Living with high potassium: Balance between nutrient acquisition and K-induced salt stress signaling

Abstract

High potassium (K) in the growth medium induces salinity stress in plants. However, the molecular mechanisms underlying plant responses to K-induced salt stress are virtually unknown. We examined Arabidopsis (Arabidopsis thaliana) and its extremophyte relative Schrenkiella parvula using a comparative multiomics approach to identify cellular processes affected by excess K and understand which deterministic regulatory pathways are active to avoid tissue damages while sustaining growth. Arabidopsis showed limited capacity to curb excess K accumulation and prevent nutrient depletion, contrasting to S. parvula which could limit excess K accumulation without restricting nutrient uptake. A targeted transcriptomic response in S. parvula promoted nitrogen uptake along with other key nutrients followed by uninterrupted N assimilation into primary metabolites during excess K-stress. This resulted in larger antioxidant and osmolyte pools and corresponded with sustained growth in S. parvula. Antithetically, Arabidopsis showed increased reactive oxygen species levels, reduced photosynthesis, and transcriptional responses indicative of a poor balance between stress signaling, subsequently leading to growth limitations. Our results indicate that the ability to regulate independent nutrient uptake and a coordinated transcriptomic response to avoid nonspecific stress signaling are two main deterministic steps toward building stress resilience to excess K⁺-induced salt stress.

The ability to decouple potassium uptake from nitrogen uptake is important for surviving high potassium-induced salinity stress in plants.

Introduction

Can excess potassium (K⁺) in soil be too much of a good thing for plants? Its role as an essential macronutrient for plants is well established (Wang and Wu, 2013). Because of the known plant growth promoting effects of K when supplemented in agricultural fields, majority of past studies have focused on K⁺ uptake into plants from low concentrations in the soil. Soil salinity is defined as the soluble salt content in soil composed of a combination of Na⁺, K⁺, and other ions (Raza et al., 2022) contrary to the majority of genetic studies exploring salinity stress on plants exclusively using Na⁺. Consequently, we barely understand excess K⁺-induced salt stress in plants. Additionally, toxicity in plants exposed to high concentrations of plant nutrients such as boron (Wang et al., 2021), copper (Lequeux et al., 2010), and nitrogen (Yoshitake et al., 2021) have been investigated, while the systematic evaluation of molecular phenotypes and associated mechanisms behind K⁺ toxicity are less studied.

Worldwide there are many soils naturally high in K⁺ leading to saline soils (Kresge et al., 1988; Duval et al., 2005; Warren, 2016). Eighty-five percent of salt-affected top soils are categorized as agriculturally unfavorable saline soils due to a combination of salts including sodium- and K-salts while only the remaining 15% are categorized as sodic soils enriched specifically in Na⁺ (FAO, 2021, 2022). Potassium is the seventh most abundant element in the continental crust and weathering of K-containing minerals is one of the major sources of high K⁺ in soils followed by excess application of fertilizer (Jobbágy and Jackson, 2001; David, 2010). The use of recycled water in agriculture is another indirect anthropogenic cause resulting in extremely high K⁺ levels in top soils. Many industrial and agricultural processing plants produce wastewater with exceptionally high K⁺ concentrations (Arienzo et al., 2009). During wastewater treatments, unlike N, P, or organic matter which are typically processed using microbial activity, K does not get reduced but rather gets concentrated due to evaporation. This often leads to K⁺ levels exceeding 100–1,000-fold compared to nutrient levels optimal for plant growth (Arienzo et al., 2009). The need to use currently abandoned or low yielding agricultural lands due to salinity and recycled wastewater for irrigation due to freshwater scarcity is imperative (IPCC, 2021). These needs cannot be addressed without foundational knowledge on how plant nutrient balance is achieved in the absence of prime agricultural land. Therefore, knowing the tolerance mechanisms against high K becomes an impending need in our quest to convert marginal lands into productive agricultural lands.

Past studies have reported that excess K⁺ severely affect growth of multiple crops and even halophytes (Eijk, 1939; Ashby and Beadle, 1957; Eshel, 1985; Matoh et al., 1986; Wang et al., 2001; Ramos et al., 2004; Richter et al., 2019; Zhao et al., 2020). These studies suggest that K⁺ may not elicit the same physiological or metabolic stresses Na⁺ does and plants may require distinct genetic pathways in addition to canonical salt response pathways (Pantha and Dassanayake, 2020) to survive high K-induced salt stress.

In this study, we aimed to identify K-induced salt stress responses and deduce the underlying cellular mechanisms plants have evolved to adapt to K-toxicity. We compared Arabidopsis (Arabidopsis thaliana), sensitive to high K⁺, to its extremophyte relative, Schrenkiella parvula that thrives in high K⁺ soils (Nilhan et al., 2008; Oh et al., 2014). We examined stress responses exhibited by the two model species to high K⁺ treatments using a multiomics approach (Supplemental Figure 1) to identify the relevant genetic pathways influential in regulating or are affected by K-toxicity. Our results revealed an extensive ionomic, metabolic, and transcriptomic reprogramming during high K stress in the stress-sensitive model while providing insights into how the extremophyte model had evolved alternative metabolic and transcriptomic adjustments to enable growth under excess K.

Results

KCl is more toxic than NaCl at the same osmotic strengths

Excess K⁺ exerted more severe growth disturbances than observed for Na⁺ at the same concentrations in both species (Figure 1; Supplemental Figures 2 and 3). The extremophyte, S. parvula, was more resilient to higher concentrations of KCl than Arabidopsis before it showed significant alterations to root, leaf, or silique development (Figure 1, B–D; Supplemental Figures 2 and 3). CO₂ assimilation was reduced in Arabidopsis in response to 150 mM KCl within 24 hours after treatment (HAT), corresponding to a reduction in total shoot carbon (Figure 1, E and F). Under long-term treatments, there was a greater reduction in total leaf area in both species with excess K⁺ than Na⁺ compared to control conditions (Figure 1D). These results suggest that despite similar levels of osmotic stress elicited by K⁺ and Na⁺ at equal concentrations, K⁺ may exert additional stresses different from Na⁺ that may initiate unique responses.

Figure 1.

KCl is more toxic than NaCl at the same osmotic strength. A, Seedlings of Arabidopsis thaliana (Ath; on the left for each plate) and S. parvula (Spa; on the right for each plate) on 1/4 MS media supplemented with 0 (control) to 200 mM NaCl and KCl. Black line on plates refers to the end of the root growth. The control plates for NaCl and KCl are same. B, Primary root length measured on 17-day-old seedlings grown under conditions used in (A) based on a 12-day treatment of NaCl or KCl. C, Average root hair length (average length taken from 10 longest root hairs) measured under the same growth conditions used in [a] for variable KCl treatments and monitored for a week. D, Leaf area measured for hydroponically grown Arabidopsis and S. parvula treated with increasing levels of NaCl and KCl until the appearance of first flower. E, Photosynthesis measured as the rate of CO₂ assimilation of the entire shoot/rosette in 25-day-old hydroponically grown plants and monitored up to 72 HAT. F, Total carbon as a % weight based on total dry mass for root and shoot tissue under conditions used in (E). G, 21-day-old plants treated for 14 days (salt applied every other day). Imaged 20 days after the completion of the treatment. Arabidopsis plants treated with KCl show more severe salt stress effects compared to the same concentration of NaCl. In Arabidopsis, 200 mM KCl treated plants did not develop any flowers (as indicated by arrowhead) while 200 mM NaCl treated plants flowered and developed siliques. In S. parvula, a similar observation was made at 400 mM KCl as indicated by yellow arrowhead. The control plants used in the NaCl and KCl treatment assays were the same for both species. Minimum five plants per condition used in B and a minimum of three plants per condition used for D and E. Asterisks indicate significant changes between the treated samples to its respective control samples (t test with P ≤ 0.05). Same letters in the bar graph indicate that those samples are not significantly different (one-way ANOVA followed by Tukey's post-hoc test, P-value ≤ 0.05). Data are presented as a mean of at least three independent biological replicates ± SD. Open circles indicate the measurement from each plant (B,C) and replicate (D–F). DAT, days after treatment; HAT, hours after treatment.

