitrophenol (p-NP) produced per gram of soil per h. Soil urease was determined according to Tabatabai and Bremner (1972) and reported as µg of NH4-N released per g of soil per h.
Soil microbiome analysis
The homogenized soil stored at -20°C (0.75 g) was used for high molecular weight metagenomic DNA extraction. The soil DNA was extracted using the Fast DNA spin kit for soil (MP biomedicals, Qbiogen, Cat. No. 6560-200) following the manufacturer's protocol. The purity of extracted DNA was checked spectrophotometrically (Nanodrop™ spectrophotometer, USA) and stored at -20°C until further use. The V3-V4 hypervariable region of the 16S rRNA gene was amplified from the samples using the primers Pro341F and Pro805R (Takahashi et al. 2014). The PCR amplicons of three technical replicates were pooled at the equimolar quantity and sequenced using the Illumina MiSeq platform at Oneomics (Tiruchirappalli, India).
The quality of raw reads such as base sequence, per tile sequence, per sequence GC content, per base N content, sequence length distance, sequence duplication, overrepresented sequence, and adapter content were checked using Fast QC to obtain high-quality reads. Adapter sequences were removed, and read pairs were quality trimmed and merged. Data processing was performed using QIIME2, version 2021.4. The quality of trimmed sequences was assessed using the deblur plugin within QIIME2. For quality control, the reads were subject to filtration, de-replication, reference‐free chimera detection, and paired‐end reads merging. Greengenes database was used to classify each identical read or Amplicon Sequence Variant (ASV) to the highest resolution. The reads classified as mitochondria and chloroplast were filtered out, and the unassigned ASVs were retained. The processed reads were clustered into operational taxonomic units (OTUs) and further explained at the species level.
Operational taxonomical units (OTUs) and taxonomical affiliation analyses of 16S rRNA gene V3-V4 regions were conducted in QIIME2 (Edwards et al. 2015). The raw sequences obtained were quality-checked using the fastq tool. Then the sequences were imported to QIIME2. The sequences were denoised via deblur tool, which merges and denoises the paired-end reads and produces the table and representative sequences. The feature table and summary were generated, providing the sequence information associated with the samples and features. A representative sequence for each OTU was screened for further annotation. The pre-trained naive Bayesian classifier (SILVA database) was used to classify each read or ASV. To get the phylogenetic relationship of all OTUs, the representative sequences were subjected to alignment using MAFFT. The informative sites were masked, and the phylogenetic tree was generated. All the data were rarefied randomly, with the smallest sample size of 5283 sequences (Supplementary data-1).
Statistical analysis
All the data collected from soil analyses were statistically analyzed in R software (Version 4.1.1) (R Core Team, Vienna, Austria). Three-way analysis of variance (ANOVA) was performed for soil biological attributes with genotype, moisture stress, and niche as factors to determine the significant difference between the samples. Tukey's honestly significant difference test (Tukey's test) at p = 0.01 was used for multiple comparisons of means. Principal component analysis (PCA) was performed for all the assessed variables of soil samples using the princomp function of the factoextra-package of R. The PCA biplot, contribution plot, and eigenvalues corresponding to the variation explained by each principal component were visualized using the viz function of factoextra.
The OTUs, taxonomy tables, and metadata (details about the samples) of microbiome analysis were imported to R software (Version 4.1.1), and further analyses were performed. Alpha diversity, representing the complexity of species diversity within a sample through the Shannon index, was employed in the data using the plot_richness function of the phyloseq package. A significance test was performed between samples using the pairwise Wilcox test. Beta diversity, representing the difference between samples in species complexity, was performed using genotype, moisture stress, and niche as three factors. Bray-Curtis distance was calculated for the samples, and the distances were ordinated as principal coordinate analysis (PCoA) in distance and plot_ordination functions of R package phyloseq. The phylum-level abundance was extracted through the tax_glom function of the phyloseq package and visualized as a stacked bar chart using ggplot (tidyverse package of R). VennDiagram package was adopted to display the distribution of unique and shared OTUs among different samples. Differential abundance analysis was performed on microbiome data to identify the differentially abundant genus between genotypes, moisture stress, and niches. The R-package DEseq2 was used to extract the differences and visualize as a heatmap using the package ComplexHeatmap.
