Recurrent evolution of heat-responsiveness in Brassicaceae COPIA elements

Abstract

Background: The mobilization of transposable elements (TEs) is suppressed by host genome defense mechanisms. Recent studies showed that the cis-regulatory region of Arabidopsis thaliana COPIA78/ONSEN retrotransposons contains heat-responsive elements (HREs), which cause their activation during heat stress. However, it remains unknown whether this is a common and potentially conserved trait and how it has evolved.

Results: We show that ONSEN, COPIA37, TERESTRA, and ROMANIAT5 are the major families of heat-responsive TEs in A. lyrata and A. thaliana. Heat-responsiveness of COPIA families is correlated with the presence of putative high affinity heat shock factor binding HREs within their long terminal repeats in seven Brassicaceae species. The strong HRE of ONSEN is conserved over millions of years and has evolved by duplication of a proto-HRE sequence, which was already present early in the evolution of the Brassicaceae. However, HREs of most families are species-specific, and in Boechera stricta, the ONSEN HRE accumulated mutations and lost heat-responsiveness.

Conclusions: Gain of HREs does not always provide an ultimate selective advantage for TEs, but may increase the probability of their long-term survival during the co-evolution of hosts and genomic parasites.

Keywords: Brassicaceae; COPIA; Evolution; Heat stress; ONSEN.

Fig. 1

Transcriptome analysis of heat-stressed A. lyrata and A. thaliana plants. a Effects of HS on ONSEN heat-responsiveness in A. thaliana and A. lyrata. Both species were stressed at 37 °C for the indicated number of hours (h) and subsequently analyzed for the amount of ONSEN transcript (log10) by RT-qPCR relative to GAPD-H transcript amounts. * significant (t-test, P <0.05) transcript enrichment relative to 0 h control. Error bars indicate standard deviation of three biological replicates. b Design of plant HS treatment for RNA-seq and representative phenotypes of control, 6 h heat-stressed at 37 °C and recovered plants. c, d Number of significantly (c) upregulated or (d) downregulated protein-coding genes after 6 h at 37 °C and 48 h recovery at non-stress conditions in both species. e, f Number of significantly upregulated (e) TEs and (f) TE families after 6 h HS and 48 h recovery. g Identification of TE groups enriched for heat-responsive copies. Retrotransposons were divided into SINE, SADHU, LINE, COPIA, and GYPSY family members. The relative enrichment of heat-activated TEs was calculated as ratio between % of all heat-activated to % of all TEs genome-wide and expressed on a log2 scale. The major heat-responsive COPIA families in (h) A. lyrata and (i) A. thaliana. The families containing a single HRE are displayed as “single copies.” j RT amino acid sequences (Additional file 12)-based phylogenetic network of selected heat-responsive (colored) and non-responsive (black) A. lyrata and A. thaliana COPIA families. The data are also provided as un-rooted three in Additional file 5: Figure S2

Fig. 2

Evolution of ONSEN heat-responsiveness. a Schematic representation of in silico reconstruction of putative HREs in ONSEN 5’ LTR in different Brassicaceae species. HRE reconstruction follows criteria proposed by [20]. Colored boxes spanning the entire height of the gray field indicate HREs found in ≥50 % of the heat-responsive copies in A. thaliana and A. lyrata or all copies in other species. The lower boxes represent less frequent (<50 %) variants. Detailed information including sequences underlying individual HREs can be found in Additional file 5: Figure S2. b Transcript levels of ONSEN elements in Brassicaceae after 6 h and 12 h at 37 °C. Quantitative PCR values were obtained using species-specific primer pairs and normalized to UBC28. Error bars indicate standard deviation of three biological replicates and * P <0.05 in Student’s t-test. c Sequence conservation over the ONSEN 5’ LTR. Species-specific 5’ LTR consensus sequences were compared to A. lyrata query using 20 bp sliding window and 7 bp minimum consensus length. The y-axis for each species shows 50–100 % sequence conservation. Regions with ≥70 % similarity (pink-filled) were considered as conserved. Red and yellow background colors indicate the A. lyrata 4P and gap HRE regions. d Reconstruction of ONSEN HRE evolution. The phylogenetic tree was developed using the CHALCONE SYNTHASE gene of each individual species. The numbers at branches indicate bootstrap values. Blue lines show species with proto-HREs, red shows those carrying 4P HREs, black shows loss of HREs in B. stricta, and gray shows the COPIA78 family in C. rubella

Fig. 3

COPIA37 and TERESTRA are novel heat-responsive COPIA families. a In silico reconstruction of putative HREs in the 5’ LTR of COPIA37 in different species. HRE classification follows criteria proposed by [20]. Colored boxes spanning the entire height of the gray field indicate HREs found in ≥50 % of the heat-responsive copies in A. thaliana and A. lyrata or all copies in other species. The lower boxes represent less frequent (<50 %) HREs. Detailed information including sequences underlying individual HREs can be found in Additional file 5: Figure S5. b Transcript levels of COPIA37 in Brassicaceae after 6 and 12 h 37 °C HS. The values were normalized to transcript levels of UBC28. Error bars indicate standard deviation between three biological replicates and * P <0.05 in Student’s t-test. c Schematic representation of A. lyrata TERESTRA (TERESTRA). LTRs are indicated in gray. Capsid protein (GAG), integrase (INT), RT, and RNAse H1 domains are shown within the light-blue-labeled TERESTRA protein-coding part. Primer binding sequence (PBS) and polypurine tract (PPT) are indicated by red boxes. d Sequence similarities within pair-wise LTR alignments between A. lyrata and A. thaliana TERESTRA, ONSEN, and COPIA46 families. More than 70 % similarity was expected for members of the same family. TERESTRA is absent in A. thaliana. e In silico reconstruction of putative HREs in the 5’ LTR of TERESTRA. The criteria were as described for Fig. 3a. Detailed information including sequences underlying individual HREs can be found in Additional file 5: Figure S9. f Transcript levels of TERESTRA in response to 6 and 12 h 37 °C HS in different Brassicaceae. The experiment was performed as described in (b). g Reconstruction of TERESTRA HRE evolution. The phylogenetic tree was developed using a chalcone synthase gene of each individual species. The numbers at the base of the branches indicate bootstrap values. Black lines show species with low efficiency HREs and red lines highlight independently evolved high efficiency HREs in A. lyrata and B. stricta. Gray lines denote species where TERESTRA could not be found. TE transcript accumulation of (h) ONSEN, (i) COPIA37, and (j) TERESTRA after 0, 6, and 12 h 37 °C HS preceded by 48 h control (no inhibitor), 10 μM 3-deazaneplanocin A (DZNep), or 40 μM zebularine treatment. Transcript amounts were normalized to UBC28 mRNA and signals from drug and heat-treated samples were recalculated as fold-changes relative to 0 h. Error bars indicate variation between two biological replicates and * P <0.05 in Student’s t-test

Fig. 4

ROMANIAT5-2 controls heat-responsiveness of APUM9 in A. thaliana. a Schematic representation of the ROMANIAT5-2 – APUM9 region in A. thaliana. The yellow block within the 3’ LTR represents a 3P/Gap heat responsive element (HRE). S position of primers for RT of the sense transcripts, A position of primers for RT of the anti-sense transcripts, F and R forward and reverse quantitative PCR primers. META1 is a transposon fragment flanking ROMANIAT5-2 3’ LTR. Silex: the orange block corresponds to the genomic fragment cloned upstream of the 4× “upstream activating sequence” (UAS, violet) and green fluorescent protein (GFP; green). b Schematic representation of the A. lyrata APUM9 locus. Reads per kilobase per million reads (RPKM) for (c) ROMANIAT5 and (d) APUM9 under control, 6 h at 37 °C HS and HS with 48 h recovery at control conditions (HS + R). * P <0.05 in t-test. e RT-PCR analysis of Silex reporter construct response to HS. NS non-stressed control plants, CS and HS control- or heat-stressed plants, respectively, +0 and +5d days of recovery at non-stress conditions, RT+ and RT– samples with and without RT, respectively. 18S rRNA transcript serves as positive control. f GFP signal in control and 24 h heat-stressed (HS2) Silex, detected after 0, 1, 2, or 5 days of recovery. Red – chlorophyll, green – GFP. g Close-up view of plants treated as described in (f). h Strand-specific RT-qPCR of APUM9 and ROMANIAT5-2 in A. thaliana after 6 h HS. i Putative HREs in ROMANIAT5 LTRs in Brassicaceae. j RT-qPCR for ROMANIAT5 in Brassicaceae after 6 and 12 h at 37 °C HS. The values were normalized to UBC28. Error bars indicate standard deviation between three biological replicates and * P <0.05 in Student’s t-test