To eliminate the possibility that plant toxicity effects at 150 mM KCl treatments were caused by excess K⁺, and not by excess Cl⁻, we further treated both species with different K-salts (KNO₃ and K₂SO₄) that presented similar osmotic strengths with constant [K⁺]. In Arabidopsis both primary root length and lateral root number consistently significantly decreased compared to control conditions under high K⁺ despite the anion used for the salt treatments (Supplemental Figure 4). Our results are consistent with the observations Zhao et al., 2020 reported for Arabidopsis grown under 100 mM K⁺ and Na⁺ salts given with Cl⁻ or SO₄²⁻ to deduce that K was more toxic compared to Na⁺ and Cl⁻ at comparable concentrations.

Overall, 150 mM KCl was sufficient to induce physiological stress responses in Arabidopsis within a 3-day period, while impacting long-term growth in S. parvula. Notably, KCl treatments at 200 mM KCl given for 2 weeks were lethal to Arabidopsis in tested conditions (Figure 1G). Fresh weight of Arabidopsis plants grown hydroponically for 28 days significantly decreased within 24 h of 150 mM KCl treatment albeit no significant change in dry weights in either species (Supplemental Figure 2B). Therefore, we wanted to ensure that this treatment duration and intensity did not cause excessive internal cell death in Arabidopsis roots, although Arabidopsis did survive long-term 150 mM KCl treatments. We assessed cell viability in treated plants using propidium iodide (Supplemental Figure 5). We observed intact cells and root tissue with new root hair and lateral root development indicative of actively growing roots in the absence of excessive cell death up to 72 h of exposure to 150 mM KCl. Therefore, we selected 150 mM KCl for our cross-species comparative-omics study to investigate K⁺ toxicity responses, when both species are expected to show active cellular responses during the early stages of the treatment at 0, 3, 24, and 72 HAT (Supplemental Figure 1).

Excess K⁺ causes nutrient depletion in Arabidopsis but not in S. parvula

Plants have evolved multiple transporters to facilitate K⁺ entry into roots rather than its exit pathways (Shabala and Cuin, 2008). Therefore, we hypothesized that high K⁺ in soils will lead to high K⁺ contents in plants and that shoots will accumulate more of the excess K⁺ in the transport sequence of soil–root–shoot. We predicted that the differential tissue compartmentalization of K⁺ will alter K-dependent cellular processes. With elemental profiling of 18 nutrients and six common toxic elements (Supplemental Methods, Supplemental Figure 1, Supplemental Table 1), we aimed to test if high K⁺ would (1) accumulate as an isolated process; (2) cause nutrient imbalances; or (3) allow entry of other toxic elements and thereby indirectly cause toxicity symptoms.

Shoot K⁺ levels were higher than levels in roots, but this difference was strikingly larger in S. parvula than in Arabidopsis under control conditions (Figure 2A). Importantly, S. parvula maintained shoot K⁺ levels comparable to its control during K⁺ treatments (Figure 2A). The total K content in whole plants was significantly higher at 72 HAT for both species although we observed there were species-specific differences in root and shoot K distributions that were time-dependent (Figure 2A; Supplemental Figure 6A).

Figure 2.

Excess K accumulation caused severe nutrient imbalance in Arabidopsis compared to S. parvula. A, K accumulated differently between Arabidopsis and S. parvula. Arabidopsis promoted shoot retention while S. parvula promoted root retention of K. B, Increased accumulation of K resulted in N depletion in Arabidopsis compared to S. parvula. C, Differential accumulation of K caused imbalances of multiple nutrients in Arabidopsis roots even in early time points and caused hardly any noticeable differences in S. parvula. D, Percent change in CNPK elemental content in root and shoot of Arabidopsis and S. parvula under high [K⁺]. Arabidopsis fails to maintain major macronutrient levels preventing depletion of those pools contrasted to S. parvula. Data are represented as mean of at least 4 (for A and C) and 3 (for B) independent replicates with ± SD given using ≥ 5–8 hydroponically grown plants per replicate. Elemental compositions were quantified using ICP-MS and total N and C was obtained using an elemental combustion system. Nitrogen and carbon was measured separately but from the same plant samples. The total elemental composition is reported after normalization for dry weight. All quantitative measurements were evaluated by one-way ANOVA followed by Tukey's post-hoc test, P-value ≤ 0.05 and same letters in the bar graph does not differ statistically significantly and asterisks indicate significant changes between the treated samples to its respective control samples. Open circles indicate the measurement from each replicate. HAT, hours after treatment.

Potassium accumulation in Arabidopsis resulted in a severe nutrient imbalance leading to the depletion of seven nutrients including nitrogen (N; Figure 2, B–D). The depletion of total N in Arabidopsis was further detected as a total decrease in the NO₃⁻ content (Supplemental Figure 6B). In line with uninterrupted root growth observed under 150 mM KCl (Figure 1, A–C), S. parvula showed a remarkable capacity in maintaining its macronutrients levels (Figure 2D). Moreover, other elements did not accumulate under excess K⁺ to suggest indirect toxicity effects in either species (Figure 2C). The overall ionomic profiles suggest that constraining K⁺ accumulation while upholding other nutrient uptake processes is necessary for K⁺ toxicity tolerance.

S. parvula is more responsive than Arabidopsis at the metabolome level to excess K⁺ stress

We obtained 472 (145 known and 327 unannotated) metabolite profiles across tissues and time points to identify metabolic processes influenced by K⁺ stress (Supplemental Table 2). The relative abundances of metabolites were highly correlated between the two species for similar conditions (Figure 3A). For downstream analysis, we only considered the known metabolites with significant abundance changes (MACs) between control and KCl treated conditions (Figure 3B).

Figure 3.

S. parvula root metabolome is more responsive than Arabidopsis and induces specific antioxidants and osmoprotectants to adapt to high K⁺ stress in both roots and shoots. A, Arabidopsis and S. parvula metabolome profiles are highly corelated with each other upon high K⁺ in both roots and shoots. Pearson correlation coefficient is calculated for 145 known metabolites (r) and all metabolites including the unannotated metabolites (r′). B, Known metabolites that significantly changed in abundance at 24 and 72 HAT. C, Overview of the temporal changes in primary metabolic pools based on known metabolite annotations. Metabolites that significantly changed in abundance at respective time points are shown followed by a “/” and the total number of metabolites counted in that category. D, Individual metabolites in each functional group mapped to represent their abundance starting at basal level (inner circles) to 24 and 72 h of exposure to high K shown in concentric outer rings. Metabolites mostly highlighted in Results and Discussion are marked with arrow heads in the outer circle in both plots. Significance test for metabolite abundance was performed with one-way ANOVA followed by Tukey's post-hoc test, P-value ≤ 0.05. Data are represented as mean of at least three independent replicates using ≥ 5 hydroponically grown plants per replicate.

We observed three distinct trends from metabolite profile comparisons (Figure 3, B–D). First, with longer stress durations, the number of MACs increased in both species. Second, S. parvula contained more MACs in roots, even though it had fewer physiological and ionomic adjustments compared to Arabidopsis under excess K⁺ (Figures 1 and 2). Arabidopsis shoots had more MACs than in S. parvula consistent with a reduction in photosynthesis seen in Arabidopsis during K⁺ toxicity (Figure 1C). Third, more MACs in S. parvula increased in abundances contrasting to the depletion of those in Arabidopsis (Figure 3D). Amino acids were dominant among S. parvula MACs compared to sugars dominant among Arabidopsis MACs (Figure 3D). Our results suggest that the two species use two distinct primary metabolite groups to respond to excess K⁺ in addition to contrasting shoot-root responses in line with having only a few MACs under shared responses between the species (Figure 3B).