Soil biological attributes
The soil organic carbon (SOC) was significantly influenced by the genotype, moisture stress, and niches. SOC was significantly higher in Norungan soils (2.4%) than Co51 irrespective of stress and niche, and SOC was reduced by 3.3 percent during drought conditions compared to normal conditions. The rhizospheric soils were observed to have an increase of 20.76 % in SOC content regardless of crop and stress. Norungan rhizospheric soil under normal conditions had the highest SOC (7.33 mg/g), followed by the Norungan rhizospheric soil under drought conditions. Co51 rhizospheric soil under normal conditions and Norungan rhizospheric soil under drought conditions were at par with each other. The least SOC content was recorded in Co51 drought bulk soil (6.03 mg/g) and Norungan normal bulk soils (6.13 mg/g) (Fig. 2A). Norungan accounted for 18.15% higher MBC than Co51 rice (Fig. 2B), drought-induced soil accounted for 3.6% more than normal soil, while the rhizosphere soil had 3.75% higher MBC than bulk soil. Norungan rhizospheric soil under normal conditions recorded a maximum MBC of 263.6 µg/g, followed by bulk soils of Norungan under moisture stress conditions (258.3 µg/g). The drought-induced rhizospheric soils of Co51 had 4.7% higher MBC (251.25 µg/g) than drought-induced rhizospheric Norungan soil (239.65 µg/g) (Fig. 2B). The flex of soil labile carbon was similar to SOC, where Norungan soils had 1.52 mg/g of labile carbon, which was 14.47% higher than Co51 crop soil (1.30 mg/g) (Fig. 2C). Among all the samples, Norungan rhizospheric normal soil had the highest labile carbon content of 2.07 mg/g followed by Norungan rhizospheric drought soil and Co51 rhizospheric normal soil. The least labile carbon release was observed in bulk soil samples of co51 under drought and normal conditions with 1.17 g/kg and 1.24 g/kg, respectively (Fig. 2C).
Norungan accounted for 6.94% higher substrate-induced respiration than Co51 (Fig. 2D), and drought-induced condition was 6.27% higher than normal. In comparison, the rhizosphere accounted for 6.9% higher than bulk soil. Norungan drought rhizosphere soil accounted for 3.93 µg of CO2/g as substrate-induced respiration rate was the highest, followed by Norungan drought bulk (3.39 µg of CO2/g) and Co51 normal rhizosphere soil (3.39 µg of CO2/g). The bulk soils of Co51 and Norungan had no apparent change in the SIR (Fig. 2D). The metabolic quotient (the base respiration being maintained by the soil microorganisms) was quantified for the soil samples and presented in Fig. 2E. Norungan accounted for significantly lower qCO2 (mean 1.08) than Co51 (mean 1.04); drought condition lowered the qCO2 than normal condition, and rhizosphere soil accounted 1.14 µg CO2/g, while bulk soil had qCO2 of 0.99 µg CO2/g, which was 15% lower. Among the rhizospheric soils under normal conditions, Norungan had a higher metabolic quotient (1.11 mg of CO2/g) than Co51 (1.16 mg of CO2 /g).
