Adaptation to coastal soils through pleiotropic boosting of ion and stress hormone levels in wild Arabidopsis thaliana

Summary

• Local adaptation in coastal areas is driven chiefly by tolerance to salinity stress. To survive high salinity, plants have evolved mechanisms to specifically tolerate sodium. However, the pathways that mediate adaptive changes in these conditions reach well beyond Na⁺.

• Here we perform a high-resolution genetic, ionomic, and functional study of the natural variation in Molybdenum transporter 1 (MOT1) associated with coastal Arabidopsis thaliana accessions. We quantify the fitness benefits of a specific deletion-harbouring allele (MOT1^DEL) present in coastal habitats that is associated with lower transcript expression and Mo accumulation.

• Analysis of the leaf ionome revealed that MOT1^DEL plants accumulate more Cu and less Na⁺ than plants with the non-coastal MOT1 allele, revealing a complex interdependence in homeostasis of these three elements. Our results indicate that under salinity, reduced MOT1 function limits leaf Na⁺ accumulation through ABA signalling. Enhanced ABA biosynthesis requires Cu. This demand is met in Cu deficient coastal soils through MOT1^DEL increasing the expression of SPL7 and the copper transport protein COPT6.

• MOT1^DEL is able to deliver a pleiotropic suite of phenotypes that enhance salinity tolerance in coastal soils deficient in Cu. This is achieved by inducing ABA biosynthesis and promoting reduced uptake or better compartmentalization of Na⁺, leading to coastal adaptation.

Introduction

Our understanding of the genomic basis of adaptive variation is increasing, but there is still very little known about the pleiotropic consequences of adaptive variation. Powerful models of plant adaptation to environmental variation in mineral nutrient and trace elements offer promising starting points to understand these pleiotropic consequences that arise from the interconnectedness of ion homeostasis networks (Busoms et al., 2015; 2018; Terés et al., 2019). In particular, ionomic approaches enable global contrasting of mineral nutrient and trace element accumulation between species and environments, allowing broad insight into the consequences of adaptive variation at particular alleles. Combined with genome-wide association (GWA) studies, ionomics has allowed a far-reaching assessment of adaptive, naturally evolved changes in mineral nutrient uptake and storage (Huang & Salt, 2016).

Using such approaches, Baxter et al. (2008) identified Molybdenum Transporter 1 (MOT1) as the causal gene driving reduced shoot Mo in two accessions of A. thaliana (Van-0 and Ler-0). Forsberg et al. (2015) showed that variation in shoot Mo content across 283 A. thaliana accessions is controlled by variation in MOT1, and at least some of this variation may be adaptive to soil Mo concentrations (Kiani et al., 2012). MOT1 belongs to the sulphate transporter superfamily, is localized to mitochondria, and regulates whole plant Mo homeostasis in both A. thaliana (Baxter et al., 2008; Tomatsu et al., 2007) and rice (Huang et al., 2019). By dissecting the variance-heterogeneity of MOT1 in detail, Forsberg et al. (2015) characterised two major polymorphisms in MOT1: a deletion in the MOT1 promoter (MOT1^DEL) and a duplication inside of a previous transposable element (MOT1^DUP). The duplication increases the expression of MOT1 and results in higher accumulation of Mo in shoots; the deletion reduces expression of the MOT1 gene and causes a reduction in whole plant Mo content.

We have identified plants harbouring the reference-like MOT1 allele (Col-0-like; MOT1^C) and plants with the two structural variants, MOT1^DEL and MOT1^DUP in numerous natural demes of A. thaliana in the NE of Spain (Busoms et al., 2015), allowing a closer look at the mechanism underlying these observations. Interestingly, the MOT1^DEL allele occurs primarily in coastal environments and based on this distribution we suggest that MOT1 loss-of-function may control tolerance to elevated soil salinity at the coast. Several studies have shown that the application of Mo and Cu to soil increases fitness under salt stress, reducing Na⁺ content and increasing both abscisic acid (ABA) biosynthesis and the ability of plants to take in essential macronutrients such as K⁺ (e.g. Eskandari & Mozaffari, 2014; Wu et al., 2017). Forsberg et al. (2015) detected an association with variation in leaf Mo, independent of MOT1, ~600 kb upstream of MOT1. Using a T-DNA knockout allele, they proposed Copper Transporter 6 (COPT6) as a promising candidate gene for this association, suggesting a concrete genetic link between Mo and Cu homeostasis.

A further connection between Cu and Mo homeostasis is given by the fact that Cu is essential for ABA biosynthesis, because Cu is needed for the synthesis of the molybdenum cofactor Moco (Schwarz & Mendel, 2006). ABA plays a key role in plant responses to abiotic stress such as drought and salinity (Zhang et al., 2006). Under stress conditions, tolerant plants accumulate ABA, inducing stress-responsive genes. The main site of Na⁺ toxicity is usually the leaf blade; therefore, the activation of genes promoting the compartmentalization of Na⁺ before it reaches the leaves is crucial for salinity tolerance (Munns & Tester, 2008). We hypothesized that MOT1 ^DEL has a role in local adaptation to coastal saline habitats by enhancing COPT6 expression, Cu uptake, Moco and ABA biosynthesis, resulting in less leaf Na⁺ accumulation thus allowing salinity stress resilience. To test this, we performed high-resolution, multiyear experiments in the field and under lab-controlled conditions, focusing on the complex interplay between Mo-Cu-Na⁺ and their transporters. This set of laboratory and field-based experiments has led us to conclude that the naturally occurring MOT1^DEL allele has an adaptive advantage in coastal conditions, providing enhanced salinity tolerance (NaCl) through the pleiotropic enhancement of both Cu and ABA levels.

