Arabidopsis

Vylíčilová, Hana, Husičková, Alexandra, Spíchal, Lukáš, Srovnal, Josef, Doležal, Karel, Plíhal, Ondřej & Plíhalová, Lucie, 2016, C 2 - substituted aromatic cytokinin sugar conjugates delay the onset of senescence by maintaining the activity of the photosynthetic apparatus, Phytochemistry 122, pp. 22-33 : 25-28

2.4. Regulation of gene expression in senescent Arabidopsis leaves

Using BAP and DMSO treatments as references, we performed genome-wide expression profiling of senescent Arabidopsis leaves treated with compounds 3 and 6 in order to better understand the regulation of senescence by ArCKs at the molecular level. Compounds 3 and 6 were selected because they exhibited the highest anti-senescence activity in the detached wheat leaf bioassay ( Table 2 ). Expression changes were monitored in detached senescent Arabidopsis leaves after 48 h of incubation in darkness with a 10 µM solution of one of the tested compounds or with DMSO alone; a complete list of DE genes from the three treatments (i.e. treatment with 3, 6 and BAP) is shown in the Supplementary Table S1. Transcriptome profiling was performed using a standardized procedure developed for the detached leaf assay, which has previously been used to investigate the genetic effects of treatment with Kin derivatives that exhibit anti-senescent activity ( Mik et al., 2011). While these Kin derivatives were effective anti-senescence agents under both light and constant dark conditions, their senescence-delaying activity was greatest in darkness.

Hierarchical clustering analysis of the resulting data sets revealed that the samples treated with specific cytokinins clustered together and exhibited low variability in their responses, as did the mock-treated wild type Arabidopsis leaf samples ( Fig. 2A ). Interestingly, the BAP-treated samples formed the most distinct group and had a gene expression profile that differed significantly from those for all of the other groups. In contrast, the gene expression profiles for the groups treated with 3 and 6 were quite similar. This is readily apparent in the heat map shown in Fig. 2B , which presents data for 8659 genes whose expression changed significantly after treatment with 3, 6 or BAP ( P value Ḉ 0.01 in at least one treatment). To limit the number of target genes, we adopted more stringent statistical criteria for identifying genes whose expression had changed significantly. Specifically, the data were RMA normalized and genes were required to have a signal ratio change log 2 of À 0.5 or Ḉ — 0.5 in addition to a P -value of <0.01; see the Methods section for details. In this way we defined a group of 1128 genes whose expression changed after treatment with 3 (of which 510 were upregulated and 618 downregulated), and 944 genes whose expression changed upon treatment with 6 (548 upregulated, 396 downregulated). These two groups overlapped extensively: there were 671 genes common to both ( Fig. 2C ; Table S2 , Supplementary data). To better understand the molecular functions of these genes, we categorized the transcripts in both groups according to their GO terms ( Ashburner et al., 2000) using categories such as ‘transcription factor activity’, ‘DNA or RNA binding’, and ‘protein binding’ ( Fig. 2D ). This analysis indicated that the affected genes were rather evenly distributed over the defined categories.

We then examined the genes whose expression changed significantly in the three datasets (i.e. in senescent leaves treated with 3, 6 or BAP) in more detail. This analysis revealed that the cytokinin derivatives had distinct modes of action to those observed for the parent free bases. In keeping with previous reports ( Brandstatter and Kieber, 1998; Rashotte et al., 2003), many cytokinin-related genes were upregulated in the BAP-treated leaves ( Fig. 3 ; Table S2 , Supplementary data). Importantly, these included the cytokinin response regulators ARR7, ARR9, ARR5, ARR6 and ARR4. Exposure to high concentrations of BAP also prompted the induction of several cytokinin dehydrogenase genes including CKX1, CKX2, CKX3 and CKX4. Upregulation of response regulators, several CKX genes, and other cytokinin response genes was also observed in senescent leaves treated with 3 or 6.

It is interesting to compare our results to those of a recent meta-analysis of microarray data reported by various laboratories, which identified a core list of cytokinin response genes ( Brenner et al., 2012). The results of both this meta-analysis and a search of the Genevestigator database (https://genevestigator.com/gv/) conducted by ourselves indicate that tZ treatment leads to rapid reprogramming of gene expression in Arabidopsis . Specifically, cytokinin response regulators such as ARR15, ARR5, ARR16, ARR7, ARR4, ARR6 and ARR9 were strongly upregulated in response to tZ treatment. Other cytokinin-responsive genes identified in the meta-analysis and database search include CKX4 and CKX5 (which code for two cytokinin dehydrogenase isoforms), AHK4 and AHK1, CRF5 or CYP735A2, and CYP82F1. This group of core cytokinin-responsive genes clearly overlaps extensively with the list of genes whose expression was altered significantly following treatment with 3, 6, or BAP ( Fig. 3 ). It is also consistent with the cytokinin bioassay results presented in the preceding section and thus confirms that the halogenated aromatic cytokinin derivatives considered in this work are indeed active cytokinins whose signaling effects partially mirror those of BAP and tZ.

Our analysis also revealed some genes that were only affected by treatment with 3 and 6, most of which were directly or indirectly linked to photosynthesis. Some of these genes are listed in Fig. 3 , in which the most significant hits are categorized according to their function in cytokinin signaling and metabolism or photosynthesis and related categories. Importantly, genes encoding components of the photosystem II light harvesting complex (LHCII), namely At 2g 05070, At 5g 54270, At 1g 44575, At 3g 01440, At 3g 55330 and At 2g 39470, were upregulated by treatment with 3 and 6. In contrast, BAP treatment had mostly negligible effects on these genes. In addition, our in silico analyses using the Genevestigator database confirmed that these genes are probably not affected by tZ treatment in Arabidopsis seedlings: treatments with 1 µm tZ solution for 30 min, 1 h, or 3 h had almost no measurable effects on the expression of any photosystem II-related gene. As mentioned above, the different modes of action of 3, 6 and BAP may not be directly attributable to differential activation of cytokinin response regulators because all three of these ligands upregulated most of the ARR genes to a similar degree. However, ARR8 and ARR15 were found to be less upregulated in samples treated with BAP than those treated with 3 or 6 ( Fig. 3 ). This may be important because ARR15 is a negative regulator of AHK4-mediated cytokinin signal transduction whose expression is particularly strong in roots ( Kiba et al., 2003). Therefore, negative regulation of the cytokinin signaling machinery may diminish some of the negative effects associated with exogenous cytokinin treatment at higher concentrations. However, we cannot exclude the possibility that the tested cytokinin derivatives may activate multiple signaling pathway(s) simultaneously, some of which may be closely related to the cytokinin pathway. One of these may be the auxin signaling pathway: we found several auxin-related genes that were differentially regulated by treatment with 3 and 6, including PIN3 and PIN5 ( Fig. 3 ). Another example is downregulation of the auxin-responsive gene At 2g 45210, which codes for senescence-associated gene 201 (SAG201), a positive regulator of senescence.

We also found a group of genes involved in chlorophyll degradation that were downregulated in response to treatment with 3 and 6 ( Fig. 3 ; Table S2 , Supplementary data). Several of these genes were previously described as leaf anti-senescent markers that respond to auxin, cytokinin and some other molecules that regulate leaf senescence ( Li et al., 2012). This group included At 4g 13250, which codes for a protein that is involved in LHCII degradation in rice, non-yellow coloring protein 1 (NYC1) ( Kusaba et al., 2007). It also included genes encoding other chlorophyll/LHCII catabolic reductases such as At 4g 22920, which codes for non-yellowing protein 1 (NYE1), and At 4g 11910, which codes for non-yellowing protein 2 (NYE2). The latter gene was also downregulated in response to BAP treatment together with At 3g 44880, which codes for Pheophorbide A oxygenase/accelerated cell death 1 (PAO/ACD1).

In conclusion, we have collected evidence that selected cytokinin derivatives have similar signaling outputs to their parent free bases in general but also exhibit selectively modulated anti-senescence activity. This modulation is primarily due to upregulation of genes coding for the subunits of LHCII and LHCI and downregulation of genes that are responsible for LHCII and chlorophyll degradation.

Figures

Fig. 2. Transcriptome reprogramming mediated by cytokinin derivatives as revealed by DNA microarray analysis (Affymetrix). Hierarchical clustering was performed for all data sets, i.e. Arabidopsis leaves treated with BAP, compound 3, compound 6, and DMSO alone (A). Clustering analysis of genes exhibiting differential expression after treatment with BAP, compound 3 or compound 6 (B).Venn diagrams of genes regulated by 3 and genes regulated by 6, showing the substantial overlap of the two sets (C). GO classification of genes regulated by 3 and 6 (D); genes were categorized into groups according to their GO terms, based on their predicted or putative functions.

Fig. 3. Selected genes identified in differential expression analyses of detached Arabidopsis leaves treated with cytokinin derivatives. Senescent leaves were incubated in darkness for 48 h in MS medium containing 10 µM 3, 6, or BAP and their gene expression profiles were compared to those for DMSO-treated control leaves of the same age. Gene expression values were calculated as log2-ratios of the signals for the treated samples relative to those for the control leaves. The heat map is color-coded using the scheme presented in the figure; genes shown in gray were not significantly affected. Only genes related to photosynthesis (specifically those related to photosystem I/II, the Calvin cycle, and chlorophyll/LHCII catabolism) and cytokinin/other plant hormone signaling, with expression changes significant at the P <0.01 level and log2 fold change values of À1 or Ḉ—1 for at least one treatment are shown.