Fluctuating selection on migrant adaptive sodium transporter alleles in coastal Arabidopsis thaliana

Significance The natural landscape contains a highly heterogeneous array of environments that drive the adaptive differentiation of populations, including adaptation to elevated salinity. Our research emphasizes an integrated genetic, physiological, and ecological approach to understand the role of naturally evolved high-affinity K+ transporter (HKT1;1) allelic variants in the adaptation of Arabidopsis thaliana populations to fluctuating salinity dynamics in nature. This information not only provides a case study fruitfully taking identification of natural variants through population demographic dynamics to molecular function but also is valuable for improving the sustainability of crop yields as the stress from salinity escalates due to increasing population pressures and global climate change.


This PDF file includes:
Supplementary text S1 to S7 Figs. S1 to S7 Tables S1 to S5 Other supplementary materials for this manuscript include the following: Datasets S1 and S2 Supplementary Information Text Section S1. DNA and RNA extractions for HKT1;1 SNP genotyping and root expression DNA was extracted from frozen leaf tissue using 50 mM TRIS (pH 9) and 5 mM EDTA (pH 8).
A total 45 cycles PCR was performed with 30 sec at 94°C, 15 sec annealing at 60°C followed by 30 sec extension at 72°C. PCR product was then digested with XhoI overnight and separated on 3% agarose gel.
Total RNA was extracted using the Qiagen RNeasy Plant Mini Kit (http://www.qiagen.com), and DNase digestion was performed during the extraction procedure according to the manufacturer's instructions. Two micrograms of total RNA were used as a template to synthesize first-strand cDNA with random hexamers using SuperScript II Reverse Transcriptase (Invitrogen Life Technologies, http://www.invitrogen.com). Quantitative real-time PCR was performed with the first strand cDNA as a template on a sequence detector system (ABI Prism 7000, Applied Biosystems) with Maxima SYBR Green qPCR Master Mixes (Thermo Scientific).

Section S3. Extraction methods for soil ionome analysis
Extraction method for soil samples consisted of a digestion with 20 mL of 1 M NH4HCO3, 0.005 M diaminetriaminepentaacetic acid, and 5 mL of pure water during 1 h of shaking on the rotary shaker at low speed. Each sample was gravity filtered through qualitative filter papers until obtaining approximately 5 mL of filtrate, which was transferred into Pyrex tubes; 0.7 mM trace grade c. HNO3 was added and digested at 115°C for 4.5 h.

Section S4. Modified high molecular weight CTAP protocol for DNA extraction
A. thaliana leaf material (0.4 g) was ground in liquid nitrogen and 10ml of CTAB DNA extraction buffer was added (Tris-HCl 100mM, CTAB 2% (w/v), NaCl 1.4M, EDTA 20mM) and 20l of Proteinase K at 20mg/ml. The mixture was incubated at 55°C for 1 hour, then cooled on ice. 0.5X volume Chloroform (Fisher Scientific) was added and inverted to mix. Samples were spun down at 3000rpm for 30 minutes. The upper phase was taken, and 1X volume of phenol:chloroform:isoamyl alcohol (25:24:1) was added and spun for 30 minutes at 3000rpm. Again, to the upper phase only, 10% volume NaOAc at 3M was added along with 2.5X volume of ice cold 100% ethanol. The tubes were inverted to mix and incubated on ice for 30 minutes. The mixture was spun down for 30 minutes at 3000rpm at 4°C. The pellet was washed in 4ml of ice cold 70% ethanol (Sigma-Aldrich). Tubes were spun for 10 minutes at 3000rpm at 4°C. The 70% ethanol wash was repeated twice more. The pellet was air dried and resuspended in 300ul nucleasefree water with 3ul RNase A at 4mg/ml. DNA concentration was checked on a QuBit Flourometer 2.0 using the QuBit dsDNA HS Assay kit.

Section S5. Data processing for whole-genome resequencing
Genome resequencing of 74 A. thaliana individuals was performed on Illumina HiSeq 2500 in paired. Following demultiplexing and removal of adaptor sequences using 'Cutadapt' standard quality trimming was performed 'Trimmomatic' (settings: LEADING:10 TRAILING:10 SLIDINGWINDOW: 4:15 MINLEN:50) [1,2]. Next, sequence data was processed to: (1) remove duplicate reads using Picard (MarkDuplicates); (2) apply a 'namefix' to the bam files using Picard (AddOrReplaceReadGroups) and (3) realign Indels using the GATK 'GenomeAnalysis' Toolkit [3]. Biallelic SNPs were identified using 'HaplotypeCaller' and genotyped using 'GenotypeGVCF' (both in GATK). Data was quality filtered using GATK SelectVariants using these parameters: QD Section S6. 10X library construction and Supernova genome assembly DNA from samples S1 (T11), S2 (JBB) and S12 (PA10) was diluted to 0.5 ng/l with EB and checked with a QuBit Flourometer 2.0 using the QuBit dsDNA HS Assay kit. The Chromium User Guide was followed as per the manufacturer's instructions (10X Genomics, CG00043, Rev A). The final library was quantified using qPCR (KAPA Library Quant kit, ABI Prism qPCR Mix, Kapa Biosystems). Sizing of library fragments was confirmed using a Bioanalyzer (High Sensitivity kit, Agilent). Samples were pooled based on molarities calculated using the two QC measurements. The S1 and S12 Chromium samples were representative of HKT1;1 LLS allele and S2 was representative of HKT1;1 HLS-1 . Therefore, we necessary to assemble separately a sample from the PA10 population that harboured the HKT1;1 HLS-2 . This was possible because the four resequenced PA10 plants were essentially clonal: we therefore pooled them together for sufficient coverage. We then used a hand-curated reference guided assembly method to create the third HKT1;1 HLS-2 locus.

Section S7. GWAS analysis
GWAS analysis was performed in two steps: first, a principle components analysis (PCA) was performed and a kinship matrix was calculated using the GAPIT package in R [4]. The "Q" matrix was determined by the PCA to account for effects due to population structure, and the kinship matrix (K) was calculated using the VanRaden algorithm [5] and the EMMA method to determine the familial relatedness. Secondly, the compressed mixed linear model (CMLM) was             Table S1. Non-synonymous changes in HKT1;1 coding region between the three versions. Fixation frequency of each SNP substitution among the total individuals of each HKT1;1 version (HKT1;1 LLS : n=62, HKT1;1 HLS-1 : n=9, HKT1;1 HLS-2 : n=6). Table S2. AraGWAS significant hits on At4G10310 (HKT1;1) gene (Na23 project) and SNP bases corresponding to our HKT1;1 alignment. Table S3. Score, number of observations and corresponding allele, estimate, location and SNP bases corresponding to our HKT1;1 alignment of the 12 first significant hits. Table S4. ANOVA (Fisher test) of fitness from inland and coastal A. thaliana demes cultivated at BLA-coastal and SCF-inland common gardens in spring of 2013 and 2014. Results including all plants and reanalysis excluding the plants harbouring the HKT1;1 HLS allele version. Table S5. Quality control parameters of DNA library preparation for the de novo genome assembly of three A. thaliana samples. Table S6. Assembly metrics of three A. thaliana genome assemblies using 10x Genomics Chromium platform and Supernova assembler.

Other supplementary material
Additional data Dataset S1 (separate file) Dataset S1. Sequencing sample information: name, deme of origin, GPS coordinates, location and HKT1;1 allelic variant.