All three factors significantly influenced soil dehydrogenase. Norungan accounted for 6.29% higher dehydrogenase than Co51, normal moisture conditions accounted for 13% more activity than drought, and rhizosphere accounted for 26% higher dehydrogenase than bulk soil. Rhizosphere soil of Norungan under normal moisture accounted for significantly highest dehydrogenase, followed by the same sample under drought and Co51-rhizosphere soil under normal conditions. The bulk soils of Co51 and Norungan accounted for the least activity (Fig. 2F). Alkaline phosphatase of soil samples collected from Norungan and Co51 under normal and drought conditions were presented in Fig. 2G. All three assessed factors significantly altered the alkaline phosphatase activity of the soil, and among the rice genotypes, Norungan accounted for the higher alkaline phosphatase than Co51; normal moisture < drought conditions; rhizosphere soil < bulk soils. The maximum phosphatase activity was recorded in Norungan rhizosphere soils grown with normal moisture conditions (2.79 µg PNP/g/h) followed by the same under drought conditions (2.50 µg PNP/g/h) and Co51 normal conditions (2.59 µg PNP/g/h) (Fig. 2G). The rice genotypes, moisture-stress, and ecological niche significantly influenced the soil urease activity. Norungan accounted for 17.87% higher urease than Co51; normal moisture conditions had 29.87% higher than drought-induced conditions, and rhizosphere soil recorded 24.64% more urease than bulk soil. Among all the three interactions, the normal moisture-regime grown Norungan rhizosphere soil recorded the maximum urease activity followed by the same condition with Co51. The rhizosphere soil of Co51 had significantly lower urease than Norungan under drought conditions (Fig. 2H).
The results of the principal component analysis of assessed soil variables for the soil samples collected from rhizosphere and bulk soils of two-different rice cultivars (Norungan and Co51) under normal and moisture-stress conditions are presented in Fig. 3. All the assessed variables except available nitrogen and available phosphorus positioned in the positive quadrant for PC1 (Fig. 3A). The PC1 (Dim1) contributed 43.6% variability, while the PC2 (Dim2) added additional variability of 16.4% to the total cumulative variability (Fig. 3B). The significance of the contribution of variables to PCs was presented in Fig. 2C. Dehydrogenase, phosphatase, SOC, urease, metabolic quotient, available phosphorus, and soil labile carbon were the significant contributors for the variability of the samples. The scoring plots showing the orthogonal positions of samples as differentiated with each factor were presented in Fig. 3D-F. The overall comparison between two different genotypes of rice did not show much difference, as both the samples over-lapped each other (Fig. 3D). When the drought was enforced, the discrimination of samples due to assessed variables was distinct, as the normal samples grouped separately from the drought-induced samples (Fig. 3E). Similarly, the rhizosphere soil samples showed clear discrimination with bulk soil, irrespective of genotypes and moisture-stress conditions (Fig. 3F). The PCA results indicated that the assessed variables discriminated the moisture-stress and ecological-niche of soils, but not the genotypes of rice.
Rhizosphere microbiome
The metagenomic DNA extracted from the rhizosphere and bulk soils of Co51 and Norungan exposed to drought and normal moisture conditions were sequenced in the V3-V4 region of the 16S rRNA gene by Illumina sequencing. In pre-processing steps after sequencing, non-targets, chimeric sequences, and low-quality bases were removed and retained, and a total of 323,913 good-quality unique reads were obtained. Using the QIIME2 pipeline, a total of 5340 operational taxonomic units (OTUs) were obtained with 97% similarity. Rarefaction curves of all the samples confirmed that sufficient coverage of OTUs was obtained in all the samples.
The Shannon index, representing the alpha diversity (within samples), showed that bulk soils had higher diversity than the rice rhizosphere soils, irrespective of genotype and stress conditions (Fig. 4A). Under normal moisture conditions, Norungan reduced its rhizosphere microbiome diversity significantly. In contrast, the Co51 showed a trivial reduction. Under drought-induced conditions, Co51 and Norungan showed significantly low diversified rhizosphere than bulk soils. When comparing normal and drought-induced conditions, the Norungan rhizosphere showed a significant difference from Co51. Bray-Curtis distance-based principal coordinate analysis indicated that the bacterial community clustered distinctly based on drought conditions. When the rice was under normal conditions, the discrimination of rhizospheres from bulk was insignificant. While drought stress led to a significant shift in the rhizosphere of Norungan and Co51 (Fig. 4B). The rice soils accounted for 33 phyla, of which 10 phyla were major abundant in all the samples. In comparison, the remaining 23 phyla had less than 1% of proportions in the total abundance (Fig. 4C). Proteobacteria (37.50%) followed by Firmicutes (16.20%) and Actinobacteria (15.59%) were the predominant phyla across the samples. In Norungan, the abundance and richness of Firmicutes decreased drastically by two-fold during drought compared to normal moisture conditions. The Bacteroides abundance was also reduced by 57% due to drought in the Norungan rhizosphere. On the other hand, Acidobacteria (51%), Actinobacteria (54%), Chlolorflexi (41%), and Proteobacteria (34%) were increased in the drought-induced rhizosphere than the normal rhizosphere of Norungan. In Co51, the Firmicutes (79%) and Bacteriodetes (170%) abundance increased due to drought, while the abundance of Acidobacteria (15%) and Proteobacteria (31%) decreased in the drought-induced rhizosphere.