Material and methods

Collection of plant and soil material

Thirty-six demes of A. thaliana were selected in Catalonia (NE of Spain). A ‘deme’ is defined as a small group of A. thaliana plants growing in relatively homogeneous ecological conditions and separated from other groups by at least 35 m. Each year (from 2013 to 2015) in each site samples were collected such that half of each collected plant was taken for ionomic analysis and the other half was used to extract DNA for MOT1 genotyping. Seeds of each individual were collected directly in the field and they were stored in packets over silica gel in a sealed box until used. For the soil elemental composition analysis, we collected three soil samples (approximately 50 grams of soil from the first 10 cm depth) at each site during the first week of May of each year. Soil samples from the coastal and inland field sites were collected twice a month during A. thaliana reciprocal transplants (from February to June of 2013, 2014, and 2015). Soil was air-dried under laboratory conditions, passed through a 2-mm sieve, and stored dry.

Soil and leaf elemental analysis

For soil analysis, we performed three independent soil analyses per site. To characterize the elemental composition of soil, analyses were performed on the 2-mm fraction samples following the extraction method described in Busoms et al. (2018). Plants from the field or the laboratory were sampled by removing 2–3 leaves (1–5 mg dry weight) and washed with 18 MΩ water before being placed in Pyrex digestion tubes. The digestion method and calibration are described in Busoms et al. (2018).

SNP genotyping and sequencing of MOT1

DNA from leaf material was extracted following the method detailed in Supporting Information Methods S1. An SSR marker (Simple Sequence Repeat) was developed based on the deletion in the promoter of MOT1 in Van-0 plants (CS1584) (Baxter et al., 2008) with forward primer 5′-CTCCGGTTATCGCGTTGTAT-3′ and reverse primer 5′-ACTGTCGCCATCAAGGTTTT-3′. We visualized, aligned, and blasted sequences from the NCBI database with Geneious 10.0.9 (Kearse et al., 2012).

Field-based reciprocal common garden experiments

After genotyping plants from 2012 and 2013 collections, we selected two demes containing plants with the MOT1^C allele and plants with the MOT1^DEL named T9 and LLO2. In March of 2014 and 2015, 100 seeds of each deme/MOT1 variant (T9^C, T9^DEL; LLO2 ^C, LLO2 ^DEL) were sown at Blanes Coastal field (41° 40’ 37,64”; 2° 48’ 3,86”) and at Santa Coloma de Farners Inland field (41° 50’ 41,04”, 2° 40’ 36,13”) into 16 x 16 cm squares. We studied the fitness of 10 plants for each deme/MOT1 variant at each site and the other 10 plants were harvested on April of 2014 and 2015 to analyse the leaf ionome. Rosette diameters were measured every week during 2 months and the number of siliques was counted at maturity. Only seed containing siliques were considered. The 2013 and 2014 reciprocal transplants are described in Busoms et al. (2015). Data were reanalysed classifying the plants as MOT1^DEL (shoot Mo²⁺ content less than 3.5 µg/g); MOT1^DUP (shoot Mo content higher than 9 µg/g) and MOT1^C (Supporting Information Dataset S1).

Salinity tolerance assays

Seeds of the mot1-2 (SALK 069683) and copt6-1 (SALK 083438) T-DNA insertion alleles and aba2-1 (CS 156) were obtained from the Arabidopsis Biological Resource Center (Alonso et al., 2003). The same demes selected for the field common garden experiment together with Col-0 and the mot1-2 mutant were used to perform irrigation, hydroponics and salt spray experiments with NaCl to test salinity tolerance and analyse leaf ionome (procedure and growth conditions specified in Supporting Information Methods S2 and S3).

Genotype data of the 1135 A. thaliana strains were downloaded from http://1001genomes.org/ (1135_snp-short-indel.vcf.gz). MOT1 allele type classification and normalized leaf values of 18 ions for the 1135 strains cultivated in common soil under control conditions can be found in Supporting Information Datasets S2 & S3. Moreover, a worldwide set of 306 natural accessions was phenotyped for leaf Na⁺ content. All the accessions are listed in Supporting Information Dataset S4 and MOT1 type is indicated. The experimental design is detailed in Supporting Information Methods S4.

