The genome of Draba nivalis shows signatures of adaptation to the extreme environmental stresses of the Arctic

Abstract

The Arctic is one of the most extreme terrestrial environments on the planet. Here, we present the first chromosome-scale genome assembly of a plant adapted to the high Arctic, Draba nivalis (Brassicaceae), an attractive model species for studying plant adaptation to the stresses imposed by this harsh environment. We used an iterative scaffolding strategy with data from short-reads, single-molecule long reads, proximity ligation data, and a genetic map to produce a 302 Mb assembly that is highly contiguous with 91.6% assembled into eight chromosomes (the base chromosome number). To identify candidate genes and gene families that may have facilitated adaptation to Arctic environmental stresses, we performed comparative genomic analyses with nine non-Arctic Brassicaceae species. We show that the D. nivalis genome contains expanded suites of genes associated with drought and cold stress (e.g., related to the maintenance of oxidation-reduction homeostasis, meiosis, and signaling pathways). The expansions of gene families associated with these functions appear to be driven in part by the activity of transposable elements. Tests of positive selection identify suites of candidate genes associated with meiosis and photoperiodism, as well as cold, drought, and oxidative stress responses. Our results reveal a multifaceted landscape of stress adaptation in the D. nivalis genome, offering avenues for the continued development of this species as an Arctic model plant.

Keywords: Arctic; Brassicaceae; adaptation; chromosome-scale assembly; linkage map.

FIGURE 1

TheD. nivalisgenome assembly. (a) Circos plot of the eight chromosomes (Dniv1–Dniv8) showing the distribution of gene annotations (blue) and LTR‐RT elements (orange) in 500 Kb windows. Ticks represent 2 Mb intervals. The inner tracks show the distribution of TIR (grey) and Helitron (red) elements in 500 Kb windows. (b) Organization of ancestral Brassicaceae genomic blocks in the eightD. nivalischromosomes (Dniv1–8) based on CCP and comparative analyses relative toA. lyrataandA. alpina. Centromere positions of chromosomes 1–3 and 8 are tentative, but supported by results in Figures S2, S3and the structure of other Arabideae species (Mandáková et al., 2020). Genomic blocks are coloured to match eight colours corresponding to eight chromosomes of the Ancestral Crucifer Karyotype (Lysak et al., 2016). (c) Relative abundance of TE superfamilies in selected species (see Table S5; Willing et al., 2015). Whereas LTR‐RT abundance is similarly elevated inD. nivalisandA. alpinarelative toA. lyrata, LINEs appear to be reduced and TIR‐CACTA elements enriched inD. nivalisrelative to both species [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 2

Syntenic relationships betweenD. nivalisandArabis alpina(a) andArabidopsis lyrata(b). TheD. nivalisgenome was aligned to the genomes ofA. alpinaandA. lyratawith NUCmer. Chromosomes are colour‐coded to match the Ancestral Crucifer Karyotype (ACK; Lysak et al., 2016) structurally resembling theA. lyratagenome (b) [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 3

Detailed CCP analysis ofD. nivalischromosomes Dniv4, Dniv5, Dniv6, and Dniv7.Arabidopsis thalianaBAC contigs were in situ hybridized to pachytene chromosome spreads inD.nivalis(insets show the same probes localized on diakinetic chromosomes or mitotic metaphase chromosomes). Arrowheads indicate the position of centromeres. Green, yellow, and red colour corresponds to fluorescence of Alexa 488, Cy3 and Texas Red, respectively. Chromosomes were counterstained with DAPI. Scale bars 10 μm

FIGURE 4

Evolutionary dynamics of abundant LTR‐RT tribes inD. nivalis. Results are based on the percent nucleotide similarity of LTR sequences among full‐length copies in the genome. Selected tribes show peaks indicative of transposition bursts that were above the average of all tribes forCopiaandGypsy(97.64% and 97.30%, respectively), distinguishing retrotransposons such as ATCOPIA95 showing ancestral proliferation from tribes such as ATCOPIA35 with a majority of recent transposition events. Summary statistics of all LTR‐RT tribes inD. nivalisandA. alpinaare provided in Table S6 [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 5

Pfam domains associated with oxidation‐reduction processes are enriched inD. nivalis. Heatmap comparing Pfam domains annotated with the BP GO term “oxidation‐reduction process” (GO:0055114) in selected Brassicaceae species. Cells are coloured by Z‐score. The dendrogram on the left represents groupings based on similar counts of selected Pfam domains. Pfam domains detailed in green are significantly enriched inD. nivalisrelative to the other nine species (Z‐score > 1.96). Pfam domains detailed in red are not significantly enriched, but have > 10 genes annotated inD. nivalis. The genomes of bothD. nivalisandR. raphinistrumcontain relatively abundant and significantly enriched Pfam domains associated with oxidation‐reduction processes, which are important in stress response signalling pathways [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 6

Gene family evolution inD. nivalis. (a) Time‐calibrated species tree of 10 Brassicaceae genomes showing branches labelled with significant gene family expansions and contractions (+gene families gained, ‐gene families lost). To the right we show the percentage of gene pairs derived from different modes of duplication in the 10 species. ForD. nivalis,gene duplication modes are shown for both the whole genome (labelledD. nivalis) and for the 2,645 genes that constitute 198 significantly expanded gene families (EGFs). Patterns of gene duplication are broadly consistent between the whole genome and the EGFs, except that transposed (TRD) duplications are less frequent (p‐value = 1.8e‐15), and proximal (PD,p‐value < 2.2e‐16) and tandem (TD,p‐value = 2.432e‐10) duplications are more frequent in EGFs than would be expected by chance (Fisher's exact test). Values for all species other thanD. nivalisare from Qiao et al. (2019), and since their analysis did not includeR. raphanistrum, the results shown are from the closely relatedRaphanus sativa. DSD, dispersed duplication; WGD, whole‐genome duplication; TRD, transposed duplication; PD, proximal duplication; TD, tandem duplication. (b and c)Select biological process (b) and molecular function (c) GO terms significantly enriched (Fisher's exact testp‐value < 0.05) inD. nivalisEFGs (in b, oxidation‐reduction process was not significant, indicated by “ns”). Terms are grouped into broad categories to simplify interpretation. Pie charts for each term show the modes of gene duplication inferred for genes annotated with these terms (DSD, WGD, TRD, PD, TD; colour scheme follows a, with unclassified in grey) [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 7

Genes under positive selection inD. nivalis. (a) Distribution of the ratios of ln likelihoods (lnL) from tests for positive selection in 15,828D. nivalisgenes. Genes with a higher proportion of nonsynonymous to synonymous substitutions have a higher lnL ratio, and those with an lnL ratio above theX ²critical value (3.84, dashed line,p‐value < 0.05,df=1) are considered significantly likely to contain codons that evolved under positive selection inD. nivalis(PSGs; see Methods). Coloured dots represent genes that are annotated with biological process (BP) GO terms of particular interest for Arctic adaptation. (b) Summary of key BP GO terms in theD. nivalisPSGs. Asterisks (*) indicate significantly enriched terms relative to the genomic background. Parent terms are in bold. (c and d) Venn diagrams showing the overlap of BP (c) and molecular function (d) GO terms resulting from analysis of Pfam domains, expanded gene families (EGF), and PSGs inD. nivalis(see also Table S10) [Colour figure can be viewed at wileyonlinelibrary.com]