We predicted that S. parvula was enriched in MACs that minimized cellular stress, while MACs in Arabidopsis were more representative of pathways disrupted by K⁺ toxicity. Galactose metabolism was significantly enriched among MACs in S. parvula roots and Arabidopsis shoots (Supplemental Figure 7A). Galactose metabolism includes several key metabolites (Supplemental Figure 7B) known to protect plants against oxidative and osmotic stresses (Taji et al., 2002; Nishizawa et al., 2008). Interestingly, these increased uniquely in S. parvula roots (Supplemental Figure 7B). Antioxidants and osmoprotectants (galactinol, raffinose, myo-inositol, sucrose, and proline; Taji et al., 2002; Gong et al., 2005; Nishizawa et al., 2008) increased in S. parvula compared to Arabidopsis (Figure 3D; Supplemental Figure 7B). Collectively, the extremophyte, S. parvula, seemed to boost the protective metabolites minimizing damage from oxidative bursts in its roots, while the more stress sensitive species, Arabidopsis, was depleting its initial pools of such protective metabolites during stress.

Arabidopsis transcriptional responses are inept against K⁺ toxicity

We examined the transcriptional response landscapes to deduce distinct cellular processes initiated by Arabidopsis and S. parvula upon excess K⁺. S. parvula showed lesser transcriptome-wide variation than Arabidopsis based on a PCA of 23,281 ortholog pairs (Figure 4A). Transcriptome similarities between tissues were consistent with high correlations between samples within a tissue/species (Supplemental Figure 8). The total number of differentially expressed genes (DEGs) was remarkably higher in Arabidopsis (9,907 in roots and 12,574 in shoots) compared to those in S. parvula (1,377 in roots and 494 in shoots) calculated using raw count value (Supplemental Figure 9, Supplemental Tables 3 and 4).

Figure 4.

Arabidopsis shows nonselective transcriptional modifications during excess K⁺ stress than S. parvula. A, Principal component analysis (PCA) of ortholog (23,281 ortholog pairs) expression between Arabidopsis and S. parvula root and shoot transcriptomes at 0, 3, 24, and 72 HAT. B, Enriched functional processes based on GO annotations associated with DEGs in Arabidopsis. The temporal sequence of enriched functions is given as 3 h specific, 3 and 24 h together with 24 h specific, 24 and 72 h shared, 72 h specific, and present at all-time points from 3–24–72 h. Only functional processes that were detected at least in two time points are shown and the processes are sorted based on their functional hierarchy when applicable. C, Arabidopsis accumulated higher levels of hydrogen peroxide (H₂O₂) and superoxide ions (O₂⁻) in leaves under excess K⁺ stress compared to S. parvula. Hydroponically grown plants were used treated similar to the plants used for the RNA-seq study and stained with DAB and NBT. D, Transcriptional profiles of selected pathways that serve as indicators of transcriptional mismanagement in Arabidopsis during excess K⁺ stress compared to S. parvula. DEGs were called using DESeq2 with a P-adj. value based on Benjamini–Hochberg correction for multiple testing set to < 0.01. Data are represented as mean of three independent replicates using ≥ 5 hydroponically grown plants per replicate.

We annotated all DEGs with a representative gene ontology (GO) function (Supplemental Tables 4 and 5) and examined their time-dependent progression in Arabidopsis (Figure 4B; Supplemental Figure 9, Supplemental Tables 6–9). The temporal transcriptomic response was the highest at 24 and 72 HAT in roots and shoots, respectively (Figure 4B). One of the primary emergent functions enriched in Arabidopsis was response to stress (Figure 4B). Within this category, the prevalent specific responses across all-time points were responses to salt, oxidative, and ionic stresses. Biotic stress-related GO terms (e.g. response to bacterium, response to chitin, and immune response) were also prominently enriched at all-time points (Figure 4B) in Arabidopsis, which seemed counterintuitive when S. parvula orthologs of those were mostly either unchanged or suppressed (Supplemental Tables 4 and 8). We examined if this was a sign of transcriptional misregulation at the peak response time inept to the extant stress experienced by Arabidopsis.

We observed three indicators to suggest that there was broad misregulation in transcriptional responses upon excess K⁺ accumulation in Arabidopsis. First, transcriptional responses were enriched for all major plant hormone pathways in shoots beyond the expected enrichment associated with ABA, suggestive of wide-ranging disruptions to hormone signaling (Figure 4B). Second, we found autophagy as the most enriched process among DEGs, accompanied by additional enriched processes including cell death and leaf senescence. Third, enriched responses to cold, heat, wounding, drought, and hypoxia suggestive of unmitigated oxidative stress were prevalent during 24–72 HAT (Figure 4B). Therefore, we assessed the ROS accumulation in Arabidopsis and S. parvula leaves during high K⁺ stress using oxidative stress markers, H₂O₂ and O₂⁻. As predicted by the transcriptional response, leaves of Arabidopsis showed severe oxidative stress compared to S. parvula (Figure 4C). While fewer DEGs at 72 HAT compared to 24 HAT implied a degree of stress acclimation in Arabidopsis (Figure 4B; Supplemental Figure 9), the prolonged ABA signaling together with broad activation of hormone pathways, autophagy, and oxidative stress serve as transcriptional molecular phenotypes to indicate inept responses in Arabidopsis. These molecular phenotypes become even more compelling when compared to their respective orthologous profiles in S. parvula which remained mostly unchanged (Figure 4D).

Carboxylic acid/carbohydrate and amino acid metabolism formed the second largest group among the most affected processes following stress responses (Figure 4B). These processes were enriched at all-time points in Arabidopsis roots and shoots consistent with the previous physiological and metabolic responses that showed primary C and N metabolism were severely affected (Figures 1, E and F, 2D, and3, C and D). Interruptions to photosynthesis at 24 and 72 HAT (Figure 1E) were aligned with the corresponding transcriptional processes enriched at the same time points in Arabidopsis shoots (Figure 4B). Contrarily, root development was detected as a transcriptionally enriched process at 3 HAT (Figure 4B), although the interruptions to root hair growth was detected at 72 HAT (Figure 1C).

S. parvula shows targeted transcriptomic responses steered toward stress tolerance

We searched for DEGs among Arabidopsis whose S. parvula orthologs also showed active responses (Figure 5A). We posited that these orthologs represented cellular processes that require active transcriptional adjustments to survive the accumulation of excess K⁺. We further predicted that the diametric inter-species transcriptional responses (i.e. genes that are induced in one species when their orthologs are suppressed in the other species) will be deleterious to the stress sensitive species, while shared responses will be beneficial yet likely underdeveloped or unsustainable to survive prolonged stress in the stress sensitive species (Figure 5A; Supplemental Figure 9 and Supplemental Table 10).

Figure 5.

S. parvula shows a confined transcriptomic response geared toward concurrent induction of abiotic stress responses and enhanced transcriptional allocation to C and N metabolism. A, The overall expression specificity and response direction of orthologs in Arabidopsis and S. parvula. Selected orthologs are DEGs at least in one time point compared to the respective control condition and then counted as a nonredundant set when all 3, 24, and 72 HAT samples were considered for total counts. B, The proportion contributing to abiotic and biotic stress stimuli within nonredundant DEGs. The effort to suppress biotic stress responses in S. parvula roots (10%) was similar to the proportional induction for biotic stress responses in Arabidopsis (9%). C, The functionally enriched processes represented by DEGs in S. parvula that responded to high K⁺. A nonredundant set from all-time points (3, 24, and 72 HAT) are given. A node in each cluster represents a GO term; size of a node represents the number of genes included in that GO term; the clusters that represent similar functions are given a representative cluster name, ID, and also shares the same color; and the edges between nodes show the DEGs that are shared between functions. All clusters included in the network have adj. P-values ≤ 0.05 with false discovery rate correction applied. More significant values are represented by darker node colors. The functional enrichment network was created using GOMCL. DEGs were called using DESeq2 with a P-adj. value based on Benjamini–Hochberg correction for multiple testing set to < 0.01. Data are represented as mean of three independent replicates using ≥ 5 hydroponically grown plants per replicate.

