Transcriptional reprogramming of nucleotide metabolism in response to altered pyrimidine availability in Arabidopsis seedlings

Abstract

In Arabidopsis seedlings, inhibition of aspartate transcarbamoylase (ATC) and de novo pyrimidine synthesis resulted in pyrimidine starvation and developmental arrest a few days after germination. Synthesis of pyrimidine nucleotides by salvaging of exogenous uridine (Urd) restored normal seedling growth and development. We used this experimental system and transcriptional profiling to investigate genome-wide responses to changes in pyrimidine availability. Gene expression changes at different times after Urd supplementation of pyrimidine-starved seedlings were mapped to major pathways of nucleotide metabolism, in order to better understand potential coordination of pathway activities, at the level of transcription. Repression of de novo synthesis genes and induction of intracellular and extracellular salvaging genes were early and sustained responses to pyrimidine limitation. Since de novo synthesis is energetically more costly than salvaging, this may reflect a reduced energy status of the seedlings, as has been shown in recent studies for seedlings growing under pyrimidine limitation. The unexpected induction of pyrimidine catabolism genes under pyrimidine starvation may result from induction of nucleoside hydrolase NSH1 and repression of genes in the plastid salvaging pathway, diverting uracil (Ura) to catabolism. Identification of pyrimidine-responsive transcription factors with enriched binding sites in highly coexpressed genes of nucleotide metabolism and modeling of potential transcription regulatory networks provided new insights into possible transcriptional control of key enzymes and transporters that regulate nucleotide homeostasis in plants.

Keywords: Arabidopsis; nucleotide metabolism; purine; pyrimidine; transcriptome.

Figure 1

Experiment design. At day 2, PALA was added to a final concentration of 1 mM. Control and PALA-treated seedlings were sampled at days 5, 7, and 9. Some cultures containing 1 mM PALA were supplemented with 1 mM Urd at day 5, then seedlings were sampled at day 7 and day 9. For qRT-PCR assays, samples were also collected at the day 5 and 7 time points, and also the day 12 time point (not shown). Developmental and biochemical phenotype and gene expression data were based upon comparisons of PALA-treated, or PALA+Urd supplemented seedlings with untreated control seedlings at each time point.

Figure 2

Developmental and biochemical phenotypes resulting from changes in pyrimidine availability in Arabidopsis seedlings. (A) Effects of PALA ± supplements on seedling growth and development; control (A), 1 mM PALA (B), 1 mM PALA supplemented with either 1 mM Ura (C) or 1 mM Urd (D) on day 5, then grown for an additional 4 days (day 9) or 7 days (day 12). Scale bar = 0.5 cm. (B) Chlorophyll contents of seedlings growing for 5, 7, or 9 days in control medium, or medium containing 1 mM PALA or 1 mM PALA supplemented with 1 mM Thd, 1 mM Ura, or 1 mM Urd on day 5. (C) DNA and RNA contents of Arabidopsis seedlings grown in control medium or 1 mM PALA for 9 days, or in 1 mM PALA for 5 days, then supplemented with 1 mM Urd for four additional days. Values are means ± S.E. for three biological replicates. Asterisks mark samples that are significantly different from the untreated control at each timepoint (*p ≤ 0.05; ** p ≤ 0.01).

Figure 3

Metabolism of [2-¹⁴C]-uracil in 9-day-old PALA-treated seedlings. The soluble fraction is free uracil or labeled metabolites in seedling tissues. The insoluble fraction represents incorporation of uracil label into nucleic acids. CO₂ is label released by uracil catabolism. Asterisks mark samples that are significantly different from the control (*p ≤ 0.05; ** p ≤ 0.01).

Figure 4

Heatmap showing clustering (A) of normalized expression values for 6,245 DE genes (FC ≥ 2). Clusters 1–8 show genes with increased expression in response to PALA treatment, with reversal by Urd supplementation, and represent a timeline of early to late responses (left to right). Genes in clusters 9–12 show genes that were repressed by PALA treatment, with Urd reversal, with early to late responses (left to right). Control, PALA treatment, and PALA + Urd treatment (C, P, PU) at days 5, 7, and 9. Statistically overrepresented GO BioProcess categories for genes in each cluster (B).

Figure 5

De novo purine and pyrimidine synthesis pathways and their subcellular compartmentation in Arabidopsis (Zrenner et al., 2006; Witte and Herde, 2020). Pyrimidine synthesis is integrated with the synthesis of Arg at the level of CP utilization by ATC and OTC. The UMP end product exerts major control over pyrimidine synthesis through allosteric regulation of ATC activity (Belin et al., 2020) and is also formed through uracil salvaging by UPRT or Urd salvaging by PNK1 (see Figure 6 ). UMP and Orn modulate CPS activities; Arg and the N-sensing regulatory protein PII control Arg synthesis at the NAGK step (Slocum, 2005). Gene expression information is integrated into pathways to show steps that are induced or repressed in response to PALA treatment. Highly coexpressed genes in the three pathways are indicated with an asterisk. PALA inhibition of ATC is shown. Abbreviations used for enzymes and transporters are listed in Supplemental Table 2 and pathway intermediates are listed in Supplemental Table 3 .

Figure 6

Pyrimidine salvaging (intracellular and extracellular) and catabolism. Pathway organization and subcellular compartmentation are according to Witte and Herde (2020). Gene expression information is integrated into pathways to show steps that are induced or repressed in response to PALA treatment. Salvaging of nucleobases derived from RNA turnover in the vacuole and extracellular sources of nucleotides are indicated. Cytosolic salvaging of Cyd (via CDA1 and Urd), Urd, and Thd and plastid salvaging pathways for Ura and Urd are shown. Induced salvaging of Cyd to CMP by UCK2 and downstream cytidylates by UMK1/4 and NDPK1 activities is not shown). Further metabolism of UMP into thymidylate, uridylate, and (deoxy)cytidylate precursors of RNA and DNA synthesis and carbohydrate metabolism are also shown. Inhibition of UMP synthesis by PALA inhibition of the de novo pathway is indicated (see Figure 5 ). Abbreviations used for enzymes and transporters are listed in Supplemental Table 2 and pathway intermediates are listed in Supplemental Table 3 .

Figure 7

Purine salvaging (intracellular and extracellular) and catabolism. Pathway organization and subcellular compartmentation are according to Witte and Herde (2020). Gene expression information is integrated into pathways to show steps that are induced or repressed in response to PALA treatment. Salvaging of Ado and Gua derived from RNA turnover in the vacuole, deoxypurine metabolism, and salvaging of inosine and hypoxanthine are not shown. Abbreviations used for enzymes and transporters are listed in Supplemental Table 2 and pathway intermediates are listed in Supplemental Table 3 .

Figure 8

Heatmap showing clustering of normalized expression values for 89 DE TF genes (FC ≥ 2). Clusters 1–4 show genes with increased expression in response to PALA treatment, compared with untreated controls. Urd supplementation reversed this response. Genes in clusters 5–6 show the opposite response to PALA and Urd. Control, PALA treatment and PALA + Urd treatment (C, P, PU) at days 5, 7, and 9. Highly overrepresented GO BioProcess categories and fold-enrichment values for each cluster are shown.

Figure 9

Integrative network graphs (iRegNet) showing five highly coexpressed upstream genetic regulators (outer nodes) for differentially expressed TF (FC ≥ 1.5) that were either induced (A) or repressed (B) by pyrimidine starvation at day 5. Edge thickness indicates strength of co-expression between genes. Induced genes with FC ≥5 on day 5 and repressed genes with FC ≥ 5 on day 7 (bold font) represent the most highly regulated early TF genes. TF with enriched targets in genes of nucleotide metabolism, representing potential downstream regulations, are indicated (blue font).

Figure 10

Hypothetical regulatory networks for transcription factors ASIL2 (A) and ABR1 (B). Edge thickness indicates strength of co-expression between genes. ASIL2 is repressed on day 5. ABR1, a potential downstream target of ASIL2 is induced by day 7. Both TF have binding sites within the nucleotide metabolism genes and may further regulate the downstream targets by induction of ethylene-responsive TF, coordinating synthesis, and salvaging of nucleotides with Arg synthesis. Models were developed based on the DE genes with statistically enriched targets for ASIL2 and ABR1 (AGI numbers; green—induced genes, yellow—repressed genes).

Figure 11

Hypothetical regulon for DE genes of pyrimidine and purine catabolism. Transcription factors WRKY70 and NAC047 are highly induced by PALA and could coordinately regulate target gene expression in both pathways, although binding sites for these TF are not statistically overrepresented in the target genes.