The occurrence of unique OTUs with more than one read in any sample was summarized to reveal the bacterial community of rice as influenced by cultivars (Co51 and Norungan), stress conditions (normal and drought), and niches (rhizosphere and bulk). The Venn diagrams showed the distribution of bacterial OTUs within and between cultivars (Fig. 5A). In Co51, 116 unique OTUs and 401 shared OTUs with bulk soils were observed in the rhizosphere under normal conditions. When the drought was induced, the Co51-rhizosphere accounted for 105 unique and 277 shared OTUs. In the case of Norungan, the normal moisture rhizosphere had 124 unique, 199 shared OTUs, and drought conditions increased the unique OTUs to 137 in the Norungan rhizosphere. The Venn diagram comparing Co51 and Norungan revealed that the unique OTUs of Co51-rhizosphere remained unchanged between normal (148) and drought (147), while Norungan had a significant reduction in unique OTUs from normal (208) to drought (132). Interestingly, the common OTUs of Co51 and Norungan increased from 161 (under normal moisture conditions) to 226 (under drought-induced conditions).
Differential abundance analysis was performed for the OTUs across the samples to identify differentially abundant taxa between groups. Among the OTUs, 11 genera belonging to 3 phyla (Bacteriodetes, Firmicutes, and Proteobacteria) were recorded to have significantly different abundance (Fig. 5B). The genera Alloprevotella, Acidaminococcus, and Sutterella, were grouped in drought conditions indicating higher abundance in drought-induced rhizosphere soils of Co51 and Hydrogenoanaerobacterium, Sedimentibacter, and Veillonella were recorded to have higher matrix value in Norungan non-rhizosphere soils. The remaining genera, viz., Brevundimonas, Bacillus, Providencia, and Morganella, formed a cluster group highly present in the samples of Norungan soils, irrespective of niche. Among the three phyla, the genus-level contribution in differential abundance was recorded as Firmicutes (60%), Bacteroidetes (20%), and Proteobacteria (20%).
Though intensive agriculture ensures food security worldwide, it affects soil health, water availability, human health, animal health, and environmental health (Jian et al. 2020; Lehmann et al. 2020). Apart from this, the crop's resilience to biotic and abiotic stresses under intensive agro-systems is low, as the cultivars and hybrids developed for the present agricultural systems have neglected the plant-associated microbiome. The root-associated microbiome and its functioning are considered significant factors affecting crop growth, yield, and resistance against biotic and abiotic stresses under sustainable agriculture (Nerva et al. 2022). The landraces recruit their rhizo-microbiome depending on the soil and environmental factors, while the high-yielding cultivars fail to form a potential host-microbiome interaction, hence unsuccessful in performing when unexpected drought, pest, and diseases occur (Chang et al. 2022). Therefore, understanding the mechanisms of microbiome interaction by landraces will help to develop interventions for the growth and fitness of high-yielding cultivars to ensure sustainability. Though several reports on the impact of cultivars and moisture-stress conditions on rice rhizosphere are available, no comprehensive report on both soil biochemical attributes and rhizo-microbiome analyses was documented yet. The present study provides vital information on rhizosphere functioning and rhizosphere-associated microbial communities of two contrasting genotypes of rice grown under the same soil conditions after a 25-days long drought. Though most of the biological attributes of rhizosphere soil got affected due to drought, the landrace Norungan assimilated resilience and accounted for less impact than the high-yielding cultivar, Co51. Likewise, the rhizosphere microbiome shift also varied between Norungan and Co51 when the crops were exposed to drought. We also identified the differential microbial taxa of each genotype that may possess specific ecological functioning in their host plants.
The comprehensive results of the present work on rhizosphere's biochemical attributes revealed that most of the assessed variables, including soil organic carbon, biomass carbon, labile carbon, and enzymes like dehydrogenase, urease, and phosphatase, were significantly higher in Norungan rhizosphere than in the Co51 rhizosphere. When the drought was introduced, all these assessed attributes decreased considerably in both genotypes' rhizosphere. However, resilience in carbon pools and biochemical attributes was observed in the Norungan rhizosphere but not in the Co51.
The soil carbon pool is composed of soil organic and inorganic carbon, which plays a vital role in the carbon cycle. Soil organic carbon equilibrium is governed by several interacting factors such as temperature, moisture, texture, quantity, and quality of organic matter, methods of organic matter application, soil tillage, cultivation methods, and cropping system (Vineela et al. 2008). In the present investigation, the rhizosphere's organic carbon had a significant impact due to genotypes and moisture stress, and Norungan > Co51 and normal > drought were the trend observed. Microbial biomass carbon, the measure of the living component of soil organic carbon, is the potential indicator representing soil's microbial activity and nutrient dynamics (Anderson and Domsch 1985). In the present work, the landrace Norungan had significantly higher biomass in the rhizosphere than Co51, indicating that the Norungan rhizosphere harbors more microbiome than Co51. The drought affected the soil MBC, irrespective of crop genotype and niche. This implies the detrimental impacts of drought on soil microbial activities. The labile carbon represents the readily available carbon to be utilized by the soil microorganisms. An increased soil labile carbon level indicates high microbial proliferation, soil health, and fertility (Ramírez et al. 2020). As the root exudates comprise a high proportion of readily available carbon substrates, the labile soil carbon would always be higher in the rhizosphere than bulk soil, irrespective of crop or genotype (Bhattacharyya et al. 2019). In the present work, the Norungan rhizosphere accounted for nearly 30% higher SLC than Co51, indicating that the landrace Norungan releases much higher levels of exudation than Co51, which was used as a trap to attract soil-borne microorganisms and thereby formed diversified and functionally active microbiome than Co51. The substrate-induced respiration and metabolic quotient of rhizosphere soils of Co51 and Norungan were compared with bulk soils and moisture stress conditions. These two respiration indices are potential sensitive indicators of soil microbial activity and are used to monitor soil microbial metabolism (Anandakumar et al. 2022; Tamilselvi et al. 2015). The substrate-induced respiration rate is high when the soil's microbial population is active. When the microbes receive adequate resources, the metabolic quotient (basal respiration) is low. In other words, a low metabolic quotient indicates the minimal energy required to maintain microbial function (Anderson and Lebepe-Mazur 2003). In the present work, the rhizosphere soils of Co51 and Norungan accounted for higher substrate-induced respiration and lower metabolic quotient than bulk soils. The difference between Norungan and Co51 in terms of respiration indices is trivial. However, when drought was induced, the Norungan had a higher response in respiration rates than Co51, indicating that the drought-responsive rhizosphere functioning is higher for landraces than cultivars. Soil enzymes, viz., dehydrogenase, urease, and phosphatase, are sensitive indicators of soil health, and the rhizosphere always accounted for higher activities than bulk soils (Xu et al. 2022). In the present investigation, the landrace Norungan rhizosphere accounted for significantly higher enzyme activities than Co51, and drought had a significantly deleterious impact on these enzymes for Co51 but less for Norungan. These results are in accordance with the other biological attributes reported in this study.
Rhizosphere functioning is essential for every plant as it regulates its nutrients and water flow. The natural plant recruits its microbiome in the rhizosphere to do these functions through a synchronized approach. However, when high-yielding cultivars are developed through breeding programs, we omitted these rhizosphere-mediated traits; hence, the rhizosphere functioning may not be as expected. Under that circumstances, the crop heavily depends on synthetic chemical fertilizers for readily available nutrients. Further, the plant cannot withstand the biotic and abiotic stresses due to the less-functional rhizosphere (Abdullaeva et al. 2021; Chang et al. 2022; Nerva et al. 2022; Ray et al. 2020; Sun et al. 2021b). In the present work, we showed that the drought-tolerant landrace Norungan and high-yielding cultivar Co51 performed more or less similar rhizosphere functioning under normal conditions. Still, when drought was enforced, the Norungan rhizosphere showed resilience and maintained its rhizosphere functioning as that of normal moisture condition. The high-yielding cultivar Co51 exhibited less rhizosphere functioning under drought conditions. The possible reasons for the drought-affected rhizosphere functioning are (1) drought-induced osmotic stress causing cell lysis and microbial death, significantly reducing the microbial biomass and its functioning (Turner et al. 2003). In addition, the root exudation of the host plant affected both qualitatively and quantitatively and led to the loss of nutrient resources for the microbiome (Williams et al. 2020). The consequences of these two effects were noticed in the present work. The microbial biomass and related variables, viz., respiration and enzymes, got reduced in drought-induced soil compared with normal moisture soil, irrespective of rhizosphere and bulk soil. Further, the functioning of the Norungan rhizosphere got plasticity in drought-induced conditions but not in the Co51 rhizosphere. Norungan grew without reducing the plant biomass and root system when the drought was introduced. Hence the rhizosphere functioning was less affected. In contrast, Co51 got affected by its root system by drought, which explained the reason for significantly reduced biological attributes. Similar consequences were reported in the rhizospheres of different grass species (Bouasria et al. 2012; de Vries et al. 2019; Sanaullah et al. 2011; Wang et al. 2021; Xiao et al. 2022; Xue et al. 2017) and maize (Zhang et al. 2021a; Zhang et al. 2021b) at drought-induced conditions.
Rice harbor diversified microflora in the rhizosphere, rhizoplane, phyllosphere, and endosphere, which includes nitrogen fixers, nitrifiers, plant growth regulators, methanogens, methane oxidizers, sulfur oxidizers, mineral solubilizers, and decomposers (Kumar et al. 2017; Prasanna et al. 2012). The soil's physico-chemical properties, climatic factors, and cultivation methods play a significant role in shaping the rice microbiome (Kim et al. 2020; Xu et al. 2020). The genotypes also play a crucial role in shaping the rice microbiome (Edwards et al. 2015). The predominant phyla reported across different rice cultivars [six cultivars, viz., Nipponebare, IR50, M109, 93-11, TOg7102, and TOg7267] are Acidobacteria, Actinobacteria, Bacteroides, Chloroflexi, Firmicutes, Gemmatimonadetes, Nitrospirae, Planctomycetes, Proteobacteria, and Verrucomicrobia (Edwards et al. 2015). The japonica rice cultivar Koral also accounted for the same groups of phyla (Hernández et al. 2015). Xiong et al. (2021) identified the common and differentially abundant rhizosphere microbiome of two cultivars and one hybrid rice. In the present work, we documented Acidobacteria, Actinobacteria, Chloroflexi, Firmicutes, Gemmatimonadetes, Nitrospirae, Proteobacteria, Rokubacteria, and Verrucomicrobia as predominant phyla in Norungan and Co51 rice rhizospheres. The results revealed that the bulk soils and rhizosphere soils shared a common microbiome but varied in abundance. This implies the recruitment of the rhizosphere microbiome from bulk soils, and we have not found any specific phyla to the rhizosphere alone. The Norungan rhizosphere accounted for a decreased proportion of proteobacteria and an increased proportion of Firmicutes due to drought, while in the Co51 rhizosphere, this change of phyla is insignificant.
