DNA-binding specificity and molecular functions of NAC transcription factors

Abstract

The family of NAC (NAM/ATAF1,2/CUC2) transcription factors has been implicated in a wide range of plant processes, but knowledge on the DNA-binding properties of the family is limited. Using a reiterative selection procedure on random oligonucleotides, we have identified consensus binding sites for two NAC proteins. The consensus sequences are similar, but not identical; both contain the core CGT[GA]. The strict consensus sequences, comprising only the most frequent base at each position, are: TTNCGTA and TTGCGTGT. In silico analysis of target promoter regions corroborated the selection results. Furthermore, NAC protein binding to the CaMV 35S promoter was shown to depend on sequences similar to the consensus of the selected oligonucleotides. Electrophoretic mobility shift assays demonstrated that NAC proteins bind DNA as homo- or heterodimers and that dimerization is necessary for stable DNA binding. The ability of NAC proteins to dimerize and to bind DNA was analysed by structure-based mutagenesis. This identified two salt bridge-forming residues essential for NAC protein dimerization. Alteration of basic residues in a loop region containing several highly conserved residues abolished DNA binding. Thus, the results presented here contribute significantly to our understanding of the specificity and molecular functions of the NAC protein DNA-binding domain.

Introduction

Life processes in plants depend on a multitude of transcription factors to regulate gene expression. The NAC proteins constitute one of the largest families of plant-specific transcription factors [1] and have been implicated in developmental programs, defence and abiotic stress responses (reviewed in [2]). This protein family is defined by an N-terminal domain called the NAC domain [3], which mediates DNA binding [4], [5], [6]. The three-dimensional structure of the NAC domain has been determined and shows a unique transcription factor fold [6]. However, the mode of DNA recognition is unknown.

The DNA-binding ability of NAC proteins was first suggested by the activation of the cauliflower mosaic virus (CaMV) 35S promoter [Hirt, original GenBank annotation and personal communication in 7]. Subsequently, a number of NAC proteins have been shown to bind fragments of this promoter [4], [5], [6], [8]. The identification of endogenous target genes of the NAC transcription factors has only recently been pursued. A number of genes are upregulated by overexpression of either of three stress-inducible Arabidopsis NAC genes, ANAC019, ANAC055 and ANAC072/RD26 [9], [10]. In a yeast one-hybrid screen, the three ANAC proteins bound a fragment of the early responsive to dehydration stress 1 (ERD1) promoter. Substitution analysis identified CACG (reverse complement: CGTG) as the core sequence recognized by the ANAC proteins in the ERD1 promoter [10].

To investigate the DNA binding of NAC proteins further, we performed PCR-assisted binding site selection experiments with two Arabidopsis NAC proteins, ANAC019 and ANAC092 (nomenclature from [11]). The consensus sequences obtained were evaluated by electrophoretic mobility shift assays (EMSAs), and analyses of putative NAC target gene promoters and the CaMV 35S promoter substantiated the results. NAC proteins were shown to bind DNA as dimers in an EMSA experiment with different-sized NAC proteins. A NAC domain truncation rendering the domain unable to dimerize also prevented DNA binding. Molecular functions of the NAC domain, dimerization and DNA binding, were examined using site-directed mutagenesis. The mutational analysis identified residues important for dimerization and a loop region involved in DNA binding.