Small Maf proteins (MafF, MafG, MafK): History, structure and function

Abstract

The small Maf proteins (sMafs) are basic region leucine zipper (bZIP)-type transcription factors. The basic region of the Maf family is unique among the bZIP factors, and it contributes to the distinct DNA-binding mode of this class of proteins. MafF, MafG and MafK are the three vertebrate sMafs, and no functional differences have been observed among them in terms of their bZIP structures. sMafs form homodimers by themselves, and they form heterodimers with cap 'n' collar (CNC) proteins (p45 NF-E2, Nrf1, Nrf2, and Nrf3) and also with Bach proteins (Bach1 and Bach2). Because CNC and Bach proteins cannot bind to DNA as monomers, sMafs are indispensable partners that are required by CNC and Bach proteins to exert their functions. sMafs lack the transcriptional activation domain; hence, their homodimers act as transcriptional repressors. In contrast, sMafs participate in transcriptional activation or repression depending on their heterodimeric partner molecules and context. Mouse genetic analyses have revealed that various biological pathways are under the regulation of CNC-sMaf heterodimers. In this review, we summarize the history and current progress of sMaf studies in relation to their partners.

Keywords: MafF; MafG; MafK; Small Maf; bZIP transcription factor.

Fig. 1

Structures of the Maf family proteins and other bZIP-type transcription factors. c-Maf, MafA, MafB and NRL constitute the large Maf family. MafG, MafK and MafF constitute the small Maf family. c-Jun and c-Fos are members of the AP-1 family. The protein regions are indicated by different colors: the transcriptional activation domain (TAD; green), extended homology region (EHR; red), CNC domain (orange), basic region (purple) and leucine zipper (blue). The lengths of amino acid sequences for human bZIP factors and C. elegans Skn-1 are provided from the NCBI database (NP_002219, NP_005243, NP_005351, NP_963883, NP_005452, NP_006168.1, NP_002350, NP_002351, NP_001155044, NP_001129495, NP_741404)

Fig. 2

Structures of mouse mafG, mafK, and mafF genomic loci. The first exons are indicated by blue boxes. The other exons are indicated by red boxes. Bent arrows indicate transcription from promoters. The first exons are differentially retained in different splice isoforms. Known cis-regulatory regions, the FXR-binding region, the Nrf2-sMaf-binding region and the hematopoietic and cardiac enhancer for MafK (HCEK) region are indicated by open circles with the names of transcription factors acting through these regions.

Fig. 3

Amino acid sequence alignments of human, mouse and chicken sMafs. SUMO consensus motifs, extended homology regions, basic regions, leucine zipper leucine residues, and phosphorylation sites are indicated by boxes. Amino acid residues not conserved among individual sMafs are underlined in the chicken sequences. MafF is less conserved among species. Tyr64 in red is conserved among basic regions of the small and large Maf families but not in the other bZIP factors. Arg57 and Asn61 in blue are not unique to the Maf family but critical for GC binding.

Fig. 4

Structure of the MafG homodimer binding to DNA. Two different views (A and B) of the crystal structure of the MafG homodimer and DNA are shown (Protein Data Bank (PDB) Code 3A5T). A symmetry axis is indicated as a red pole in the center of MafG homodimers. Oligonucleotide sequences are shown at each location.

Fig. 5

Cis-acting consensus sequences recognized by CNC and sMaf. (A) Alignment of the CRE, C-MARE, TRE, T-MARE, NF-E2, and ARE/EpRE motifs. De novo motifs revealed by ChIP-Seq studies (Nrf2, p45NF-E2, and Bach1) are shown below [modified data from (Hirotsu et al., 2012b; Warnatz et al., 2011; Fujita et al., 2013) are shown]. (B) Binding mode of Skn-1, CNC, Bach, and sMaf proteins. Although Skn-1 binds to DNA as a monomer (left box), CNC and Bach proteins bind to DNA as heterodimers with sMafs (right box). Although CNC-sMaf and Bach-sMaf heterodimers preferentially bind to CsMBE, sMaf homodimers preferentially bind to MARE. Asterisk: the exact binding sequences for Jun-sMaf and Fos-sMaf heterodimers have not yet been determined.

Fig. 6

Genetic analyses of CNC, Bach, and sMaf mutant mice. (A) In the absence of sMaf, CNC and Bach proteins cannot bind to DNA to exert their functions. It is expected that phenotypes observed in CNC and Bach mutant mice are also observed in sMaf mutant mice. (B) Summary of phenotypes observed in sMaf single, double, and triple knockout mutant mice. Note that the phenotype observed in the CNC and Bach mutant mice are mostly observed in sMaf mutant mice depending on the magnitudes of sMaf loss-of-functions. When similar phenotypes are reported in the analyses of CNC or Bach mutant mice, the compositions of CNC-sMaf heterodimers are shown with their illustrations.