Characterization of a maize G-box binding factor that is induced by hypoxia

G-box cis-acting DNA sequence elements are present in the promoter region of a number of signal-inducible plant genes. In many cases this motif is essential for gene expression. Maize nuclear extracts contain a protein complex that binds specifically to the G-box sequence. Previously, a protein called GF14 was described that is physically associated with the G-box binding complex, but is not a DNA-binding factor in and of itself. This paper reports the isolation of a cDNA encoding a maize G-box binding factor (GBF). The deduced amino acid sequence indicates that maize GBF1 is a basic region-leucine zipper protein. GBF1 binds to the G-box element with specificity similar to that of the binding activity in nuclear extracts. Furthermore, maize GBF1 and the factor detected in nuclear extract are identical in their molecular weight and are immunologically related. GBF1 mRNA accumulates rapidly in hypoxically induced maize cells prior to the increase in Adh1 mRNA levels. Taken together with results that indicate that GBF1 binds to the hypoxia-responsive promoter of maize Adh1, these observations suggest that GBF1 may be one of the factors involved in the activation of Adh1.     

In vivo and in vitro characterization of protein interactions with the dyad G-box of the Arabidopsis Adh gene.

Expression of the alcohol dehydrogenase (Adh) and ribulose-1,5-bisphosphate carboxylase small subunit (RbcS) genes of higher plants is cell-type-specific and environmentally inducible. However, the tissues in which these two genes are expressed, their modes of induction, and their protein functions are quite distinct. Adh is expressed in non-green tissue, induced by anaerobiosis, and repressed in leaves. RbcS is only expressed in green tissue. An 8-base pair G-box element (5'-CCACGTGG-3') is associated with light-induced expression of RbcS and chalcone synthase. The same sequence is also present in the 5'-flanking region of Arabidopsis thaliana Adh, and this sequence is associated with a trans-acting factor in vivo. We report here that in vitro Adh G-box binding activity is present in crude whole cell extracts of both cell culture and leaves of Arabidopsis. The authenticity of in vitro Adh G-box binding is supported by in vivo and in vitro dimethylsulfate footprinting. A clear in vivo Adh G-box footprint occurs in cell cultures, but comparable in vivo binding to the Adh G-box does not occur in leaves. Therefore, there does not appear to be a direct correlation between the presence of the G-box factor in a tissue and its binding to the Adh G-box.     

Heterodimerization between light-regulated and ubiquitously expressed Arabidopsis GBF bZIP proteins.

The promoters of a variety of plant genes are characterized by the presence of a G-box (CCACGTGG) or closely related DNA motifs. These genes often exhibit quite diverse expression characteristics and in many cases the G-box sequence has been demonstrated to be essential for expression. The G-box of the Arabidopsis rbcS-1A gene is bound by a protein, GBF, identified in plant nuclear extracts. Here we report the isolation of three Arabidopsis thaliana cDNA clones encoding GBF proteins referred to as GBF1, GBF2 and GBF3. GBF1 and GBF2 mRNA is present in light and dark grown leaves as well as in roots. In contrast, GBF3 mRNA is found mainly in dark grown leaves and in roots. The deduced amino acid sequences of the three cDNAs indicate that each encodes a basic/leucine zipper protein. In addition, all three proteins are characterized by an N-terminal proline-rich domain. Homodimers of the three proteins specifically recognize the G-box motif, with GBF1 and GBF3 binding symmetrically to this palindromic sequence. In contrast, GBF2 binds to the symmetrical G-box sequence in such a way that the juxtaposition of the protein and the DNA element is clearly asymmetric and hence distinct from that observed for the other two proteins. The fact that GBF1, GBF2 and GBF3 possess both distinct DNA binding properties and expression characteristics prompt us to entertain the notion that these proteins may individually mediate distinct subclasses of expression properties assigned to the G-box. Furthermore, we demonstrate that GBF1, GBF2 and GBF3 heterodimerize and these heterodimers also interact with the G-box, suggesting a potential mechanism for generating additional diversity from these GBF proteins.     

DNA binding activity of the Arabidopsis G-box binding factor GBF1 is stimulated by phosphorylation by casein kinase II from broccoli.