Response to stress was the largest functional group representative of diametric responses showing induction in Arabidopsis compared to suppression in S. parvula (Supplemental Figure 9). Most of the subprocesses in this cluster were associated with biotic stress (Supplemental Table 10). Therefore, we further assessed this transcriptional divergence between the two species as a proportional effort invested in biotic versus abiotic stress out of total nonredundant DEGs within each species (Figure 5B). The effort to suppress biotic stress in S. parvula roots (10%) was similar to the proportional induction for biotic stress in Arabidopsis (9%; Figure 5B). Contrarily, orthologs that were suppressed in Arabidopsis, but induced in S. parvula roots were associated with ion transport and cell wall organization (Supplemental Figure 9A). These transcriptional adjustments support the physiological response observed for S. parvula where uncompromised root growth was coincident to uninterrupted nutrient uptake during exposure to excess K⁺ (Figures 1B and2C). The orthologs suppressed in Arabidopsis but induced in S. parvula shoots were enriched in carboxylic acid and amine metabolism and transport functions (Supplemental Figure 9B). Collectively, the antithetical transcriptomic, metabolic, ionomic, and physiological responses between the two species support the stress-resilient growth of S. parvula distinct from Arabidopsis (Figures 1–5).

The orthologs that showed shared inductions (156 in roots and 199 in shoots) were largely represented by abiotic stress responses (Supplemental Figure 9 and Supplemental Table 11). The orthologs with shared suppression (149 in roots and 84 in shoots) were enriched in biotic stress responses in roots and photosynthesis in shoots. However, suppressed orthologs associated with biotic stress in Arabidopsis (0.004%) were minimal based on a proportional effort compared to that in S. parvula (10%; Figure 5B).

Over 50% of orthologs differently expressed in response to excess K⁺ in S. parvula roots and ∼30% in shoots showed unique expression trends different from Arabidopsis (Figure 5A). We postulated that S. parvula activates decisive transcriptional regulatory circuits that are either absent (i.e. S. parvula-specific responses) or organized differently (i.e. diametric responses) than in Arabidopsis when responding to excess K⁺ stress.

The overall transcriptomic response of S. parvula encapsulates induction of more targeted salt stress responses than that of Arabidopsis, including oxidative stress responses, sugar and amino acid metabolism, and associated ion transport, with concordant induction in growth promoting processes and transcriptional resource recuperation by suppressing biotic stress responses (Figure 5C). The transcriptional effort to facilitate growth amidst excess K⁺ accumulation in tissues is reflected by induced transcripts involved in cell wall biogenesis, RNA processing, and development along with concurrent suppression for rapid growth limiting processes such as cell wall thickening and callose deposition (Supplemental Table 11).

Balance between differential expression of genes encoding K⁺ transporters

The ability in S. parvula to curb K⁺ accumulation and prevent the depletion of major nutrients (Figure 2; Supplemental Figure 6) led us to hypothesize that genes encoding ion transporters are differentially expressed to prevent nutrient imbalance during excess K⁺ stress. Moreover, ion transport was among the most represented functions within and between species transcriptome comparisons (Figures 4B and5C; Supplemental Figure 9A). We predicted that the genes coding for transporters which allow K⁺ into roots and upload to the xylem or phloem would be primary targets for down-regulation in the effort to restrain K⁺ accumulation, while inducing those that aid in vacuolar sequestration.

We first searched for genes that encoded K⁺ transporters that showed statistically significantly different basal level abundances (Figure 6A) or/and responses to high K⁺ (Figure 6B). We categorized those into four transport routes: (1) limit entry into roots, (2) promote efflux from roots, (3) constrain long-distance transport between root and shoot, and (4) enhance sequestration into vacuoles. We then assessed routes that were potentially weakened in Arabidopsis and/or alternatively regulated in S. parvula.

Figure 6.

Differential expression of K⁺ transporters in Arabidopsis and S. parvula. A, Basal level expression comparison of orthologs between Arabidopsis and S. parvula in roots and shoots. The dash diagonal line marks identical expression in both species and solid lines represent a two-fold change in one species compared to the other species. Transporters/channels with >two-fold change basal differences between the species are labeled. B, The temporal expression of functionally established K⁺ transporters and channels in Arabidopsis and their S. parvula orthologs upon high K⁺ treatment. C, Key K⁺ transporters and channels differently regulated between roots and shoots in Arabidopsis and S. parvula. DEGs at each time point were called using DESeq2 compared to 0 h with a P-adj. value based on Benjamini–Hochberg correction for multiple testing set to < 0.01. Data are represented as mean of three independent replicates using ≥ 5 hydroponically grown plants per replicate.

1. Limit entry into roots: The most suppressed gene (13-fold reduction) in Arabidopsis under excess K⁺ is the HIGH-AFFINITY K⁺ TRANSPORTER 5 (HAK5; Gierth et al., 2005), which encodes a major transporter for K⁺ uptake during low K⁺ availability (Figure 6B). Under sufficient K⁺ levels, the low affinity transporter, ARABIDOPSIS K TRANSPORTER 1 (AKT1) with POTASSIUM CHANNEL 1 (KC1) is activated (Gierth et al., 2005; Wang et al., 2016). HAK5 and AKT1 are activated by the protein kinase complex, CALCINEURIN B-LIKE PROTEINS 1 and 9 (CBL1&9)/CBL-INTERACTING PROTEIN KINASE 23 (CIPK23; Ragel et al., 2015). Under excess K⁺ stress both transporter complexes and their core regulatory unit in Arabidopsis roots were suppressed (Figure 6B and C; Supplemental Figure 10A). Interestingly, the orthologous transporters in S. parvula roots were unchanged but the orthologs of the interacting kinase complex were suppressed (Figure 6B; Supplemental Figure 10A). The transcriptional effort to limit entry of K⁺ into roots during K⁺ stress is further exemplified by the suppression of RELATED TO AP2 11 (RAP2.11), a transcriptional activator of HAK5 (Kim et al., 2012) and the induction of AUXIN RESPONSE FACTOR 2 (ARF2), a repressor of HAK5 transcription (Zhao et al., 2016), in Arabidopsis roots (Figure 6C). Similarly, ARF2 was induced in S. parvula roots. Other K⁺ transporters associated with K⁺ uptake into roots were suppressed in Arabidopsis [e.g. K⁺ UPTAKE PERMEASE 6 (KUP6) and KUP8] under excess K⁺ stress (Figure 6B). Alternatively, S. parvula roots showed a concerted transcriptional suppression of multiple genes encoding cyclic nucleotide gated channels, CYCLIC NUCLEOTIDE GATED CHANNEL 3 (CNGC3)/10/12/13 (Figure 6, B and C). This suggested limitations to nonselective uptake of K⁺ into roots (Gobert et al., 2006; Guo et al., 2008).

2. Promote efflux from roots: K⁺ efflux transporters in plants that exclusively extrude excess K⁺ from roots to soil are unknown. In S. parvula induced expression levels for GATED OUTWARDLY-RECTIFYING K⁺ CHANNEL (GORK; a K⁺ outward rectifying channel; Ivashikina et al., 2001) and SALT OVERLY SENSITIVE 1 (SOS1; Na⁺ exporter; Shi et al., 2000) in roots are notable (Figure 6; Supplemental Figure 11). Induction of GORK is known to cause K⁺ leakage from roots under biotic and abiotic stresses in plants leading to programmed cell death (Demidchik et al., 2014). Arabidopsis did not induce the expression of GORK in roots at any of the tested time points. S. parvula shows induction of GORK in roots without the corresponding transcriptomic signatures for cell death as expected in Arabidopsis (Figures 6, B, C and4, B–D; Supplemental Figure 6). We also note that the basal expression of GORK in S. parvula roots is higher than in Arabidopsis roots (Figure 6A; Supplemental Figure 11). SOS1, the antiporter with the highest Na⁺ efflux capacity in roots, is known for its Na⁺ specificity (Oh et al., 2009). Therefore, induction of SOS1 in S. parvula under high K⁺ stress (Supplemental Figure 11) is likely an effort to counterbalance the increasing osmotic stress due to elevated K⁺ by exporting available Na⁺ from roots. This explanation fits with Na⁺ being the only ion depleted in S. parvula roots during excess K⁺ (Figure 2C).

