Nearby transposable elements impact plant stress gene regulatory networks: a meta-analysis in A. thaliana and S. lycopersicum

Abstract

Background: Transposable elements (TE) make up a large portion of many plant genomes and are playing innovative roles in genome evolution. Several TEs can contribute to gene regulation by influencing expression of nearby genes as stress-responsive regulatory motifs. To delineate TE-mediated plant stress regulatory networks, we took a 2-step computational approach consisting of identifying TEs in the proximity of stress-responsive genes, followed by searching for cis-regulatory motifs in these TE sequences and linking them to known regulatory factors. Through a systematic meta-analysis of RNA-seq expression profiles and genome annotations, we investigated the relation between the presence of TE superfamilies upstream, downstream or within introns of nearby genes and the differential expression of these genes in various stress conditions in the TE-poor Arabidopsis thaliana and the TE-rich Solanum lycopersicum.

Results: We found that stress conditions frequently expressed genes having members of various TE superfamilies in their genomic proximity, such as SINE upon proteotoxic stress and Copia and Gypsy upon heat stress in A. thaliana, and EPRV and hAT upon infection, and Harbinger, LINE and Retrotransposon upon light stress in S. lycopersicum. These stress-specific gene-proximal TEs were mostly located within introns and more detected near upregulated than downregulated genes. Similar stress conditions were often related to the same TE superfamily. Additionally, we detected both novel and known motifs in the sequences of those TEs pointing to regulatory cooption of these TEs upon stress. Next, we constructed the regulatory network of TFs that act through binding these TEs to their target genes upon stress and discovered TE-mediated regulons targeted by TFs such as BRB/BPC, HD, HSF, GATA, NAC, DREB/CBF and MYB factors in Arabidopsis and AP2/ERF/B3, NAC, NF-Y, MYB, CXC and HD factors in tomato.

Conclusions: Overall, we map TE-mediated plant stress regulatory networks using numerous stress expression profile studies for two contrasting plant species to study the regulatory role TEs play in the response to stress. As TE-mediated gene regulation allows plants to adapt more rapidly to new environmental conditions, this study contributes to the future development of climate-resilient plants.

Keywords: Gene regulation; Plant genomes; Regulatory networks; Stress; Transposable elements.

Fig. 1

Positioning and abundance of TEs nearby protein-coding genes. A) Three genomic positionings of TEs relative to genes were considered: “upstream” contains TEs within 1 kb upstream of the gene, “downstream” contains those within 1 kb downstream of the gene, and “intron” contains TEs within introns. Only the transposon boundary closest to the gene was considered for this classification. B) Well-defined TE superfamilies (TEFs) adjacent to protein-coding genes in A. thaliana and the proportion of gene-proximal TEs they hold. ‘Rest’ indicates the sum of Mariner, ATDNA12T3_2, Tc1 and ATREP18 TEFs. C) Well-defined TE superfamilies (TEFs) adjacent to protein-coding genes in S. lycopersicum and the proportion of gene-proximal TEs they hold. ‘Rest’ indicates the sum of Mariner, Retrotransposon, Helitron and TRIM_LARD TEFs

Fig. 2

Fold enrichment of differentially expressed genes near specific TE superfamilies (TEFs) upon stress for A) A. thaliana and B) S. lycopersicum. The Chi-squared test was conducted separately for up- and downregulated genes with TEFs upstream, downstream or within introns. The intensity from yellow to red reflects the enrichment score with values between 1 and 4.5, as compared to all differentially expressed genes near all TEFs in that specific genomic positioning and stress condition. The significance of enrichment is indicated within the tiles: * = FDR adjusted p-value < 0.05, ** = FDR adjusted p-value < 0.01, *** = FDR adjusted p-value < 0.001. We additionally filtered out significant results for which the observed number of differentially expressed genes near a TEF was less than 5 and the expected number was less than 2. Only TEFs, stress conditions and genomic positionings for which a valid enrichment was found are shown. A. thaliana: heat_B = 1 h incubation at 44 °C - leaves, paraquat_A = spray with 25 μM paraquat, photorespiratory_mutant_B = SHORT_ROOT (shr) mutant – 24 h photorespiratory stress, proteasome_inh_A = 100 μM proteasome inhibitor MG132, proteasome_mutant_B = rpn-10 mutant – RPN10 is a subunit of the 26S proteasome, salt_heat_A = 150 mM NaCl for 15 days + 1 h incubation at 44 °C – leaves. S. lycopersicum: hormone_B = 48 h after treatment with ACC (ethylene precursor), infection_necrotrophic_A = infection by Colletotrichum gloeosporioides - leaves, infection_necrotrophic_C = infection by Pseudomonas syringae pv. tomato DC3000 - leaves, infection_viral_A = infection by Tomato yellow leaf curl virus - leaves, light_A = constant shade – shoot apical meristem / leaf primordia, light_B = constant sun – shoot apical meristem / leaf primordia, light_C = sun to shade – shoot apical meristem / leaf primordia, light_D = constant sun - shoot apical meristem / leaf primordia, stress_tolerance_A = male-sterile, stress tolerant mutant

Fig. 3

TE-mediated heat and proteotoxic stress gene regulatory network for A. thaliana. Copia elements in upstream regions and Gypsy elements in introns of heat-responsive genes recruited specific regulatory factors. Also, Gypsy elements within introns and downstream regions and SINE within introns of proteoxic stress-responsive genes hosted cis-regulatory motifs targeted by specific TFs. We can distinguish several regulons, related to the different TEF-differentially expressed genes associations from left to right: SINE/proteasome mutant targeted by ARR18 (grey), SINE/proteasome inhibitor targeted by REF6 (green), Copia/salt-heat targeted by BPC1, ZML2, REF6, NAC6, TCP and RAMOSA1 (orange), Copia/heat targeted by BPC1, ZML2, REF6, NAC6, TCP and RAMOSA1, in addition to HSFB2A and S1FA3 (red), Gypsy downstream/proteasome mutant targeted by HSFB2A, S1FA3, TCP16, AT2G01818, DREB/CBF, GBF3 and MYC4 (darkblue), Gypsy intron/proteasome mutant targeted by HSFB2A, S1FA3, NAC, MYC4 and ILR3 (lightblue), Gypsy/heat targeted by DREB/CBF (purple) and Gypsy/salt-heat targeted by DREB/CBF (pink)