For the complementation study, we quantified the shoot Mo, Na⁺ and Cu content (µg/g dry weight) of 6 plants of each type: Col-0, mot1-2, copt6-1, T9^C, T9^DEL, LLO2^C and LLO2^DEL and F1s from crosses between T9^C x T9^DEL, LLO2^DEL x LLO2^C and T9^DEL x LLO2^DEL cultivated in common soil irrigated with 0.5-strength Hoagland solution (pH 6.0) with 50 mM NaCl for 4 weeks.

Expression (qRT-PCR) analysis

RNA from leaf and root material was extracted following the method described in Supporting Information Methods S1. Primers used for MOT1 (At2g25680), COPT6 (At2g26975), SPL7 (At5g18830), CNX6 (At2g43760), CNX1 (At5g20990) and ABA3 (At1g16540) transcript quantification are detailed in Table S1. For normalization across samples, the expression of the ACTIN 1 gene (At2g37620) was analysed. For each sample, the average value from triplicate real-time PCRs was used to estimate transcript abundance. Data were analysed using the SDS software (Applied Bio-systems version 1,0). Ct values were determined based on efficiency of amplification. The mean Ct values were normalized against Actin1 gene and Ct values calculated as (Ct_Gene- Ct_Actin1). The root and leaf expression of MOT1 was calculated with the 2^−ΔCt method. The relative expression of each target gene was calculated with the 2^−ΔΔCt method (NaCl treatment – Control).

ABA quantification

ABA concentration (nmol·L⁻¹) was quantified by indirect enzyme-linked assay (ELISA) using a Phytodetek ABA test Kit (Agdia, Elkhart, IN, USA) on 3 plants per treatment from T9 (Mix), T1 (MOT1^DEL) and PA10 (MOT1^C) demes along with Col-0 and mot1-2, copt6-1 and aba2-1 mutants cultivated in common soil irrigated with 0 and 50 mM of NaCl for 2 weeks.

Statistical analysis

For the hierarchical clustering, we generated a progressive alignment of 262 SNPs for the MOT1 gene tree and 37,574 SNPs for the whole-genome tree from 74 plants from Busoms et al. (2018). Pairwise genetic distance between individuals and between demes was calculated using the Maximum Likelihood statistical method and Jukes and Cantor substitution model using JMP 13.0 (SAS, 2016). Mean-standardized values (−1< value >1) of elemental contents of soil and leaf material were used to represent the radar plots and compare between MOT1 allelic variants (Supporting Information Dataset S2). One-way or multivariate ANOVA was used to test for significant differences between means of fitness, elemental contents of soil and leaf material, gene expression and geostatistical variables. To test for correlations between two variables a Bivariate Fit was conducted. To perform multiple comparisons of group means we used Tukey’s HSD (Supporting Information Dataset S5) using JMP 13.0 (SAS, 2016).

Results and Discussion

Previous work has provided evidence that A. thaliana populations from NE Spain harbour substantial genetic variability and adaptive variation to elevated salinity in coastal soils at high geographic and genomic resolution (Busoms et al., 2015; 2018). Here we build on these works, focussing on novel structural variation in Molybdenum Transporter MOT1 and the adaptive consequences of this variation.

MOT1 allelic variation and distribution

Genome-wide association studies indicate that the MOT1 locus exhibits contrasting effects on leaf Mo accumulation through different natural alleles (Forsberg et al., 2015; Baxter et al., 2008). Tomatsu et al. (2007) and Baxter et al. (2008) first detected a 53 bp deletion on the MOT1 promoter associated with reduced MOT1 expression and decreased shoot Mo content (MOT1^DEL). Forsberg et al. (2015) also found independent polymorphisms which were associated with increased MOT1 expression and elevated shoot Mo concentration: (i) two forms of a 330 bp long duplication in a transposable element sequence AT2TE47050 (MOT1^DUP); and (ii) two SNP markers, one located ~25 kb downstream (MOT1-SNP1) and another located ~600 kb upstream of the MOT1 coding locus (MOT1-SNP2).

To determine which MOT1 alleles are present in wild populations of A. thaliana adapted to high soil salinity and naturally heterogenous soils in the north-east (NE) coast of Spain (Busoms et al., 2015; 2018), we performed a high-resolution survey over 3 consecutive years (2013-2015). We collected, genotyped and quantified shoot Mo content from 36 demes (small groups of plants growing in relatively homogeneous ecological conditions; see first section of Material and Methods) in a region roughly 60 km across (Fig. 1a). This yielded 25 demes containing only plants with the MOT1 Col-0 reference-like allele, MOT1^C (Fig. 1a: black-coastal and grey-inland dots), while the rest of the demes harboured either mixtures of the other alleles or only MOT1^DEL (Fig. 1a: green dots). Additionally, there were 3 demes located in the north of the area that contained mixtures of plants with either MOT1^C or MOT1^DUP alleles (Fig. 1a: red/black dots). The overall distribution of MOT1 alleles was non-random across the study area: while we did not detect a correlation with soil Mo levels (r²=0.087), we detected a significant association between the presence of MOT1^DEL and the distance to the sea (ChiSq= 324,83; p<0.001), with the demes containing only MOT1^DEL allele being close to the coast (Fig. 1b), suggesting a potential adaptive value of this allele in coastal environments. Moreover, we did not detect MOT1^DEL plants in any inland deme, suggesting a possible cost of the allele under inland conditions.

