Selective activation of the developmentally regulated Ha hsp17.6 G1 promoter by heat stress transcription factors

Abstract

Using two well-characterized heat stress transcription factors (Hsfs) from tomato (Lycopersicon peruvianum; LpHsfA1 and LpHsfA2), we analyzed the transcriptional activation of the Ha hsp17.6 G1 promoter in sunflower (Helianthus annuus) embryos. In this system, we observed transient promoter activation only with LpHsfA2. In contrast, both factors were able to activate mutant versions of the promoter with improved consensus Hsf-binding sites. Exclusive activation by LpHsfA2 was also observed in yeast (Saccharomyces cerevisiae) without other Hsfs and with a minimal Cyc1 promoter fused to the Ha hsp17.6 G1 heat stress cis-element. Furthermore, the same promoter mutations reproduced the loss of activation selectivity, as observed in sunflower embryos. The results of in vitro binding experiments rule out differential DNA binding of the two factors as the explanation for the observed differential activation capacity. We conclude that the specific sequence of this heat stress cis-element is crucial for Hsf promoter selectivity, and that this selectivity could involve preferential transcriptional activation following DNA binding. In sunflower embryos, we also observed synergistic transcriptional activation by co-expression of LpHsfA1 and LpHsfA2. Mutational analyses of the Ha hsp17.6 G1 promoter, combined with in vitro binding assays, suggest that mixed oligomers of the two factors may be involved in promoter activation. We discuss the relevance of our observations for mechanisms of developmental regulation of plant heat stress protein genes.

Figure 1

Chimeric genes with the WT and mutant Ha hsp17.6 G1 HSE sequences. Top, Chimeric gene promoter context for the β-glucuronidase (GUS; −533::GUS, Carranco et al., 1999) and LacZ fusions used in this work. The arrows indicate the Ha hsp17.6 G1 and the Cyc1 promoters. The sequences of the Ha hsp17.6 G1 HSE complex are indicated with a black box. The corresponding, unmodified nucleotide sequences for this HSE (WT) are shown (from −110 to −81 from the transcription initiation site of Ha hsp17.6 G1). The core sequences for the HSE pentanucleotide repeats are marked in bold face with crucial DNA contact positions (G or C) underlined. These core repeats are numbered, below the sequence, from 1 to 5. Nucleotide substitutions in the different mutant HSEs are indicated with lowercase letters. The different mutant sequences were substituted for the WT HSE in each chimeric gene promoter context as described in “Materials and Methods.”

Figure 2

Transcriptional activation in sunflower embryos. Plant material was particle bombarded with a luciferase reference plasmid, the indicated chimeric GUS reporter gene (WT, Mut0, or MutP), and the Hsf effector plasmid combination shown in the top left corner: (−), no effector plasmid; (+), effector plasmids bombarded simultaneously. We show average reporter GUS activities normalized with luciferase (GUS/LUC, as in Prieto-Dapena et al., 1999), with bars indicating the ses. Data correspond to at least five independent experiments, with each plasmid combination repeated at least 25 times. Nonsignificant differences in reporter gene activities are indicated (P ≥ 0.05). Values for the significant statistical differences mentioned in the text were as follows: for the WT gene with effector plasmid(s) versus without effector plasmid(s), with LpHsfA2, F = 139.6, P = 0.0001; and with LpHsfA1 + LpHsfA2, F = 429.1, P = 0.0001. The induced level obtained with LpHsfA1 + LpHsfA2 did not differ from that of with LpHsfA1 M3 + LpHsfA2, F = 0.25, P = 0.61. In addition, the values obtained with LpHsfA1 + LpHsfA2 were significantly higher than the addition of reporter activities separately obtained with each Hsf (F = 27.35, P = 0.001). For the mutP gene with effector plasmid(s) versus without effector plasmid(s): with LpHsfA2, F = 41.4, P = 0.0001; and with LpHsfA1, F = 9.3, P = 0.021.

Figure 3

Transcriptional activation in yeast. Reporter β-galactosidase activity measured in yeast strains transformed with plasmids containing the indicated HSE (WT, Mut0, or MutP) is shown. These RSY4 strains (Boscheinen et al., 1997) contained one (LpHsfA1 or LpHsfA2) or two (LpHsfA1 + LpHsfA2) tomato Hsfs. The values for two independent transformants per plasmid combination were averaged, and the ses are represented with bars. Bottom, The expression level of the Hsfs determined by western analysis of protein extracts from each yeast strain using the immune sera indicated to the right. Left, Data for the single Hsf strains with each reporter plasmid (LpHFSA1 above, LpHsfA2 below); right, data for the double Hsf strain (with the WT or MutP plasmid); in this case, LpHsfA1 and LpHsfA2 were sequentially detected using the same protein extracts.

Figure 4

In vitro binding experiments. Two labeled DNA fragments that contained synthetic (oHSE) or natural HSE sequences from Ha hsp17.6 G1 (WT) were subjected to mobility shift assays in agarose gels. Binding reactions contained the probe indicated on top (0.13 ng of labeled DNA) and the components were summarized at the bottom, including unlabeled oligonucleotide fragments used as specific or unspecific competitors (oHSE and oHSEm, respectively; see “Material and Methods”), as well as different amounts of protein extracts from yeast strains containing LpHsfA1 or LpHsfA2. Optimal and limiting amounts of total protein used for each extract were experimentally determined (see “Results”): LpHsfA1, limiting = 10 μg (+), and optimal = 20 μg (+). In the case of LpHsfA2, 40 μg (+) was used as the optimal amount of total protein. Higher amounts of this extract were used in some reactions (+*, 70 μg). Ticks to the left mark the position of the LpHsfA2 homo-oligomeric complexes. The arrows to the right mark the position of the different mixed complexes observed with increasing amounts of the LpHsfA2 extract (lane 8, 40 μg of protein, and lane 10, 70 μg of protein). Bottom, Western immunodetection of LpHsfA1 and LpHsfA2 using 25 μg of total protein from the respective extracts used in the binding reactions.

Figure 5

Effect of the HSE mutations on synergistic activation of the Ha hsp17.6 G1 promoter by LpHsfA2 and LpHsfA1. Bars represent average GUS activity determined in tobacco protoplasts transformed with different combinations of reporter and effector plasmids. Results correspond to three independent experiments for each plasmid combination, with GUS activity measured in triplicate in each experimental repetition. The chimeric genes, indicated by WT, Mut1, Mut2, Mut3, and Mut5, correspond to the natural and mutant HSE denominations of the GUS fusions in Figure 1. Each reporter gene was transformed without (−) and with the effector plasmids for the Hsfs indicated in the upper right corner. Induction is the ratio between activities obtained with and without effector plasmid(s). Statistical values for relevant comparisons of observed induced activities: WT with LpHsfA2 versus no effector plasmid (F = 119.6, P = 0.001). WT with LpHsfA1 + LpHsfA2 versus the addition of the activities separately obtained with each Hsf (F = 19.6, P = 0.0001). WT versus Mut1 with LpHsfA1 + LpHsfA2 (F = 0.047, P = 0.83). WT versus Mut2 (F = 33.7, P = 0.0001), Mut3 (F = 58.7, P = 0.0001), or Mut5 (F = 20.6, P = 0.0001), in all cases with LpHsfA1 + LpHsfA2. Statistical significance as in the legend of Figure 2. Panels below depict the verification, by western immunodetection, of expression levels for LpHsfA1 and LpHsfA2 in each experimental combination. For combinations simultaneously expressing LpHsfA1 and LpHsfA2, the same protein samples were used for sequential detection of both Hsfs (marked with asterisks).