3. Constrain long-distance transport between roots and shoots: The long-distance transport of K⁺ via xylem loading is mediated by STELAR K⁺ OUTWARD RECTIFIER (SKOR), NITRATE TRANSPORTER 1.5 (NRT1.5), and KUP7 in Arabidopsis (Gaymard et al., 1998; Han et al., 2016; Li et al., 2017). SKOR and NRT1.5 were suppressed in Arabidopsis roots as predicted. However, KUP7 showed induction in Arabidopsis roots at 72 HAT concordantly when K⁺ accumulation was observed in shoots (Figures 2A and6B). Contrastingly, none of these transporters were differently regulated in S. parvula roots. AKT2 is the dominant channel protein regulating long-distance transport via loading and unloading to the phloem (Dreyer et al., 2017). It too is statistically significantly suppressed in Arabidopsis shoots, but unchanged in S. parvula roots and shoots (Figure 6, B and C).

4. Enhance sequestration into vacuoles: Vacuolar [K⁺] is spatiotemporally regulated primarily by Na⁺, K⁺/H⁺ antiporters, NA⁺/H⁺ EXCHANGER 1 (NHX1) and NHX2 and secondarily with higher selectivity for K⁺ by NHX4 (Bassil et al., 2019). The transcriptional signal to promote K⁺ sequestration in Arabidopsis roots or shoots is unclear with mixed regulation among NHXs compared to a more coordinated co-expression in S. parvula shoots (Figure 6, B and C). Furthermore, Arabidopsis induced CA²⁺ ACTIVATED OUTWARD RECTIFYING K⁺ CHANNEL (KCO) genes encoding K ⁺ -selective vacuolar channel known to release K⁺ from vacuoles to the cytosol (Voelker et al., 2006) whereas, S. parvula, suppressed KCO orthologs implying K⁺ sequestration (Figure 6, B and C). Such an attempt is further reinforced by the suppression of tonoplast localized nonselective cation channels CNGC19 and CNGC20 (Yuen and Christopher, 2013) in S. parvula roots.

Concordant to decreased photosynthesis, Arabidopsis shoots represent a molecular phenotype suggestive of closed stomata via an induction of GORK together with a suppression of guard cell localized POTASSIUM CHANNEL IN ARABIDOPSIS THALIANA 1 (KAT1)/KAT2 (Figure 6, B and C; Ivashikina et al., 2001; Szyroki et al., 2001). Such a molecular phenotype is absent in S. parvula. Additionally, a sweeping array of differentially regulated aquaporins and calcium signaling genes were apparent in Arabidopsis compared to limited orthologous responses in S. parvula (Supplemental Figure 10). This reinforces our overall depiction of the stress response in S. parvula to reflect a more restrained response during excess K ⁺.

Excess K⁺-induced nitrogen starvation in Arabidopsis avoided in S. parvula

The reduction in total nitrogen, nitrate, and amino acids in Arabidopsis while those either increased or remained unchanged in S. parvula (Figures 2 and3; Supplemental Figure 6); followed by suppression of amine metabolism-associated genes in Arabidopsis when those were induced in S. parvula (Figures 4B and5C; Supplemental Figure 9B) necessitated further examination on how excess K⁺ may alter N-metabolism in plants. Under low [K ⁺_soil], N uptake in the form of nitrate is tightly coupled to K uptake and translocation within the plant. Many of the K and N transporters or their immediate post-transcriptional activators are co-regulated at the transcriptional level (Coskun et al., 2017). We searched for specific transcriptomic cues to determine how N transport was interrupted under high K⁺, leading to a deficiency in physiological processes needed to maintain growth or creating a shortage of protective metabolites against oxidative and osmotic stress.

The dual affinity nitrate transporter, NITRATE TRANSPORTER 1.1 (NRT1.1/NPF6.3/CHL1) is the main NO₃⁻ sensor/transporter accounting for up to 80% of NO₃⁻ uptake from roots (Feng et al., 2020). Within 3 HAT and onward, NRT1.1 in Arabidopsis roots is down-regulated (Figure 7A). At low [NO₃⁻_soil], NRT1.1 is activated by CIPK23 to function as a high-affinity NO₃⁻ transporter (Coskun et al., 2017). In Arabidopsis (and not in S. parvula) roots, CIPK23 is concurrently suppressed with the main K-uptake system formed of HAK5 and AKT1-KC1 (Figure 6C). This potentially limits the N content in roots within 24 HAT (Figure 2B). Correspondingly, Arabidopsis roots activated N starvation signals by inducing the expression of genes encoding high-affinity NO₃⁻ transporters, NRT2.1 and NRT2.4 (O’Brien et al., 2016) despite sufficient N in the growth medium (Figure 7A). Therefore, Arabidopsis showed a molecular phenotype of K⁺-induced N-starvation.

Figure 7.

Excess K⁺ induced nitrogen starvation and suppressed associated N-assimilation pathways in Arabidopsis compared to S. parvula. A, Nitrogen uptake and distribution was severely affected in Arabidopsis upon high K⁺. The expression changes associated with major nitrogen transporters in Arabidopsis that regulate root uptake and long-distance transport of N are summarized. B, Coordinated transcriptional regulation showing high K⁺-induced suppression of nitrogen assimilation and efforts to accumulate glutamate-derived osmoprotectants. The arrows in front of heatmap blocks indicate the direction of the reaction. C, The primary metabolite pools derived from glutamate in roots and shoots. DEGs at each time point were called using DESeq2 compared to 0 h with a P-adj. value based on Benjamini–Hochberg correction for multiple testing set to < 0.01. Significance test for metabolite abundances was performed with one-way ANOVA followed by Tukey's post-hoc test, P-value ≤ 0.05. Data are represented as mean of at least three independent replicates ± SD (≥ 5 hydroponically grown plants per replicate). Open circles indicate the measurement from each replicate. TCA, tricarboxylic acid; HAT, hours after treatment.

The long-distance transport from root to shoot via xylem loading of NO₃⁻ in roots is primarily regulated via NRT1.5 (NPF7.3) which is a NO₃⁻/K⁺ cotransporter (Li et al., 2017). In Arabidopsis (and not in S. parvula) roots, NRT1.5 was suppressed possibly in an attempt to limit excess K⁺ accumulation in shoots, but consequently depriving NO₃⁻ in shoots (Figures 2B and7A). High K⁺-induced N-starvation in Arabidopsis was further reflected by its additional transcriptional effort to remobilize NO₃⁻ internally. For example, NRT1.7 and NRT1.8 were induced to promote translocation of NO₃⁻ from old to young leaves and from xylem back into roots, respectively, while NRT1.9, NRT1.11, and NRT1.12 were suppressed to restrict transport via phloem in shoots (Figure 7A; O’Brien et al., 2016). The transcriptional regulatory emphasis on NRT1.8 is outstanding during excess K⁺, given that it is the highest induced gene (104-fold and 73-fold at 24 and 72 HAT, respectively) in the entire Arabidopsis transcriptomic response (Figure 7A; Supplemental Figure 11, Supplemental Table 4). Interestingly, NRT1.8 is triplicated in S. parvula (Oh and Dassanayake, 2019), possibly allowing additional regulatory flexibility to redistribute NO₃⁻ via the xylem back to the roots. In agreement with our transcriptomic results, the nitrate content was also reduced significantly upon 150 mM KCl treatment for 72 h in Arabidopsis but not in S. parvula (Supplemental Figure 6).

NH₄⁺ provides another N source and the growth medium included 0.2 mM NH₄⁺ compared to 1.4 mM NO₃⁻. Therefore, we expected to see transcriptional induction of NH₄⁺ transporters to compensate for excess K⁺-induced N-starvation in Arabidopsis roots. NH₄⁺ and K⁺ transport are known to be antagonistically regulated (Coskun et al., 2017). The high-affinity NH₄⁺ transporters (AMTs) are inhibited by CIPK23 (Straub et al., 2017). Counterintuitive to our expectations, AMT1; 1/2/3, which account for >90% of ammonium uptake into roots (Yuan et al., 2007), were co-suppressed in Arabidopsis (Figure 7A).