Fig. 1. Sampling, MOT1 alleles studied, genetic structure, and relationship with soil Mo concentrations.

(a) Geographic location in Catalonia, NE Spain of the 36 demes of Arabidopsis thaliana studied, with the 19 sequenced demes named. (b) Distribution of each MOT1 allele as a function of distance to the sea and soil Mo concentration. The relation between soil Mo and distance to the sea (dotted line) is not significant (r² = 0.087). Black dots: coastal demes consisting of MOT1^C; grey: inland demes consisting of MOT1^C; green: coastal demes consisting of MOT1^DEL; black/green: demes harbouring both MOT1^C and MOT1^DEL; black/red dots: demes harbouring both MOT1^C and MOT1^DUP; and the two field sites Blanes (coastal, blue rhombus) and Santa Coloma de Farners (Inland, orange rhombus). (c) Sequence alignment of the MOT1 locus for MOT1^C, MOT1^DEL, MOT1^DUP alleles and mot1-2 mutant (SALK_069683). Genomic positions, MOT1 CDS, AT2TE47050 and polymorphisms are indicated. Neighbour-joining cladogram of MOT1 allele (based on 262 SNPs) (d) and the whole-genome fourfold degenerate (putatively neutral) sites (37,574 SNPs) (e) among 74 A. thaliana individuals from our study region in NE Spain. In MOT1, the highest probability for the value of K=3 classifies and the three versions of the MOT1 allele: MOT1^C (black), MOT1^DEL (green), MOT1^DUP (red). The same colours have been maintained on the whole-genome cladogram, adding an asterisk (*) to individuals with the MOT1^C allele from demes containing plants with the MOT1^DEL allele.

We analysed the MOT1 locus in 74 resequenced individuals from 19 demes chosen to represent the full range of ecotypic and genetic variation throughout this area (Dataset S1 in Busoms et al. (2018)). We also performed MOT1 transcript expression analysis and quantified leaf Mo levels across selected demes (Supporting Information Fig. S1). In MOT1^DEL individuals we observed the 53 bp deletion, 13 bp upstream from the transcription start site, we also detected an additional non-synonymous change in the coding region (Fig. 1C), and we confirmed the low leaf Mo content and the reduced expression of MOT1 in both roots and leaves (Supporting Information Fig. S1). Individuals with the MOT1^DUP allele have a 322 bp duplication inside of the transposable element AT2TE47050 (Fig. 1c), a higher expression of MOT1 especially in roots, and plants accumulate higher levels of Mo in their leaves (Supporting Information Fig. S1). Although Mo leaf content in MOT1^DEL individuals is reduced, plants do not show symptoms of Mo deficiency: Mo levels are within the normal range (0.4 – 1.0 μg/g) for Brassicaceae species (Mengel et al., 2001).

To gain an understanding of the population genomic relationships among plants harbouring each of the three major MOT1 alleles, we generated a Neighbour-Joining (NJ) cladogram of the MOT1 locus from our 74 resequenced individuals and compared it to the genome-wide cladogram. This group consisted of 22 individuals with MOT1^DEL; 6 with MOT1^DUP; and 46 with the MOT1^C allele (Fig. 1d, e). As predicted, the MOT1 locus cladogram differentiates the MOT1 alleles into three distinct clades (Fig. 1d). Except for the LLO-1 individual that genetically clusters with the individuals from the A5 deme (suggesting a rare case of migration due to the demes being more than 40 km apart, previously discussed in Busoms et al. (2018)), Fig. 1e shows that genomes harbouring MOT1^DEL or MOT1^DUP are nested among MOT1^C sisters, indicating that different MOT1 alleles have recombined onto common genetic backgrounds.

MOT1 influence on coastal adaptation

Despite our dense sampling, we could only find the MOT1^DEL allele in demes located less than 3 km from the coast, suggesting that this allele may have a role in adaptation to coastal environments. To test this idea, we reanalysed data from reciprocal transplants performed in coastal and inland field sites in 2013 and 2014 (Busoms et al., 2015), considering the status of the MOT1 allele version in each individual plant. Fig. 2a shows that plants harbouring MOT1^DEL are consistently better adapted to coastal conditions, often producing over twice as many siliques when grown at the coast and outperforming the rest of the coastal demes.

Fig. 2. Local adaptation of MOT1 alleles and global elemental accumulation analysis.

