Evolution of Salt Tolerance in Arabidopsis Thaliana on Siliceous Soils Does Not Convey Tolerance to Saline Calcareous Soils

Purpose Alkaline salinity constrains crop yield. Previously, we found local adaptation of Arabidopsis thaliana demes to saline-siliceous soils (pH ≤ 7) and to non-saline carbonate soils. However, any natural population of A. thaliana was localized on saline-alkaline soils. This suggests that salinity tolerance evolved on saline-siliceous soils may not confer tolerance to alkaline salinity. This hypothesis was explored by addressing physiological and molecular responses to saline-alkaline conditions of A. thaliana demes differing in salinity and carbonate tolerance. Methods A. thaliana native to saline-siliceous soils (G3), to non-saline carbonate soils (G1), or to soils with intermediate levels of these factors (G2) were cultivated in common gardens on saline-siliceous or saline-calcareous substrate. Hydroponics and irrigation experiments conrmed the phenotypes. Growth, mineral concentrations, genome differences, and expression of candidate genes were assessed in the different groups. Results G3 performed best on saline-siliceous soil and in hydroponics with salinity (pH 5.9). However, G3 was more sensitive to saline-alkaline conditions than G1 and G2. Fitness under saline-alkaline conditions was G2 > G1>G3 and G2 best maintained ion homeostasis under alkaline salinity. Whole genome scan did not differentiate among the groups, while distinctive patterns for FRO2, NINJA, and CCB4 were found and conrmed by qPCR. Conclusion In A. thaliana, salinity tolerance evolved on saline-siliceous soils does not provide tolerance to alkaline salinity. Plants from soils with intermediate conditions (G2) have more plasticity to adapt to alkaline salinity than those locally adapted to these individual stress factors. Higher expression of NINJA and CCB4 may contribute to this better adaptation.


Introduction
Saline stress is one of the majors constrains in agriculture. Soil salinization is frequently co-occurring with alkalinity, especially under arid and semi-arid climates, in which high evapotranspiration rates move an excess of soluble ions like Na + , Ca 2+ , Mg 2+ , K + , CO 3 2− and HCO 3 − to the super cial soil layers.
Moreover, soil salinization may destroy soil structure and cause changes in the biological activity (Singh, 2021). Besides natural causes, human activities like crop irrigation are adding more soluble salts to the soils. Including both the natural and anthropogenic origin, it is estimated that 831 million hectares of the Earth are covered by saline-alkaline soils (FAO, 1973). On these soils, plants need to cope with the deleterious effects of both high pH and excess of salts. Salinity causes both ion toxicity and osmotic effects. Plant adaptive mechanisms towards NaCl salinity are well established. Sodium enters by NNCs (non-speci c channels) reducing the membrane potential. A huge cost of energy is spent to remove salt from the cytosol, store it in the vacuoles, or return it to the apoplast by Salt Overly Sensitive 1 (SOS1). The High A nity Potassium transporter 1 (HKT1) is an e cient transporter retrieving Na + from the xylem and limiting Na + transport to the shoots (Sandhu et al., 2017). Weak alleles of HKT1 favor leaf Na + accumulation, which may be an advantage under uctuating and moderate saline conditions through contribution to ABA signaling and osmotic adjustment (Busoms et al., 2018).
The presence of high carbonate and bicarbonate levels affects soil micronutrient availability, especially Fe and Zn (Marschner 1995); moreover, it alters the root cell membrane potential further enhancing nutrient de ciencies. Carbonate/bicarbonate stress causes an imbalance in the plant carbon metabolism due to dark xation of CO 2 /HCO 3 − . The observed enhancement of the production of organic acids and an inhibition of iron translocation from roots to shoots may be a consequence (Alhendavi et al., 1997). Better mobilization of Fe from alkaline soil by root exudation of catechol groups bearing phenolics has been identi ed as a key process in tolerance to alkaline soils in A. thaliana (Terés et al., 2018). Higher expression of FRO2 and IRT1 conveys tolerance to iron de ciency (Connolly et al., 2003). Tolerance to alkaline salinity has been less explored. However, several studies The mechanisms behind these synergistic effects are still not clearly established. Double stress has a strong in uence on the root architecture. In comparison to NaCl stress alone, the root architecture of Lotus tenuis changed to a more herringbone phenotype under mixed salt-alkaline stress (Paz et al., 2012). Under alkaline-saline conditions a strong inhibition of nutrient uptake, especially nitrate and phosphate, in addition to micronutrient de ciencies and a high Na + /K + ratio may injure sensitive plants more than saline stress caused by NaCl under neutral or slightly acidic conditions (Gao et al., 2014). As under saline stress alone, the capacity to accumulate compatible solutes is a key feature of tolerance to saline alkaline conditions. Enhancement of proline (Shi and Sheng, 2005), betaine, and soluble sugar concentrations are quantitively the most important changes of osmotically active substances .
Halophytes have evolved extreme tolerance to saline conditions. Geochemical modelling using species distribution models for Australian grass species suggests a correlation between salt and alkalinity tolerance (Saslis-Lagoudatis et al., 2015). In fact, many well-known halophytes have evolved under extreme saline -alkaline conditions (Akhani, 2006 There is relatively little information on the mechanisms underlying tolerance differences to saline stress under alkaline conditions in crop plants. In Oryza sativa, relative expressions of HKT1 and SOS1 were compared between two contrasting lines differing in alkaline-saline stress tolerance. Tolerant lines were able to more e ciently limit Na + accumulation in the leaves by higher activation of both genes in comparison to sensitive lines. Furthermore, Fe acquisition and rhizosphere acidi cation genes were highly induced to maintain plant nutrient homeostasis (Chuamnakthong et al., 2019). Exploring tolerance mechanisms in natural populations of the model plant A. thaliana locally adapted to moderate levels of salinity and alkalinity may provide insights into the underlying genetic background that will be useful in crop breeding.  Busoms (2015) predicted the lack of A. thaliana in saline-alkaline soil from this region and several surveys in different years con rmed this model prediction (unpublished data). The distribution of A. thaliana is disrupted at coastal locations with saline, lime-rich soils. This indicates that tolerance to saline alkaline eld conditions has still not evolved in this species.
Nonetheless, the presence of tolerance mechanisms to moderate saline or alkaline conditions as individual stress factors in our A. thaliana populations makes these plants an ideal material for exploring the involvement of these tolerance mechanisms to alkaline salinity resistance.
Thus, the objective of this study was to examine the contribution of individual saline and alkaline tolerance mechanisms to the tolerance to the double stress. For this purpose, we analyzed germination, growth, and reproductive tness of A. thaliana demes differing in saline and alkaline tolerance.  (Sonmez et al., 2008). Texture, water holding capacity (WHC) and organic matter was determined following the methods described by Porta (2006). Calcium carbonate content (%) was measured according to Loeppert et al., (1996).
To determine the available mineral nutrient concentrations, 5 grams of soil were dried at 60°C for 48 hours in 50-mL Falcon tubes. Each sample was diluted to 6.0 mL with DTPA-NH 4 and analyzed for B, Ca, Co, Cu, Fe, K, Mg, Mn, Mo, Na, Ni, P, S, and Zn on an ELAN-DRCe ICP-MS instrument (PerkinElmer, Sciex). National Institute of Standards and Technology (NIST) traceable calibration standards (ULTRA Scienti c) were used for calibration (Soltanpour and Schwab, 1977).

Plant ionome
Plant material was dried for 4 days at 60 ºC. Approximately 0.1g was used to perform and open-air digestion in Pyrex tubes using 0,7 mL concentrated HNO 3 at 110 ºC for 5 h in a hot-block digestion system (SC154-54-Well Hot Block™, Environmental Express, SC, Charleston, USA). The concentrations of the following elements (Ca, K, Mg, Na, P, S, B, Mo, Cu, Fe, Mn and Zn) were determined by inductively coupled plasma optical emission spectroscopy ICP-OES (Thermo Jarrell-Ash, model 61E Polyscan, England) (Soltanpour and Schwab, 1977).

Greenhouse experiment
In January 2019, 19 A. thaliana demes and Col-0 were sown in contrasted saline and saline-alkaline soil. Five seeds of each deme were sown in 30 pots (15 of each soil type) and distributed randomly in the greenhouse. Two weeks after germination, seedlings were thinned out so that only one plant per pot was left. Irrigation was applied twice a week. Every week, pictures of the entire rosette were taken. Siliques number was counted at plant maturity. Air temperature, humidity and sun incidence were monitored throughout the experiment.
Salinity-alkalinity tolerance assays Irrigation experiment Plants from A1 (G1), T6 (G3), LG5 (G2), V1 (G2)V3 (G2) demes and Col-0 (REF) were cultivated individually in square pots of 10-cm diameter in sterilized quartz sand. Sterilized seeds were sown on wet soil and the pots were covered with polyvinyl chloride lm until the seedlings had germinated. Pots with germinated seedlings were placed in a growth chamber with 12-h-light/12-h-dark photoperiod, an irradiance of 150 mmol m −2 s −1 , and a constant temperature of 22°C. Plants were watered with ¼ strength Hoagland at neutral pH 5.9 solution every 2 to 3 days. After 2 weeks, seedlings of each deme were split into 4 groups and different treatments were applied (n=8). To avoid osmotic shock, the treatment solutions were gradually increased to achieve the nal treatment conditions after one week. The expression of the target gene relative to the expression of the reference gene was calculated using the 2 − ΔΔCt method (Livak and Schmittgen, 2001).

Statistical analysis
For the hierarchical clustering of the Catalan A. thaliana demes, we generated a progressive alignment of 37574 SNPs for the whole-genome tree of 75 plants from (Busoms et al., 2018). For the clustering of CCB4, NINJA, FRO2 and SOS1 genes the alignment was generated with 134, 25, 37 and 28 SNPs respectively. Pairwise genetic distance between individuals and between demes was calculated using the Maximum Likelihood statistical method and the Jukes-Cantor substitution model.
Data normality was checked for all phenotypes and non-normal data were transformed before applying any parametrical tests. Mean-standardized values ( 1< value >1) of elemental contents of soil and leaf material were used to represent the radar plots and compare each group.

Results
Distribution of native A. thaliana demes according to climate, soil and genetic characteristics The A. thaliana demes used in this study occur spontaneously in the North-East of the Iberian Peninsula (Fig. 1A), which corresponds to the southern edge of the distribution area of this species (Krämer, 2015). Almost all Catalan A. thaliana populations are placed in the same climatic region which corresponds to the Mediterranean coastal climate. Coastal and inland demes have been distinguished based on the salinity concentrations in the soils (Busoms et al., 2015). However, the distribution of A. thaliana demes is clearly interrupted in the saline alkaline areas of the coast (dark green areas in Fig. 1A). Soils in these areas have high salinity, high pH and high carbonate contents ( Supplementary Fig. S2 B-D). Coastal populations only occur on siliceous substrate ( Supplementary Fig. S2A). Previous work has provided evidence that Arabidopsis populations from this region harbor substantial genetic variability and adaptive variation to elevated salinity (Busoms et al., 2015(Busoms et al., , 2018  ; and G3-plants from soils with low CaCO 3 contents and high salinity (0.65% CaCO 3 , 129 Na + mg g -1 ) (Fig. 1B).
Native soil parameters monitored during 2013 to 2015 (Busoms, 2015) were used as input to perform a Principal Component Analysis (PCA). Soil Na, Mg, Mo, chlorides and sulphates groups G3 demes, while Fe, Zn, K, P, pH, CaCO 3 and WHC differentiate G1 demes (Fig. 1C). Correlations between the different parameters of the native soils (pH, OM, WHC and nutrients) can be found in Supplementary Dataset S3. Fig. 1D represents the mean relative values of available mineral nutrients in the native soil of each group.
G1 soil was characterized by higher Na + , Mg and Cl concentrations, while higher K + concentrations were found in soils from G2 and G3. Phosphorus, Zn and Ni concentrations were higher in G3 than in G2 with intermediate levels of carbonate and salinity (Supplementary Dataset S3).
Hierarchical clustering of the whole-genome tree of 74 individuals did not reveal a clear population structure de ning the three established groups (Fig. 1E). The native soils characteristics do not provide a genome-wide differentiation and several lineages are present in different habitats. Despite we do not observe genotypic differentiation by geography, the distance between some demes of the G2 (around 60 km apart) explains why they cluster in two different nodes.

Reciprocal soil transplant experiment
To reveal potential differences in tness among demes in response to salinity on alkaline or siliceous substrate, a reciprocal soil transplant experiment was established in a greenhouse. Plants were grown from seeds in pots lled either with saline-siliceous or saline-alkaline soil. Both soils had a similar sandyloam texture and salinity levels, but clearly differ in pH (6.5 vs. 8.4) and carbonate content (2% vs. 18%) ( Supplementary Fig. S3A). The nutrient availability of both soils was also considerable different. Higher Fe, Zn, Mn, and Mg concentrations were found in the saline-siliceous soil, while Na + and K + concentrations were similar in both (Supplementary Fig. S3B and Dataset S4).
Plant tness was assessed measuring rosette diameter of 5 weeks old plants ( Fig. 2A) and the number of siliques per plant at maturity (Fig. 2B). On saline-siliceous soil, plants of G3, native to saline siliceous soils, performed better than plants from G1 or G2 ( Fig. 2A, B). Contrastingly, on saline-alkaline soil best performance was observed for G2 plants, native to soils with intermediate levels of salinity and alkalinity. Plants adapted to saline-siliceous soils suffered severe reduction of growth and number of siliques when cultivated on saline-alkaline soil ( Fig. 2A, B and Supplementary Dataset S5). The plasticity index based on rosette diameter shows that plasticity is a major factor responsible for the better growth of G2 plants on alkaline-saline soil. Contrastingly, G3 plants were maladapted to alkaline-saline conditions, while G1 plants revealed low plasticity (Fig. 2C).
Plants from different groups also differed in the relative leaf ion concentrations when grown on the contrasting soils (Fig. 2E, F and Supplementary Dataset S6). Plants adapted to saline-siliceous soils (G3) were able to maintain higher K + concentrations on saline-siliceous soil than plants from G1 and G2, while concentrations of Mo and P were considerably lower in G3 plants (Fig. 2F). On the saline-alkaline soil, G2 plants maintained higher leaf concentrations of all analyzed nutrients (Fig. 2F). Sodium concentrations were similar in G2 and G3 plants; only G1 plants restricted better the Na transport to the leaves.

Saline-alkaline stress experiments
To further con rm the differential responses of the A. thaliana demes to saline-alkaline conditions, germination tests and hydroponic cultures were performed. Under control conditions (no NaCl, pH 5.9), the germination rates determined on agar plates did not differ among the three groups. Moderate salinity (50 mM NaCl) applied at slightly acidic pH (pH 5.9) somewhat reduced the germination rates in all three groups to similar extent. Contrastingly, under non-saline alkaline conditions (10 mM NaHCO 3 , pH 8.3) germination rates were higher in demes native to soils with the highest carbonated level (G1) and lowest in demes locally adapted to salinity on siliceous soil (G3). The saline-alkaline treatment (40 mM NaCl + 10 mM NaHCO 3 ) almost completely inhibited germination of G3 plants, while in G1 and G2 demes germination rates were similar to those observed with NaCl salinity under slightly acidic pH conditions (Fig. 3A). The germination rates of the demes under the alkalinity treatment were positively correlated to the soil carbonate content in their native habitats, but negatively related to the corresponding soil Na + contents. A negative correlation between germination rate and native soil Na + was also observed for the alkaline-saline treatment (Supplementary Dataset S7).
For growth performance under saline alkaline conditions, two levels of stress were assayed: a moderate stress level with 40 mM NaCl + 10 mM NaHCO3 (pH 8.3) and a strong level with 60 mM NaCl + 15 mM NaHCO 3 (pH 8.3). The strong treatment reduced root and leaf length in all three groups (Fig. 3B). Under the moderate saline-alkaline stress plants from G1, and G3 also suffered inhibition of root and leaf elongation, while G2 plants were hardly affected (Fig. 3B, Supplementary Dataset S7). These results con rm the better tolerance of G2 plants to saline-alkaline stress.

Irrigation experiments and molecular analyses
Further experiments using sand cultures irrigated with different treatment solutions con rmed the superior behavior of G2 plants under saline-alkaline conditions and the best performance of G3 plants under saline slightly acidic pH conditions ( Fig. 4A; Supplementary Fig. S4). A. thaliana Col-0 plants were included here as a reference. Tolerance to the saline-alkaline treatment of the Col-0 plants was similar to that of G2; only the relative rosette diameter was more inhibited in Col-0.
Based on a previous GWAS performed with the A. thaliana HapMap cohort cultivated in the same salinealkaline soil used here for the reciprocal transplant experiment (unpublished data), we explored the candidate genes associated with leaf Na + and Fe concentrations. We generated a neighbor-joining (NJ) cladogram for each locus and we selected NINJA (Novel Interaction of JAZ) and CCB4 (Cofactor Assembly of Complex C) because these were the only genes that clustered the majority of G2 plants in the same node (Fig. 4F, H). Moreover, we included in the analysis two genes, SOS1 (Salt Overly Sensitive 1) and FRO2 (Ferric Reduction Oxidase 2), with a well-established function in tolerance to salinity and Fe de ciency, respectively. The cladogram of SOS1 revealed no genomic relationship between groups (Fig. 4B) but FRO clustering grouped all G2 plants and one G1 deme (SCF) in the same node (Fig. 4D), suggesting that FRO2 can have a role in saline-alkaline tolerance.
Quantitative PCR analysis of the selected genes performed 10 days after exposure to the different treatment solutions con rmed that SOS1 was enhanced by salinity stress in all plants (Fig. 4C). However, in T6 plants (G3) this gene was highly expressed in the plants submitted to salinity but not in the ones exposed to saline-alkaline stress. Interestingly, in the rest of plants SOS1 expression was equally high in the plants submitted to the saline-alkaline treatment (Fig. 4C). Regarding FRO2, the alkalinity treatment enhanced the expression in all plants, but the salt-alkaline exposure only increased the expression of FRO2 in the G2 plants (Fig. 4E). A similar pattern was observed for the CCB4 gene. CCB4 was signi cantly enhanced by saline-alkaline stress only in G2 plants (Fig. 4G). Alkalinity also raised the expression of CCB4 in Col-0 and A1 (G1) plants, suggesting that this gene is activated due to the presence of bicarbonate. It is well-known that abiotic stresses activate plant defense hormones like jasmonic acid (JA). NINJA encodes part of a repressor complex that negatively regulates JA signaling (Acosta et al., 2013). We detected that NINJA was enhanced in G2 plants under all the treatments but especially under alkaline salinity (Fig. 4I).

Local adaptation to soil conditions in NE Catalonia
In NE Catalonia A. thaliana demes have a particular distribution. In previous studies we have shown that soil Na + drives divergent selection of coastal demes with higher tolerance to salinity (Busoms et al., 2015). These demes (G3) occur all on siliceous saline soils. Soil carbonate (pH >7) is a further factor for local adaptation of A. thaliana at non-saline inland sites (G1 demes) (Terés et al., 2018) (Fig. 1A). Based on soil characteristics and phenotype data, here we de ned a third group of plants (G3) from habitats with intermediate soil levels of Na + and carbonate (Fig. 1B). However, no A. thaliana demes were found under the harsh conditions of coastal saline-alkaline habitats with high Na + and carbonate concentrations. This is illustrated by the fact that the distribution of A. thaliana demes on the Catalonian coast is restricted to saline-siliceous soils on plutonic rocks and interrupted characteristically on coastal calcareous dolomites, limestones, sandstones and gypsum (Fig. S2A). This distribution indicates that in this region local adaptation to strong saline-alkaline conditions has still not evolved in A. thaliana.
Plants able to adapt to alkaline-saline conditions must e ciently manage multiple stress factors: high Na + , low osmotic potential, low availability of micronutrients, especially Fe and Zn, and imbalance of carbon metabolism due to dark xation of inorganic carbon . PCA analysis using physico-chemical characteristics of the native soils of the A. thaliana demes analyzed here con rm the clear separation of the three plant groups (Fig. 1C). Chloride content, sulfate and Mg, Mo and Na are correlated, and they overlap with the coastal populations, while Ca, Zn and P concentrations are correlated and are characteristic for the alkaline soils. Differences in the relative availability of soil nutrients among the habitats further characterize the three plants groups (Fig. 1D).
Although soil Na + and carbonate concentrations clearly distinguish G1 and G3 demes and we previously have shown that these soil factors drive local adaptation in A. thaliana demes in this region (Busoms et al., 2015;Terés et al., 2018), the whole-genome tree does not reveal a clear strati cation by soil of the 3 groups (Fig. 1E). Contrastingly phenotype differences were evident both in greenhouse common gardens and hydroponics (Figs. 2 and 3).
Phenotype differences under saline-alkaline stress Although A. thaliana is unable to colonize the saline-alkaline soils of the Catalonian coast, plants from all three groups were able to grow in common gardens on such alkaline saline-soil under greenhouse conditions. The lack of competition in the experimental pots and the regular irrigation to eld capacity provided less stressful conditions. A main obstacle for coastal demes from siliceous soils to colonize the nearby saline alkaline soils seems to be the strong inhibition of the germination rate (Fig. 3A). Synergistic effects of salinity and alkalinity on germination have previously been reported in alfalfa cultivars differing in alkaline-saline tolerance (Zhang et al., 2017a); germination and early growth stages are also the most sensitive in wheat (Lin et al., 2012). All A. thaliana demes suffered inhibition of germination on alkaline-saline substrate. G3 plants were most severely inhibited, while G1 and G2 plants had similar germination rates (Fig. 3A).
Subsequent growth of germinated transplanted seedlings to hydroponics revealed better performance of G2 plants under saline-alkaline conditions. This was especially visible for the mild alkaline-saline treatment (40 mM NaCl + 10 µM NaHCO 3 ) which did not cause any reduction in rosette diameter and root length in the G2 plants native to soils with intermediate levels of salinity and alkalinity (Fig. 3B).
Contrastingly, plants native to non-saline alkaline soil (G1) and those native to saline-siliceous soils (G3) suffered from growth inhibition even under the mild alkaline-saline stress treatment. Comparison of relative growth values (Fig. 4A) clearly demonstrates the best performance of G2 under saline-alkaline conditions in contrast to G3 plants, with best growth under saline, low pH conditions.
Higher performance of G2 under saline-alkaline stress was also con rmed by the common garden study ( Fig. 2A-B). Under saline conditions, G3 plants were best maintaining high K + /Na + tissue ratios (Fig. 2F).
However, under saline-alkaline conditions G3 plants had the lowest relative K + concentrations.
Contrastingly, G2 plants not only maintained high K + /Na + ratios but also had higher relative concentrations of Fe and Zn, micronutrients that are often limiting growth of plants under alkaline conditions (Riaz et al., 2020). Maintenance of a high K + /Na + ratio is critical for salinity tolerance (Rubio et al, 2020). Alkaline salinity has an especially severe inhibitory effect on this parameter (Lin et al., 2012).
Our results suggest that alkalinity may hamper the transport mechanisms that allow salt tolerant plants to maintain high K + /Na + ratios under non-alkaline saline conditions.
Transcript expression of selected genes SOS1 is required for salinity tolerance by maintaining Na + and K + homeostasis in plants exposed to salinity (Shi et al., 2000). Extreme increases of the Na + /K + ratio in sos1 rice mutants have been observed, especially in the xylem parenchyma (El Mahi et al., 2019). In our A. thaliana demes, SOS1 expression was enhanced by both saline and saline-alkaline treatments in all demes to similar extent. Exceptions were T6 from G3 and V3 from G2, which had considerably higher SOS1 expression under non-alkaline salinity (Fig. 4C). Contrastingly, in saline tolerant rice higher SOS1 expression has been reported under pH 8 than under pH 7 (Chummnathong et al., 2019). The A. thaliana T6 deme has evolved on saline, slightly acidic soil and the highest expression of SOS1 was obtained in slightly acidic saline solutions simulating conditions of the native habitat. The lower expression under alkaline conditions suggests that alkaline salinity is either less perceived by these plants or that other constraints imposed by the alkaline-saline conditions inhibit the full expression of SOS1. SOS1 gene expression is regulated by a complex signaling network (Ji et al., 2013). Recently it has been shown that AtbZIP62 negatively regulates both SOS1 expression (Rolly et al., 2020) and systemic acquired pathogen defense probably mediated by salicylate; but an interplay with alkaline stress has not been shown so far.
FRO2, coding for ferric-chelate reductase plays an important role in plants iron acquisition both in roots and in leaves. The enzyme is responsible for the reduction of FeIII prior to transport into the root or leaf cells and its activity is substantially enhanced under Fe de ciency (Connolly et al., 2003). Alkalinity enhanced FRO2 expression more in roots of carbonate tolerant than carbonate sensitive soybean varieties (Waters et al., 2018). Contrastingly, FRO2 expression is decreased in chlorotic leaves of sensitive orange trees exposed to carbonate (Martínez-Cuenca et al., 2017). Under the alkaline non-saline conditions of this study, FRO2 expression was enhanced in all demes; G2 plants achieved the highest values (Fig. 4E) Under saline alkaline stress, transcription of NINJA was characteristically enhanced in all demes, except in T6 adapted to non-alkaline salinity (Fig. 4I). This G3 deme showed extremely low expression of NINJA under all treatments. NINJA is a key factor in jasmonate signaling; the protein associates with the transcription factor ERF19 inhibiting its transcriptional role (Huang et al., 2019). Interestingly, jasmonate has been found to inhibit the expression level of FRO2 (Maurer et al., 2011). We selected NINJA for this expression analysis as a previous GWA study revealed an association between NINJA and the variance in leaf Na + concentrations (unpublished results). Results here indicate that a constitutively low expression of NINJA is associated with low Na + leaf concentrations under saline treatments in T6. Non-alkaline salinity increased NINJA expression in salt sensitive A1 of the G1 group and in Col-0, but not in plants of G2 (Fig. 4I). In these plants tolerant to alkaline-salinity, NINJA expression was substantially enhanced only under saline -alkaline conditions. NINJA, by assembly with FRS proteins, may act as a repressor of glucosinolate (GS) biosynthesis (Fernández-Calvo et al., 2020). In alkaline tolerant A. thaliana exposed to bicarbonate we have previously observed upregulation of GS biosynthesis related genes and we proposed that enhanced GS production may be a way to retrieve excessive organic carbon produced by Under saline-alkaline stress the optimization of an e cient acquisition and use of nutrients is essential.
Iron is especially critical (Li et al., 2016). Under these conditions, G2 plants were able to maintain better leaf Fe concentrations (Fig. 2F). Adjustment of number of photosynthetic cells to the available Fe and preferential allocation of Fe into the thylakoid membranes seem essential for avoidance of chlorosis development (Terry and Zayed, 1995). Heme proteins are a major sink for chloroplast Fe (Kroh and Pilon, 2020). Cytochrome b 6 f, one of these heme proteins, is a major target for Fe de ciency (Hantzis et al., 2018). Cytochrome b 6 f plays a key role not only in the lineal electron transfer between PSII and PSI, but also is crucial in cyclic electron transport, in avoidance of ROS damage, and acting as redox sensor (Malone et al., 2021). Under alkaline-saline stress, here we found a considerable upregulation of CCB4 expression speci cally in plants of G2 (Fig. 4G). CCB4 is a key protein in the formation of the functional cytochrome b 6 f as it is required for incorporation of heme into the apoprotein (Schöttler et al., 2015). A key role for CCB4 in alkaline saline tolerance of A. thaliana is further supported by the observation that the neighbor joining cladogram for CCB4 clustered all G2 demes in the same group (Fig. 4F). Moreover, the truncated CCB4 allele found in T6 may explain the lower expression of CCB4 in these plants ( Supplementary Fig. S5).

Plasticity
Local adaptation is a xed genetic variation that provides an advantage to one speci c climatic and/or edaphic condition (Valladares et al. 2014;Josephs 2018). In NE Catalonia we have previously shown that G1 and G3 demes are locally adapted to non-saline alkaline soil and to non-alkaline saline soils, respectively. Here we found superior performance of G2 plants under moderate alkaline-saline conditions in greenhouse and hydroponic experiments. However, these demes native to habitats with intermediate levels of carbonate and NaCl are not locally adapted to the harsh natural conditions of the alkaline-saline soils of the Catalonian coast. The analysis of plasticity index in the three plant groups clearly show that plasticity is a major factor responsible for the better growth of G2 plants on alkaline-saline soil (Fig. 2C).
Contrastingly, G3 plants were maladapted to alkaline-saline conditions, while G1 plants revealed low plasticity. Plasticity is favored under conditions of large gene ow, while local adaptation is enhanced under limited gene ow (Scheiner, 2013). The demes adapted to non-alkaline salinity on the Catalonian coast are still under divergent selection (Busoms et al., 2015). It is likely that genes with unfavorable in uence under alkaline salinity are present in G3 demes adapted to non-alkaline salinity. This view is In conclusion, Arabidopsis thaliana distribution in the north-east of Spain is disrupted in coastal areas with saline alkaline soils. Under less severe controlled conditions, demes native to sites with intermediate levels of Na + and CaCO 3 (G2) were observed to be the most tolerant to alkaline salinity. The high sensitivity of G3 plants to saline-alkaline stress suggests that the evolved molecular mechanism conveying tolerance to salinity on siliceous substrates can be detrimental under alkaline conditions. G2 plants with higher plasticity under saline alkalinity stress are able to maintain both higher germination rates and better ion homeostasis under alkaline-saline conditions. The higher expression of several key genes, such as FRO2, CCB4 and NINJA, seem to play a key role in the better tolerance to alkaline -saline stress in G2. This nding is especially relevant for breeding programs considering salinity tolerance in crops destined to saline soils on limestone parent material.
Declarations Figure 2 Soil reciprocal transplant performance of A. thaliana groups. Mean ± SE of (A) growth (rosette diameter, RD) and (B) tness (number of siliques produced) of A. thaliana plants cultivated in saline (BLA site) and saline-alkaline (ESC site) soil in semi-controlled conditions (UAB greenhouse). Letters indicate signi cant differences (Tukey's HSD, P < 0.05) between A. thaliana demes groups (G1=blue, G2=green, G3=red). (C) Mean growth ± 95% con dence intervals of G1, G2 and G3 plants cultivated in saline and saline-alkaline soil in a greenhouse. Normalized difference of 12 elements in the leaves of plants from the three groups cultivated in (D) saline soil or (E) saline-alkaline soil in a greenhouse. Elements exhibiting signi cant differences (according to a t-test) are marked with an asterisk (*, P<0.05). Figure 3