To study the phosphorylation of one of the G-box binding factors from Arabidopsis (GBF1), we have obtained large amounts of this protein by expression in Escherichia coli. Bacterial GBF1 was shown to be phosphorylated very efficiently by nuclear extracts from broccoli. The phosphorylation activity was partially purified by chromatography on heparin-Sepharose and DEAE-cellulose and was characterized. It showed the essential features of casein kinase II activity: utilization of GTP in addition to ATP as a phosphate donor, strong inhibition by heparin, preference for acidic protein substrates, salt-induced binding to phosphocellulose, and salt-dependent deaggregation. The very low Km value for GBF1 (220 nM compared to approximately 10 microM for casein) was in the range observed for identified physiological substrates of casein kinase II. Phosphorylation of GBF1 resulted in stimulation of the G-box binding activity and formation of a slower migrating protein-DNA complex. The conditions of this stimulatory reaction fully corresponded to the properties of casein kinase II, in particular its dependence on the known phosphate donors. The DNA binding activity of the endogenous plant GBF was shown to be reduced by treatment with calf alkaline phosphatase; this reduction was diminished by addition of fluoride and phosphate or incubation in the presence of casein kinase II and ATP.     

Intracellular localization of GBF proteins and blue light-induced import of GBF2 fusion proteins into the nucleus of cultured Arabidopsis and soybean cells

The G-box is an important regulatory element found in the promoters of many different genes. Four members of an Arabidopsis gene family encoding basic leucine zipper proteins (GBFs) which bind the G-box have previously been cloned. To study GBFs, a polyclonal antibody was raised against GBF1 expressed in bacteria. This antibody also recognized GBF2 and GBFS. Immunoblot analysis of nuclear and cytoplasmic fractions from Arabidopsis and soybean (SB-M) cell cultures indicated that over 90% of proteins detected with anti-GBF1 were cytoplasmic. Electrophoretic mobility shift assays indicated that over 90% of G-box binding activity was cytoplasmic. DMA affinity chromatography demonstrated that each protein detected with anti-GBF1 specifically bound the G-box. To study individual GBFs, DNA constructs fusing GBF1, GBF2 and GBF4 to GUS were made and assayed by transient expression in SB-M protoplasts. Of GUS:GBF1 proteins, 50–62% were localized in the cytoplasm under all conditions tested, while 97% of GUS:GBF4 was localized in the nucleus. By contrast, whereas about 50% of GUS:GBF2 was found in the cytoplasm of dark-grown cells, over 80% of this protein was found in the nucleus in cells cultured under blue light. Deletion analysis of GBF1 identified a region between amino acids 112 and 164 apparently required for cytoplasmic retention. These results suggest the intriguing possibility that limitation of nuclear access may be an important control on GBF activity. In particular, GBF2 is apparently specifically imported into the nucleus in response to light.     

TGA1 and G-box binding factors: two distinct classes of Arabidopsis leucine zipper proteins compete for the G-box-like element TGACGTGG.

Regulatory elements containing the sequence ACGT are found in several plant promoters and are recognized by various basic/leucine zipper (bZIP) proteins. The Arabidopsis G-box binding factor 1 (GBF1), initially identified by its ability to bind to the palindromic G-box (CCACGTGG), also interacts with the TGACGT motif if this hexamer sequence is followed by either the dinucleotide GG--as found in the Hex motif of the wheat histone 3 promoter--or GT. Here we describe the isolation of an Arabidopsis bZIP protein, denoted TGA1, that also recognizes ACGT-containing sequences. However, TGA1 differs from members of the GBF family in the spectrum of base pair permutations flanking the ACGT sequence that are required for DNA binding. TGA1 primarily requires a TGACG motif and preferentially binds to those pentamers that are followed by a T residue. We show that although both TGA1 and GBF1 bind to the Hex motif (TGACGTGG), this binding can be distinguished on the basis of their specific DNA-protein contacts. Furthermore, TGA1 also differs from members of the GBF family in that it apparently does not form heterodimers with any member of this family.     

Mutation of either G box or I box sequences profoundly affects expression from the Arabidopsis rbcS-1A promoter.