Fitness (mean ± SE of silique number) of (a) Arabidopsis thaliana plants from inland demes harbouring MOT1^C (grey), from coastal demes harbouring MOT1^C (black), MOT1^DEL (green) or MOT1^DUP (red) alleles cultivated at Coastal and Inland common gardens in 2013 (n=359) and 2014 (n=400); (b) plants from LLO2 and T9 demes harbouring MOT1^C (black) or the MOT1^DEL (green) alleles cultivated at Coastal and Inland common gardens in 2014 (n=40) and 2015 (n=60). Significant differences are indicated with letters (between alleles) or with asterisks (between sites). Normalized difference of 17 elements in (c) soils from coastal (blue) and inland (orange) common gardens collected in the 3 years of cultivations (n= 60); (d) leaves of plants harbouring the MOT1^C (black), MOT1^DEL (green) or MOT1^DUP (red) allele from 2013 (n=293) and 2014 (n=390) reciprocal transplants; and (e) leaves of plants from LLO2 and T9 demes harbouring the MOT1^C (black) or the MOT1^DEL (green) allele from 2014 (n=40) and 2015 (n=40) reciprocal transplants. Elements exhibiting significant differences (t-test, p<0.05) are marked with an asterisk (*).

To isolate the effect of the MOT1 allele and homogenise genetic backgrounds, we selected two mixed demes with plants harbouring MOT1^{C or DEL} alleles (T9^C, T9^DEL; LLO2^C, LLO2^DEL) for further reciprocal transplant experiments at the same coastal and inland sites. In both years (2014 and 2015), plants with the MOT1^C allele from both demes were less fit than plants with the MOT1^DEL allele when grown at the coast. However, no difference was observed between plants with the two alternative alleles at the inland site (Fig. 2b), confirming coastal-specific adaptation of MOT1^DEL harbouring plants. We have been unable to identify plants harbouring the MOT1^DEL in our inland sampling, suggesting a potential fitness cost for this allele in inland environments. However, given our benign common garden field conditions, it is perhaps not surprising that we do not detect a fitness cost over a single generation. That said, we do see differences in fitness proxies between plant harbouring each MOT1 allele as a function of salinity challenge (Fig. 3a, b), with a lack of sodium challenge associated with lower fitness in the MOT1^DEL allele, supporting the idea of a fitness effect. These results underscore a strongly environmental context-dependence on fitness estimations of naturally-evolved alleles.

Fig. 3. Interplay of MOT1 and salinity tolerance.

(a) Fresh biomass (g) of Arabidopsis thaliana plants being exposed to either 0, 50 or 100 mM of NaCl in hydroponic solution during 3 weeks and (b) fitness (silique number produced) of plants cultivated in potting mix soil under controlled conditions and irrigated with 0, 50 and 100 mM of NaCl until maturity. Data represents the mean ± SE of plants with the MOT1^C (black), plants with the MOT1^DEL (green) allele from LLO2 and T9 demes and mot1-2 (yellow) and Col-0 (blue) individuals. Data represent the mean ± SE of 10 plants per accession and asterisks indicate significant differences (t-test, * p<0.05, ** p<0.005, *** p<0.0005). Mean ± SE of (c) relative dry weight (DW treatment/DW control) and (d) survival days after a high salinity treatment of plants with the MOT1^C allele (black), plants with the MOT1^DEL allele (green), Col-0 (blue) and mot1-2 (yellow) plants cultivated in potting mix soil irrigated with 0 (Control), 75 mM (Salt) or 150 mM (High-Salt) of NaCl and sprayed with 0 (CS) or 150 mM (SS) of NaCl for 2 weeks. Bars represent the mean ± SE and boxplots represent the median and quartile range of 6 plants per accession. Letters indicate significant differences (Tukey’s HSD, p<0.05).

During the 3 years of the reciprocal transplants, we monitored the soil mineral nutrient and trace element content at each field site (Supporting Information Dataset S2). Inland and coastal soils consistently differed in Na⁺, Mg, P, Zn, Cu, Mo and Rb concentrations (Fig. 2c). However, the accumulation of these elements in plant tissues did not fully reflect the soil quantities. Instead, a difference emerged specifically related to Mo. At all sites MOT1^DUP plants had high shoot Mo, while MOT1^DEL plants had not only lower Mo, but also decreased Na⁺ (Fig. 2d). MOT1^DEL plants were also able to uptake more Cu, P and K⁺ than MOT1^C plants under coastal conditions (Fig. 2d, e). These results indicate that the allelic variation present in MOT1 not only results in reduced Mo accumulation in the leaves, but this variation is also linked to mechanisms that allow MOT1^DEL plants to exclude Na⁺. Furthermore, we can affirm that this low Mo concentration in MOT1^DEL plants does not adversely affect their normal development as these plants had better or equal growth and reproductive fitness compared to MOT1^C plants when cultivated under natural field conditions, as measured by silique number (Fig. 2b).

MOT1 interaction with Na⁺ and Cu homeostasis

To confirm that the MOT1^DEL allele can mediate a salinity resistance phenotype whilst maintaining low Mo and Na⁺ contents and high Cu content in leaves, we performed controlled environment experiments. We modulated NaCl concentrations in both soil and hydroponics using the same demes we had cultivated in the field (T9^C, T9^DEL; LLO2^C, LLO2^DEL). Without NaCl addition, plants containing MOT1^C or MOT1^DEL alleles grew similarly. After treatment with 50 mM or 100 mM NaCl for 3 weeks, growth of all plants was reduced. However, plants with the MOT1^DEL allele or the mot1-2 mutant grew significantly better than plants with the MOT1^C allele in high NaCl conditions (Fig. 3a). This enhanced salinity tolerance linked to low MOT1 expression was also reflected in fitness. Both the mot1-2 mutant and MOT1^DEL plants cultivated in the same soil produced more siliques than Col-0 and MOT1^C plants when they were irrigated with NaCl at any concentration (Fig. 3b). Ionomic data from these experiments confirmed that plants with the MOT1^DEL allele subjected to salinity stress accumulate more Cu and less Na⁺ in their leaves than plants with the MOT1^C allele under the same conditions (Supporting Information Fig. S2 & Dataset S2).

We have previously established that coastal adaptation in the A. thaliana demes we are studying here is driven primarily by adaptation to elevated soil salinity by assessing the fitness of plants grown in a controlled environment in excavated coastal and inland soil (Busoms et al., 2015). However, we were interested to additionally understand if the MOT^DEL can also confer elevated salinity tolerance to exposure of NaCl applied as a spray to simulate salt spray from the sea. We therefore cultivated the same demes from the previous experiments in potting mix soil and 3 weeks after sowing we started 5 different treatments: (1) Control-CS (plants irrigated and sprayed with 0 mM of NaCl), (2) Salt-CS (plants irrigated with 75 mM and sprayed with 0 mM of NaCl), (3) Control-SS (plants irrigated with 0 mM and sprayed with 150 mM of NaCl), (4) Salt-SS (plants irrigated with 75 mM and sprayed with 150 mM of NaCl), and (5) High-Salt-SS (plants irrigated with 150 mM and sprayed with 150 mM of NaCl). After 2 weeks of treatment, plants from the first four treatments were harvested, dried and weighed. Plants from the fifth treatment were visually inspected to record survival. Compared to control plants, coastal demes with the MOT1^DEL allele showed high tolerance to soil salinity, salt spray and the combination of both (Fig. 3c,d). Col-0 plants were highly sensitive to salt spray, dying a few days after the start of the High-Salt-SS treatment. In contrast, mot1-2 plants grew less than MOT1^DEL individuals but always performed better than Col-0 in all the salinity treatments, exhibiting a similar tolerance than plants with the MOT1^C allele (Fig. 3c,d).

To functionally confirm the association between the MOT1 allele and salinity in a broad diversity panel, we challenged a set of 306 natural accessions (14 with the MOT1^DEL allele) with 50 mM NaCl five weeks after sowing, and further added 25 mM NaCl each following week for two weeks. We measured leaf Na⁺ content before and after the treatment. This confirmed on a much broader sampling that MOT1^DEL plants accumulate less Na⁺ in aerial tissues when Na⁺ solution levels are high (Fig. 4a).

Fig. 4. Allelic effects on ion content.

(a) Mean ± SE of Na⁺ content in leaves of plants from 306 Arabidopsis thaliana European accessions (14 MOT1^DEL (green); 292 MOT1^C (black)) cultivated in control conditions during 5 weeks and leaves from the same plants after being irrigated with NaCl during 4 weeks. Asterisks indicate significant differences between MOT1 alleles (t-test, ** p<0.005). (b) Normalized difference of 17 elements in leaves of 953 accessions from the 1135 A. thaliana worldwide collection classified between plants harbouring the MOT1^C (black), the MOT1^DEL (green) or the MOT1^DUP (red) allele. (c) Mean ± SE of shoot Mo, Na⁺ and Cu content (μg · g⁻¹ dry weight) of Col-0, T9^C, T9^DEL, LLO2^C and LLO2^DEL parental plants, F1 plants from crosses between T9^C x T9^DEL, LLO2^DEL x LLO2^C and T9^DEL x LLO2^DEL and mot1-2 and copt6-1 mutants cultivated in potting mix soil irrigated with 50 mM of NaCl for 4 weeks. Boxplots represent the median and quartile range of 6 plants per accession and letters indicate significant differences (F-test, p<0.05).

To further test whether the MOT1 deletion is related to Na⁺ exclusion, we analysed the leaf ionome profiles of 953 accessions from the 1135 A. thaliana collection (Campos et al., 2021) after growth in non-saline potting mix (Supporting Information Dataset S2). We classified them into three groups (MOT1^C: black; MOT1^DEL: green; MOT1^DUP: red) and located them on the on-line map (Supporting Information Fig. S3a). Geographical variables (‘Country’ and ‘Distance to sea’), and 18 estimated soil parameters based on Lucas 2009/2012 topsoil data (Panagos et al., 2012) were tested. We could discern no clear association with the presence of MOT1 allelic variation and these variables (Supporting Information Fig. S3b). However, leaf Mo and Na⁺ content were significantly lower in plants with the MOT1^DEL allele when cultivated under common conditions (Fig. 4b), indicating a world-wide effect of the presence of these MOT1 alleles on the accumulation of Mo and Na⁺. We also observe elevated K⁺ associated with MOT1^DEL.

The main interaction between Mo and Cu occurs via the synthesis of the Mo containing cofactor MoCo, where COPT transporters play a role. In particular, the Cu transporter COPT6 has been associated with MOT1 (Forsberg et al., 2015; Billard et al., 2014). To examine whether polymorphisms in MOT1 and COPT6 independently affect Mo and Cu levels in planta, we compared Mo and Cu levels in mot1-2 and copt6-1 plants cultivated in common soil irrigated with 50 mM NaCl for 3 weeks (Fig. 4c). In both mot1-2 and copt6-1 mutants leaf Mo contents are low in comparison to the wild type Col-0. Moreover, the mot1-2 mutant exhibited elevated leaf Cu levels, suggesting a potential link between these two genes (Fig. 4c).

To further confirm that the MOT1 locus is responsible for the ionomic changes detected in the Catalonian demes from NE Spain, we performed a complementation test, crossing LLO2^DELxLLO2^C, T9^CxT9^DEL and T9^DELxLLO2^DEL. F1 and parental plants were cultivated together with mot1-2 and copt6-1 mutants. F1 plants from the LLO2^DELxLLO2^C, T9^CxT9^DEL crosses contained intermediate levels of shoot Mo, Na⁺ and Cu relative to the parents (Supporting Information Dataset S2), indicating complementation. Only the F1s from the T9^DELxLLO2^DEL cross had significantly lower shoot Mo and Na⁺ (p<0.005) and higher Cu, similar to the MOT1^DEL parents (Fig. 4c), confirming that T9^DEL and LLO2^DEL are allelic with the recessive loss of function MOT1 allele.

MOT1 link to ABA signalling and salinity tolerance

Soil ionome analyses showed that our coastal soils, apart from being rich in Na⁺ due to the marine influence (Busoms et al., 2015), contain lower levels of Cu compared to inland soils (Fig. 5a). Such levels are not considered critical for normal plant growth but we detected that leaf Cu content in the majority of plants with the MOT1^C allele growing at the coast is low (between 1 and 7 μg/g), levels that may indicate Cu deficiency (Marschner, 1995). However, coastal plants with the MOT1^DEL allele are able to accumulate more Cu inside the plant (Fig. 5b). Expression of SPL7 and COPT6 are elevated in plants containing MOT1^DEL suggesting that these genes are playing a role in this elevated leaf Cu (Fig. 5c). This is also supported by the enhanced leaf Cu and SPL7 expression we observe in the mot1-2 mutant (Fig. 4c; Fig. 5c). SPL7 activates the transcription of multiple genes involved in Cu homeostasis in A. thaliana (Yamasaki et al., 2009) and the COPT6 plasma membrane transporter has previously been shown to be induced under Cu deficiency conditions (Garcia-Molina et al., 2013; Puig, 2014; Peñarubia et al., 2015). In consequence, MOT1^DEL plants can maintain leaf Cu levels within the adequate range despite growing in soils with lower Cu concentrations, in contrast to MOT1^C plants.

Fig. 5. Stress signalling and MOT1.

(a) Mean ± SE of soil Na⁺, Cu and Mo content of 9 soil samples per deme and (b) mean ± SE of leaf Na⁺, Cu and Mo content of Arabidopsis thaliana plants collected from their natural habitat classified as coastal demes harbouring the MOT1^C allele (black), coastal demes harbouring the MOT1^DEL allele (green), demes harbouring the MOT1^DUP allele (red) and inland demes harbouring the MOT1^C allele (grey). (c) Relative transcript expression (treatment (50 mM NaCl) vs control (0 mM NaCl)) of SPL7, COPT6, CNX6, CNX1 and ABA3 in shoots of triplicate bioreps of plants cultivated in hydroponics under control conditions or exposed to 50 mM of NaCl for 2 weeks. Data represent the mean ± SE of 3 plants per accession. (D) ABA content (nmol · L⁻¹) of plants cultivated in common soil irrigated with 0 and 50 mM of NaCl for 2 weeks. Colour code of each accession is described into the figure. Boxplots represent the median and quartile range of 3 plants per accession. Letters indicate significant differences (Tukey’s HSD, p<0.05).

COPT6 is localized mainly in the vasculature of green tissues where it can facilitate Cu redistribution (Garcia-Molina et al., 2013) and also plays an important role on ABA biosynthesis. ABA regulates SPL7 and subsequently affects the expression of its COPT targets (Carrió-Seguí et al., 2016). Noting the potential relationship between COPT6 and MOT1 (Forsberg et al., 2015), we hypothesized MOT1^DEL may mediate a degree of salinity tolerance at the coast due to enhancement of Cu and ABA response. We therefore quantified the expression of the main genes involved in MoCo biosynthesis, a critical cofactor in ABA biosynthesis, and ABA levels in plants with various MOT1 alleles submitted to salt stress.

The formation of active Mo containing enzymes depends not only on the availability of Mo but also on the presence of the two metals Fe and Cu. It is assumed that Cu binds to the dithiolate group of MPT at the end of step 2 of MoCo biosynthesis (Mendel, 2007). In this second step, MPT is initially formed by the incorporation of two sulfur atoms into cPMP in a reaction catalyzed by MPT synthase, a heterotetrameric complex of Cnx6 and Cnx7 subunits. The two Cnx7 small subunits are responsible for S transfer into the complex but the MPT formation is mainly regulated by CNX6 (Ide et al., 2011). In the next step, MPT is adenylated by Mo insertase CNX1, forming MPT-AMP. In the last step, MPT-AMP is transferred to the Cnx1 E-domain where Mo is inserted by replacing Cu, forming MoCo which is finally allocated to Mo apoenzymes (Kaufholdt et al., 2013), such as aldehyde oxidase (AO). ABA3 is an enzyme required for the enzymatic activity of AO, the enzyme essential for the biosynthesis of bioactive ABA (Watanabe et al., 2018). We observed that salt stress increased the expression of CNX6, CNX1 and ABA3 in all genotypes but they were strongly upregulated in the mot1-2 mutant and MOT1^DEL plants (Fig. 5c). Consequently, ABA levels were also higher in MOT1^DEL plants after salinity treatment (Fig. 5d), supporting the hypothesis that MOT1^DEL allele provides an enhanced ABA response to saline challenge.

We found that ABA levels in copt6-1 mutants were significantly lower in all treatments (Fig. 5d), supporting the role of Cu in ABA biosynthesis. Under salinity stress, plants harbouring the MOT1^DEL allele showed the highest induction of ABA biosynthesis (Fig. 5d). This may allow MOT1^DEL plants to maintain their normal growth even after being treated with 100 mM of NaCl (Supporting Information Fig. S4b).

At the cellular level, both Mo and Cu are required for MoCo biosynthesis but the amount of each element can act either positively or negatively during this process (Peñarubia et al., 2015). It has been reported that CNX1 binds to the cytoskeleton (Schwarz et al., 2000) and could have a role as a Mo or Cu sensor, interacting with MOT1 or COPT transporters to regulate Mo and Cu concentration in the cell (Tejada-Jimenez et al., 2009). Sulfurylated MoCo is a cofactor of ABA–AO that functions in the last step of ABA biosynthesis and increased ABA levels being involved in responses to salinity has been well documented (e.g. Park et al., 2016). Therefore, the crosstalk between Mo-Cu-Na⁺ via ABA synthesis may explain the greater tolerance to salinity of our plants harbouring the MOT1^DEL allele (Fig. 6).

Fig. 6. Model for the salinity tolerance strategy of MOT1^DEL plants adapted to coastal conditions.

ABA, abscisic acid.

Conclusions

Complex interactions occur among the homeostatic controls of various nutrients in plants (Kearse et al., 2012). However, a comprehensive understanding of the interactions between pathways integrating the homeostasis of key elements is still lacking, especially as they apply to adaptive contexts in nature. Here we focused on allelic variation detected on the A. thaliana Molybdenum Transporter MOT1, in particular on the most prevalent variant, MOT1^DEL, which is characterized by low MOT1 expression and reduced leaf Mo content. We explore the adaptive value of this low expression in the context of the linked phenotype of altered Na⁺ and Cu homeostasis. Within our study area in NE Spain, A. thaliana plants harbouring the MOT1^DEL allele could only be found near the salinity-rich coast. Reciprocal transplant experiments demonstrated that plants with the MOT1^DEL allele are better adapted to these coastal conditions than those with the normal-functioning MOT1^C allele. Moreover, salt stress challenge experiments demonstrated that the MOT1^DEL allele is associated with enhanced salinity tolerance. Apart from Na⁺, leaf Cu content was the only element that was consistently elevated under salinity challenge. This underscores the relationship between COPT6 and MOT1 and the importance of Cu in MoCo and ABA biosynthesis. We accordingly suggest that MOT1^DEL has an adaptive advantage in coastal conditions, due to pleiotropic enhancement of both Cu and ABA levels.

Supplementary Material

DataSetS1

Dataset S1:MOT1 allele type classification and leaf Mo content of the plants cultivated in the reciprocal transplants.

DataSetS3

Dataset S3:MOT1 allele type classification and coordinates of the worldwide set of 1135 A. thaliana natural accessions.

DataSetS4

Dataset S4:MOT1 allele type classification and leaf Na⁺ content of a worldwide set of 306 A. thaliana natural accessions.

supinfo

Methods S1. DNA and RNA extractions for MOT1 genotyping and target genes expression.

Methods S2. Irrigation and hydroponic experiments procedures.

Methods S3. Methods S3 Salt spray and survival experiment design.

Methods S4. Salinity test design of 306 A. thaliana world-wide accessions.

Table S1: Primers list for transcript quantification of target genes.

Fig. S1: Leaf Mo content and roots and shoots MOT1 expression.

Fig. S2: Ionomic profiles of MOT1 alleles exposed to salinity.

Fig. S3: Map and geostatistical analysis of the 1212 world-wide A. thaliana accessions classified by its MOT1 allele type.

Fig. S4: Growth of plants with different MOT1 alleles exposed to soil salinity and salt spray.

DataSetS5

Dataset S5: Statistical analysis of all the comparisons performed in this study.

DataSetS2

Dataset S2: Soil and leaf ionome analysis from all the experiments conducted in this study.