Biotic Stress-Induced Priming and De-Priming of Transcriptional Memory in Arabidopsis and Apple

Abstract

Under natural growth conditions, plants experience various and repetitive biotic and abiotic stresses. Salicylic acid (SA) is a key phytohormone involved in the response to biotic challenges. Application of synthetic SA analogues can efficiently prime defense responses, and leads to improved pathogen resistance. Because SA analogues can result in long-term priming and memory, we identified genes for which expression was affected by the SA analogue and explored the role of DNA methylation in this memorization process. We show that treatments with an SA analogue can lead to long-term transcriptional memory of particular genes in Arabidopsis. We found that subsequent challenging of such plants with a bacterial elicitor reverted this transcriptional memory, bringing their expression back to the original pre-treatment level. We also made very similar observations in apple (Malus domestica), suggesting that this expression pattern is highly conserved in plants. Finally, we found a potential role for DNA methylation in the observed transcriptional memory behavior. We show that plants defective in DNA methylation pathways displayed a different memory behavior. Our work improves our understanding of the role of transcriptional memory in priming, and has important implication concerning the application of SA analogues in agricultural settings.

Keywords: DNA methylation; benzo(1,2,3)thiadiazole-7-carbo-thioic acid S-methyl ester (BTH); epigenetics; plant protection; transcriptomics.

Figure A1

Growth repression effect of BTH on grafted and in vitro apple plants. (A) Total number of internodes of grafted apple plants treated six times with BTH (1 mM) or water. Two months after grafting, plants were treated every three days. Number of internodes was counted three days after the last treatment. Bars represent the mean of at least two biological replications. (B) Pictures of in vitro grown apple plantlets. After propagation plantlets were grown for 4 weeks after propagation on media supplemented with or without 1 mM BTH. Error bars show ± SE of the mean. Significant differences according to Student’s t-test results: **, p < 0.01.

Figure A2

Different expression value plot of all Class B DETs, and the comparison ww vs. bf confirming the de-priming expression profile. Log₂ expression values of DETs of the two sub-categories in Class B (transgressive DETs excluded), as well as their different expression value in the comparison ww vs. bf, were plotted.

Figure A3

Gene ontology analysis of DETs in Class A and Class B. (A) PANTHER overrepresentation test of DETs in Class A. Gene ontology (GO) was available for 1131 out of the 1801 DETs. DETs correlated with the regulation of gene expression and epigenetics (GO:0040029) were overrepresented, with 4.85 fold enrichment compared to the Arabidopsis reference. (B) Overrepresentation test of Class B. Gene ontology for 645 of the 1058 DETs was available. GOs were enriched for disaccharide metabolic process, sulfur compound metabolic process, response to stress and stimulus (GO: 0005984, GO: 0006790, GO:0006950, GO:0050896) were overrepresented in this category. Gene ontology overrepresentation test was probed using the PANTHER webtool, using a Fisher’s exact test (Mi et al., 2017) [64].

Figure A4

Class B DETs tended to be depleted in gene body methylation. Based on previously published data (Cokus et al., 2008) [44] genes were classed for the presence of DNA methylation in a promoter region (2 kb upstream, 5’ methylation), body of the genes (gene body methylation), and methylation downstream of the transcription end site (3’ methylation).

Figure 1

Benzo(1,2,3)thiadiazole-7-carbo-thioic acid S-methyl ester (BTH) induces short- and long-term defense responses in Arabidopsis. (A) Production of reactive oxygen species (ROS), measured in RLU (relative light units), in wild-type Arabidopsis leaf discs (Col-0), treated with 1 µM flg22, 1 mM BTH or without elicitor (control). Graphs display averages of 12 replications. (B) Quantification of callose deposition. The bars represent the mean of 4 replications. (C) Localization of callose deposition by aniline blue staining. (D) Temporal order of applied Arabidopsis treatments. The first treatment was applied 7 dag (days after germination), the third treatment 13 dag. (E) Quantification of fresh weight of 21 dag old Arabidopsis plants. Plants were previously sprayed three times with water or BTH (1 mM), according the scheme shown in (D). (F) Pictures of 21 dag plants previously sprayed with water or BTH (1 mM). Error bars show ± SE of the mean. Significant differences according to Student’s t-test results: *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Figure 2

Arabidopsis transcriptome analysis set-up and results. (A) Experimental set-up of applied treatments. The first treatment was applied to young Arabidopsis plants 7 dag (days after germination), the second 10 dag and plants were harvested at 21 dag. (B) List of applied treatments after 7 and 10 dag as well as the sample name. (C) List of examined microarray comparisons and number of DETs. (D) Venn diagram showing DETs of all comparisons, and common DETs within the examined comparisons.

Figure 3

Common differentially expressed transcripts resulting from flg22 and BTH treatments. (A) Scatter plot of the log₂ expression values of Class A DETs showing ww vs. bw on the X axes and ww vs. wf on the Y axes. The expression profile is also shown as a heatmap (upper left). DETs of ww vs. bw are shown in the first column and DETs from ww vs. wf in the second column. Up-regulated DETs are indicated in green and down-regulated DETs in red. (B) Class A was divided into Class A up by BTH and Class A down by BTH subcategories. The graph represents the percentage of non-transgressive transcripts expressed in sense and antisense in both sub-categories, as well as transcripts that showed a transgressive expression pattern (sense and antisense separated) and therefore did not follow the global expression trend.

Figure 4

DETs of Class B show an inversion of gene expression profile following a second stress. (A) Scatter plot of transcripts present in the Class B DETs. The category contains DETs of the comparison ww vs. bw (X axes) and bw vs. bf (Y axes). Class B DETs are also shown with a heatmap (ww vs. bw iin the first column, and bw vs. bf in the second column) (B) Classification of transcripts present in Class B. The 1058 transcripts were compared to DETs of ww vs. bf. Results show 80.5% of the transcripts were not differently expressed by the combination of both treatments, and are considered as de-primed. 12.6% were differently expressed and therefore primed, and 6.9% did not follow the global expression trend in Class B. (C) Distribution of sense and antisense transcripts in the sub-categories Class B up by BTH and Class B down by BTH. The sub-category Class B up by BTH contains 42.9% non-transgressive antisense transcripts.

Figure 5

Whole genome expression profiling and applied comparisons in apple. (A) Experimental set up of apple plantlet treatments. Apple plantlets were treated the first time 14 dap (days after propagation) followed by one additional treatment 17 dap. Leave tissue was harvested 31 dap. (B) Combinations of applied treatments and sample names. (C) List of examined microarray comparisons and number of DETs.

Figure 6

De-priming in apple. A: scatter Plots of common transcripts present in the comparisons ww vs. bw (X axes) and bw vs. bf (Y axes). A heatmap confirms the expression profile B: The common transcripts are compared to ww vs. bf. 82.7% are de-primed, 8.2% stay primed and 9.1% are transgressive and do not follow the global trend. C: The common DETs are divided into the sub-categories up by BTH and down by BTH.

Figure 7

Expression of AMY1 serves as marker for de-priming of transcription. (A) Set-up of treatments applied to Arabidopsis plants. The first treatment was applied 7 dag (days after germination), followed by two additional treatments. Plants were harvested 21 dag. (B) Expression profile of AMY1 by different sequences of treatments of wild type plants. (C,D) The same treatment orders were applied to nrpd1-3 and met1-3. All expression values were normalized to that of the gene ACR12 (AT5G04740). Bars represent the mean of at least four biological replications. Error bars show ± SE of the mean. Significant differences according to Student’s t-test results: *, p < 0.05; ***, p < 0.001; ns: not significant.

Figure 8

Proposed model of de-priming effect on plant fitness. Here we suggest a model of a possible beneficial effect of transcriptional de-priming. If plants are not exposed to a first priming treatment, the second treatment might cause stronger deficits on plant fitness (red line) in comparison to plants that have been primed (blue line). However, priming reflects a fitness costs for plants due to the induced defense response and the maintenance of the primed transcription and/or epigenetic memory. The second treatment could cause a less pronounced fitness cost in comparison to the un-primed plants. We propose that the de-priming of a DET subset could lead to an additional positive effect on plant fitness, by fine tuning the plant defense response and returning non-beneficial DETs to a basic expression level (dashed green line).