Deciphering the Molecular Mechanisms Underpinning the Transcriptional Control of Gene Expression by Master Transcriptional Regulators in Arabidopsis Seed

Abstract

In Arabidopsis (Arabidopsis thaliana), transcriptional control of seed maturation involves three related regulators with a B3 domain, namely LEAFY COTYLEDON2 (LEC2), ABSCISIC ACID INSENSITIVE3 (ABI3), and FUSCA3 (ABI3/FUS3/LEC2 [AFLs]). Although genetic analyses have demonstrated partially overlapping functions of these regulators, the underlying molecular mechanisms remained elusive. The results presented here confirmed that the three proteins bind RY DNA elements (with a 5'-CATG-3' core sequence) but with different specificities for flanking nucleotides. In planta as in the moss Physcomitrella patens protoplasts, the presence of RY-like (RYL) elements is necessary but not sufficient for the regulation of the OLEOSIN1 (OLE1) promoter by the B3 AFLs. G box-like domains, located in the vicinity of the RYL elements, also are required for proper activation of the promoter, suggesting that several proteins are involved. Consistent with this idea, LEC2 and ABI3 showed synergistic effects on the activation of the OLE1 promoter. What is more, LEC1 (a homolog of the NF-YB subunit of the CCAAT-binding complex) further enhanced the activation of this target promoter in the presence of LEC2 and ABI3. Finally, recombinant LEC1 and LEC2 proteins produced in Arabidopsis protoplasts could form a ternary complex with NF-YC2 in vitro, providing a molecular explanation for their functional interactions. Taken together, these results allow us to propose a molecular model for the transcriptional regulation of seed genes by the L-AFL proteins, based on the formation of regulatory multiprotein complexes between NF-YBs, which carry a specific aspartate-55 residue, and B3 transcription factors.

Figure 1.

Characterization of the B3 AFL-binding sites. A, Logos representing the DNA-binding specificities of ABI3, FUS3, and LEC2 as identified in vitro. Sequences presented for FUS3 and ABI3 correspond to the top-scoring 8-mers obtained for these proteins in protein-binding microarray assays; their corresponding position weight matrices (PWMs) were used for logo representations. The motif presented for LEC2 was derived from high-throughput SELEX experiments using MEME (see “Materials and Methods”). Scores are in arbitrary units (bits). B, Positions of the different cis-regulatory elements identified in the OLE1 promoter. The RYL elements identified using PWMs obtained in A are highlighted in blue, whereas GBL elements identified using PLACE (http://www.dna.affrc.go.jp/PLACE/index.html) are highlighted in red. Numbers placed above the DNA sequence indicate nucleotide positions from the translational start site.

Figure 2.

Functional dissection of the OLE1 promoter in planta. A, At top is a schematic representation of the RYL and GBL elements identified within the OLE1 promoter sequence. A series of 5′ deletions was generated, and translational fusions to the uidA gene were prepared. The corresponding transgenic embryos were assayed for GUS activity at the torpedo, bent-cotyledon, and maturing stages. The length of the promoter tested is indicated at left. B and C, Mutations of the RYL (B) and GBL (C) elements were generated in the context of the 257-bp OLE1 promoter, and translational fusions to the uidA gene were prepared. The corresponding transgenic embryos were assayed for GUS activity. D, Combinations of mutations affecting GBL and RY elements were generated in the context of the 257-bp OLE1 promoter, and translational fusions to the uidA gene were prepared. The corresponding transgenic embryos were assayed for GUS activity. From left to right are the name of the mutagenized promoter under study, a schematic representation of the promoter with mutagenized elements in black, representative photographs of stained embryos at three different stages of the maturation process, and a bar graph showing the partition of staining intensities (from 0 = colorless to 4 = intense staining) among a population of n independent transformants.

Figure 3.

Synergistic effects of LEC2, LEC1, and ABI3 on the activation of ProOLE1 in moss protoplasts. Transient expression assays were carried out in moss protoplasts. GFP activity was measured in protoplasts transfected with the ProOLE1_-257:GFP reporter and constructs allowing the expression of B3 AFL (LEC2, ABI3, or FUS3) or LEC1, alone or in combination. Means of at least three replicates ± sd are presented. AU, Arbitrary units; NS, nonsignificant. Asterisks indicate significant differences. *P value 5%, ***P value 0.1%. Statistical analysis is provided in Supplemental Table S3.

Figure 4.

Enhanced activation of ProOLE1 by NF-YB proteins is specific to the LEC1 type. Transient expression assays were carried out in moss protoplasts. GFP activity was measured in protoplasts transfected with the ProOLE1_-257:GFP reporter and constructs allowing the expression of LEC2, ABI3, and LEC2 and ABI3 together with different NF-YB proteins, namely LEC1, NF-YB7 (At2g13570), and LEC1m (LEC1 D55K). Means of at least three replicates ± sd are presented. AU, Arbitrary units; NS, nonsignificant. Asterisks indicate a significant difference. ***P value = 0.001. Statistical analysis is provided in Supplemental Table S3.

Figure 5.

Importance of the RYL and GBL cis-regulatory elements in the activation of proOLE1 by the L-AFLs in moss protoplasts. At top is a schematic representation of the RYL (RYL1–RYL3) and GBL (GBL1–GBL3) elements identified in the minimal OLE1 promoter. At bottom are results from transient expression assays carried out in moss protoplasts with different mutagenized versions of the OLE1 promoter. GFP activity was measured in protoplasts transfected with the ProOLE1:GFP reporter and constructs allowing the expression of LEC2, ABI3, or a combination of LEC2, LEC1, and ABI3. From left to right are the name of the mutagenized promoter under study, a schematic representation of the promoter with mutagenized elements in black, and a bar graph showing fluorescence intensities in moss protoplasts expressed as a ratio of the intensity measured with and without expression of the L-AFLs. Statistical analysis is provided in Supplemental Table S3.

Figure 6.

Interaction of LEC1, LEC2, and NF-YC2 in pull-down experiments. Coimmunoprecipitation experiments were performed using native or mutated LEC1 (LEC1m D55K) as a bait and LEC2 as a prey in the presence (or absence) of NF-YC2 and of a proOLE1:GFP construct. HA- and Myc-tagged proteins were coexpressed in Arabidopsis protoplasts and unraveled by western blot using anti-Myc and anti-HA antibodies, respectively. INPUT, 2% of the full protein extract; IP, immunoprecipitated sample. Asterisks indicate heavy (black) and light (red) chains of the antibody used for immunoprecipitation.

Figure 7.

Model for the transcriptional regulation of OLE1 expression by the L-AFL proteins. In this model, LEC2 and ABI3 transcription factors directly and independently bind DNA and have a synergistic effect, together with LEC1, on the activation of the ProOLE1 promoter. This may be due to their cooperating in a multiprotein complex, as shown by pull-down assays that unraveled the occurrence of LEC1/LEC2/NF-YC2 ternary complexes. bZIPs may play a role in this complex, although they were not tested in this study. D, Asp involved in the activity of LEC1; blue bars, RYL element; red bars, GBL element.