Microbe-mediated drought mitigation of crops through rhizosphere harnessing is a complementary approach to reducing crop loss due to drought (de Vries et al. 2020). When the drought is enforced on a crop, compartment-specific reconstruction of microbiota occurs in rice (Santos-Medellín et al. 2017). A predominant increase of Actinobacteria, Chloroflexi, and Firmicutes due to drought was reported in rice across different genotypes (Santos-Medellín et al. 2017; Santos-Medellín et al. 2021; Zhang et al. 2018). In the present work, we observed that the Norungan rhizosphere drastically reduced its Firmicutes due to drought, while the Co51 rhizosphere remained unchanged. In contrast, the Proteobacteria got enhanced abundance due to drought in Norungan, while the Co51 accounted for the reduced abundance of Proteobacteria. We also reported that the rhizosphere microbiome of rice remained the same when the genotypes were grown under normal conditions. Still, when drought-induced, the landrace and cultivar responded differentially in accommodating their root microbiome, especially the recruitment of Proteobacteria, Actinobacteria, Firmicutes, and Acidobacteria. Among the two genotypes, the landrace Norungan showed significant resilience in its rhizosphere functioning, assuming that the microbiome recruitment during drought is more prominent than in the high-yielding cultivar. High-level of conservation and microbiome shift by Norungan suggest that the recruitment strategy of landraces, wild species, and native plants in rhizosphere microbiome assemblage is better than high-yielding cultivars and hybrids (Santos-Medellín et al. 2021).
Our study affords a comprehensive description of the effect of drought stress on rhizosphere functioning and root-associated microbiomes of two contrasting rice genotypes. Our results revealed that the rice landrace Norungan mitigates the drought through its microbiome recruitment, thereby maintaining its rhizosphere and growth as normal moisture conditions. Hence, it is suggested from our study that the rhizosphere functioning and rhizo-microbiome analyses may be included as essential traits in breeding programs for developing drought-tolerant rice varieties. In addition, the drought-responsive microorganisms with plant-growth promotion and drought mitigation from Norungan can be explored for developing synthetic microbial community inoculants as a mimic to the natural rhizosphere microflora for sustainable rice production.
Author Contributions
The experiments were conceived and designed by NSA and DB. The samples were processed, and the experiments were performed by NSA. Statistical analyses of the data were done by NSA and DB. The manuscript was prepared by NSA and DB.
Acknowledgment
Nunna Sai Aparna Devi acknowledges the Department of Science and Technology, New Delhi, India, for INSPIRE fellowship to carry out the present work [DST/INSPIRE/03/2018/001687].
Conflict of interest
The authors declare no conflict of interest.
Data availability
The sequences (V3-V4 region of 16S rRNA gene) obtained from this study were deposited to the NCBI database under Sequence Read Archive (SRA) with project number PRJNA900492. Link: www.ncbi.nlm.nih.gov/bioproject/PRJNA900492
Table 1 Characteristics of two contrasting rice genotypes used in the present investigation
Trait |
Norungan |
Co51 |
Type |
Landrace |
Cultivar |
Parentage |
- |
ADT 43/ PR 272-1745 |
Special trait |
Drought tolerant |
High yielding |
Grain type |
Long, dark brown thick |
Medium slender |
Kernel |
Dark brown |
White |
Root system |
Deep penetrating |
Thin hairy root |
Variety |
Coarse |
Fine |
Flowering (days) |
88 |
75-80 |
Height (cm) |
104-110 |
90-100 |
Panicle length (cm) |
19 |
23-28 |
Grains/ panicle |
150 |
250-300 |
1000 seed weight (g) |
34 |
16 |
Duration (days) |
125-140 |
105-110 |
Crop |
Rainfed |
Rainfed |
Stature |
Short/medium (100-115 cm) |
Semi-dwarf (90-100 cm) |
Basal leaf |
Green with purple lines |
Green |
Resistance |
Plant hopper and leaf hopper |
Moderate resistance against blast, brown plant hopper, green leaf hopper |
Yield (kg/ha) |
2810 |
6623 |
Season |
September-October, May-June |
June-July / September-October |
Reference |
Keerthivarman et al. (2019) |
Robin et al. (2019) |