Drought, exacerbated by climate change, poses a significant threat to crop yields. Plants and their associated microbiota respond to decreasing soil water content, impacting plant fitness. However, a comprehensive understanding of the interplay between plant and microbial responses remains elusive, hindering targeted efforts to improve crop drought resistance. Numerous studies highlight the beneficial roles of Actinobacteria and Proteobacteria, as well as fungal endophytes and mycorrhizal fungi, in enhancing plant tolerance to drought and salinity stresses. These microbes employ diverse mechanisms, including modulation of plant stress genes, reduction of ethylene levels (via ACC deaminase), stimulation of osmolyte production, and epigenetic modification. Plants also exhibit direct responses to water stress through various genetic, molecular, and physiological mechanisms. The holobiont concept, encompassing the host and its microbiota as a single evolutionary unit, provides a valuable framework for studying adaptation to environmental stresses. This study hypothesizes that the transcriptomic response of the wheat holobiont to decreasing soil water availability will be predominantly microbial, reflecting the dynamic nature of the microbiome compared to the plant genome.

The literature extensively documents the contribution of plant-associated microbes to drought tolerance. Actinobacteria and Proteobacteria, known for producing ACC deaminase, have shown promise in improving plant resistance to drought and salinity. Fungal endophytes and mycorrhizal fungi also play crucial roles in enhancing plant performance under abiotic stress, mainly through improved water use efficiency. Microbes isolated from drought-prone environments frequently exhibit enhanced drought resistance properties when associated with plants. The mechanisms involved are varied, including the modulation of plant stress genes, ethylene reduction via ACC deaminase, stimulation of osmolyte production through bacterial volatile organic compounds, and plant epigenetic response alteration. While the hologenome theory emphasizes the significance of microbial communities in host biology, debates persist regarding its exact definition and implications. The current research builds on these previous findings, aiming to quantify the relative contribution of the plant host and its microbiome to transcriptomic changes under drought conditions.

A field experiment was conducted using rainout shelters to manipulate precipitation levels (25%, 50%, 75%, and 100% of natural rainfall) for two wheat genotypes (*Triticum aestivum* cv. AC Nass and *Triticum turgidum* spp. durum cv. Strongfield). The experiment was replicated across six blocks. Rhizosphere soil and root samples were collected from the Strongfield cultivar (drought-tolerant) on July 26, 2017, at the peak growing season. Soil water content was measured by weighing samples before and after drying. Total RNA was extracted from rhizosphere soil (using the RNeasy PowerSoil Total RNA Kit) and roots (using the RNeasy Plant Mini Kit), treated with DNase, and sequenced using Illumina HiSeq4000 (2 x 100 bp paired-end reads). Libraries for rhizosphere samples employed microbial ribosome subtraction, whereas root samples used a poly-dT reverse transcription approach. The raw data (BioProject accession PRJNA880647) underwent quality control using Trimmomatic and DUK. Reads were co-assembled using Megahit, transcripts were predicted using Prodigal, and annotated following JGI guidelines, including KEGG ortholog assignment. Abundance profiles were determined by mapping reads against contigs (using BWA mem), and differential abundance analysis was performed between the 100% and 25% precipitation treatments using EBSeq (FDR = 0.05). Statistical analyses were conducted in R.

The experiment successfully established a significant difference in soil water content (SWC) across the four precipitation treatments (p = 0.000367), with the lowest SWC at 11% (25% precipitation) and the highest at 23% (100% precipitation). Analysis focused on the two extreme treatments (25% and 100%). A total of 1,069,108,624 clean reads were obtained, assembled into 1,269,055 transcripts. In the roots, 42,001 transcripts (3.7%) were differentially abundant (DA) between treatments, with fungi accounting for 55.41% of these, mostly more abundant under 25% precipitation. Plants showed a contrasting pattern, with most DA transcripts being less abundant under 25% precipitation. In the rhizosphere, 21,765 transcripts (1.82%) were DA, with bacteria representing 65.14%, a majority being more abundant in the 25% precipitation treatment. Higher-level taxonomic analysis (phylum/class) and functional annotation (COG categories) revealed that specific taxa and functions were over- or under-represented among DA transcripts. Analysis of the top 50 most abundant DA transcripts in roots revealed a strong representation of Agaricomycetes (Coprinopsis cinerea), mostly with increased abundance under 25% precipitation, related to amino acid and carbohydrate transport. In contrast, plant transcripts were mostly less abundant under lower water conditions. In rhizosphere, the top 50 DA transcripts showed mainly Proteobacteria, many less abundant in the 25% condition. A comparison of shared DA transcripts between roots and rhizosphere identified 513 positive and 47 negative DA transcripts, with Actinobacteria heavily represented among the positive transcripts.

The results strongly support the hypothesis that the microbial partners exhibit a more significant transcriptomic response to decreasing soil water content than the wheat plant. This aligns with previous work on the willow holobiont, showing stronger fungal responses to soil contamination. The dynamic and plastic nature of the microbial metagenome, capable of rapid modification through changes in community composition, recruitment of new members, or horizontal gene transfer, contributes to this pronounced microbial response. Plant responses are limited to altered gene expression. The observed changes in plant gene expression could influence root exudation, potentially mediating the microbial response. The relatively mild drought stress in this experiment (SWC around 12% at lowest), and use of a drought-resistant wheat variety could explain the lack of a more pronounced plant response. Future research could compare responses in sensitive plant holobionts under more extreme conditions. The challenge of disentangling the contribution of microbial community shifts and gene expression changes to the observed metatranscriptomic patterns is acknowledged. However, the combined shifts likely impact holobiont adaptation to drought. The overrepresentation of Actinobacteria among positive DA transcripts in the rhizosphere contrasts with the underrepresentation of Proteobacteria and Acidobacteria, consistent with their differential water stress sensitivity. The abundance of amino acid and carbohydrate transport and metabolism genes in roots suggests a role in osmolyte production, a key adaptation to drought.

This study demonstrates that the microbial component of the wheat holobiont is the primary driver of the transcriptomic response to decreasing soil water content. The abundance of Actinobacteria and genes involved in osmolyte production are highlighted. The microbiome's dynamic response and beneficial potential highlight its central role in adaptation to water stress. Future research should focus on understanding the specific contributions of individual microbial taxa and genes to plant drought tolerance, and on manipulating the microbiome to enhance resilience in crop plants.

The study focused on a single drought-tolerant wheat genotype and a limited range of drought stress conditions. The difficulty of separating the effects of microbial community shifts and gene expression changes on metatranscriptomic profiles is acknowledged. The mild drought stress applied might not fully capture the plant's response under more extreme conditions.