We next checked whether the N assimilation pathway from NO₃⁻ to glutamine via NH₄⁺ was also suppressed in A. thaliana. Indeed, the genes encoding nitrate reductase (NIA1/NR1, NIA2/NR2) and nitrite reductase (NIR) were coordinately down-regulated in Arabidopsis under high K⁺ (Figure 7B). In angiosperms, the main assimilation point of inorganic N to organic compounds is the GS-GOGAT (glutamine synthetase-glutamate synthase) pathway which is tightly coupled to the N and C metabolic state of the plant (O’Brien et al., 2016). The cytosolic GLN1; 2 and plastidial GLN2 (coding GS enzymes) together with GLT and GLU1 (coding GOGAT enzymes) were suppressed (Figures 2D and7B). Contrastingly, the induction of GLN1; 1 and GLN1; 3 together with GLU2 especially in Arabidopsis shoots may reflect an effort to recycle N under N-starved conditions (Figure 7B).

We predicted that the suppression of the N-assimilation pathway would be reflected in the change in primary metabolites derived from glutamate in Arabidopsis. We checked if Arabidopsis had weakened resources to mount appropriate defenses against osmotic and oxidative stress coincident to the depletion of metabolites directly derived from glutamate that are osmolytes and antioxidants (Figure 7C). Both species showed a coordinated effort to accumulate proline and its immediate precursors via induction of key proline biosynthesis genes (Figure 7B, P5CS1/2, P5CR). However, only S. parvula was able to significantly accumulate proline during exposure to excess K⁺ (Figure 7C). Proline has dual functions as an osmoprotectant and an antioxidant (Hayat et al., 2012). We see similar pronounced efforts in increasing antioxidant capacity via GABA and beta-alanine (Figure 7, B and C), concordant to increased synthesis of raffinose and myo-inositol against osmotic stress in S. parvula (Supplemental Figure 7B). Overall, S. parvula is able to accumulate carbon and nitrogen-rich antioxidants and osmoprotectants by maintaining N uptake from roots and N-assimilation pathways independently from the suppressed K-uptake pathways. Contrastingly, the two processes were jointly suppressed in Arabidopsis leading to the depleted N resources (Figure 2B; Supplemental Figure 6) and, in turn, failure to accumulate C and N-rich protective metabolites (Figure 7C).

A previous study showed that orthologous single-copy genes between Arabidopsis and S. parvula had generally high sequence similarity even between their syntenic promoter sequences and these were less likely to show differential expression in response to salt stress (Oh et al., 2014). Given the differential expression between some of the main K and nitrate transporter genes identified in the current study, we further checked if their putative promoter sequences lost sequence similarity to noticeable levels. The putative promoter sequences were 2 kb upstream syntenic sequences between gene pairs from translation start site. Several key transporter genes including NRT1.1 and NRT1.5 that were down-regulated in Arabidopsis but not differently expressed in S. parvula showed clear loss of sequence similarity in their promoters between the two species compared to some key transporters that did not show differential expression in both species (Supplemental Figure 12).

Co-expressed gene clusters indicate stress preparedness in S. parvula

We generated co-expressed clusters using 14,318 root and 14,903 shoot ortholog pairs of which, we identified five root and three shoot clusters (Supplemental Figure 13, RC1–5 and SC1-3, respectively; Supplemental Figure 13 and Supplemental Table 12). In three root co-expressed clusters, Arabidopsis orthologs showed a maximum response at 24 HAT, while S. parvula showed constitutive responses (Supplemental Figure 13A, RC1-3). These clusters largely represented transcripts associated with stress responses, C and N metabolism, transport, and root development we discussed earlier (Figure 4). The fourth and fifth clusters (Supplemental Figure 13A, RC4-5, 203 ortholog pairs), where S. parvula showed a response, comprised functionally uncharacterized genes (37%) that could not be summarized into representative processes. This highlights the extent of functional obscurity of genes that respond to specific ionic stresses minimally characterized in Arabidopsis (Supplemental Figure 13A, RC5), and the previously unexplored regulatory modes detected in orthologs of closely related species whose functional assignment may have been overlooked due to the lack of responses in Arabidopsis (Supplemental Figure 13A, RC4). In all three co-expressed shoot clusters, Arabidopsis again showed a peak response at 24 HAT, while S. parvula orthologs showed constitutive expression (Supplemental Figure 13B). The enriched functions in shoot clusters largely overlapped to include stress responses and C and N metabolic processes.

S. parvula showed constitutive expression in all clusters (9,633 orthologs) except in RC4 and RC5, while Arabidopsis showed constitutive expression only in RC4 (76 orthologs; Supplemental Figure 13). Overall, these co-expressed clusters between Arabidopsis and S. parvula demonstrate the transcriptome-level stress preparedness in S. parvula to facilitate growth and development during excess K⁺ stress.

Discussion

Salt tolerance mechanisms against high K⁺ are largely unknown compared to the collective understanding for high Na⁺ tolerance in plants. Our results demonstrate that high K⁺ is more deleterious than Na⁺ given at the same external concentrations (Figure 1). Previous studies support this observation noting that excess KCl caused more severe stress symptoms than Na-salts (Eijk, 1939; Ashby and Beadle, 1957; Eshel, 1985; Matoh et al., 1986; Cramer et al., 1990; Wang et al., 2001; Ramos et al., 2004; Richter et al., 2019; Zhao et al., 2020). The canonical adaptations described for salt tolerance mechanisms associated with NaCl-induced salt stress (Pantha and Dassanayake, 2020) are insufficient to explain adaptations required for KCl-induced salt stress.

The extremophyte, S. parvula, amidst high K⁺ can sustain its growth and development; compartmentalize excess K in roots instead of in shoots; maintain uninterrupted nutrient uptake; increase its antioxidants and osmoprotectants; decouple transcriptional regulation between K and N transport; and coordinately induce abiotic stress response pathways along with growth promoting pathways (Figure 8). Contrastingly, the more stress-sensitive plant, Arabidopsis shows, interrupted growth; excessive accumulation of K⁺ in roots and shoots; depletion of essential nutrients; depletion of N-containing metabolites; and sweeping transcriptomic adjustments suggesting initiation of autophagy, ROS accumulation, induction of both abiotic and biotic stresses, and responses to all major hormone pathways (Figure 8). Based on our comparative analyses, we propose two deterministic steps in the overall stress response sequence to survive high K⁺ stress.

Figure 8.

Processes deduced from the multiomics study, physiological assessments, and inferences made for confirmed gene functions in Arabidopsis that collectively suggest major differences between stress-resilient and stress-sensitive growth under high potassium-induced salinity stress.

Surviving K toxicity by avoiding N starvation

K is a macronutrient and plants have evolved many functionally redundant transporters to uptake K⁺ into roots and redistribute within plants (Shabala and Cuin, 2008). When external [K⁺] exceed physiologically optimal conditions, it is not surprising that the immediate response from both plants was to suppress expression of K⁺ transporters that primarily control K⁺ influx into roots (Figure 6). Additionally, S. parvula down-regulated nonselective CNGCs that may be permeable to K⁺ in roots. Several CNGCs are reported to allow Na⁺ or K⁺ transport and have been implicated in their functions during Na-induced salt stress by controlling Na influx into roots. However, their functional and spatiotemporal specificity remains largely unresolved (Dietrich et al., 2020) and needs to be determined before evaluating how selected CNGCs may be involved in limiting excess K influx under K-induced salt stress.

Arabidopsis further seems to suppress long-distance transport of K⁺ via NRT1.5 that co-transports NO₃⁻ and K⁺ (Li et al., 2017). NO₃⁻ is transported as a counterion with K⁺ in root to shoot translocation as described by the “Dijkshoorn–Ben Zioni model” (Dijkshoorn et al., 1968; Zioni et al., 1971; Coskun et al., 2017). The suppression of NRT1.5 limits NO₃⁻ remobilization in plants (Chen et al., 2012). This interference to N transport within the plant is compounded by the transcriptional co-suppression of NRT and AMT transporters known to limit N intake from soil (Figure 7; Tegeder and Masclaux-Daubresse, 2018). This creates an N-starved condition for Arabidopsis not observed for S. parvula (Figure 2 ;Supplemental Figure 6). During limiting K⁺ conditions, N-uptake is down-regulated to prevent excess N-induced toxicity in plants as a favorable mechanism to adapt to K ⁺ -starvation (Armengaud et al., 2004). This interdependent N and K transport and regulation favorable at low [K ⁺_soil] appear to be detrimental at high [K⁺] as it creates an antagonistic pleiotropic effect (condition-dependent traits that can cause positive as well as negative impacts).

We observed a significant induction of NRT1.8/NPF7.2 suggestive of an effort to reimport N from the stele in Arabidopsis (Li et al., 2010). However, this transcriptional effort did not cascade to the ionomic level (Figure 2, B and D; Supplemental Figure 6). N remobilization via induction of NRT1.8 while concurrently suppressing NRT1.5 (Figure 7A) during N starvation is regulated by ethylene-jasmonic acid signaling together with low N-sensing by nitrate reductase (Chen et al., 2012; Zhang et al., 2014). Both ethylene and jasmonic acid signaling are among the enriched differently regulated transcriptional processes in Arabidopsis (Figure 4B). Interestingly, S. parvula appears to have a more flexible and effective regulatory capacity to allow N-uptake decoupled from restricted K-uptake and it does not suppress internal remobilization of NO₃⁻ and K⁺ via NRT1.5 (Figures 6 and7). This may prevent S. parvula from experiencing a high-, K-induced N-starvation.

The depletion of N uptake in Arabidopsis further cascades into depletion of primary metabolites containing N (Figure 3D) with a concomitant transcriptional suppression observed in N assimilation via the GS-GOGAT pathway (Figure 7; O’Brien et al., 2016; Ji et al., 2019). This not only creates a shortage of essential primary metabolites required for growth and development but also depletes essential antioxidants and osmolytes to defend against the mounting oxidative and osmotic stresses (Figures 3, 4, B–D, 7, and8)., High K in the growth medium is known to exert osmotic and oxidative stress (Osakabe et al., 2013; Zheng et al., 2013). This creates an overall need to boost osmotic and antioxidant defense systems to successfully survive high K⁺ toxicity.

Synergistic transcriptional and metabolic resource allocation to increase osmoprotectants during high K⁺ is much more pronounced in S. parvula than in Arabidopsis (Figures 3 and7). Proline accumulation resulting from increased synthesis and reduced catabolism have been widely shown as a key adaptation during salt stress (Kishor et al., 1995; Gong et al., 2005; Kant et al., 2006; Kumar et al., 2010; Hayat et al., 2012).

K and N regulate phosphorus uptake (Coskun et al., 2017; Maeda et al., 2018; Cui et al., 2019), while K⁺ toxicity can induce P-starvation (Ródenas et al., 2019). Arabidopsis experienced severe shortages of multiple key nutrient depletions (Figure 2 and Supplemental Figure 6), which S. parvula seemingly avoided by having independent regulatory capacity of K and N uptake (Figures 6–8). Therefore, we propose that the ability to regulate independent K⁺ uptake is a key deterministic step toward building resilience to excess K⁺.

Avoiding transcriptional misregulation of K⁺ signaling

Multiple hormonal pathways use K⁺ for developmental, biotic stress, and abiotic stress signaling (Zhang et al., 2014; Hauser et al., 2017; Shabala, 2017). Canonical mechanisms involving K⁺ signaling for growth are based on sensing external K⁺ at low or favorable conditions. When supplied with toxic levels of K⁺, Arabidopsis induced nonselective hormone signaling pathways inept to the extant stress (Figure 4) indicative of transcriptional misregulation. Therefore, we propose that the capacity to avoid transcriptional misregulation of K⁺ signaling is the second major deterministic step in surviving high K⁺ stress. If unavoided, it can lead to systemic damage via activation of ROS and autophagy pathways, as demonstrated by Arabidopsis with its increased ROS accumulation perhaps resulting from induction of futile biotic stress responses or unmitigated oxidative stresses induced by high K⁺ detrimental especially at a nutrient starved environment (Figure 4). ROS accumulation combined with autophagy are associated with abiotic and biotic stress responses and developmental processes (Liu et al., 2005; Thompson et al., 2005; Lv et al., 2014; Pantha and Dassanayake, 2020). However, uncontrolled initiation of autophagy signifies a failed stress response strategy (Floyd et al., 2015). In Arabidopsis shoots, autophagy is the most enriched transcriptional pathway. The collective transcriptional signal enriched for lipid catabolism, protein degradation, DNA repair, cell death, and leaf senescence (Das and Roychoudhury, 2014; Figure 4B) likely preceding phenotypic changes expected under long-term exposure to high to excess K⁺ further indicates the maladaptive stress response shown by Arabidopsis contrasted against a preadapted state observed for S. parvula (Figures 5 and8). Previous transcriptome and metabolome characterizations from extremophytes including S. parvula have shown similar stress-ready states for other abiotic stresses (Kant et al., 2006; Lugan et al., 2010; Oh et al., 2014; Wang et al., 2021). Future studies targeting long-term exposure to excess K⁺ complementing the current study will be needed to critically evaluate adaptive traits that are deterministic for plant survival in high K⁺ environments.

In conclusion, upon exposure to high K⁺, plants undergo physiological, metabolic, and transcriptional changes and a subset of those changes lead to stress adaptive traits while the other responses are indicative of failed molecular phenotypes unable to meet the increasing systemic toxicity exerted by excess K⁺ accumulation. The deterministic steps whether a plant would be able to survive K-induced salt stress or descend into unmitigated stress responses were primarily dictated by the ability to regulate K uptake independent from other nutrient uptake pathways while avoiding deleterious signaling processes. This decoupled regulation of K transport and stress signaling can be targeted to design improved crops that are better able to dynamically adjust to a wide array of soils or irrigation water sources with different salt compositions increasingly comprised high K⁺ in freshwater-limited environments.

Materials and methods

Detailed methods are given in Supplemental Methods.

Plant material

Arabidopsis (Arabidopsis thaliana; Col-0) and S. parvula (Lake Tuz) were grown in controlled environments hydroponically, on plates, or in soil at 23°C, 12/12 h light/dark, and 60% relative humidity (Wang et al., 2021). Multiomics data were generated from plants grown in 1/5th Hoagland's solution which included 1.2 mM K⁺ to support plant growth in all conditions. Supplemental KCl (150 mM) treatments provided excess K⁺ beyond its expected range serving as a nutrient (0.1–6 mM; Ashley et al., 2006). Tissues were harvested at 0, 3, 24, and 72 HAT for ionomic, metabolomic, and transcriptomic profiling (Supplemental Figure 1). All samples were harvested at the same time and were 28-day-old when used for –omics data generation with 3–5 biological replicates for each treatment.

Physiological assessments

Root growth assessments were made using plate-grown seedlings scanned and processed with ImageJ (Ferreira and Rasband, 2012). Propidium iodide stained roots were imaged using a Leica SP8 confocal microscope following Fulcher and Sablowski (2009). Reproductive stage assessments were taken from soil grown plants treated with 0–400 mM NaCl or KCl given for at least 2 weeks. CO₂ assimilation rates were measured on hydroponically grown plants using an infrared gas analyzer (LI-6400XT, LI-COR, Lincoln, USA). Leaves were stained for H₂O₂ and O₂⁻ using 3,3′-diaminobenzidine and nitroblue tetrazolium as described in Jabs et al. (1996) and Daudi and O’Brien (2012). Nitrate quantification was conducted on plate-grown 5-day-old seedlings treated with 0, 100, 150 mM KCl for 72 h as detailed in Supplemental Methods.

Elemental profiling

Samples were processed following Ziegler et al. (2013) to quantify K, P, Mg, S, Ca, Al, Zn, Co, Ni, Fe, Se, Cu, B, Mn, Mo, As, Rb, Cd, Na, Li, and Sr via inductively coupled plasma mass spectrometry (Baxter et al., 2014). Carbon and nitrogen were quantified using a Costech 4,010 elemental analyzer (Costech, Valencia, USA). Significant differences were calculated based on one-way ANOVA, followed by Tukey's post-hoc test (P ≤ 0.05) using agricolae and visualized as heatmaps using pheatmap (R packages, v. 1.0.12; Kolde, 2012).

Metabolite profiling

Nontargeted high throughput metabolite profiling was conducted on frozen samples using gas chromatography-mass spectrometry (Fiehn et al., 2008). Additional details on metabolite annotation and quantification are given in Supplemental Methods. Significance tests were performed as described for ionomics and visualized using circlize, R (Gu et al., 2014). MACs between control (0 HAT) and KCl treated conditions were calculated with one-way ANOVA followed by Tukey's post-hoc test, P-value ≤ 0.05.

RNA-seq analyses

Total RNA was extracted from triplicates of root and shoot tissues harvested at 0, 3, 24, and 72 HAT with 150 mM KCl (Supplemental Figure 1). Strand-specific RNA-seq libraries were sequenced on Illumina HiSeq4000. Reads uniquely mapped to Arabidopsis TAIR10 or S. parvula (Dassanayake et al., 2011) v2.2 gene models (https://phytozome-next.jgi.doe.gov/; Supplemental Table 3) were counted for identifying DEGs based on DESeq2 (Love et al., 2014) at P-adj. < 0.01 comparing each time point with the control sample (0 HAT). Orthologous gene pairs were identified using CLfinder-OrthNet (Oh and Dassanayake, 2019) and further refined as described in Wang et al. (2021). Median normalized expression of orthologs was used to determine temporal co-expression clusters based on fuzzy k-means clustering (Gasch and Eisen, 2002). Enriched functional groups were identified using BinGO (Maere et al., 2005; P-adj. ≤ 0.05) and further summarized to nonredundant functional groups using GOMCL (Wang et al., 2020).

Accession numbers

Sequences of all genes associated with this article can be found with relevant accession numbers available in The Arabidopsis Information Resource database and listed in Supplemental Table 4. Additionally, genes highlighted in this article can be found under accession numbers: AKT1, AT2G26650; AKT2/3, AT4G22200; AMT1; 1, AT4G13510; AMT1; 2, AT1G64780; AMT1; 3, AT3G24300; ARF2, AT5G62000; ATKCO1; 1, AT5G55630; CBL1, At4g17615; CBL9, At5g47100; CIPK23, At1g30270; CNGC10, AT1G01340; CNGC12, AT2G46450; CNGC13, AT4G01010; CNGC19, AT3G17690; CNGC20, AT3G17700; CNGC3, AT2G46430; GLN1.1/GSR1, AT5G37600; GLN1.2/GSR2, AT1G66200; GLN1.3, AT1G17290; GLT1, At5g53460; GLU1, At5g04140; GLU2, At2g41220; GORK, AT5G37500; GS2/GLN2, AT5G35630; HAK5, AT4G13420; KAT1, AT5G46240; KC1/KAT3, AT4G32650; KCO2, AT5G46370; KUP6, AT1G70300; KUP7, AT5G09400; KUP8, AT5G14880; NHX1, AT5G27150; NHX2, AT3G05030; NIA1/NR1, AT1G77760; NIA2/NR2, AT1G37130; NPF1.1/NRT1.12, AT3G16180; NPF1.2/NRT1.11, AT1G52190; NPF1.2/NRT1.11, AT1G52190; NPF2.13/NRT1.7, At1G69870; NPF2.9/NRT1.9, AT1G18880; NPF6.3/NRT1.1, AT1G12110; NPF7.2/NRT1.8, AT4G21680; NPF7.2/NRT1.8, AT4G21680; NPF7.3/NRT1.5, AT1G32450; NRT2.1, AT1G08090; NRT2.4, AT5G60770; P5CR, AT5G14800; P5CS1, AT2G39800; P5CS2, AT3G55610; RAP2.11, AT5G19790; SKOR, AT3G02850; SOS1, AT2G01980.

Data availability

The RNA-seq reads generated in this study are available at NCBI-SRA database under BioProject PRJNA63667. Mapped transcripts from this study for S. parvula can be browsed at https://www.lsugenomics.org/genome-browser.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Methods. Additional methods information.

Supplemental Figure S1 . Sampling scheme for the ionome, metabolome, and transcriptome experiments performed in this study.

Supplemental Figure S2 . KCl is more toxic than NaCl at the same osmotic strength shown by salt treatments extended to two weeks.

Supplemental Figure S3 . Root traits showing KCl is more toxic than NaCl at the same osmotic strength.

Supplemental Figure S4 . KCl, KNO₃, and K₂SO₄ showed similar growth responses at comparable osmotic strengths.

Supplemental Figure S5 . Cell viability during 150 mM KCl treatments.

Supplemental Figure S6 . Total potassium and nitrate content in Arabidopsis and S. parvula.

Supplemental Figure S7 . Carbohydrate and nitrogen metabolism related processes were enriched in S. parvula roots.

Supplemental Figure S8 . Overview of transcriptomes in each condition tested in response to high K stress in Arabidopsis and S. parvula roots.

Supplemental Figure S9 . The majority of differentially expressed genes (DEGs) show species specific responses followed by tissue and response time specificity.

Supplemental Figure S10 . Genes coding for calcium signaling and aquaporins are differently regulated during high K⁺ stress.

Supplemental Figure S11 . Normalized expression of GORK, SOS1, HKT1, and NRT1.8 in indicated tissues of Arabidopsis and S. parvula under 150 mM KCl treatments.

Supplemental Figure S12 . Comparison of putative promoter sequences for selected K and Nitrogen transporter genes discussed in the current study.

Supplemental Figure S13 . Co-expressed ortholog gene modules highlight stress preparedness toward high K⁺ stress in S. parvula orthologs that are constitutively expressed compared to induction or suppression of orthologs in Arabidopsis.

Supplemental Table S1 . Elemental quantification of Arabidopis and Schrenkiella parvula root and shoot.

Supplemental Table S2 . Relative abundance of metabolites for Arabidopsis and S. parvula roots and shoots sample.

Supplemental Table S3 . Number of total reads and percentage of uniquely mapped reads to Arabidopsis (TAIR10) or S. parvula v2.2 gene models for root and shoot transcriptomes under high K ⁺ .

Supplemental Table S4 . List of differentially expressed genes (DEGs) in 3, 24, and 72 hours after treatment (HAT) in Arabidopsis and S. parvula root and shoot.

Supplemental Table S5 . Enriched biological processes for a nonredundant set of induced and suppressed DEGs for Arabidopsis roots and shoots sample under 150 mM KCl.

Supplemental Table S6 . Enriched biological processes for time point-specific DEGs (3, 3&24 + 24, 24&72, 72, and 3&24&72) for Arabidopsis roots sample under 150 mM KCl.

Supplemental Table S7 . Enriched biological processes for time point-specific DEGs (3, 3&24 + 24, 24&72, 72, and 3&24&72) for Arabidopsis shoots sample under 150 mM KCl.

Supplemental Table S8 . Enriched biological processes for time point-specific (3, 3&24, 24, 24&72, 72, and 3&24&72) induced and suppressed DEGs for Arabidopsis roots sample under 150 mM KCl.

Supplemental Table S9. Enriched biological processes for time point-specific (3, 3&24, 24, 24&72, 72, and 3&24&72) induced and suppressed DEGs for Arabidopsis shoots sample under 150 mM KCl.

Supplemental Table S10 . Enriched biological processes for diametric responses (i.e. genes that are induced in one species when their orthologs are suppressed in the other species) in Arabidopsis and S. parvula roots and shoots sample under 150 mM KCl.

Supplemental Table S11 . Enriched biological processes for a nonredundant set of induced and suppressed DEGs for S. parvula roots and shoots sample under 150 mM KCl.

Supplemental Table S12 . Normalized gene expression clusters of ortholog pairs (OP) between Arabidopsis and S. parvula in roots and shoots sample.

Author contributions

P.P. conducted experiments and performed data analyses. D.L. supervised and assisted with measuring CO₂ assimilation rates. D.H.O. provided bioinformatics assistance. M.D. developed the experimental design and supervised the overall project. P.P. and M.D. interpreted results and wrote the article with input from all co-authors who revised and approved the final manuscript.

Funding

This work was supported by the US National Science Foundation awards MCB-1616827 and IOS-EDGE-1923589, US Department of Energy BER-DE-SC0020358, and the Next-Generation BioGreen21 Program of Republic of Korea (PJ01317301) awarded to M.D. P.P. was supported by an Economic Development Assistantship award from Louisiana State University.