Bioinformatic cis-element analyses performed in Arabidopsis and rice disclose bZIP- and MYB-related binding sites as potential AuxRE-coupling elements in auxin-mediated transcription

Abstract

Background: In higher plants, a diverse array of developmental and growth-related processes is regulated by the plant hormone auxin. Recent publications have proposed that besides the well-characterized Auxin Response Factors (ARFs) that bind Auxin Response Elements (AuxREs), also members of the bZIP- and MYB-transcription factor (TF) families participate in transcriptional control of auxin-regulated genes via bZIP Response Elements (ZREs) or Myb Response Elements (MREs), respectively.

Results: Applying a novel bioinformatic algorithm, we demonstrate on a genome-wide scale that singular motifs or composite modules of AuxREs, ZREs, MREs but also of MYC2 related elements are significantly enriched in promoters of auxin-inducible genes. Despite considerable, species-specific differences in the genome structure in terms of the GC content, this enrichment is generally conserved in dicot (Arabidopsis thaliana) and monocot (Oryza sativa) model plants. Moreover, an enrichment of defined composite modules has been observed in selected auxin-related gene families. Consistently, a bipartite module, which encompasses a bZIP-associated G-box Related Element (GRE) and an AuxRE motif, has been found to be highly enriched. Making use of transient reporter studies in protoplasts, these findings were experimentally confirmed, demonstrating that GREs functionally interact with AuxREs in regulating auxin-mediated transcription.

Conclusions: Using genome-wide bioinformatic analyses, evolutionary conserved motifs have been defined which potentially function as AuxRE-dependent coupling elements to establish auxin-specific expression patterns. Based on these findings, experimental approaches can be designed to broaden our understanding of combinatorial, auxin-controlled gene regulation.

Figure 1

Phylogram of GH3 homologs from various species and their associated promoter cis-element organisations. The nearest neighbours of the well-characterized soybean GH3 (Gm05g21680) [36] from several plant species were compiled and the corresponding predicted protein sequences were rooted to Physcomitrella patens PpGH3-1 (at the bottom) to create a neighbour-joining phylogram. The −1000 bp promoter sequences of the corresponding GH3 genes were manually co-plotted, in 5’ to 3’ orientation onto the phylogram. The location of specific ZRE, MRE and AuxRE cis-elements (see Table 1) are given using the presented colour code. The indicated boxes divide the phylogram into the ZRE-rich (upper section) and ZRE-poor (lower section) clade. The average number of cis-elements per promoter within the clades (ZRE rich/ZRE poor) is given next to the motif colour code. The position of the soybean GH3 (GmGH3) and its related Arabidopsis homolog (AtGH3.3), used in this study is highlighted in grey. The indicated scale reflects the number of amino acid substitutions per site.

Figure 2

Output of the applied randomization algorithm.A) Background motif abundance of the GRE motif. The parameter “number of promoters with a motif” was exemplarily determined for the GRE motif in several randomized promoter datasets (1000 random sets of 304 genes) and its distribution is visualised in the given histogram. Experimental datasets which exhibit a significant enrichment or depletion (for e.g. the GRE motif) should display a respective significant shift in their position in the background distribution. B) Background motif density of the GRE motif. The parameter “motif counts per promoter” was exemplarily determined for the GRE motif in several randomized promoter datasets (1000 random sets of 304 genes) and the average number of motif counts per promoter is visualised in the given histogram. Experimental datasets which exhibit a significant enrichment or depletion should display a respective significant shift in their position in the background distribution. C) Background dataset composition. Illustration of the number of times a promoter was randomly drawn to participate in the reference dataset. The algorithm draws individual promoters from genomic datasets only once or twice (average from 1000 dataset randomizations), indicating that only very limited redundancy is present in the background dataset modelling. D) False-positive error rate estimations for parameter I. The error rates for genomic frequent and infrequent motifs used in this study are given. It was calculated for each motif in differently sized random datasets (50, 200 and 1000). The approach performed well on all dataset sizes, however becomes inaccurate if the background distribution of the motif (e.g. MRE2) is not approximately Gaussian. The given false-positive error frequency is the number of p-value calls (x-axis) under a set α value (y-axis) observed in 1000 random calculation repetitions.

Figure 3

Motif enrichment or depletion of specific cis-elements in auxin-responsive promoters from A. thaliana. Statistical significant motif enrichment or depletion of specific ZRE, MRE/MYC and AuxRE related cis-elements (Table 1) in early (0.5 – 1 h post auxin treatment) and late (3 h post-treatment) auxin-responsive promoters from Arabidopsis thaliana. A) Motif enrichment or depletion for individual motifs in promoters of auxin-regulated genes. B) Significant enrichment or depletion of bipartite motif modules comprising ZRE, MRE/MYC and AuxRE cis-elements in promoters of auxin-regulated genes. The embedded motifs have a variable but maximal spacing of 100 bps. C) The significance level, which is defined by the determined Bonferroni corrected p-values, is displayed as colour-scale. Enriched motifs or modules with respect to parameter I are illustrated in shades of blue, whereas depleted motifs/modules are given in shades of red. Motif abundance that is significantly altered with respect to parameter II is coloured green for motif density enrichment and purple for motif density depletion.

Figure 4

Motif enrichment or depletion of specific cis-elements in auxin-responsive promoters from rice. Statistical significance of motif enrichment or depletion for specific ZRE, MRE/MYC and AuxRE related cis-elements in auxin-responsive promoters from Oryza sativa, 1–3 h after auxin treatment. A) Motif enrichment or depletion for individual motifs in promoters of auxin-regulated genes. B) Significant enrichment or depletion of bipartite motif modules comprising ZRE, MRE/MYC and AuxRE cis-elements in promoters of auxin-regulated genes. The embedded motifs have a variable, but maximal spacing of 100 bps. C) The significance level scale for the Bonferroni corrected p-values of parameters I and II is adapted from Figure 3.

Figure 5

Molecular characterisation of the GRAUX-module. A) GRE, MRE and AuxRE cis-elements within the −300 bp AtGH3.3 promoter region. Framed sequences indicate the positions of the GRE-, MRE- and AuxRE motifs within the AtGH3.3 promoter close to the transcriptional start site (TATA-box is underlined). The present cis-elements are serially numbered. The grey highlighted sequence represents the AtGH3.3 promoter region used as synthetic auxin-responsive GRAUX-module promoter construct. B) Expression profile of the synthetic ProGRAUX: GUS reporter constructs. Number of multimerisations and fold induction values are indicated. C) Auxin inducibility of the ProAtGH3.3 derived GRAUX(4x)-module reporter construct and its mutational derivates. A schematic view of the transfected reporter constructs is given. Mutated cis-elements are indicated by X. White coloured bars represent transfected, mock treated (DMSO) and black coloured bars NAA treated (0.25 μM, 16 h) samples. Presented results were obtained from transient protoplast transfection assays. Given are the mean GUS/NAN values (± SD) from 3 independent experiments. Different letters denote significant differences (p ≤ 0.05; one-way ANOVA followed by Fisher post-hoc test) between the used constructs and treatments. Fold induction values resulting from auxin application are given