A deletion analysis of the Arabidopsis thaliana rbcS-1A promoter defined a 196 bp region (-320 to -125) sufficient to confer light-regulated expression on a heterologous Arabidopsis alcohol dehydrogenase (Adh) reporter gene in transgenic Nicotiana tabacum (tobacco) leaves. This region, which contains DNA sequences I, G and GT boxes, with homology to other ribulose-1,5-bisphosphate carboxylase small subunit (RBCS) gene promoter sequences, directed expression independent of orientation and relative position in the Adh promoter. Site-specific mutagenesis of these conserved sequences and subsequent expression analysis in transgenic tobacco showed that both G box and I box mutations in the context of the full (-1700 to +21) rbcS-1A promoter substantially reduced the expression of Adh and beta-glucuronidase (GUS) reporter genes. The G box has previously been shown to specifically bind in vitro a factor isolated from nuclear extracts of tomato and Arabidopsis. This factor (GBF) is distinct from the factor GT-1 which binds to adjacent GT boxes in the pea rbcS-3A promoter. Multiple mutations in putative Arabidopsis rbcS-1A promoter GT boxes had no pronounced affect on expression, possibly due to a redundancy of these sites. Experiments in which rbcS-1A promoter fragments were fused to truncated 35S CaMV (cauliflower mosaic virus) promoter--GUS reporter constructs showed that cis-acting CaMV promoter elements could partially restore expression to G-box-mutated rbcS-1A sequences.     

A cell-specific enhancer of the mouse alpha 1-antitrypsin gene has multiple functional regions and corresponding protein-binding sites.

We have previously described the isolation and characterization of genomic clones corresponding to the mouse alpha 1-antitrypsin gene (Krauter et al., DNA 5:29-36, 1986). In this report, we have analyzed the DNA sequences upstream of the RNA start site that direct hepatoma cell-specific expression of this gene when incorporated into recombinant plasmids. The 160 nucleotides 5' to the cap site direct low-level expression in hepatoma cells, and sequences between -520 and -160 bp upstream of the RNA start site functioned as a cell-specific enhancer of expression both with the alpha 1-antitrypsin promoter and when combined with a functional beta-globin promoter. Within the enhancer region, three binding sites for proteins present in hepatoma nuclear extracts were identified. The location of each site was positioned, using both methylation protection and methylation interference experiments. Each protein-binding site correlated with a functionally important region necessary for full enhancer activity. These experiments demonstrated a complex arrangement of regulatory elements comprising the alpha 1-antitrypsin enhancer. Significant qualitative differences exist between the findings presented here and the cis-acting elements operative in regulating expression of the human alpha 1-antitrypsin gene (Ciliberto et al., Cell 41:531-540, 1985; De Simone et al., EMBO J. 6:2759-2766, 1987).     

Differential interactions of promoter elements in stress responses of the Arabidopsis Adh gene.

The Adh (alcohol dehydrogenase, EC 1.1.1.1.) gene from Arabidopsis thaliana (L.) Heynh. can be induced by dehydration and cold, as well as by hypoxia. A 1-kb promoter fragment (CADH: -964 to +53) is sufficient to confer the stress induction and tissue-specific developmental expression characteristics of the Adh gene to a beta-glucuronidase reporter gene. Deletion mapping of the 5' end and site-specific mutagenesis identified four regions of the promoter essential for expression under the three stress conditions. Some sequence elements are important for response to all three stress treatments, whereas others are stress specific. The most critical region essential for expression of the Arabidopsis Adh promoter under all three environmental stresses (region IV: -172 to -141) contains sequences homologous to the GT motif (-160 to -152) and the GC motif (-147 to -144) of the maize Adh1 anaerobic responsive element. Region III (-235 to -172) contains two regions shown by R.J. Ferl and B.H. Laughner ([1989] Plant Mol Biol 12: 357-366) to bind regulatory proteins; mutation of the G-box-1 region (5'-CCACGTGG-3', -216 to -209) does not affect expression under uninduced or hypoxic conditions, but significantly reduces induction by cold stress and, to a lesser extent, by dehydration stress. Mutation of the other G-box-like sequence (G-box-2: 5'-CCAAGTGG-3', -193 to -182) does not change hypoxic response and affects cold and dehydration stress only slightly. G-box-2 mutations also promote high levels of expression under uninduced conditions. Deletion of region I (-964 to -510) results in increased expression under uninduced and all stress conditions, suggesting that this region contains a repressor binding site. Region II (-510 to -384) contains a positive regulatory element and is necessary for high expression levels under all treatments.