Skip to main content
NCBI home page
eLife logoLink to eLife
. 2024 Jun 21;12:RP84747. doi: 10.7554/eLife.84747

Arabidopsis transcriptome responses to low water potential using high-throughput plate assays

Stephen Gonzalez 1,†, Joseph Swift 1,†,✉, Adi Yaaran 2, Jiaying Xu 1, Charlotte Miller 1, Natanella Illouz-Eliaz 1, Joseph R Nery 3, Wolfgang Busch 1, Yotam Zait 2, Joseph R Ecker 1,3,4,✉
Editors: Dominique C Bergmann5, Jürgen Kleine-Vehn6
PMCID: PMC11192529  PMID: 38904663

Abstract

Soil-free assays that induce water stress are routinely used to investigate drought responses in the plant Arabidopsis thaliana. Due to their ease of use, the research community often relies on polyethylene glycol (PEG), mannitol, and salt (NaCl) treatments to reduce the water potential of agar media, and thus induce drought conditions in the laboratory. However, while these types of stress can create phenotypes that resemble those of water deficit experienced by soil-grown plants, it remains unclear how these treatments compare at the transcriptional level. Here, we demonstrate that these different methods of lowering water potential elicit both shared and distinct transcriptional responses in Arabidopsis shoot and root tissue. When we compared these transcriptional responses to those found in Arabidopsis roots subject to vermiculite drying, we discovered many genes induced by vermiculite drying were repressed by low water potential treatments on agar plates (and vice versa). Additionally, we also tested another method for lowering water potential of agar media. By increasing the nutrient content and tensile strength of agar, we show the ‘hard agar’ (HA) treatment can be leveraged as a high-throughput assay to investigate natural variation in Arabidopsis growth responses to low water potential.

Research organism: A. thaliana

Introduction

As climate change advances, improving crop drought tolerance will be key for ensuring food security (Godfray et al., 2010; Battisti and Naylor, 2009). This has led to intense research at the molecular level to find novel loci and alleles that drive plant responses to drought conditions. Such investigations benefit from simple assays that can reproduce drought phenotypes at both the physiological and molecular levels. While some researchers use soil-based assays, these are cumbersome. For example, extracting intact root systems from the soil is difficult, and reproducing the rate at which water evaporates from the soil can be challenging (Dubois and Inzé, 2020). In light of this, chemical agents such as polyethylene glycol (PEG), mannitol, or salt (NaCl) are often employed to induce drought stress. When present in aqueous or agar media, they allow precise and dose-dependent control of water potential (Claeys et al., 2014; Hohl and Schopfer, 1991). When exposed to these media types, plants exhibit the hallmarks of drought physiology, such as reduced growth rate, reduced stomatal conductance, and increased leaf senescence (Claeys et al., 2014; Munns, 2002; Jisha and Puthur, 2014). While each of these methods lower water potential and thus induce a drought stress, each method exerts additional and distinct effects. For example, NaCl not only induces osmotic stress, but can cause salt toxicity (Munns, 2002). Since it is not metabolized by most plants mannitol is considered less toxic (Dubois and Inzé, 2020), however evidence suggests it may act as a signaling molecule (Hohl and Schopfer, 1991; Trontin et al., 2014). Since both NaCl and mannitol can enter the pores of plant cell walls (Verslues et al., 2006; Juenger and Verslues, 2023), they can induce plasmolysis, a process that does not typically occur under mild water deficit (Verslues et al., 2006). Due to its higher molecular weight, PEG treatment avoids this, and instead causes cytorrhysis (Verslues et al., 2006), a physiology more common under drought settings (Juenger and Verslues, 2023; van der Weele et al., 2000).

The unique impacts PEG, mannitol, and NaCl have on plant physiology may also extend to the level of gene expression. Indeed, a broad spectrum of transcriptional changes are documented in response to low water potential, which may be attributed to the specific method employed (Claeys et al., 2014; Zeller et al., 2009; Kreps et al., 2002). Here, we examine the transcriptional responses to PEG, mannitol, and NaCl in Arabidopsis, and compare these responses to those elicited when plants are exposed to vermiculite drying. Furthermore, we explore a new approach for reducing water potential.

Comparing differential gene expression responses elicited by PEG, mannitol, and NaCl treatment to vermiculite drying

To understand the impact PEG, mannitol, and NaCl treatment have on gene expression, we first tested their physiological effects across a range of doses. To this end, we grew Arabidopsis seedlings on agar plates supplemented with Linsmaier & Skoog (LS) nutrients for 14 days on three different doses of each stress type. Dose ranges were chosen based on published literature, and ranged from mild to severe stress levels (Dubois and Inzé, 2020; Claeys et al., 2014; van der Weele et al., 2000). As the dose of each stress type increased, the media’s water potential significantly decreased in a dose-dependent manner (Pearson, p<8.5 × 10–5). Across the doses tested, we found that each stress type’s impact on water potential was not statistically different from one other (ANCOVA post hoc, p>0.05). In response to these treatments, we found shoot biomass significantly decreased in a dose-dependent manner (Pearson, p<2 × 10–6) (Figure 1A–C, Figure 1—figure supplement 1, Supplementary file 1).

Figure 1. Benchmarking the impact different stress assays have on Arabidopsis gene expression.

(A) 22-day-old Arabidopsis growth on plates under either 1.67× hard agar (HA), 20% polyethylene glycol (PEG), 100 mM mannitol, or 75 mM NaCl treatments. (B) Water potential measurements of treatment media (n=3–4). (C) Dry weight of 22-day-old Arabidopsis seedlings under different doses of each stress treatment (n=11–12). (D) Number and intersect of differentially expressed genes (DEGs) that are dose-responsive to each stress treatment within root and shoot tissue. (E–I) Heatmaps displaying the top 50 most significant upregulated or downregulated genes in response to (E) HA, (F) PEG, (G) mannitol, (H) NaCl, and (I) vermiculite drying in the Arabidopsis root (n=2–3 biological replicates). Key genes and membership of Gene Ontology (GO) Terms for ‘response to stress’, ‘response to chemical stimulus’, or ‘response to ABA stimulus’ are indicated. ABA, abscisic acid.

Figure 1.

Figure 1—figure supplement 1. Plant growth responses to stress assays.

Figure 1—figure supplement 1.

(A–C) Images of 22-day-old Arabidopsis seedlings grown under different doses of each agar stress assay.
Figure 1—figure supplement 2. Shoot gene expression responses to each stress assay are dose-responsive.

Figure 1—figure supplement 2.

Heatmap displaying the top 50 most significant upregulated or downregulated genes in shoots in response to (A) hard agar (HA), (B) polyethylene glycol (PEG), (C) mannitol, (D) NaCl, and (E) vermiculite drying (n=2–3 biological replicates). Key genes and membership of Gene Ontology (GO) Terms for ‘response to stress’, ‘response to chemical stimulus’, or ‘response to ABA stimulus’ indicated. ABA, abscisic acid.
Figure 1—figure supplement 3. Overlapping differentially expressed genes (DEGs) responsive to different assay types.

Figure 1—figure supplement 3.

Overlap of dose-responsive differentially expressed genes in shoot (A) and root (B) in response to either hard agar (HA), polyethylene glycol (PEG), mannitol, or NaCl (replicated from Figure 1). Number of upregulated or downregulated dose-responsive genes in response to each treatment type in shoot (C) and root (D). Overlapping gene sets in (E) shoot or (F) root tissue (permutation test, p<0.001).
Figure 1—figure supplement 4. Treating vermiculite-grown Arabidopsis plants to mild drought stress.

Figure 1—figure supplement 4.

(A) Field capacity measurements of vermiculite as water evaporated over a 5-day period (n=6–12). (B) Shoot dry weight of Arabidopsis rosettes as they grew either under well-watered conditions or drought conditions over a 5-day period (t-test, n=12). (C) Images of plants after 5 days of water stress. (D) Seed yield resulting from Arabidopsis plants after drought recovery (t-test, n=50).
Open in a new tab

Genes that change their expression in response to an environmental signal often do so in a dose-responsive manner (Claeys et al., 2014; Swift et al., 2020). In light of this, we sought to discover genes whose expression was dose-responsive to the amount of PEG, mannitol, or NaCl applied. By identifying genes that were stress responsive across a range of doses, we ensured such genes responded to and were directional with the stress as a whole, and not induced or repressed at an individual dose. Taking this approach, we sequenced root and shoot bulk transcriptomes by RNA-seq, and associated each gene’s expression with the dose of stress with a linear model. To ensure we captured steady-state differences in gene expression, and avoided those that were transient, we sequenced root and shoot transcriptome profiles after 14 days of stress exposure (Dubois and Inzé, 2020; Munns, 2002; Nikonorova et al., 2018). By these means, we found hundreds of genes that were dose-responsive to each treatment within root and shoot tissue (Figure 1E–H, Figure 1—figure supplement 2, Supplementary file 2) (adj. p<0.05). We found that a portion of these dose-responsive genes were shared across treatments, suggesting a common response to low water potential (Figure 1D, Figure 1—figure supplement 3). Conversely, we also found a portion of dose-responsive genes were unique to each stress type.

Next, we wanted to compare these different methods of lowering water potential to a pot-based water deficit assay. To perform this experiment, we subjected mature Arabidopsis plants grown in pots on vermiculite supplemented with LS media to a mild water stress by withholding water for 5 days. During this period, field capacity (FC) reduced from 100% to 41%. This treatment led to a reduction in plant biomass (t-test, p=1.8 × 10–3), as well as seed yield (t-test, p=1.2 × 10–4), but did not induce visible signs of senescence or wilting (Figure 1—figure supplement 4, Supplementary file 3). We assayed root and shoot gene expression responses each day during water loss by RNA-seq. We observed a dose-dependent relationship between a decrease in FC and gene expression responses in both roots and shoots, identifying 1949 differentially expressed genes in roots and 1792 in shoots (DESeq, adj. p<0.01) (Figure 1I, Figure 1—figure supplement 2, Supplementary file 2). We ensured these genes’ expression patterns recovered upon rewatering (Figure 1I). We note that while vermiculite has greater aeration than soil, we found that the genes differentially expressed in roots in response to vermiculite drying largely agreed with a previous report detailing transcriptional responses to soil drying (Lozano-Elena et al., 2022; Figure 2—figure supplement 1). We also found differentially expressed genes responsive to vermiculite drying agreed with those responsive to transient treatment with abscisic acid (ABA), a stress hormone whose levels rise in response to water deficit (Claeys et al., 2014; Figure 2—figure supplement 2).

To assess how PEG, mannitol, and NaCl treatments compared to the vermiculite drying response described above, we overlapped genes found differentially expressed in each experiment. For shoot tissue, we found genes that were differentially expressed during vermiculite drying overlapped significantly with genes that were differentially expressed by either PEG, mannitol, and NaCl treatments (Fisher’s exact test, adj. p<0.05). Additionally, there was 88–99% directional agreement within these overlaps, indicating that genes induced or repressed by vermiculite drying were similarly induced or repressed by low water potential treatments on agar (Figure 2A and B). Along these lines, across all conditions we saw differential expression of the desiccation associated genes RESPONSE TO DESSICATION 20;29B (RD20;29B) (Msanne et al., 2011; Takahashi et al., 2000), the osmo-protectant gene DELTA1-PYRROLINE-5-CARBOXYLATE SYNTHASE 1 (P5CS1), and ABA signaling and biosynthesis genes HOMEOBOX 7 (HB7) (Valdés et al., 2012), and NINE-CIS-EPOXYCAROTENOID DIOXYGENASE (NCED3) (Tan et al., 2003; Figure 2—figure supplement 3). We note that while we observed agreement in the direction of gene expression across assays, there were differences in the amplitude of gene expression (Figure 2—figure supplement 3). This may be due to confounding factors, such as differences in the ranges of water potential tested (Figure 1B), or through comparing seedlings grown on plates with mature Arabidopsis plants grown on vermiculite.

Figure 2. Comparing hard agar (HA), polyethylene glycol (PEG), mannitol, and NaCl gene expression responses to vermiculite drying.

(A) Heatmap displaying genes differentially expressed in response to vermiculite drying in shoot or root tissue compared to their dose-responsive expression within each plate-based assay. Level of ‘directional agreement’ (i.e. differentially expressed in the same direction) found within each assay reported. (B) Overlap analysis of genes found differentially expressed due to vermiculite drying, compared to those found differentially expressed within each dose of PEG, mannitol, NaCl, or HA assays in both shoot and root (Fisher’s exact test, adj. p<0.05). (C–D) Expression patterns of HOMEOBOX12 (HB12) and DROUGHT HYPERSENSITIVE 2 (DRY2) across each assay in root tissue (n=2-3). (E) Shoot area of seedlings grown under increasing doses of HA, agar, or nutrient concentrations (n=19). (F) Number and percent overlap of genes found differentially expressed in response to increasing doses of HA, agar, or nutrient concentrations with those differentially expressed in response to vermiculite drying. (G) Total shoot area of Arabidopsis accessions grown under either 1× or 2× HA treatment (n=5–12). (H) Images of Arabidopsis Trs-0 or UKSE06-325 accessions grown on either 1× or 2× HA treatment.

Figure 2.

Figure 2—figure supplement 1. Comparing gene expression responses to vermiculite drying and polyethylene glycol (PEG) treatment with previous studies.

Figure 2—figure supplement 1.

(A) Intersect analysis of root or shoot genes found differentially expressed in response to vermiculite drying within this study, and genes found differentially expressed in response to soil drying by Lozano-Elena et al., 2022, or by Wilkins et al., 2010 (permutation test, p<0.001). (B–C) Intersect analysis of genes found differentially expressed in response to PEG treatment in shoot (B) or root (C) in this study, with those found differentially expressed in response to PEG treatment by Wong et al., 2019, and Wang et al., 2021 (permutation test, p<0.001). (D–E) Heatmap displaying direction of shoot (D) or root (E) differentially expressed in response to vermiculite drying (this study) and Wilkins et al., 2010, or Lozano-Elena et al., 2022, respectively. Directional agreement with this study’s vermiculite drying response indicated. (F–G) Heatmap displaying direction of genes differentially expressed in response to PEG treatment across each study. We note that both Wong et al. and Wang et al. assess transcriptomic responses of whole seedlings (both root and shoot), and thus we compare our shoot (F) and root (G) data separately. (E) Examining the 119 genes that were differentially expressed in response to drought (this study), PEG treatment (this study), and PEG treatment reported in Wong et al.
Figure 2—figure supplement 2. Comparing abscisic acid (ABA)-induced differential expression to vermiculite drying and hard agar (HA)-induced gene expression patterns.

Figure 2—figure supplement 2.

(A) Overlap analysis of genes found differentially expressed in response to vermiculite drying, compared to those within each dose of either transient ABA treatment, polyethylene glycol (PEG), mannitol, NaCl, or HA assays in both root and shoot (Fisher’s exact test, adj. p<0.05). (B) Heatmap displaying genes differentially expressed under vermiculite drying in root tissue compared to their dose-responsive expression within each stress assay. Direction of gene expression agreement with vermiculite-drying responsive gene expression (i.e. ‘directional agreement’) indicated.
Figure 2—figure supplement 3. Gene expression profiles of individual genes.

Figure 2—figure supplement 3.

(A–E) Expression patterns of individual genes under doses of each assay in shoot tissue: (A) RD29B, (B) RD20, (C) HB7, (D) NCED3 and (E) P5CS1. (F–H) Expression patterns of individual genes under doses of each assay in root tissue: (F) GCL1, (G) RD21, and (H) RHS18.
Figure 2—figure supplement 4. Physiological measurements of Arabidopsis seedlings in response to hard agar (HA) treatment.

Figure 2—figure supplement 4.

(A) Measurement of primary root growth rate across 8 days of growth under 1× HA (no treatment) and 2.5× HA conditions (n=16, t-test p). (B) Shoot water potential measurements of seedlings grown under different HA media doses (n=3, Pearson p=0.009). (C) Measurement of maximum quantum yield of photosystem II (PSII) (Fv/Fm) under different HA media doses (n=4, t-test p).
Figure 2—figure supplement 5. Comparing the separate effects of nutrient concentration and agar concentration on seedling growth.

Figure 2—figure supplement 5.

Image of Arabidopsis seedlings grown on either (A) 1× hard agar (HA) (i.e. 1× Linsmaier & Skoog [LS], 2% agar) or (B) 2.5× HA, which increased both nutrient and agar concentrations to 2.5× and 5%, respectively. (C) Image of seedlings grown on an increased 2.5× nutrient concentration (without a change in agar concentration). (D) Image of seedlings grown on an increased 5% agar concentration (without a change in nutrient concentration). (E) Water potential measurements of media presented in (A–D) (n=3). (F) Intersection of differentially expressed genes responsive to either agar concentration, nutrient concentration, HA treatment, or drought stress (permutation test, p<0.001).
Figure 2—figure supplement 6. Associating hard agar’s (HA) impact on shoot size with plant fitness.

Figure 2—figure supplement 6.

Comparing the impact HA treatment has on shoot size of 20 different Arabidopsis accessions to the change in their fitness found under drought conditions in the field, as reported in Exposito-Alonso et al., 2019.
Figure 2—figure supplement 7. The volume of hard agar (HA) has minimal impact on gene expression.

Figure 2—figure supplement 7.

Image of Arabidopsis seedlings grown on either (A) 75 mL of 1× HA, (B) 30 mL of 2.5× HA, or (C) 75 mL of 2.5× HA. Number of genes found differentially expressed is reported for comparisons (A) and (B), as well as (B) and (C) (DESeq, adj. p<0.01). Of the 29 genes found differentially expressed between (B) and (C), 13 are found in comparison (A) and (B). (E) Heatmap of genes found differentially expressed in either comparison.
Open in a new tab

In root tissue, we found greater variability in transcriptomic responses to the different methods of lowering water potential. In particular, we found a number of genes that were upregulated during vermiculite drying were downregulated by PEG, mannitol, and NaCl treatments (and vice versa) (Figure 2A). This trend persisted when we assessed genes found differentially expressed at each discrete dose of stress (Figure 2B). For example, 27% of PEG dose-responsive genes shared the same direction of expression seen in vermiculite drying responses. We note that previously published PEG transcriptome datasets largely agreed with our own (Figure 2—figure supplement 1). Such differential regulation in comparison to vermiculite drying is exemplified by the expression of genes such as HOMEOBOX 12 (HB12) (Valdés et al., 2012; Figure 2C), GRC2-LIKE 1 (GCL1) (Gao et al., 2007), and RESPONSE TO DEHYDRATION 21 (RD21) (Koizumi et al., 1993; Figure 2—figure supplement 3). We found genes downregulated by PEG are over-represented in the ‘monooxygenase activity’, and ‘oxygen binding’ Gene Ontology (GO) Terms (p<1 × 10–15, Supplementary file 4). Mannitol and NaCl held a 48% and 57% agreement in gene expression direction with vermiculite drying respectively. Examples of genes that followed this pattern of differential regulation in mannitol and NaCl treatments were DROUGHT HYPERSENSITIVE 2 (DRY2) (Posé et al., 2009; Figure 2D) and ROOT HAIR SPECIFIC 18 (RHS18) (Ponce et al., 2022). NaCl-responsive GO Terms included a specific downregulation of ‘phosphorous metabolic processes’ (p=5.2 × 10–6), suggesting that the roots were changing phosphate levels in response to NaCl, a process known to help maintain ion homeostasis (Miura et al., 2011). For mannitol, we observed a specific downregulation of ‘cell wall organization or biogenesis’ and ‘microtubule-based processes’ (p<7.8 × 10–3) (Supplementary file 4).

Examining differential gene expression responses to ’hard agar’ (HA) treatment

In addition to examining PEG, mannitol, and NaCl transcriptional responses, we also tested a new way of lowering water potential on an agar plate. We hypothesized that we could induce stress by increasing both the agar and nutrient concentration. We called this media ‘hard agar’ (HA), and by testing three different doses (1.25×, 1.67×, and 2.5× fold increase in both agar and LS concentration, where 1.0× is 2% agar and 1× LS), found that it limited plant shoot dry weight and media water potential in a similar way to PEG, mannitol, and NaCl treatment (Figure 1A–C). Additionally, we found that HA treatment limited Arabidopsis primary root growth rate, shoot water potential, and photosynthesis efficiency (Figure 2—figure supplement 4). At the molecular level, RNA-seq revealed 1376 and 1921 genes that were dose-responsive to the level of HA stress in roots and shoots respectively (Figure 1E, Figure 1—figure supplement 2). We found that these gene expression responses overlapped significantly with those found differentially expressed in response to vermiculite drying (Fisher’s exact test, p<1 × 10–32, 87% directional agreement) (Figure 2A and B). HA’s impact can be seen in the gene expression responses of HB12 (Figure 2C), GCL1, and RD21 (Figure 2—figure supplement 3).

An increase in nutrient concentration can induce salt-like stress while increasing agar concentration will increase tensile stress (Verger et al., 2018). We tested each of these variables separately to understand the role each played in eliciting the gene expression responses found in the HA assay. To do this, we repeated our HA dose experiment, but now increasing only the concentration of LS nutrients (1×, 1.25×, 1.67×, and 2.5×) or the concentration of agar (2%, 2.5%, 3.3%, and 5%) (Figure 2E). We found a significant decrease in shoot area size in response to an increase in nutrient concentration (Pearson p=1.7 × 10–7) and agar concentration (Pearson p=3.9 × 10–13), where the latter more closely phenocopied the effect of HA (Figure 2E, Figure 2—figure supplement 5, Supplementary file 5). Since the increase in nutrient concentration alone was responsible for changing media water potential, the phenotypic response to increased agar concentration was not in response to a lower water potential (Figure 2—figure supplement 5). Next, we examined the transcriptional responses underlying nutrient and agar responses by sequencing root tissue across each dose tested. Through linear modeling, we found 1043 genes and 938 genes that were dose-responsive to nutrient or agar concentration, respectively. Then, we investigated how these genes compared to those found in vermiculite drying responses. We found that genes differentially expressed in response to an increase in agar or nutrient concentration overlapped 12% and 17% of vermiculite drying responsive gene expression respectively (permutation test, p<0.05) (Figure 2F, Figure 2—figure supplement 5, Supplementary file 6). However, we found genes differentially expressed in response to HA treatment led to a higher overlap (26 %), suggesting that both nutrient and agar concentration contribute to the similarity between HA treatment and vermiculite drying.

Finally, we tested if our HA assay was sensitive enough to detect phenotypic variability. To achieve this, we grew 20 different Arabidopsis ecotypes on 2× HA (4% agar, 2× LS), where ecotypes were selected from a previous drought study that assessed fitness in a common garden experiment (Exposito-Alonso et al., 2019). By comparing the total shoot area after 3 weeks of growth, we found that our assay revealed variability in shoot growth responses (Figure 2G and H, Supplementary file 7). Furthermore, we found that the greater the impact HA had on reducing an accession’s relative shoot size, the better the accession’s fitness was, as measured under field conditions (Exposito-Alonso et al., 2019) (Spearman p=0.04, Figure 2—figure supplement 6). This is likely because smaller shoot systems have a better chance of survival and reproduction (Skirycz et al., 2011). This suggests that our assay may be useful for screening for novel drought-associated loci among a wider group of accessions or mutants.

In summary, our investigation has assessed the shared and unique impacts of agar-based low water potential treatments on gene expression. We also compared these effects with the expression patterns observed during vermiculite drying. We found each plate-based assay generated similar responses in shoot tissue, but more varied responses in root tissue. We note that our comparative analysis focuses largely on transcriptomic responses in Arabidopsis. We suggest investigating gene expression responses in other species as future work. Here, we also introduce another method for lowering water potential. By increasing nutrient and agar concentration, our HA approach also induced gene expression responses comparable to vermiculite drying. We describe how to make this media within the Materials and methods.

Materials and methods

HA stress assay

Arabidopsis seedlings were grown on vertical plates for 8 days under short-day conditions (8 hr light, 21°C, 150 µmoles light) on agar media (1× LS (Cassion LSP03) media, 1% sucrose, 2% agar, pH 5.7). We note LS media is identical to Murashige & Skoog media in inorganic salt content, but lacks glycine, nicotinic acid, and pyridoxine HCl. After 8 days, plants were transferred to HA plates. The 1.0× plate consisted of 2% and 1× LS media, with no sucrose (pH 5.7) at a final volume of 75 mL. Subsequent doses of increased nutrient and agar concentration (1.25×, 1.67×, and 2.5× fold increase) were made by preparing the same media but reducing the amount of water present. For example, the 1.25×, treatment plate contained 60 mL of 2.5% agar and 1.25× LS media. We note that the volume of HA itself has minimal impact gene expression responses (Figure 2—figure supplement 7). On day 14, 2 hr after subjective dawn, shoot and root samples were flash-frozen (6 plants per replicate). In total, we collected 16 samples for RNA-seq analysis (2 organs, 2–3 biological replicates, 3 treatment levels). We also collected a non-treated control set (2 biological replicates).

To test different Arabidopsis accessions on HA, plants were sown on either 1× or 2× treatments as described above, however supplemented with 0.5% or 1% sucrose respectively to encourage germination. Seedlings were grown for 3 weeks under short-day conditions in before imaging plates in duplicate (n=2–5 plants per plate) (Supplementary file 7). Shoot area was calculated from images using Plant Growth Tracker (GitHub - https://github.com/jiayinghsu/plant-growth-tracker; Xu, 2022).

Vermiculite drying assay

Arabidopsis seedlings were grown on vertical plates for 17 days under short-day conditions (8 hr light, 21°C, 150 µmoles light) on agar media (1× LS, 1% sucrose, 2% agar, pH 5.7), before transfer to vermiculite saturated with 0.75× LS media. We note at the timing of transfer lateral root formation had begun. Plants were then grown on vermiculite at 100% field capacity (FC) for 12 days (8 hr light, 21°C, 150 µmoles light). On the 13th day, the first time point was sampled (4.5 hr after subjective dawn) where tissue was flash-frozen in liquid nitrogen. After this, excess aqueous solution was drained from each pot, and then each pot was calibrated to 1× FC. Plant tissue was harvested each day on subsequent days at the same time of day. Each day, pots were weighed to measure extent of evaporation. By these means, FC was measured (Figure 1—figure supplement 4). After the 5th day sample was taken, water was re-added to the remaining pots to an excess of 1× FC. ~15 plants were sampled per time point. In total, we harvested 78 tissue samples for RNA-seq (3 biological replicates, 2 organ types, 7 days, 2 treatments). Plants were then left to grow under long-day conditions until flowering. Seeds were harvested, dried, and weighed (n=50 plants per treatment).

PEG stress assay

Arabidopsis seedlings were grown on vertical plates for 8 days under short-day conditions (8 hr light, 21°C, 150 µmoles light) on agar media (1× LS, 1% sucrose, 2% agar, pH 5.7), before transfer to PEG media of varying concentrations. PEG media plates were prepared by dissolving crystalline 6000 MW PEG into freshly autoclaved 1× LS media pH 5.7 and pouring 50 mL of PEG media solution onto 1× LS, 2% agar, media plates (pH 5.7), letting the PEG solution diffuse into the solid media overnight, then pouring off excess and transferring seedlings to PEG infused media plates as described in van der Weele et al., 2000. Plants were grown under three different treatments (12%, 20%, and 28% PEG solution wt/vol) for 14 days. On day 14, 2 hr after subjective dawn, shoot and root samples were flash-frozen (6 plants per replicate). In total, we collected 16 samples for RNA-seq analysis (2 organs, 2–3 biological replicates, 3 treatment levels).

Mannitol and NaCl osmotic stress assays

Arabidopsis seedlings were grown on vertical plates for 8 days under short-day conditions (8 hr light, 21°C, 150 µmoles light) on agar media (1× LS, 1% sucrose, 2% agar, pH 5.7), before transfer to either mannitol or salt (NaCl) media of varying concentrations. Mannitol and NaCl media plates were prepared by adding respective volume of stock solution to 1× LS, 2% agar, pH 5.7 media before autoclaving. Plants were grown under three different treatments of mannitol or NaCl (50 mM, 100 mM, and 200 mM for mannitol, 30 mM, 75 mM, and 150 mM for NaCl) for 14 days. On day 14, 2 hr after subjective dawn, shoot and root samples were flash-frozen (6 plants per replicate). In total, for either mannitol or NaCl treatment experiments, we collected 18 samples for RNA-seq analysis (2 organs, 3 biological replicates, 3 treatment levels).

ABA exogenous treatment assay

Arabidopsis seedlings were grown on vertical plates for 8 days under short-day conditions (8 hr light, 21°C, 150 µmoles light) on agar media (1× LS, 1% sucrose, 2% agar, pH 5.7), before transfer to 1× LS, 2% agar, pH 5.7 control media and grown for 14 days. On day 14, ABA solutions of 1 µM, 5 µM, and 10 µM were prepared from 10 mM ABA dissolved in ethanol stock, as well as a mock treatment solution containing 0.1% ethanol concentration. 30 min after subjective dawn, 15 mL of each solution was dispersed onto the roots of the seedlings. After 1 min of treatment, the ABA solution was removed from the plates, and the plates returned to the growth chamber. 2 hr after subjective dawn, shoot and root samples were flash-frozen (6 plants per replicate). In total, we collected 8 samples for RNA-seq analysis (root tissue only, 2 biological replicates, 4 conditions).

Osmotic potential measurements

The water potential of media was determined considering it equivalent to the osmotic potential (Ψs). Osmotic potential was measured using a vapor pressure osmometer (Model 5600, ELITech Group; Puteaux, France). Readings were taken from melted agar media constituted with the particular stress type. Osmolality readings for each sample obtained were converted to megapascals (MPa) using the equation Ψs = -CRT, where C is the molar concentration, R is the universal gas constant, T is the temperature in Kelvin. We note that to measure the water potential of PEG treatment media, we infiltrated the PEG solution into plates as described above, and then melted the PEG-infiltrated agar for measurement with the osmometer. We assessed the osmotic potential of shoot tissue 2 weeks after transplanting the seedlings to HA media. After immersion in liquid nitrogen 3 shoots were placed into 0.5 mL tubes and centrifuged to extract the tissue sap. The osmotic potential (Ψs) of the extracted sap was determined using a vapor pressure osmometer.

Chlorophyll fluorescence measurements

Chlorophyll fluorescence was assessed in eight seedlings of each plate using the Walz PAM IMAGING PAM M-series IMAG-K7 (MAXI) fluorometer. For every experiment, leaves were pre-conditioned in the dark for 1 hr. The maximum quantum yield of photosystem II (PSII) (Fv/Fm) was calculated using the formula:

Fv/Fm=(Fm−Fo)/Fm

where Fv is the variable fluorescence, Fm is the maximal fluorescence following 1 hr of dark adaptation, and F0 is the minimal fluorescence level of a dark-adapted leaf when all PSII reaction centers are open.

Root growth rate measurements

Arabidopsis seedlings were grown on vertical plates for 8 days under short-day conditions (8 hr light, 21°C, 150 µmoles light) on agar media (1× LS, 1% sucrose, 2% agar), before transfer to 2.5× HA treatment plates as described above. Root images were acquired every 2 days for a total of 8 days using scanners. Primary root length, defined as the length (scaled to cm) from hypocotyl base to root tip, was quantified using ImageJ. For each treatment we screened 4 plates, with each plate holding 4 individual plants.

RNA extraction and library preparation

Plant tissue was crushed using the TissueLyser (Agilent) and RNA extracted using RNeasy Mini Kit (QIAGEN). Number of biological replicates per library ranged between RNA quality was assessed using TapeStation High Sensitivity RNA assay (Agilent). 0.5–1 µg of total RNA proceeded to library preparation, where libraries were prepared using TruSeq stranded mRNA kit (Illumina). Resulting libraries were sequenced on the NovaSeq 6000 (Illumina) with 2×150 bp paired-end read chemistry. Read sequences were aligned to the Arabidopsis TAIR10 genome using HISAT2 (Kim et al., 2019), and gene counts called using HT-seq (Anders et al., 2015), by relying on Araport11 annotation (Cheng et al., 2017). Normalized counts can be found in Supplementary file 2. For each organ, libraries from all experiments were normalized together before calling differential expression.

Statistical analysis

To detect differential expression in our drought assay on vermiculite, we called differential expression using a linear model using the DESeq2 LRT function to associate a change in FC with change in gene expression. The same statistical approach was used to associate a change in a gene’s expression to changes in dose of HA, PEG, mannitol, and NaCl, as well as changes in agar concentration, nutrient concentration, and volume of agar used. Resulting model p-values were adjusted to account for false discovery (p-value<0.05). The complete list of differentially expressed genes for each experiment can be found in Supplementary file 2 and Supplementary file 6. Pairwise differential gene expression was called using DESeq2 (Love et al., 2014). Specifically, for plate-based assays, we called differential expression by comparing the control treatment to each treatment dose, using an adjusted p-value threshold of 0.05. Overlap analyses were performed using Fisher’s exact tests, with an adjusted p-value threshold of 0.05. The background for these intersects was all expressed genes within the respective organ. Permutation tests and GO Term enrichment analyses were performed in VirtualPlant (Katari et al., 2010), with all expressed genes within the respective organ used as background.

Acknowledgements

We thank Renee Garza for critical reading of the manuscript. JS is an Open Philanthropy awardee of Life Science Research Foundation, as well as recipient of the Pratt Industries American-Australian Association Scholarship. JRE is an Investigator of the Howard Hughes Medical Institute.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Joseph Swift, Email: jswift@salk.edu.

Joseph R Ecker, Email: ecker@salk.edu.

Dominique C Bergmann, Stanford University, United States.

Jürgen Kleine-Vehn, University of Freiburg, Germany.

Funding Information

This paper was supported by the following grant:

  • Howard Hughes Medical Institute to Joseph R Ecker.

Additional information

Competing interests

No competing interests declared.

Author contributions

Conceptualization, Data curation, Formal analysis, Validation, Investigation, Methodology, Writing - original draft, Project administration, Writing - review and editing.

Conceptualization, Data curation, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing.

Data curation, Investigation, Methodology.

Software, Methodology.

Investigation, Methodology.

Software, Methodology.

Data curation.

Supervision, Methodology.

Supervision, Methodology.

Conceptualization, Supervision, Funding acquisition, Investigation, Methodology.

Additional files

Supplementary file 1. Plant physiological measurements.
elife-84747-supp1.xlsx (21.2KB, xlsx)
Supplementary file 2. Differentially expressed genes and normalized counts in hard agar (HA), polyethylene glycol (PEG), mannitol, NaCl, or vermiculite drying experiments.
elife-84747-supp2.xlsx (16.6MB, xlsx)
Supplementary file 3. Vermiculite drying assay measurements.
elife-84747-supp3.xlsx (23.5KB, xlsx)
Supplementary file 4. Gene Ontology (GO) Term enrichment of differentially expressed genes.
elife-84747-supp4.xlsx (224.3KB, xlsx)
Supplementary file 5. Shoot area of seedlings grown under different agar and nutrient concentrations.
elife-84747-supp5.xlsx (14.8KB, xlsx)
Supplementary file 6. Differentially expressed genes and normalized counts in response to changes in nutrient or agar concentration.
elife-84747-supp6.xlsx (3.4MB, xlsx)
Supplementary file 7. Shoot area of different Arabidopsis accessions grown on hard agar (HA) media.
elife-84747-supp7.xlsx (15.4KB, xlsx)
MDAR checklist

Data availability

Raw sequencing data can be found at the National Center for Biotechnology Information Sequence Read Archive (accession number PRJNA904764). Normalized read counts and raw phenotypic datasets can be found in the Supplementary Material.

The following dataset was generated:

Gonzalez S, Swift J, Ecker J. 2022. Mimicking genuine drought responses using a high throughput plate assay. NCBI BioProject. PRJNA904764

The following previously published dataset was used:

Lozano-Elena F, Fábregas N, Coleto-Alcudia V, Caño-Delgado AI. 2018. Transcriptomic study of Arabidopsis roots overexpressing the brassinosteroid receptor BRL3, in control conditions and under severe drought. NCBI Gene Expression Omnibus. GSE119382

References

  1. Anders S, Pyl PT, Huber W. HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics. 2015;31:166–169. doi: 10.1093/bioinformatics/btu638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Battisti DS, Naylor RL. Historical warnings of future food insecurity with unprecedented seasonal heat. Science. 2009;323:240–244. doi: 10.1126/science.1164363. [DOI] [PubMed] [Google Scholar]
  3. Cheng C-Y, Krishnakumar V, Chan AP, Thibaud-Nissen F, Schobel S, Town CD. Araport11: a complete reannotation of the Arabidopsis thaliana reference genome. The Plant Journal. 2017;89:789–804. doi: 10.1111/tpj.13415. [DOI] [PubMed] [Google Scholar]
  4. Claeys H, Van Landeghem S, Dubois M, Maleux K, Inzé D. What is stress? Dose-response effects in commonly used in vitro stress assays. Plant Physiology. 2014;165:519–527. doi: 10.1104/pp.113.234641. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Dubois M, Inzé D. Plant growth under suboptimal water conditions: early responses and methods to study them. Journal of Experimental Botany. 2020;71:1706–1722. doi: 10.1093/jxb/eraa037. [DOI] [PubMed] [Google Scholar]
  6. Exposito-Alonso M, Burbano HA, Bossdorf O, Nielsen R, Weigel D. Natural selection on the Arabidopsis thaliana genome in present and future climates. Nature. 2019;573:126–129. doi: 10.1038/s41586-019-1520-9. [DOI] [PubMed] [Google Scholar]
  7. Gao Y, Zeng Q, Guo J, Cheng J, Ellis BE, Chen J-G. Genetic characterization reveals no role for the reported ABA receptor, GCR2, in ABA control of seed germination and early seedling development in Arabidopsis. The Plant Journal. 2007;52:1001–1013. doi: 10.1111/j.1365-313X.2007.03291.x. [DOI] [PubMed] [Google Scholar]
  8. Godfray HCJ, Beddington JR, Crute IR, Haddad L, Lawrence D, Muir JF, Pretty J, Robinson S, Thomas SM, Toulmin C. Food security: the challenge of feeding 9 billion people. Science. 2010;327:812–818. doi: 10.1126/science.1185383. [DOI] [PubMed] [Google Scholar]
  9. Hohl M, Schopfer P. Water relations of growing maize coleoptiles : Comparison between mannitol and polyethylene glycol 6000 as external osmotica for adjusting turgor pressure. Plant Physiology. 1991;95:716–722. doi: 10.1104/pp.95.3.716. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Jisha KC, Puthur JT. Halopriming of seeds imparts tolerance to NaCl and PEG induced stress in Vigna radiata (L.) Wilczek varieties. Physiology and Molecular Biology of Plants. 2014;20:303–312. doi: 10.1007/s12298-014-0234-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Juenger TE, Verslues PE. Time for a drought experiment: Do you know your plants’ water status? The Plant Cell. 2023;35:10–23. doi: 10.1093/plcell/koac324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Katari MS, Nowicki SD, Aceituno FF, Nero D, Kelfer J, Thompson LP, Cabello JM, Davidson RS, Goldberg AP, Shasha DE, Coruzzi GM, Gutiérrez RA. VirtualPlant: a software platform to support systems biology research. Plant Physiology. 2010;152:500–515. doi: 10.1104/pp.109.147025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kim D, Paggi JM, Park C, Bennett C, Salzberg SL. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nature Biotechnology. 2019;37:907–915. doi: 10.1038/s41587-019-0201-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Koizumi M, Yamaguchi-Shinozaki K, Tsuji H, Shinozaki K. Structure and expression of two genes that encode distinct drought-inducible cysteine proteinases in Arabidopsis thaliana. Gene. 1993;129:175–182. doi: 10.1016/0378-1119(93)90266-6. [DOI] [PubMed] [Google Scholar]
  15. Kreps JA, Wu Y, Chang H-S, Zhu T, Wang X, Harper JF. Transcriptome changes for Arabidopsis in response to salt, osmotic, and cold stress. Plant Physiology. 2002;130:2129–2141. doi: 10.1104/pp.008532. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biology. 2014;15:550. doi: 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Lozano-Elena F, Fàbregas N, Coleto-Alcudia V, Caño-Delgado AI. Analysis of metabolic dynamics during drought stress in Arabidopsis plants. Scientific Data. 2022;9:90. doi: 10.1038/s41597-022-01161-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Miura K, Sato A, Ohta M, Furukawa J. Increased tolerance to salt stress in the phosphate-accumulating Arabidopsis mutants siz1 and pho2. Planta. 2011;234:1191–1199. doi: 10.1007/s00425-011-1476-y. [DOI] [PubMed] [Google Scholar]
  19. Msanne J, Lin J, Stone JM, Awada T. Characterization of abiotic stress-responsive Arabidopsis thaliana RD29A and RD29B genes and evaluation of transgenes. Planta. 2011;234:97–107. doi: 10.1007/s00425-011-1387-y. [DOI] [PubMed] [Google Scholar]
  20. Munns R. Comparative physiology of salt and water stress. Plant, Cell & Environment. 2002;25:239–250. doi: 10.1046/j.0016-8025.2001.00808.x. [DOI] [PubMed] [Google Scholar]
  21. Nikonorova N, Van den Broeck L, Zhu S, van de Cotte B, Dubois M, Gevaert K, Inzé D, De Smet I. Early mannitol-triggered changes in the Arabidopsis leaf (phospho)proteome reveal growth regulators. Journal of Experimental Botany. 2018;69:4591–4607. doi: 10.1093/jxb/ery261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Ponce OP, Torres Y, Prashar A, Buell R, Lozano R, Orjeda G, Compton L. Transcriptome profiling shows a rapid variety-specific response in two Andigenum potato varieties under drought stress. Frontiers in Plant Science. 2022;13:1003907. doi: 10.3389/fpls.2022.1003907. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Posé D, Castanedo I, Borsani O, Nieto B, Rosado A, Taconnat L, Ferrer A, Dolan L, Valpuesta V, Botella MA. Identification of the Arabidopsis dry2/sqe1-5 mutant reveals a central role for sterols in drought tolerance and regulation of reactive oxygen species. The Plant Journal. 2009;59:63–76. doi: 10.1111/j.1365-313X.2009.03849.x. [DOI] [PubMed] [Google Scholar]
  24. Skirycz A, Vandenbroucke K, Clauw P, Maleux K, De Meyer B, Dhondt S, Pucci A, Gonzalez N, Hoeberichts F, Tognetti VB, Galbiati M, Tonelli C, Van Breusegem F, Vuylsteke M, Inzé D. Survival and growth of Arabidopsis plants given limited water are not equal. Nature Biotechnology. 2011;29:212–214. doi: 10.1038/nbt.1800. [DOI] [PubMed] [Google Scholar]
  25. Swift J, Alvarez JM, Araus V, Gutiérrez RA, Coruzzi GM. Nutrient dose-responsive transcriptome changes driven by Michaelis-Menten kinetics underlie plant growth rates. PNAS. 2020;117:12531–12540. doi: 10.1073/pnas.1918619117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Takahashi S, Katagiri T, Yamaguchi-Shinozaki K, Shinozaki K. An Arabidopsis gene encoding a Ca2+-binding protein is induced by abscisic acid during dehydration. Plant & Cell Physiology. 2000;41:898–903. doi: 10.1093/pcp/pcd010. [DOI] [PubMed] [Google Scholar]
  27. Tan B-C, Joseph LM, Deng W-T, Liu L, Li Q-B, Cline K, McCarty DR. Molecular characterization of the Arabidopsis 9-cis epoxycarotenoid dioxygenase gene family. The Plant Journal. 2003;35:44–56. doi: 10.1046/j.1365-313x.2003.01786.x. [DOI] [PubMed] [Google Scholar]
  28. Trontin C, Kiani S, Corwin JA, Hématy K, Yansouni J, Kliebenstein DJ, Loudet O. A pair of receptor-like kinases is responsible for natural variation in shoot growth response to mannitol treatment in Arabidopsis thaliana. The Plant Journal. 2014;78:121–133. doi: 10.1111/tpj.12454. [DOI] [PubMed] [Google Scholar]
  29. Valdés AE, Overnäs E, Johansson H, Rada-Iglesias A, Engström P. The homeodomain-leucine zipper (HD-Zip) class I transcription factors ATHB7 and ATHB12 modulate abscisic acid signalling by regulating protein phosphatase 2C and abscisic acid receptor gene activities. Plant Molecular Biology. 2012;80:405–418. doi: 10.1007/s11103-012-9956-4. [DOI] [PubMed] [Google Scholar]
  30. van der Weele CM, Spollen WG, Sharp RE, Baskin TI. Growth of Arabidopsis thaliana seedlings under water deficit studied by control of water potential in nutrient-agar media. Journal of Experimental Botany. 2000;51:1555–1562. doi: 10.1093/jexbot/51.350.1555. [DOI] [PubMed] [Google Scholar]
  31. Verger S, Long Y, Boudaoud A, Hamant O. A tension-adhesion feedback loop in plant epidermis. eLife. 2018;7:e34460. doi: 10.7554/eLife.34460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Verslues PE, Agarwal M, Katiyar-Agarwal S, Zhu J, Zhu J-K. Methods and concepts in quantifying resistance to drought, salt and freezing, abiotic stresses that affect plant water status. The Plant Journal. 2006;45:523–539. doi: 10.1111/j.1365-313X.2005.02593.x. [DOI] [PubMed] [Google Scholar]
  33. Wang Y, Fang Z, Yang L, Chan Z. Transcriptional variation analysis of Arabidopsis ecotypes in response to drought and salt stresses dissects commonly regulated networks. Physiologia Plantarum. 2021;172:77–90. doi: 10.1111/ppl.13295. [DOI] [PubMed] [Google Scholar]
  34. Wilkins O, Bräutigam K, Campbell MM. Time of day shapes Arabidopsis drought transcriptomes. The Plant Journal. 2010;63:715–727. doi: 10.1111/j.1365-313X.2010.04274.x. [DOI] [PubMed] [Google Scholar]
  35. Wong MM, Bhaskara GB, Wen TN, Lin WD, Nguyen TT, Chong GL, Verslues PE. Phosphoproteomics of Arabidopsis Highly ABA-Induced1 identifies AT-Hook-Like10 phosphorylation required for stress growth regulation. PNAS. 2019;116:2354–2363. doi: 10.1073/pnas.1819971116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Xu J. Plant growth Tracker. swh:1:rev:ea600a2528fc0917a739c561f94c2c3d5b1ca892Software Heritage. 2022 https://archive.softwareheritage.org/swh:1:dir:da853e03b52a7cce89790cde6d7b48fb45b6a78b;origin=https://github.com/jiayinghsu/plant-growth-tracker;visit=swh:1:snp:7cc9405ddd2da5da5ca03c04506b9cbeac303079;anchor=swh:1:rev:ea600a2528fc0917a739c561f94c2c3d5b1ca892
  37. Zeller G, Henz SR, Widmer CK, Sachsenberg T, Rätsch G, Weigel D, Laubinger S. Stress-induced changes in the Arabidopsis thaliana transcriptome analyzed using whole-genome tiling arrays. The Plant Journal. 2009;58:1068–1082. doi: 10.1111/j.1365-313X.2009.03835.x. [DOI] [PubMed] [Google Scholar]

eLife assessment

Dominique C Bergmann 1

This work critically evaluates several widely-used assays of transcriptional responses to water limitation in Arabidopsis grown on defined agar-solidified media and, finding inconsistent responses in root transcriptome responses, introduces a new 'hard agar' assay with more consistent responses. The work is valuable as a simple and alternative experimental system that would enable high-throughput genetic screening (and GWAS) to assess the impacts of environmental perturbations on transcriptional responses in various genetic backgrounds. Within this scope, the work is solid, though the debate about whether field-level physiological inferences can be made from such assays remains.

Reviewer #2 (Public Review):

Anonymous

This manuscript describes new methodology to study low water potential (drought) stress responses in agar plates. They devote considerable effort in comparing transcriptome data among various previously published experimental systems, examining how different approaches of reducing water potential impact the Arabidopsis root and shoot transcriptome. Each method purported to reduce water potential in plate-grown seedlings has a different effect on Arabidopsis root transcriptome responses, which is problematic for the field. In this reviewer's view, differences in transcriptome are not as important, and often not as informative as measurement of physiological parameters, which they do very little of in their study.

The focus on transcriptome data to the almost complete exclusion of other types of data is a symptom of a broader over-emphasis on the transcriptome that is quite prevalent in plant science now. We measure transcriptomes because we can, not because it is inherently the most informative thing to do. The important thing is protein amount, and even more so protein activity/function, which we know has an imperfect, at best, correlation with transcript level. This reviewer acknowledges that using Arabidopsis transcriptomics is a commonly employed method, and as such, the outcomes of this study will hold value for a broad audience, even if largely as a cautionary tale. If transcriptomics is used to identify candidate genes for future investigations, an approach that has had some success, then appropriate cautions should be taken in translating expectations about gene, protein, and phenotypic responses in field conditions.

Reviewer #3 (Public Review):

Anonymous

This work compares transcriptional responses of shoots and roots harvested from four plate-based assays that aim to simulate drought and from plants subjected to water deficit in pots using the model plant Arabidopsis thaliana with the goal to select a plate-based assay that best recapitulates transcriptional changes that are observed during water-deficit in pots. For the plate-based assays polyethylene glycol (PEG), mannitol, and sodium chloride (salt) treatments were used as well as a 'hard agar' assay which was newly developed by the authors. In the 'hard agar' assay, less water was added to the solid components of the media leading to an increase in agar strength and nutrient concentration. Plants in pots were grown on vermiculite with the same nutrient mix as used in the plates and drought was induced by withholding watering for five days.

The authors observed a good directional agreement of differential expressed genes for shoots between the plate assays on the vermiculite drying experiment. However, less directional agreement was observed for differential expressed genes of roots, except for their newly developed 'hard agar' assay which had good directional agreement. Testing whether the increase in agar strength or more concentrated nutrients are attributed to this, they found that both factors contributed to the effect of the 'hard agar'. Arabidopsis ecotypes that showed a stronger reduction in shoot size when grown on the 'hard agar' tended to have a lower fitness according to an external study which may indicate that the 'hard agar' assay simulates physiological relevant conditions.

The work highlights that transcriptional responses for simulated drought on plates and drought caused by water deficit are highly variable and dependent on the tissues that are observed. The authors demonstrate that transcriptomics can be used to select a suitable plate assay that most closely recapitulates drought through water deficit for plants grown in pots. Interestingly their newly developed 'hard agar' assay provides an alternative to traditional plate-based assays with improved directional agreement of differential expressed genes in roots in comparison to plants experiencing water deficit in vermiculite. It is promising that the impact of 'hard agar' on the shoot size of 20 diverse Arabidopsis accessions shows some association with plant fitness under drought in the field. Their methodology could be powerful in identifying a better substitute for plate-based high-throughput drought assays that have an emphasis on gene expression changes.

eLife. 2024 Jun 21;12:RP84747. doi: 10.7554/eLife.84747.3.sa3

Author response

Stephen Gonzalez 1, Joseph Swift 2, Adi Yaaran 3, Jiaying Xu 4, Charlotte Miller 5, Natanella Illouz-Eliaz 6, Joseph R Nery 7, Wolfgang Busch 8, Yotam Zait 9, Joseph R Ecker 10

The following is the authors’ response to the previous reviews

eLife assessment

This work is an attempt to establish conditions that accurately and efficiently mimic a drought response in Arabidopsis grown on defined agar-solidified media - an admirable goal as a reliable experimental system is key to conducting successful low water potential experiments and would enable high-throughput genetic screening (and GWAS) to assess the impacts of environmental perturbations on various genetic backgrounds. The authors compare transcriptome patterns of plant subjected to water limitation imposed with different experimental systems. The work is valuable in that it lays out the challenges of such an endeavor and points out shortcomings of previous attempts. There was concern, however, that a purely gene expression-based approach may not provide sufficient physiologically relevant information about plant responses to drought, and therefore, despite improvements from a previous version, the new methodology championed by this work remains inadequate.

Molecular biologists who study drought stress must make choices about which assays to use in their investigation. Serious resources and effort are put into their endeavor, and choice of assay matters. Our manuscript’s goal was largely practical: to guide molecular biologists employing transcriptomics in their choice of drought stress assay, and thus help ensure their work will discover transcriptional signatures of importance, and not those that may be an artifact from lowering water potential using chemical agents on agar plates.

We examine how different approaches of reducing water potential impact the Arabidopsis root and shoot transcriptome. Our manuscript shows that each method of reducing water potential has a different effect on Arabidopsis root transcriptome responses. We acknowledge that drought stress induces a complex physiological response, and can vary depending on the method used. However, by comparing across assays, we find instances where a gene is downregulated by low water potential in one assay, and upregulated by low water potential in another assay. We feel it is only natural to question why this could be, and to hypothesize that it may be caused by secondary effects caused by the way low water potential is imposed. We note that comparative transcriptomics has been a standard approach for decades. We take it as the reviewer’s opinion that it may not be insightful, but it does not factually impact our findings.

Reviewer #2 (Public Review):

This manuscript purports to develop a new system to study low water potential (drought) stress responses in agar plates. They make numerous problematic comparisons among transcriptome datasets, particularly to transcriptome data from a vermiculite drying experiment which they inappropriately present as representing an authentic "drought response" to the exclusion of all other data. For some reason, which the reviewer cannot fully understand, the authors seem intent on asserting the superiority of their experimental system to all others. They do not succeed in this and such an effort is ultimately a disservice to the field of drought research as a whole.

While they devote considerable effort in comparing transcriptome data among various experimental systems, the potentially more informative experiment at the end of the manuscript of testing growth responses of a number of Arabidopsis accessions is only done for their "LW" system. The focus of this manuscript on transcriptome data to the almost complete exclusion of other types of data which is a symptom of a broader over-emphasis on transcriptome that unfortunately is quite prevalent in plant science now. It is worth reminding that for protein coding genes, which constitute the vast majority of genes, transcriptome data is a proxy measurement. The really important thing is protein amount, and even more so protein activity/function, which we know has an imperfect, at best, correlation with transcript level. We measure transcriptomes because we can, not because it is inherently the most informative thing to do. The author's quixotic quest to see if the transcriptomes of different stress treatments match is of limited value and further diminished by their misleading presentation of one particular transcriptome data set (from their vermiculite drying experiments) as somehow a special data set that everything else must be evaluated against. This study sheds no new light on how to do relevant drought (low water potential) experiments in the lab.

Although the reviewer acknowledges that the authors have made some effort to respond to previous comments, the fundamental flaws remain and the present version of this study is little improved from the first submission.

One challenge faced by the drought community is establishing consensus regarding the definition of drought itself. According to the criteria followed by the reviewer, any method leading to a reduction in water potential qualifies as drought stress. However, the findings presented in this manuscript demonstrate that transcriptional responses in roots vary considerably across five different methods of reducing water potential. This indicates that beyond responding to a change in water potential itself, root transcriptomes will also respond to the specific way low water potential is introduced. We believe this variability is of interest to the drought research community.

Of the five methods we explore, we hold the view that the gene expression changes induced by vermiculite drying as the most analogous to the expression signatures Arabidopsis would exhibit in response to low water potential in the natural environment. In contrast, we posit that Arabidopsis grown on agar plates - where the root system is exposed to air and light, and where water potential is lowered using chemical agents - may contain gene expression signatures plant molecular biologists may not find particularly relevant. However, we acknowledge that this is our opinion, and will make this more explicit on our revised text.

More broadly, we believe that the reviewer’s observation regarding the ‘over-emphasis’ on transcriptomics that is prevalent within the plant science community justifies, rather than diminishes, the work presented here. If transcriptomics is a commonly employed method, then we anticipate that the outcomes of this study will hold value for a broad audience. Such researchers are likely not only using transcriptomics as a proxy measure for protein abundance, as the reviewer suggests, but also because it is one of the more straightforward genomic techniques biologists can use to identify candidate genes that may be chosen for further scrutiny.

Reviewer #3 (Public Review):

Comments on revised version:

Specific previous criticisms that were addressed are:

(1) that gene expression changes were only compared between the highest dose of each stress assay. In the revised version, the authors changed their framework and are now using linear modelling to detect genes that display a dose response to each specific treatment. I agree that this might be a more robust approach to selecting genes that are specific to a certain treatment.

(2) that concentrations of PEG, mannitol, NaCl, and the "low water" agar which were chosen are not comparable in regards to their specific osmotic component. I appreciate that the authors measured the osmotic potential of each treatment. It revealed that both PEG and NaCl at their highest concentration had a much more negative osmotic potential compared to the other treatment. The authors claim that using ANCOVA they did not detect any significant differences between the treatments (lines 113, 114). I do believe that ANCOVA is not the appropriate test in this case. ANCOVA has an assumption of linearity, while the dose response between concentration and osmotic potential is non-linear. This is particularly evident for PEG (Steuter AA. Water potential of aqueous polyethylene glycol. Plant Physiol. 1981 Jan;67(1):64-7. doi: 10.1104/pp.67.1.64.). Since the treatments are not the same at the highest level, I think this could have effects on the validity of comparisons by linear model. One approach could be to remove the treatment level with the highest concentration and compare the results or adjust the treatments to the same osmolarity.

(3) that only two biological replicates were collected for RNA sequencing which makes it impossible to know how much variance exists between samples. The authors added a third replicate in the revised version for most treatments. However, some treatments still have only two replicates, which cannot be easily seen from the text or the figure. I would prefer that those differences are pointed out.

(4) that the original manuscript did not explore what effect the increase of agar and nutrient concentration in the "low water" agar had on water potentials. The authors conducted additional experiments showing that changes in water potential were exclusively caused by changes in the nutrient concentration (Figure 2-figure supplement 5; lines 222-224). However, the increase in agar strength had also some effect on gene expression. While this is not further discussed in the text, I believe this effect of agar on gene expression could be similar to root responses to soil compaction.

(5) That the lower volume of media in the "low water" agar could have an effect on plants. The authors compared these effects in Figure 2-figure supplement 7. They claim that "different volumes of LW agar media do not play a significant part in modulating gene expression". While I can see that they detected 313 overlapping DEGs, there were still 146 and 412 non-overlapping DEGs. The heatmap in subpanel E also shows that there were differences in particular in the up-regulated genes. My conclusion would be that the change in volume does play a role and this should be a consideration in the manuscript.

We thank the reviewer for their suggestions. We plan to resubmit the manuscript reflecting the requested changes. Specifically, we will:

- We will detail more thoroughly the effects of agar volume on gene expression changes elicited by LW agar treatment.

- We will investigate whether the tensile stress introduced by hard agar is similar to soil compaction by an analysis with existing literature.

- Assess more rigorously the suitability of the ANCOVA model for assessing water potential changes of different media types.

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Data Citations

    1. Gonzalez S, Swift J, Ecker J. 2022. Mimicking genuine drought responses using a high throughput plate assay. NCBI BioProject. PRJNA904764
    2. Lozano-Elena F, Fábregas N, Coleto-Alcudia V, Caño-Delgado AI. 2018. Transcriptomic study of Arabidopsis roots overexpressing the brassinosteroid receptor BRL3, in control conditions and under severe drought. NCBI Gene Expression Omnibus. GSE119382

    Supplementary Materials

    Supplementary file 1. Plant physiological measurements.
    elife-84747-supp1.xlsx (21.2KB, xlsx)
    Supplementary file 2. Differentially expressed genes and normalized counts in hard agar (HA), polyethylene glycol (PEG), mannitol, NaCl, or vermiculite drying experiments.
    elife-84747-supp2.xlsx (16.6MB, xlsx)
    Supplementary file 3. Vermiculite drying assay measurements.
    elife-84747-supp3.xlsx (23.5KB, xlsx)
    Supplementary file 4. Gene Ontology (GO) Term enrichment of differentially expressed genes.
    elife-84747-supp4.xlsx (224.3KB, xlsx)
    Supplementary file 5. Shoot area of seedlings grown under different agar and nutrient concentrations.
    elife-84747-supp5.xlsx (14.8KB, xlsx)
    Supplementary file 6. Differentially expressed genes and normalized counts in response to changes in nutrient or agar concentration.
    elife-84747-supp6.xlsx (3.4MB, xlsx)
    Supplementary file 7. Shoot area of different Arabidopsis accessions grown on hard agar (HA) media.
    elife-84747-supp7.xlsx (15.4KB, xlsx)
    MDAR checklist
    elife-84747-mdarchecklist1.docx (99.8KB, docx)

    Data Availability Statement

    Raw sequencing data can be found at the National Center for Biotechnology Information Sequence Read Archive (accession number PRJNA904764). Normalized read counts and raw phenotypic datasets can be found in the Supplementary Material.

    The following dataset was generated:

    Gonzalez S, Swift J, Ecker J. 2022. Mimicking genuine drought responses using a high throughput plate assay. NCBI BioProject. PRJNA904764

    The following previously published dataset was used:

    Lozano-Elena F, Fábregas N, Coleto-Alcudia V, Caño-Delgado AI. 2018. Transcriptomic study of Arabidopsis roots overexpressing the brassinosteroid receptor BRL3, in control conditions and under severe drought. NCBI Gene Expression Omnibus. GSE119382

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES