Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2016 May 23;6:26160. doi: 10.1038/srep26160

High light intensity plays a major role in emergence of population level variation in Arabidopsis thaliana along an altitudinal gradient

Antariksh Tyagi 1, Amrita Yadav 1, Abhinandan Mani Tripathi 1,2, Sribash Roy 1,2,a
PMCID: PMC4876511  PMID: 27211014

Abstract

Environmental conditions play an important role in the emergence of genetic variations in natural populations. We identified genome-wide patterns of nucleotide variations in the coding regions of natural Arabidopsis thaliana populations. These populations originated from 700 m to 3400 m a.m.s.l. in the Western Himalaya. Using a pooled RNA-Seq approach, we identified the local and global level population-specific SNPs. The biological functions of the SNP-containing genes were primarily related to the high light intensity prevalent at high-altitude regions. The novel SNPs identified in these genes might have arisen de novo in these populations. In another approach, the FSTs of SNP-containing genes were correlated with the corresponding climatic factors. ‘Radiation in the growing season’ was the only environmental factor found to be strongly correlated with the gene-level FSTs. In both the approaches, the high light intensity was identified as the primary abiotic stress associated with the variations in these populations. The differential gene expression analysis between field and controlled condition grown plants also showed high light intensity as the primary abiotic stress, particularly for the high altitude populations. Our results provide a genome-wide perspective of nucleotide variations in populations along altitudinal gradient and their putative role in emergence of these variations.

The study of genetic variations that impart selective advantages to an organism in its natural habitat is one of the principal goals of ecological genomics. It aims at integrating the genomic changes with the environmental conditions to which an organism is exposed1. This environment-driven selection is one of the most important factors behind population differentiation which functions by imposing variable selection pressure among the populations of climatically different regions. Patterns of local adaptation are expected to emerge when selection is spatially variable and strong enough as compared to other acting evolutionary forces such as genetic drift, mutation and migration2. The natural genetic variations between the populations of a species are important resources for adaptation to their respective natural environment and to cope with the changing climatic conditions3. Over a period of time these variations, in a geographically isolated population, may even lead to speciation4,5.

Identification of the SNP loci under selection allows understanding the process of the emergence of among-population variations. The correlations between FSTs and environmental variables is recently being applied to identify signatures of natural selection3,4. The advantage of this approach is that it leads to a more straightforward evidence of selection in response to the local climate. Additionally, it also yields fewer false positives6 as compared to other approaches such as the outlier approach7.

The genetic variations in a population can be classified as shared and population specific (or private) SNPs8. The population-specific variants are considered to be recent in their origin as compared to that are shared with other populations9 and thus, may have originated de novo after colonization of a location. Although the de novo mutations are known to be largely deleterious, but if advantageous they can sweep through the population4. Recently, NGS of pooled individuals from a population is being applied for the discovery of SNPs and to determine their allele frequency in populations3,10. This method provides a cost effective alternative to the sequencing of a large number of individuals from a population11. Although this method is being primarily applied with whole genome pools, recently, transcriptome pools have also been used for such studies and found to be reliable in case of designs with a large number of individuals in the pool10. Moreover, from an evolutionary point of view coding sequences are the key genomic regions to look for the signatures of selection, as they directly influence the protein function12. However, selection may also act on SNPs in non-coding regions such as promoters, enhancers or small RNAs where they affect gene expression.

In addition to genetic variations, plants respond and adapt to environmental conditions through plastic changes in their phenotype. Differential gene expression of abiotic stress related genes may help the plants to adapt to the external environment13. Modulation of gene expression has a central role in the persistence of an organism under short- and long-term environmental changes14. Conversely, an observation of differential gene expression of abiotic stress related genes can indicate a corresponding difference in environmental conditions of treatment groups.

Arabidopsis thaliana is an established model in the field of plant molecular genetics15. Moreover, the species has also been used for local adaptation studies16,17. A. thaliana populations originating from different altitudes are particularly suitable for studying selection pressures imposed by climatic variables due to the rapid changes in environmental conditions over short geographical distances18. Changes in altitude cause variations in atmospheric temperature, pressure and solar light intensity7,19. These natural abiotic stresses are also known to cause several phenotypic changes in the highland populations18,20.

The A. thaliana populations of Indian West Himalaya inhabit a unique mountainous habitat which ranges from subtropical to temperate climate zones. These populations follow a summer-annual life history20. Previously, these populations were morphologically20 and genetically21 characterized by our group. In the present study, we performed pooled RNA-sequence (RNA-Seq) of four natural populations from four different altitudes (700 m a.m.s.l. to 3400 m a.m.s.l.). Specifically, we asked, what are the SNPs specific to these populations at local and global level? What are the population-specific SNP-containing genes among the populations? What are the SNP-containing genes strongly associated with climatic variables? Are the functions of these genes related to the high-altitude stresses? Additionally, we performed differential gene expression analysis to identify the differences in the transcriptional response of populations between field and controlled conditions. These differences are expected to indicate the abiotic stresses experienced by the populations in the field conditions. Using the two different approaches we attempt to identify the genetic mechanisms accounting for among-population patterns of variations which might have emerged in these populations due to selection under different environmental conditions.

Results

Quality and mapping of RNA-Seq reads

Four A. thaliana populations were collected from different altitudinal regions of the West Himalaya (Fig. 1 and Supplementary Table S1). The number of paired-end reads, generated from the cDNA libraries of these populations varied from 21.5 m (San, FD) to 34.3 m (Deh, FD). The results of read quality and mapping are shown in Supplementary Table S2.

Figure 1. Geographical locations of the sample collection sites.

Figure 1

Open in a new tab

(a) Elevation map of India and adjoining region (b) Elevation map of West Himalaya showing the sample collection sites (Yellow discs). Figures in parentheses indicate altitude in m a.m.s.l. (meters above mean sea level). See Supplementary Table S1 for the details of geographical co-ordinates, habitat and growing season of the four populations. The open access map was obtained from DIVA-GIS (http://www.diva-gis.org/Data).

Calling of SNPs and their patterns in population

The SNP coverage (against reference genome) of the FD and CC grown plants and that of their combined dataset is shown in Supplementary Figure S1. A total of 340258 SNPs were identified in the four populations. 11.21% of these were of synonymous, and 7.38% were non-synonymous amino acid variants. The various patterns of SNPs were consistent in each population (Supplementary Table S3). There was a significant decrease in the SNP density in populations with the increase in altitude (Pearson’s r = 0.82, p < 0.001). The Transition/Transversion ratio (1.37) was consistent among all the four populations.

In all the populations a large proportion (83%) of the amino acid changes were of the non-deleterious type and only 14.7–15.7% were of the deleterious type. The rest of the amino acid changes did not fall in either of the categories (Fig. 2). In the non-deleterious category, the percentage of known changes varied from 83.6% (Mun) to 95% in (Chi). Whereas, in the deleterious category, the known changes varied from 59.8% (Mun) to 83.5% (Chi) (Fig. 2). It was notable that in all the populations, the percentage of the known changes was higher in the non-deleterious category as compared to that of the deleterious category.

Figure 2. Percentage of non-deleterious and deleterious SNPs (including known and novel) in the four populations.

Figure 2

Open in a new tab

NA indicates the SNPs which did not fall in either of the two categories. Count of SNPs; Deh = 40484, Mun = 40742, San = 40860, Chi = 28610.

A large proportion of SNP alleles were found to be fixed (allele frequency = 1) in all the four populations (55.9%, 65.1%, 49.1% and 77.3% in Deh, Mun, San and Chi, respectively) (Supplementary Fig. S2a). The percentage frequency distribution pattern of the allele frequencies was highly consistent among the populations when only the shared changes among the four populations were considered (Supplementary Fig. S2b). The percent fixation in the shared SNPs was higher (80%–86.8%) in all the populations as compared to all SNPs. The percentage frequency distribution of the gene level SNP density was found to be highly consistent among the four populations. The highest density was observed in the 400–500 BP bin followed by a gradual decrease (Supplementary Fig. S3). The average allele frequency of all SNP positions ranged from 0.74 (San) to 0.89 (Chi), and that of local-level population-specific SNPs ranged from 0.49 (Deh) to 0.67 (Mun). The allele frequency of global-level population-specific SNPs was also lower than that of all SNPs in all the four populations (Supplementary Fig. S4).

Functional annotation of local-level population-specific SNP-containing genes

There were 8188, 8237, 8245 and 6271 genes containing non-deleterious but non-synonymous SNPs in Deh, Mun, San and Chi populations, respectively. Amongst these, 5016 genes were common in all the four populations. There was no difference in the enriched GO-terms (biological processes) among the four populations (Supplementary data D1). Deh, Mun, San and Chi had 300, 558, 219 and 212 local-level population-specific SNP-containing genes respectively (Fig. 3). The GO-term enrichment analysis with local-level population-specific SNP-containing genes resulted in a variety of GO-terms (biological process) enriched in the four populations (Fig. 3 and Supplementary Dataset D2). The Mun and Chi populations showed a significant enrichment of genes related to stress responses. In Mun 85 genes (15.2%) were abiotic stress responsive, out of which 50 (58.8%) were light/radiation responsive (Supplementary Table S4). In Chi ‘pigment metabolic process’ (14 genes) (Supplementary Table S5) and ‘glucosinolate metabolic process’ (9 genes) were found to be significantly enriched GO-terms that could be related to the known abiotic stresses prevailing at high altitudes. Deh showed no GO-term enrichment and San showed no stress responsive GO-term. In San ‘catabolic process’ was the only significantly enriched GO-term, which included two genes related to protection against high light intensity stress viz. FAR-RED ELONGATED HYPOCOTYLS 3 (FHY3) and ULTRAVIOLET HYPERSENSITIVE 1 (UVH1). The local-level population-specific SNP-containing genes of the combined dataset of San and Chi did not show any significant enrichment in abiotic stress related categories.

Figure 3. Local-level population-specific SNP-containing genes.

Figure 3

Open in a new tab

Venn diagram showing the count of the local-level population-specific SNP-containing genes in the four populations. The top GO-terms enriched in the four populations are shown. 5016 SNP-containing genes were common among the four populations. See Supplementary data D1 and D2 for the complete GO-term enrichment results and list of genes under the enriched GO-terms.

Functional annotation of global-level population-specific SNP-containing genes

There were 137, 252, 110 and 70 global-level population-specific SNP-containing genes in Deh, Mun, San and Chi respectively. The GO-term enrichment analysis of these genes showed a similar profile of GO-terms as that of local-level population-specific SNP-containing genes (Supplementary Dataset D2). In all the populations, almost half of the genes under the enriched GO-terms were contributed by global-level population-specific SNP-containing genes. Moreover, there was a complete loss in the significance of the enriched GO-terms from the local-level population-specific genes when analyzed without the genes under the enriched GO-terms in global-level population-specific genes. For example, removal of the seven global-level population-specific SNP-containing genes under the ‘pigment metabolic process’ GO-term from the 212 genes of local-level SNP-containing genes resulted in no significantly enriched biological process in Chi.

FST Comparisons between populations

The above-stated SNPs were identified against the reference genome of A. thaliana. However, to correlate the genomic variations with the climatic variables, the SNPs are required to be called between the populations. Therefore, expressed genome-wide SNPs were called against all pairs of populations and their pairwise gene level FSTs were determined.

A total of 29373, 35929 and 36306 SNPs were found in the pairwise comparisons between populations of FD, CC and combined dataset respectively. These SNPs were represented in 13352, 14200 and 9796 genes in the respective datasets. Out of the SNP-containing genes, 7145, 8712 and 8195 were found to have at least one significant (−log10 P-value ≥ 1.301) pairwise FST comparison in the Fischer’s exact test in the respective datasets. We called these genes as ‘highly variable SNP-containing genes’ and were used to correlate with climate data in the partial Mantel test. Out of these highly variable SNP-containing genes in all the tree data sets, ~10% had no known function in the TAIR10 annotations. The remaining genes showed significant enrichment of ‘response to stimulus’ (~32% genes) and ‘response to abiotic stimulus’ (~15% genes) in all the three datasets.

There was a strong correlation between the allele frequencies of the individual FD pools and their CC counterparts with an average FST of; 0.028 ± 0.0003 SE (Deh), 0.039 ± 0.0004 SE (Mun), 0.035 ± 0.0003 SE (San) and 0.029 ± 0.0005 SE (Chi). The percentage frequency distribution of the FST values showed that a majority of FST values lied in the lower range (<0.02) in all the three datasets (Fig. 4). However, in the case of Deh-Chi, Mun-San and Mun-Chi comparisons, a large percentage of FST values were greater than 0.98. These results indicated that Deh and Chi were the most genetically distinct whereas the San and Chi were the closest among the four populations. The phylogenetic analysis of the four populations also showed that San and Chi were the closest, and Deh and Chi were the most distal populations (Fig. 5).

Figure 4.

Figure 4

Open in a new tab

Percentage frequency distribution of pairwise FSTs among the four populations in (a) Field (FD) (b) controlled condition (CC) grown and (c) combined dataset of FD and CC populations. The trend lines for the pairwise comparisons of the six population pairs are shown in different colours. x-axis; pairwise FST, y-axis; % frequency.

Figure 5. Phylogenetic tree based on SNPs.

Figure 5

Open in a new tab

Majority-rule consensus tree of 500 un-rooted neighbour joining trees constructed using the FST matrix of the Field (FD) and Controlled Conditions (CC) grown population pools. The pairwise FST matrix was randomly re-sampled 500 times. Values on the nodes indicate the percentage bootstrap support.

Environmental association of SNP-containing genes

Initially, 26 bio-climatic factors were considered for environmental association analysis with highly differentiated SNP-containing genes. However, only five amongst them were found to be uncorrelated (r < 0.8) viz. temperature seasonality (TS), temperature annual range (TAR), radiation seasonality (RS), mean precipitation in the growing season (MPGS) and mean radiation in the growing season (MRGS). Further, the bio-climatic variable RS was excluded from analysis as its variation was very small among the four locations (0.19 to 0.22). Results of PCoA analysis of these variables indicated that amongst the four populations, Chitkul and Sangla had the smallest whereas, Chitkul and Dehradun had the largest difference in bio-climatic factors (Fig. 6 and Supplementary Fig. S5).

Figure 6. PCoA of Bio-climatic variables.

Figure 6

Open in a new tab

Principle Co-ordinate Analysis (PCoA) of the four populations using five uncorrelated bio-climatic factors; MPGS, mean precipitation in growing season; MRGS, mean radiation in growing season; RS, radiation seasonality; TS, temperature seasonality; TAR, temperature annual range. See Supplementary Fig. S5 for the scree plot of the co-ordinates.

The 95% quantile of correlation (threshold ‘r’) determined by simulation analysis varied for the four bio-climatic variables (Table 1). Overall 380, 700 and 629 genes passed the adopted stringency criteria of having the minimum threshold and a significant P-value in FD, CC and combined datasets. Out of these 54, 120 and 108 genes had a GO-term classification of ‘response to abiotic stress’ in the respective datasets. When the GO-term enrichment analysis was performed for these genes, only the bio-climatic factor ‘mean radiation in growing season’ (MRGS) showed an enrichment in its closely related GO-category i.e. ‘response to radiation’ with 11, 5 and 7 genes in FD, CC and combined datasets respectively (Table 2 and Supplementary data D3, D4, D5). Thus, these genes were identified as having clear association with the local environment.

Table 1. Four un-correlated bio-climatic variables and their 95% effect size cut-off in field (FD), controlled condition (CC) grown and combined populations with their closely related GO-terms.

Bio-climatic variable effect size r(FD, CC, Combined) Closely related GO-terms
Temperature seasonality (TS) 0.787, 0.799, 0.800 response to temperature stimulus(GO:0009266), response to cold(GO:0009409), response to heat(GO:0009408)
Temperature annual range (TAR) 0.803, 0.808, 0.807 response to temperature stimulus(GO:0009266), response to cold(GO:0009409), response to heat(GO:0009408)
Mean precipitation in growing season (MPGS) 0.813, 0.819, 0.823 response to water(GO:0009415), response to water deprivation(GO:0009414)
Mean radiation in growing season (MRGS) 0.808, 0.808, 0.796 response to radiation(GO:0009314), response to light stimulus(GO:0009416)
Open in a new tab

Table 2. List of genes having strong correlation in the three-step procedure of stringency filtration post environmental correlation tests.

Sample AGI code Gene name (Abbreviation) NO. SNPs FST Deh vs. Mun FST Deh vs. San FST Deh vs. Chi FST Mun vs. San FST Mun vs. Chi FST San vs. Chi
FD AT5G35320 – 10 0.43 0.45 0.48 0.14 0.14 0.21
AT5G05690 CONSTITUTIVE PHOTOMORPHOGENIC DWARF (CPD) 8 0.60 0.94 1.00 0.19 0.18 0.26
AT5G05520 – 7 0.49 0.86 0.93 0.16 0.20 0.15
AT1G27650 (ATU2AF35A) 5 0.08 0.98 1.00 0.89 0.91 0.02
AT2G37678 FAR-RED ELONGATED HYPOCOTYL 1 (FHY1) 5 0.49 0.67 0.80 0.10 0.15 0.36
AT2G26930 4-(CYTIDINE 5′-PHOSPHO)-2-C-METHYL-D-ERITHRITOL KINASE (CDPMEK) 4 0.09 0.80 0.83 0.49 0.51 0.01
AT2G46970 PHYTOCHROME INTERACTING FACTOR 3-LIKE 1 (PIL1) 4 0.45 0.80 0.94 0.55 0.69 0.15
AT3G23990 HEAT SHOCK PROTEIN 60 (HSP60) 3 0.44 0.66 0.93 0.04 0.22 0.15
AT4G16780 HOMEOBOX PROTEIN 2 (HB-2) 3 0.27 0.99 1.00 0.42 0.44 0.01
AT5G06460 UBIQUITIN ACTIVATING ENZYME 2 (UBA 2) 3 0.51 0.83 0.97 0.31 0.55 0.18
AT1G78420 – 2 0.58 0.72 0.75 0.09 0.12 0.01
CC AT4G36220 FERULIC ACID 5-HYDROXYLASE 1 (FAH1) 11 0.68 0.66 0.97 0.63 0.97 0.12
AT4G33650 DYNAMIN-RELATED PROTEIN 3A (DRP3A) 8 0.65 0.40 0.94 0.43 0.99 0.26
AT1G78420 – 2 0.35 0.23 0.73 0.28 0.83 0.23
AT4G25570 (ACYB-2) 2 0.03 0.95 1.00 0.84 0.88 0.01
AT5G20520 WAVY GROWTH 2 (WAV2) 1 0.01 0.05 0.03 0.03 0.01 0.01
Combined AT5G43470 RECOGNITION OF PERONOSPORA PARASITICA 8 (RPP8) 15 0.26 0.46 0.39 0.26 0.13 0.10
AT4G16420 HOMOLOG OF YEAST ADA2 2B (ADA2B) 6 0.50 0.70 0.93 0.49 0.77 0.16
AT1G74310 HEAT SHOCK PROTEIN 101 (HSP101) 4 0.35 0.88 0.80 0.50 0.40 0.03
AT5G23120 HIGH CHLOROPHYLL FLUORESCENCE 136 (HCF136) 4 0.29 0.59 0.76 0.43 0.61 0.04
AT2G26150 HEAT SHOCK TRANSCRIPTION FACTOR A2 (HSFA2) 3 0.59 0.99 1.00 0.86 0.87 0.00
AT2G46340 SUPPRESSOR OF PHYA-105 1 (SPA1) 3 0.58 0.83 0.94 0.54 0.69 0.07
AT1G15820 LIGHT HARVESTING COMPLEX PHOTOSYSTEM II SUBUNIT 6 (LHCB6) 2 0.0027 0.0056 0.0051 0.0018 0.0005 0.0009
Open in a new tab

Gene-level pairwise FST of the four population is shown for the strongly correlated genes in field (FD), controlled condition (CC) grown and combined populations samples.

Patterns of global gene expression in Arabidopsis thaliana populations and validation of differentially expressed genes (DEGs)

The above analyses indicated that the SNP-containing genes were primarily related to light intensity. To further ascertain if light intensity was the major abiotic stress in these locations, we performed differential gene expression (DGE) analysis among these populations. The number of down-regulated unigenes ranged from 37 (Chi-CC vs. Chi-FD) to 105 (Deh-CC vs. Deh-FD) and that of up-regulated ranged from 124 (Deh-CC vs. Deh-FD) to 250 (Chi-CC vs. Chi-FD) (Supplementary Table S6). Amongst the DEGs, on an average 96.1% (92.3 to 100%) unigenes had known functions in the TAIR10 database (Supplementary Table S6). The results of DGE between populations and their GO-term enrichments are provided in Supplementary Table S7a–h. The results indicated that the number of genes and the GO terms related to high light intensity were increasing with altitude.

The DGE values of a few randomly selected genes were further validated by qRT-PCR. The log2 fold-change values resulting from qRT-PCR were highly correlated with that of NGS, with overall of r = 0.91 and P-value < 0.0001 (Supplementary Fig. S6). In individual comparisons, the correlation value ranged from 0.85 to 0.98 and was highly significant in all the cases (P-value < 0.0001).

Discussion

We report genome-wide polymorphisms in the coding region of West Himalayan Arabidopsis thaliana populations and their associations with the climatic variables of their natural habitat. The four populations were geographically proximal but colonized diverse habitats with distinct environmental conditions, primarily due to the differences in the altitude of their location. The climatic condition of West Himalayan region range from sub-tropical to temperate22. There are several reports on high altitude plant adaptations5,18,23. But, unlike the previous reports, the populations studied here represent a greater altitudinal gradient and a wider climatic range.

The pooling strategy has widely been used for estimation of SNP alleles in populations using NGS3,24. The strategy has been reported to be as accurate as sequencing individuals11,25. The pooling strategy has also been employed by using RNA samples (RNA-Seq). Using transcriptome sequencing of 10 bank voles, Konczal et al.26 showed that the pooled RNA-Seq exhibits accuracy comparable with that of a pooled genome re-sequencing, provided the variation in expression level is accounted for. Moreover, Westram et al.10 showed that pooled RNA-Seq is reliable for variant allele frequency estimation, given a large number of individuals are included in the pool. RNA-Seq targets only the expressed genes, thereby providing a straightforward understanding of the underlying biological processes12. Our strategy of pooling a large number of samples, as well as employing methods such as exclusion of nucleotide positions not present in any of the populations and sub-sampling to an even number of reads helped to minimize the effect of DGE. Moreover, combining the field and controlled condition grown samples also minimized the effect of DGE in addition to providing an increased read depth. Further, the patterns of SNP distribution was found to be highly consistent in terms of depths and alternative frequency distribution indicating a consistently high-quality sequencing across the samples.

We employed a two-way approach to determine the genes that might provide adaptive fitness to populations in a local habitat. First, the population-specific SNPs were identified in each population against the A. thaliana reference genome and then the genes containing them were functionally categorized. Our hypothesis was that the population-specific SNPs may have arisen de novo in a particular population or may have migrated from a nearby population where they existed as selectively neutral loci in very low and undetectable frequencies4. Such polymorphisms may account for the among-population variations and if advantageous in local climate, can cause selective sweeps in populations27 and can be linked to the fitness of a population28. In the second approach, the gene-level pairwise FSTs of the populations were compared with the corresponding pairwise differences between bio-climatic variables. This approach is based on a presumption that an unusually strong correlation between the allele frequency differentiation and environmental variables indicates a potential identification of a candidate locus29. Such a strong correlation may also account for the population-level variations. However, such correlation may not always imply the causation for adaptation. The integration of genomic and ecological data may not be straightforward as it is difficult to distinguish between demography and selection as a cause of observed genetic variation30.

When all the SNP-containing genes were analyzed, no variation was found in enriched functional categories between the populations. This was due to the large percentage of SNP-containing genes which were shared among the populations (61.2 to 80.0%). A large number of shared SNP-containing genes also indicate a high gene-flow between the populations. However, in the case of local-level population-specific SNP-containing genes, the functional enrichment was found to be markedly different between all the four populations. While the lowest altitude population, Deh did not retrieve any abiotic stress related GO-term, the high altitude populations of Mun and Chi showed enrichment of GO-terms related to high altitude stresses. Importantly, the enriched light-stress responsive genes containing global-level population-specific SNPs appeared to provide a fitness advantage to these populations, as there was observed a complete loss of significance in the GO-term enrichment analyses without these genes. For example, the functional categories enriched in Mun were ‘response to light stimulus’, ‘response to radiation’ and ‘response to red or far red light’. Interestingly, almost half of the observed genes in this functional category contained variants that were global-level population-specific. For example, CULLIN4 (CUL4) is responsible for repression of photomorphogenesis; light harvesting complex of photosystem II 5 (LHCB5) encoding CHLOROPHYLL A/B BINDING PROTEIN CP26 of the antenna system of the photosynthetic apparatus and PHOTOTROPIN 2 (PHOT2) that functions as a blue-light photoreceptor. In Chi, there were seven genes related to pigment metabolism which contained global-level population-specific SNPs. For example, ZETA-CAROTENE ISOMERASE (Z-ISO), a carotene isomerase; CALRETICULIN-3 (CRT-3), which has been linked to anthocyanin accumulation in conjugation with elf18, Chlorophyllase (CLH-1) involved in chlorophyll degradation. Plant pigments such as carotenoids and anthocyanins are known to play an essential role in photoprotection by mechanisms such as scavenging of reactive oxygen species or by absorbing UV radiation31. In addition to the pigment metabolism related genes, ‘glucosinolate metabolic process’ was also found to be overrepresented by SNP-containing genes in Chi. Glucosinolates, a class of secondary metabolites in Brassicaceae, have been shown to have a role in plant stress responses including extreme temperatures and high light intensities32.

Unlike Mun and Chi, despite being a high altitude population, San did not show any enrichment in the population-specific SNPs containing genes related to the high altitude abiotic stresses. However, two highly characterized genes, FHY3 and UVH1 were found to contain population-specific SNPs in San. The FHY3 transcription factor is a key component in the light signaling pathways such as binding of COP1 promoter in response to UV-B33, PHYA signaling pathways34 and in the regulation of chlorophyll biosynthesis35. UVH1 is a DNA repair gene and a homolog of yeast RAD1. The product of RAD1 is a component of repair-endonucleases and confers resistance to UV radiation by excision repair in response to DNA damage36. The non-enrichment of abiotic stress related GO-terms in San could have been due to its geographical proximity to Chi (24 Km downstream on Baspa river). The San population might have got dispersed from Chitkul and due to genetic drift, lost the alleles having the fitness advantage in Chitkul. The milder climatic conditions in Sangla could be a possible explanation for this observation.

Further, the population-specific SNPs had a lower allele frequency as compared to the all-SNPs dataset, suggesting their comparatively recent and de novo appearance in these populations9. Our earlier study has shown that these populations have a long evolutionary history in the West Himalayan region and the time to their most common ancestor (TMRCA) was found to be about 0.45 million years ago21. During this long evolutionary history, the ancestral populations colonized different climatic regions. Subsequently, novel variants might have arisen in response to the selective forces and the beneficial (non-deleterious) among them have swept the populations.

In the second approach, out of the highly correlated SNP-containing genes only 0.75 to 1.37% were found to be strongly correlated with the four uncorrelated bio-climatic factors in the three datasets. Moreover in the combined dataset, the small FST observed between the field and the controlled condition pools in all the populations confirmed that the variation in gene expression did not confound the allele frequency estimations. The highly stringent procedure allowed us to identify the most relevant SNP-containing genes associated with the four non-redundant bio-climatic variables. These climatic variables have previously been reported to mold plant adaptation3,7,16,17. However, out of the four bio-climatic factors only ‘mean radiation in the growing season’ (MRGS) resulted in its closely related GO-term enrichment i.e. ‘response to radiation’, represented by 11, 5 and 7 genes in FD, CC and combined datasets, respectively. The three datasets retrieved the same GO-term for the strongly correlated genes, indicating a role of increasing light intensity in population level variation along the altitudinal gradient. Although the other three bio-climatic variables had strongly correlated genes, none of these retrieved a closely related GO-terms. This was as expected because being an annual with a short life cycle, the climatic conditions during the growing period (three months) has a major effect on the plant. During this period, the solar radiation intensity and its spectrum remains largely unaffected and varies between different altitudes. The effect of increasing light intensity with altitude was also corroborated by the results of DGE analysis. The GO-term enrichment analysis showed that light intensity related genes were primarily expressed in FD samples and, this pattern was more prominent with increasing altitude of populations. Therefore, the enrichment in the correlated GO-category in a radiation-related bio-climatic factor was not surprising. Among the other three selected bio-climatic variables, ‘temperature seasonality’ and ‘temperature annual range’ being annual factors might have a small effect on the development of an annual plant like A. thaliana. The ‘mean radiation in growing season’ was the only highly variable bio-climatic factor among the four sites. Thus, the solar radiation of the growing season might play a major role in the population level variation of these populations.

Most of the strongly correlated genes were well characterized. For example, in FD, CONSTITUTIVE PHOTOMORPHOGENIC DWARF (CPD) is known for their role in UV-B acclimation of plants37, and FAR-RED ELONGATED HYPOCOTYL 1 (FHY1) is regulated by phytochrome A (phyA) and is involved in balancing the light signaling38. In the CC sample, the strongly correlated gene FERULIC ACID 5-HYDROXYLASE 1 (FAH1) is involved in synthesis of sinapate esters, well known to function as plant sunscreens from UV-radiation39. In the combined dataset, the strongly correlated gene SUPPRESSOR OF PHYA-105 1 (SPA1) is a regulator of Phytochrome-A. SPA1 interacts with CONSTITUTIVE PHOTOMORPHOGENESIS (COP1) and regulates photomorphogenesis in plants38,40. Another highly characterized gene, a transcriptional adaptor (ADA2B) is a component of Histone acetyltransferase GCN5 and is involved in plant adaptation to abiotic stresses including high light intensity stress40. HIGH CHLOROPHYLL FLUORESCENCE (HCF136) is a stabilizer of PSII and its mutants have been reported to be devoid of PSII activity41. A heat shock protein (HSP101) and a heat shock transcription factor (HSFA2) were also among the candidate genes. HSPs are generated in response to high light intensity induced production of ROS42 and most of the HSPs are known to respond in high light stresses43. It is notable that a heat shock protein (HSP60) of FD sample was also found to be strongly correlated.

Our study may be limited due to the use of interpolated climate data rather than actual field measurements. Being remotely located these sites are presently not equipped with meteorological stations (except in Dehradun). However, interpolated climatic data has been widely used in studies related to genetic diversity, ecological niche modeling, adaptive divergence etc.3,16,44,45. The two strategies employed here viz. identification of population-specific SNP-containing genes and correlation of FSTs with climatic variables retrieved similar kind of GO-terms (i.e. high light intensity). However, the strongly variable genes identified by the two methods were entirely different. This was expected because an SNP having variable allele frequency can never be population-specific and vice versa. Another limitation in our study was the use of a small number of populations to detect the environmentally variable genes. Generally, a large number of populations are required to interpret the environment guided variations in the populations such as in Mendez-Vigo et al.5. However, a comparable number of populations have also been used for identification of genetic variations such as in Fischer et al.3. Yet, we emphasise on the need to take cautions to interpret these findings in general. However, efforts are being made to sequence more populations from this region.

Our two-way approach led to the identification of genes having a primary biological function related to high light intensity stress. High light intensity is one of the major stresses to which high altitude plants are exposed, as evident from the DGE analysis. Moreover, the environmental factors during the growth period of an annual plant are more important as compared to the annual averages. Thus, the selection of individuals in response to solar radiation stress during growing season of A. thaliana might play a major role in the emergence of genetic variation as compared to the annual factors. Further experiments involving functional validation of the genes may provide a deeper insight into their role in population-level variation, particularly in high altitude. Moreover, a number of SNP-containing genes had unknown functions. The SNPs in these genes may also contribute in the emergence of population-level variations in West Himalayan A. thaliana.

Methods

Collection of samples

Leaf samples were collected from four different populations in the West Himalaya. The populations were named with the first three letters of the name of nearest town or village, viz. Dehradun (Deh, 700 m a.m.s.l.), Munsiyari (Mun, 2000 m a.m.s.l.), Sangla (San, 2600 m a.m.s.l.) and Chitkul (Chi, 3400 m a.m.s.l.) (Fig. 1 and Supplementary Table S1). Collections were performed at a stage defined as, plants with first flower open (principal growth stage 6.00)46 to plants having no more than 10 green siliques. To minimize the effect of microclimate, areas in each site were selected which maximally represented plants of almost uniform growth and microclimate. A total 50 individuals were randomly selected for each population from five such distinct patches. A minimum distance of two meters was maintained between any two plants. One to two rosette and cauline leaves were collected from each plant and stored in the RNAlater™ solution (Ambion, USA). Seeds of matured plants were collected from each population at the senescence/seed shedding stage. These seeds were grown under controlled conditions of 22 °C, 140 μmol/m2/sec light and 16/8 hour day/night cycle. The leaves were collected from the stage as described above for the FD individuals and were stored in RNAlater™ solution for 24 hours before the extraction of total RNA.

Extraction of total RNA and sequencing

Total RNA was extracted from the leaf tissue of ten plants representing one patch from the field (FD) as well as from the controlled conditions (CC) grown plants using mirVana™ Kit (Ambion, USA). The quantity and quality of the RNA was measured by QIAxpertTM (Qiagen, USA) and 2100 BioAnalyzer™ (Agilent Technologies, USA), respectively. Equal amounts of RNA from each patch of a population were pooled together and considered as representative of the respective populations. The pooled RNA was treated with DNase (Ambion TURBO DNA-free DNase kit). cDNA libraries for the eight pools (four each of FD and CC) were prepared using the TruSeqTM RNA Sample Prep Kit v2. (Illumina, USA) with 1 μg of total RNA., Paired-end sequencing was performed using the Illumina HiSeq 1000 platform.

Quality control of raw reads, mapping and calling of SNPs against the reference genome

Raw reads with a minimum of 80% bases with a quality score greater than 30 were selected. The filtered reads were mapped against the A. thaliana reference genome TAIR1047, followed by mapping using Tophat48. The Ambiguously mapped reads were filtered using Picard-tools v1.119.

To call the SNPs, the FASTQ formatted sequence files of FD and CC samples from each population were combined. Mapped bam alignment files were used to call SNPs against the A. thaliana reference genome (TAIR10) using FreeBayes v0.0.349. The parameters used for calling variants were, ploidy = 200; minimum minor allele frequency = 0.05; minimum coverage = 20. Finally, to remove the biasness due to DGE between different datasets, the nucleotide positions that were not present in any of the populations were excluded. The called SNPs were annotated using the SIFT 4G (siftdb.org) tool. SIFT 4G classifies amino acid variants into synonymous and non-synonymous. Within the non-synonymous variants, it predicts the deleterious and non-deleterious types of amino acid changes. Variants that were not yet reported (as on 31st Mar 2015) in the NCBI dbSNP (ncbi.nlm.nih.gov/SNP) database were identified as novel variants.

Identification of local- and global-level population-specific SNP-containing genes and their GO-term enrichment analysis

The deleterious and synonymous variants were excluded from further analysis. The remaining non-deleterious but non-synonymous SNPs were classified into two groups of genes. First, the genes containing SNPs that were present in a particular population but absent in other three local populations (local-level population-specific SNP-containing genes); second, such local-level population-specific SNP-containing genes which contained SNPs that were also not reported in the dbSNP (global-level population-specific SNP-containing genes). Our earlier study indicated that the San and Chi populations are phylogenetically very close21. Therefore, considering these two as a single population, a dataset of population-specific SNP-containing genes was also included in the analysis. To determine the patterns of SNPs among populations several statistical measures viz. the frequency distribution of alternative frequency, the frequency distribution of read depths and average AF of SNPs under local and global population-specific SNPs were estimated.

GO-term enrichment of all the non-synonymous SNP-containing genes and the population-specific SNP-containing genes was performed using AgriGO50. The significance for each GO-identifier was computed with the Fisher’s exact test and P-value ≤ 0.05. To adjust for multiple comparisons, significant enrichment was accepted if the corresponding false discovery rate (FDR) was ≤0.05 in Yekutieli’s multi-test adjustment method.

Analysis of the climate data

Bio-climatic variables of the four locations were derived from the interpolated climatic data. The data was extracted from the WorldClim51 database at the highest available resolution of 30″ (second) and collected over a 50 year period (Year, 1950 to 2000). The solar radiation intensity is an important climatic variable along altitudes, but its data is currently not available from the WorldClim database. Therefore, the data for four radiation-related bio-climatic factors was extracted from the CliMond52 database at the highest available resolution of 10′ (minute) and collected over a 30 years period (Year, 1961 to 1991). It is important to assess the effect of climatic factors prevailing only during the growing season in these annual populations. Therefore, we derived three additional bio-climatic factors viz. mean temperature (MT), mean precipitation (MP) and mean radiation (MR) of the growing season (GS) of three months in the respective locations. Thus, overall 26 bio-climatic variables were considered for the association analysis (Supplementary Table S8).

Pairwise correlation analysis (Pearson’s r) was conducted among these 26 bio-climatic variables to exclude the highly correlated factors (r ≥ 0.8). To illustrate climatic variations between the sampling locations, Principal Co-ordinate analysis (PCoA) was performed.

Estimation of the gene-level allele frequency differentiation between populations (FST)

Pairwise FST analysis of the FD and CC samples was performed using the PoPoolation2 program53. The SNPs were called using Samtools followed by the filtering of indels. To minimize the confounding effect of DGE on allele frequency estimation, the allele count data was sub-sampled to an even target coverage of 10 per variant position while excluding upper 2% most highly expressed reads following Westram et al.10. Random sub-sampling cycles were run in 50 replicates using sub-sampling with replacement strategy. Subsequently, the allele frequency at each position was averaged to get an unambiguous synchronized SNP dataset. FSTs at gene level were calculated using the sliding window strategy with a pool size of 100. Fisher’s exact test was performed to test the significance of the FSTs. To identify the patterns of FSTs between populations, frequency distribution of gene level FSTs was determined in bins of 0.01.

The gene-level FST was also estimated for the combined data of FD and CC samples at an even target coverage of 20 per variant position and pool size of 200, following the above protocol. Further, to determine the effect of DGE on the SNP allele frequency, the position level FSTs were determined between individual pools.

Estimation of overall population differentiation (FST)

The population level FST between all the pairs of populations in the combined data and individual samples (FD and CC) was calculated using Genepop v454. A majority-ruled (MJ) consensus tree was constructed by random re-sampling the population level FST matrix 500 times using PHYLIP55.

Correlation between pairwise FST and Environmental factors

Genes with at least one significant FST comparison we called as ‘highly variable SNP-containing genes’. There were 7145, 8712 and 8195 highly variable SNP-containing genes in FD, CC and combined dataset respectively. These genes were used to correlate with the five selected environmental factors using the partial Mantel’s test (PMT). The pairwise FST matrix (Dx) of a gene was correlated with the distance matrix of four uncorrelated environmental factors (Dy) individually while controlling for the overall FST (Ds, population structure) between the populations by following Fischer et al.3. The PMT runs of these genes were performed using the parametric Pearson’s r correlation in the vegan R package by permuting Dx 1000 times.

Identification of genes strongly correlated with environmental variables

To identify genes with a strong correlation with environmental variables, a three-step stringency cut-off procedure was employed. First, the effect-size ‘r’ was determined as a threshold criterion for the correlation values between FST and bio-climatic variables. FST values between 0 and 1 were randomly generated to create 7145, 8712 and 8195 simulated matrices corresponding to the same number of highly correlated genes in FD, CC and combined dataset respectively. For this, we first performed PMTs between the simulated pairwise FST values (Dx) (random values) and the five environmental factors (Dy) (real values), while controlling for the population structure using genome-wide pairwise FST (Ds) (real values). The 95% quantile of each bio-climatic factor was used as the threshold for r. Above this threshold value, a specific SNP locus was considered to be associated with the respective environmental factor. Second, genes having the r-values larger than the threshold but insignificant (P-value > 0.05) were discarded. Third, only those genes which fell under ‘response to abiotic stimulus’ GO-term were selected. This allowed us to exclude genes associated with the housekeeping cellular processes, biotic factors and the genes of unknown functions. Finally, GO-term enrichment of genes under each bio-climatic variable was performed. The genes under an enriched GO-term were considered as strongly correlated if the GO-term was closely related to its corresponding bio-climatic factor (Table 1).

Analysis of DEGs in four populations and their qRT-PCR validation

DGE analysis of four different populations was performed to determine DEGs of FD populations as compared to their respective CC populations (all FD vs. CC). In all the four comparisons the CC population was taken as the denominator for the determination of fold-change in gene expression.

The read counts per unigene were estimated using the HTseq-count program56. HTseq-count pre-processes the RNA-Seq data for DGE analysis by counting the number of reads mapped to a gene. The generated read counts were used for the identification of DEGs using the R bioconductor package DESeq57. DESeq infers differential signal between a pair of RNA-Seq samples by using a model based on the negative binomial distribution. Finally, DESeq expresses the results in the form of log2 fold-change with their corresponding P-values and false discovery rates (FDR/ Q-value). Unigenes having a minimum of two-fold log2 fold-change value and a Q-value ≤ 0.05 in the pairwise comparisons of the samples were selected for further analysis.

The GO-term enrichment analysis of the significantly expressed unigenes was performed using the Singular Enrichment Analysis (SEA) implemented in the AgriGO online tool. Genes were annotated with the locus identifier information from TAIR10. Significance for each GO-identifier was computed with the Fisher’s exact test. To adjust for multiple comparisons, significant enrichment was accepted if the corresponding Q-value ≤ 0.05 in Yekutieli’s multi-test adjustment method. The lists of GO-terms generated from the results of high-throughput experiments are large and often highly redundant and thus, are difficult to interpret. Therefore, the REVIGO web server58 was used to remove this redundancy and summarize the significantly enriched GO-terms in a graphical form which was easier to understand and interpret.

qRT-PCR was performed to validate the gene expression results generated from NGS platform. Ten to fifteen differentially expressed genes (down- and up-regulated) from each sample were randomly selected (Supplementary Table S9). The primers for qRT-PCR were designed using the NCBI primer blast tool59 with primer melting temperatures (Tm) of 57 °C to 63 °C, product size of 90 to 120 bp and a maximum Tm difference of 3 °C between the forward and reverse primers. The total RNA from three biologically independent population pools (50 plants each) of FD and CC samples was extracted as described above. Additionally, three technical replicates of each biological replicate were also used. The integrity of RNA was checked by agarose gel electrophoresis. High quality RNA was selected and quantified by using the QIAxpert™ microfluidic UV/VIS spectrophotometer (Qiagen, USA). The RNA was reverse transcribed to cDNA using the SuperScript™ III First-Strand Synthesis Kit (Invitrogen, USA) according to the manufacturer’s protocol. The PCR master mixture was setup by mixing 5 μl DyNAmo Flash SYBR Green (Thermo) (2X), 1 μl cDNA, 1 μl forward primer (5 pm/μl), 1 μl reverse primer (5 pm/μl), and milli-Q water up to 10 μl. qRT-PCR was performed with cycling conditions: initial denaturation at 95 °C for 10 min, followed by denaturation at 95 °C for 20 sec, 40 cycles of annealing and extension together at 60 °C for 60 sec. The amplification reaction was performed using the ABI 7500 real-time PCR system (Applied Biosystems, USA). The threshold cycle (Ct) value for each gene was quantified, normalized by Ct value of internal control gene, GLYCERALDEHYDE-3-PHOSPHATE DEHYDROGENASE (GADPH). The relative expression was calculated using the 2−ΔΔCt method60. The relative expression values were then converted to log2 fold-change. The differential expression values from qRT-PCR were correlated with that of NGS using the Pearson’s correlation test.

Additional Information

Accession codes: The RNA-Seq FASTQ files were submitted to NCBI sequence record archive (SRA) with the accession number: SRA347035.

How to cite this article: Tyagi, A. et al. High light intensity plays a major role in emergence of population level variation in Arabidopsis thaliana along an altitudinal gradient. Sci. Rep. 6, 26160; doi: 10.1038/srep26160 (2016).

Supplementary Material

Supplementary Information
srep26160-s1.pdf (315.5KB, pdf)
Supplementary Data D1
srep26160-s2.xls (6.9MB, xls)
Supplementary Data D2
srep26160-s3.xls (39.5KB, xls)
Supplementary Data D3
srep26160-s4.xls (43KB, xls)
Supplementary Data D4
srep26160-s5.xls (102.5KB, xls)
Supplementary Data D5
srep26160-s6.xls (110KB, xls)

Acknowledgments

Authors are thankful to the Director, CSIR-National Botanical Research Institute for providing facilities for the study, Council of Scientific and Industrial Research (CSIR), New Delhi for fellowships provided to A.T. and University Grants Commission, New Delhi, to A.M.T. The authors are also thankful to the Centre for cellular and Molecular Platforms (C-CAMP), India, for carrying out the RNA-Seq job. The authors are thankful to CSIR for funding this study under the project: SIMPLE (BSC-109). The authors are thankful to Sumit Bag for his assistance in the simulation analysis.

Footnotes

Author Contributions S.R. conceived the experiments; A.T., A.Y. and A.M.T. conducted the experiments; S.R., A.T. and A.Y. analysed the results; S.R. and A.T. wrote the manuscript.

References

  1. Ungerer M. C., Johnson L. C. & Herman M. A. Ecological genomics: understanding gene and genome function in the natural environment. Heredity 100, 178–183, doi: 10.1038/sj.hdy.6800992 (2008). [DOI] [PubMed] [Google Scholar]
  2. Blanquart F., Kaltz O., Nuismer S. L. & Gandon S. A practical guide to measuring local adaptation. Ecology letters 16, 1195–1205, doi: 10.1111/ele.12150 (2013). [DOI] [PubMed] [Google Scholar]
  3. Fischer M. C. et al. Population genomic footprints of selection and associations with climate in natural populations of Arabidopsis halleri from the Alps. Molecular ecology 22, 5594–5607, doi: 10.1111/mec.12521 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Savolainen O., Lascoux M. & Merila J. Ecological genomics of local adaptation. Nature reviews. Genetics 14, 807–820, doi: 10.1038/nrg3522 (2013). [DOI] [PubMed] [Google Scholar]
  5. Mendez-Vigo B., Pico F. X., Ramiro M., Martinez-Zapater J. M. & Alonso-Blanco C. Altitudinal and climatic adaptation is mediated by flowering traits and FRI, FLC, and PHYC genes in Arabidopsis. Plant physiology 157, 1942–1955, doi: 10.1104/pp.111.183426 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. De Mita S. et al. Detecting selection along environmental gradients: analysis of eight methods and their effectiveness for outbreeding and selfing populations. Molecular ecology 22, 1383–1399, doi: 10.1111/mec.12182 (2013). [DOI] [PubMed] [Google Scholar]
  7. Kubota S. et al. A Genome Scan for Genes Underlying Microgeographic-Scale Local Adaptation in a Wild Arabidopsis Species. PLoS genetics 11, e1005361, doi: 10.1371/journal.pgen.1005361 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Choudhury A. et al. Population-specific common SNPs reflect demographic histories and highlight regions of genomic plasticity with functional relevance. BMC genomics 15, 437, doi: 10.1186/1471-2164-15-437 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Stephens J. C. et al. Haplotype variation and linkage disequilibrium in 313 human genes. Science 293, 489–493, doi: 10.1126/science.1059431 (2001). [DOI] [PubMed] [Google Scholar]
  10. Westram A. M. et al. Do the same genes underlie parallel phenotypic divergence in different Littorina saxatilis populations? Molecular ecology 23, 4603–4616, doi: 10.1111/mec.12883 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Schlotterer C., Tobler R., Kofler R. & Nolte V. Sequencing pools of individuals - mining genome-wide polymorphism data without big funding. Nature reviews. Genetics 15, 749–763, doi: 10.1038/nrg3803 (2014). [DOI] [PubMed] [Google Scholar]
  12. Piskol R., Ramaswami G. & Li J. B. Reliable identification of genomic variants from RNA-seq data. Am. J. Hum. Genet. 93, 641–651, doi: 10.1016/j.ajhg.2013.08.008 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Whitman D. & Agrawal A. A. in Phenotypic plasticity of insects: mechanisms and consequences (eds Whitman Douglas & Ananthakrishnan T. N.) (Science Publishers, 2009). [Google Scholar]
  14. Lopez-Maury L., Marguerat S. & Bahler J. Tuning gene expression to changing environments: from rapid responses to evolutionary adaptation. Nature reviews. Genetics 9, 583–593, doi: 10.1038/nrg2398 (2008). [DOI] [PubMed] [Google Scholar]
  15. Arabidopsis Genome I. Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 408, 796–815, doi: 10.1038/35048692 (2000). [DOI] [PubMed] [Google Scholar]
  16. Hancock A. M. et al. Adaptation to climate across the Arabidopsis thaliana genome. Science 334, 83–86, doi: 10.1126/science.1209244 (2011). [DOI] [PubMed] [Google Scholar]
  17. Fournier-Level A. et al. A map of local adaptation in Arabidopsis thaliana. Science 334, 86–89, doi: 10.1126/science.1209271 (2011). [DOI] [PubMed] [Google Scholar]
  18. Montesinos-Navarro A., Wig J., Pico F. X. & Tonsor S. J. Arabidopsis thaliana populations show clinal variation in a climatic gradient associated with altitude. The New phytologist 189, 282–294, doi: 10.1111/j.1469-8137.2010.03479.x (2011). [DOI] [PubMed] [Google Scholar]
  19. Körner C. The use of ‘altitude’ in ecological research. Trends Ecol. Evol. 22, 569–574, doi: 10.1016/j.tree.2007.09.006 (2007). [DOI] [PubMed] [Google Scholar]
  20. Singh A. et al. Morphological trait variations in the west Himalayan (India) populations of Arabidopsis thaliana along altitudinal gradients. Curr. Sci. 108, 2213 (2015). [Google Scholar]
  21. Tyagi A. et al. Genetic diversity and population structure of Arabidopsis thaliana along an altitudinal gradient. AoB PLANTS 8, doi: 10.1093/aobpla/plv145 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kottek M., Grieser J., Beck C., Rudolf B. & Rubel F. World Map of the Koppen-Geiger climate classification updated. Meteorol. Z. 15, 259 (2006). [Google Scholar]
  23. Luo Y., Widmer A. & Karrenberg S. The roles of genetic drift and natural selection in quantitative trait divergence along an altitudinal gradient in Arabidopsis thaliana. Heredity 114, 220–228, doi: 10.1038/hdy.2014.89 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kofler R., Betancourt A. J. & Schlotterer C. Sequencing of pooled DNA samples (Pool-Seq) uncovers complex dynamics of transposable element insertions in Drosophila melanogaster. PLoS genetics 8, e1002487, doi: 10.1371/journal.pgen.1002487 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Rellstab C., Zoller S., Tedder A., Gugerli F. & Fischer M. C. Validation of SNP allele frequencies determined by pooled next-generation sequencing in natural populations of a non-model plant species. PloS one 8, e80422, doi: 10.1371/journal.pone.0080422 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Konczal M., Koteja P., Stuglik M. T., Radwan J. & Babik W. Accuracy of allele frequency estimation using pooled RNA-Seq. Molecular ecology resources 14, 381–392, doi: 10.1111/1755-0998.12186 (2014). [DOI] [PubMed] [Google Scholar]
  27. Smith J. M. & Haigh J. The hitch-hiking effect of a favourable gene. Genet. Res. 23, 23–35 (1974). [PubMed] [Google Scholar]
  28. Sjostrand A. E., Sjodin P. & Jakobsson M. Private haplotypes can reveal local adaptation. BMC genetics 15, 61, doi: 10.1186/1471-2156-15-61 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Coop G., Witonsky D., Di Rienzo A. & Pritchard J. K. Using environmental correlations to identify loci underlying local adaptation. Genetics 185, 1411–1423, doi: 10.1534/genetics.110.114819 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Stapley J. et al. Adaptation genomics: the next generation. Trends Ecol Evol 25, 705–712, doi: 10.1016/j.tree.2010.09.002 (2010). [DOI] [PubMed] [Google Scholar]
  31. Taiz L. & Zeiger E. Plant physiology, 5th edition. 135–137 (Sinauer Associates). [Google Scholar]
  32. Del Carmen Martinez-Ballesta M., Moreno D. A. & Carvajal M. The physiological importance of glucosinolates on plant response to abiotic stress in brassica. Int J Mol Sci 14, 11607–11625, doi: 10.3390/ijms140611607 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Huang X. et al. Arabidopsis FHY3 and HY5 positively mediate induction of COP1 transcription in response to photomorphogenic UV-B light. The Plant cell 24, 4590–4606, doi: 10.1105/tpc.112.103994 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Ouyang X. et al. Genome-wide binding site analysis of FAR-RED ELONGATED HYPOCOTYL3 reveals its novel function in Arabidopsis development. The Plant cell 23, 2514–2535, doi: 10.1105/tpc.111.085126 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Wang W. et al. A pair of light signaling factors FHY3 and FAR1 regulates plant immunity by modulating chlorophyll biosynthesis. J Integr Plant Biol, doi: 10.1111/jipb.12369 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Liu Z., Hossain G. S., Islas-Osuna M. A., Mitchell D. L. & Mount D. W. Repair of UV damage in plants by nucleotide excision repair: Arabidopsis UVH1 DNA repair gene is a homolog of Saccharomyces cerevisiae Rad1. Plant J. 21, 519–528 (2000). [DOI] [PubMed] [Google Scholar]
  37. Sheerin D. J. et al. Light-activated phytochrome A and B interact with members of the SPA family to promote photomorphogenesis in Arabidopsis by reorganizing the COP1/SPA complex. The Plant cell 27, 189–201, doi: 10.1105/tpc.114.134775 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Zuo Z., Liu H., Liu B., Liu X. & Lin C. Blue light-dependent interaction of CRY2 with SPA1 regulates COP1 activity and floral initiation in Arabidopsis. Curr. Biol. 21, 841–847, doi: 10.1016/j.cub.2011.03.048 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Landry L. G., Chapple C. C. & Last R. L. Arabidopsis mutants lacking phenolic sunscreens exhibit enhanced ultraviolet-B injury and oxidative damage. Plant Physiol 109, 1159–1166 (1995). [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Kaldis A., Tsementzi D., Tanriverdi O. & Vlachonasios K. E. Arabidopsis thaliana transcriptional co-activators ADA2b and SGF29a are implicated in salt stress responses. Planta 233, 749–762, doi: 10.1007/s00425-010-1337-0 (2011). [DOI] [PubMed] [Google Scholar]
  41. Plücken H., Müller B., Grohmann D., Westhoff P. & Eichacker L. A. The HCF136 protein is essential for assembly of the photosystem II reaction center in Arabidopsis thaliana. FEBS Lett. 532, 1–2 (2002). [DOI] [PubMed] [Google Scholar]
  42. Scarpeci T. E., Zanor M. I. & Valle E. M. Investigating the role of plant heat shock proteins during oxidative stress. Plant Signal Behav. 3, 856–857 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Bechtold U. et al. Impact of chloroplastic- and extracellular-sourced ROS on high light-responsive gene expression in Arabidopsis. J. Exp. Bot. 59, 121–133, doi: 10.1093/jxb/erm289 (2008). [DOI] [PubMed] [Google Scholar]
  44. Russell J. et al. Genetic diversity and ecological niche modelling of wild barley: refugia, large-scale post-LGM range expansion and limited mid-future climate threats? PloS one 9, e86021, doi: 10.1371/journal.pone.0086021 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Wilczek A. M., Cooper M. D., Korves T. M. & Schmitt J. Lagging adaptation to warming climate in Arabidopsis thaliana. Proc. Natl. Acad. Sci. USA 111, 7906–7913, doi: 10.1073/pnas.1406314111 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Boyes D. C. et al. Growth stage-based phenotypic analysis of Arabidopsis: a model for high throughput functional genomics in plants. The Plant cell 13, 1499–1510 (2001). [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Lamesch P. et al. The Arabidopsis Information Resource (TAIR): improved gene annotation and new tools. Nucleic Acids Res. 40, D1202–1210, doi: 10.1093/nar/gkr1090 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Trapnell C., Pachter L. & Salzberg S. L. TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 25, 1105–1111, doi: 10.1093/bioinformatics/btp120 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Garrison E. & Marth G. Haplotype-based variant detection from short-read sequencing. arXiv preprint arXiv:1207.3907 (2012).
  50. Du Z., Zhou X., Ling Y., Zhang Z. & Su Z. agriGO: a GO analysis toolkit for the agricultural community. Nucleic Acids Res. 38, W64–70, doi: 10.1093/nar/gkq310 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Hijmans R. J., Cameron S. E., Parra J. L., Jones P. G. & Jarvis A. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25, 1965–1978 (2005). [Google Scholar]
  52. Kriticos D. J. et al. CliMond: global high‐resolution historical and future scenario climate surfaces for bioclimatic modelling. Methods Ecol. Evol. 3, 53–64 (2012). [Google Scholar]
  53. Kofler R., Pandey R. V. & Schlotterer C. PoPoolation2: identifying differentiation between populations using sequencing of pooled DNA samples (Pool-Seq). Bioinformatics 27, 3435–3436, doi: 10.1093/bioinformatics/btr589 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Rousset F. genepop’007: a complete re-implementation of the genepop software for Windows and Linux. Mol Ecol Resour 8, 103–106, doi: 10.1111/j.1471-8286.2007.01931.x (2008). [DOI] [PubMed] [Google Scholar]
  55. Felsenstein J. PHYLIP-phylogeny inference package (version 3.2). cladistics 5, 163–166 (1989). [Google Scholar]
  56. Anders S., Pyl P. T. & Huber W. HTSeq–a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169, doi: 10.1093/bioinformatics/btu638 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Anders S. & Huber W. Differential expression analysis for sequence count data. Genome biology 11, R106, doi: 10.1186/gb-2010-11-10-r106 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Supek F., Bosnjak M., Skunca N. & Smuc T. REVIGO summarizes and visualizes long lists of gene ontology terms. PloS one 6, e21800, doi: 10.1371/journal.pone.0021800 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Ye J. et al. Primer-BLAST: a tool to design target-specific primers for polymerase chain reaction. BMC bioinformatics 13, 134, doi: 10.1186/1471-2105-13-134 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Livak K. J. & Schmittgen T. D. Analysis of relative gene expression data using real-time quantitative PCR and the 2−DD CT Method. Methods 25, 402–408, doi: 10.1006/meth.2001.1262 (2001). [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Information
srep26160-s1.pdf (315.5KB, pdf)
Supplementary Data D1
srep26160-s2.xls (6.9MB, xls)
Supplementary Data D2
srep26160-s3.xls (39.5KB, xls)
Supplementary Data D3
srep26160-s4.xls (43KB, xls)
Supplementary Data D4
srep26160-s5.xls (102.5KB, xls)
Supplementary Data D5
srep26160-s6.xls (110KB, xls)

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES