Evolution of the C4 phosphoenolpyruvate carboxylase promoter of the C4 species Flaveria trinervia: the role of the proximal promoter region

Background The key enzymes of photosynthetic carbon assimilation in C4 plants have evolved independently several times from C3 isoforms that were present in the C3 ancestral species. The C4 isoform of phosphoenolpyruvate carboxylase (PEPC), the primary CO2-fixing enzyme of the C4 cycle, is specifically expressed at high levels in mesophyll cells of the leaves of C4 species. We are interested in understanding the molecular changes that are responsible for the evolution of this C4-characteristic PEPC expression pattern, and we are using the genus Flaveria (Asteraceae) as a model system. It is known that cis-regulatory sequences for mesophyll-specific expression of the ppcA1 gene of F. trinervia (C4) are located within a distal promoter region (DR). Results In this study we focus on the proximal region (PR) of the ppcA1 promoter of F. trinervia and present an analysis of its function in establishing a C4-specific expression pattern. We demonstrate that the PR harbours cis-regulatory determinants which account for high levels of PEPC expression in the leaf. Our results further suggest that an intron in the 5' untranslated leader region of the PR is not essential for the control of ppcA1 gene expression. Conclusion The allocation of cis-regulatory elements for enhanced expression levels to the proximal region of the ppcA1 promoter provides further insight into the regulation of PEPC expression in C4 leaves.


Background
About 90% of terrestrial plant species, including major crops such as rice, soybean, barley and wheat, assimilate CO 2 via the C 3 pathway of photosynthesis. Ribulose-1,5bisphosphate carboxylase/oxygenase (Rubisco) acts as the primary CO 2 -fixing enzyme of C 3 photosynthesis, but its ability to use O 2 as a substrate instead of CO 2 results in the energy-wasting process of photorespiration. The photosynthetic C 4 cycle represents an addition to the C 3 pathway which acts as a pump that accumulates CO 2 at the site of Rubisco so that the oxygenase activity of the enzyme is inhibited and photorespiration is largely suppressed. C 4 plants therefore achieve higher photosynthetic capacities and better water-and nitrogen-use efficiencies when compared with C 3 species [1]. C 4 photosynthesis is characterized by the coordinated division of labour between two morphologically distinct cell types, the mesophyll and the bundle-sheath cells. The correct functioning of the C 4 cycle depends upon the strict compartmentalization of the CO 2 assimilatory enzymes into either mesophyll or bundle-sheath cells [2]. Phosphoenolpyruvate carboxylase (PEPC), which serves as the actual CO 2 pump of the C 4 pathway, is specifically expressed in the mesophyll cells of C 4 leaves. This enzyme is not an unique feature of C 4 species; other PEPC isoforms with different catalytic and regulatory properties are found in both photosynthetic and non-photosynthetic tissues of all plants where they participate in a variety of metabolic processes, e.g. replenishment of citric acid cycle intermediates and regulation of guard cell movement [3].
The polyphyletic origin of C 4 photosynthesis suggests that the photosynthetic C 4 isoforms of PEPC have evolved independently several times from non-photosynthetic C 3 isozymes [4]. During the evolution of C 4 PEPC genes from ancestral C 3 genes, changes in expression strength and organ-and cell-specific expression patterns must have occurred. While C 4 PEPC genes are highly expressed in the mesophyll cells of the leaf, the C 3 isoform genes are only moderately transcribed in all plant organs [5][6][7][8].
To investigate the molecular evolution of a C 4 PEPC gene we are using the genus Flaveria (Asteraceae) as a model system. This genus includes C 4 and C 3 as well as C 3 -C 4 intermediate species [9,10] and thus provides an excellent system for studying the evolution of the C 4 photosynthetic pathway [11]. Previous studies on the ppcA1 gene of F. trinervia, encoding the C 4 isoform of PEPC, revealed that the strong mesophyll-specific expression is largely regulated at the transcriptional level and that the available 2188 bp (with reference to the AUG start codon of the ppcA1 reading frame) of the 5' flanking sequences contain all the essential cis-regulatory elements for high and mesophyll-specific expression [12]. Two parts of the ppcA1 promoter of F. trinervia, a proximal region (PR) up to -570 in combination with a distal region (DR) from -1566 to -2141, are sufficient to direct a high mesophyll-specific expression of a β-glucuronidase (GUS) reporter gene in transgenic F. bidentis (C 4 ) plants [13]. The orthologous, 2538 bp comprising ppcA1 promoter of the C 3 species F. pringlei displays only weak activity in all interior leaf tissues in transgenic F. bidentis, but fusion of the C 4 -DR to this C 3 PEPC promoter leads to a confinement of GUS expression to the mesophyll [13]. Analysis of the C 4 -DR revealed that the 41-bp module MEM1 (mesophyll expression module 1) is responsible for the C 4 -character-istic spatial expression pattern of the ppcA1 gene of F. trinervia. Furthermore, it was shown that a high level of expression in the mesophyll requires an interaction of the C 4 -DR with the C 4 -PR. This suggests that quantity elements for an elevated expression of the C 4 PEPC gene are located within the PR of the 5' flanking sequences [13].
Using the yeast one-hybrid system, Windhövel and colleagues [14,15] identified four different proteins which bind to the PR of the ppcA1 promoter of F. trinervia, but not to the corresponding part of the ppcA1 promoter of F. pringlei. These proteins (named FtHB1 to FtHB4) belong to the class of zinc finger homeodomain proteins (ZF-HD). Two regions of the C 4 -PR specifically interact with the FtHB proteins in vitro: an intron sequence within the 5' untranslated leader region and a DNA fragment that is located upstream of the putative TATA-box. To the latter one, the FtHB proteins showed a much lower binding affinity [14]. Homeobox proteins are known to act as transcriptional regulators of eukaryotic gene expression [16][17][18], and the fact that the FtHB homeobox proteins interact specifically with the PR of the ppcA1 promoter of F. trinervia makes them prime candidates for transcription factors that are involved in the establishment of the C 4characteristic expression pattern of the C 4 ppcA1 gene.
In this study we have investigated the role of the proximal promoter region of the ppcA1 gene of F. trinvervia with regard to its high and mesophyll-specific expression by transgenic analyses in the closely related C 4 species F. bidentis. We demonstrate that the proximal promoter region of the ppcA1 gene contains cis-regulatory elements that determine promoter strength. Furthermore, we show that the deletion of an intron located in the 5' untranslated segment of ppcA1 does not alter promoter activity in transgenic F. bidentis.

Experimental strategy
We are interested in elucidating the molecular events that are crucial for the evolution of the high and mesophyllspecific expression of the C 4 phosphoenolpyruvate carboxylase gene (ppcA1) of the C 4 plant F. trinervia. In this study we focus on the proximal promoter region (PR) of the ppcA1 gene with respect to its function in establishing the C 4 -characteristic expression pattern. We performed a comparative analysis of three different promoter-GUS fusion constructs (Fig. 1) in transgenic F. bidentis plants. F. bidentis is a close relative to F. trinervia, but in contrast to F. trinervia this C 4 species is transformable by Agrobacterium tumefaciens mediated gene transfer [19] and was therefore chosen for these experiments.
Construct ppcA-PR Ft -DR(+) Ft served as a reference because it was already known from previous experiments that a combination of the distal (DR) and the proximal (PR) promoter regions was sufficient to direct a high and mesophyll specific expression of a GUS reporter gene in F. bidentis [13]. To find out if the PR of the C 4 ppcA1 promoter contains quantity elements conferring high expression in the mesophyll cells we designed construct ppcA-PR Fp -DR(+) Ft . Here, the C 4 -PR was exchanged for its counterpart from the orthologous ppcA1 gene of the C 3 species F. pringlei. Deletion of the intron sequences in the 5' untranslated segment of promoter construct ppcA-PR Ft -DR(+) Ft resulted in the formation of construct ppcA-PR Ft -∆Intron-DR(+) Ft . Thereby a putative binding site for the ZF-HD proteins FtHB1 to FtHB4 [14] was removed from the C 4 -PR. Hence, this chimeric promoter-GUS fusion could answer the question whether the intron-located putative binding site of the FtHB proteins is necessary for the establishment of the C 4 -specific ppcA1 expression pattern.

The proximal region of the ppcA1 promoter of F. trinervia harbours cis-regulatory elements for a high level of PEPC expression in the mesophyll
Gowik et al. [13] assumed that the PR of the ppcA1 promoter of F. trinervia comprises cis-regulatory determinants conferring high levels of expression in mesophyll cells of C 4 leaves. To examine whether the PR actually harbours such quantity elements we analyzed the GUS expression patterns of constructs ppcA-PR Ft -DR(+) Ft and ppcA-PR Fp -DR(+) Ft ( Fig. 1) in transgenic F. bidentis.
In F. bidentis plants that had been transformed with promoter construct ppcA-PR Ft-DR(+) Ft , GUS expression was exclusively detected in the mesophyll cells of the leaves ( Fig. 2A). This observation shows that the DR and PR of the ppcA1 promoter together are sufficient for a high and mesophyll-specific expression of the linked GUS reporter gene and therefore confirms the results obtained by Gowik et al. [13]. Replacement of the C 4 -PR by the corresponding region from the ppcA1 promoter of F. pringlei (construct ppcA-PR Fp -DR(+) Ft ) did not cause any alteration in the cellular GUS expression pattern when compared to ppcA-PR Ft -DR(+) Ft ; GUS activity was still restricted to the mesophyll compartment (Fig. 2B). However, both chimeric promoters differed greatly in transcriptional strength. Quantitative GUS assays revealed that promoter activity was decreased by a factor of 15 when the C 4 -PR was substituted for the C 3 -PR (Fig. 2D). This clearly demonstrated that the C 4 -characteristic transcription-enhancing cis-regulatory elements must be located within the proximal region of the ppcA1 promoter of F. trinervia. The low expression level of construct ppcA-PR Fp -DR(+) Ft could be the result of an absence of transcription-enhancing cis-regulatory elements in the C 3 -PR, but it might also be caused by problems in the interaction of the C 4 -DR and the C 3 -PR.

The intron in the C 4 -PR is not required for the establishment of a C 4 -specific expression pattern of the ppcA1 gene of F. trinervia
The 5' untranslated region of the ppcA1 gene of F. trinervia contains an intron between positions -209 and -40 (+1 refers to the starting point of translation). Introns are of prominent importance for the molecular evolution of eukaryotic genomes by facilitating the generation of new genes via exon-shuffling and by providing the possibility to create multiple proteins from a single gene via alternative splicing [20][21][22]. Furthermore, it has been shown that introns can affect many different stages of gene expression, including both transcriptional and post-transcriptional mechanisms [22][23][24].
Here, we wanted to investigate whether the first intron of the ppcA1 gene of F. trinervia is essential for establishing the C 4 -characteristic expression pattern. We therefore Schematic presentation of the promoter-GUS fusion constructs used for the transformation of Flaveria bidentis (C 4 ) Figure 1 Schematic presentation of the promoter-GUS fusion constructs used for the transformation of Flaveria bidentis (C 4 ).
deleted the intron sequences from the C 4 -PR in construct ppcA-PR Fp -DR(+) Ft , resulting in the formation of construct ppcA-PR Ft ∆Intron-DR(+) Ft (Fig. 1). The histochemical analysis of transgenic F. bidentis plants demonstrated that the ppcA-PR Ft ∆Intron-DR(+) Ft promoter was exclusively active in the mesophyll cells of the leaves (Fig. 2C). The quantitative examination of GUS activity (Fig. 2D) also revealed no significant differences between ppcA-PR Ft ∆Intron-DR(+) Ft (6,5 nmol MU/(mg*min)) and ppcA-PR Ft -DR(+) Ft (5,9 nmol MU/(mg*min)). These data suggest that the 5' located intron of ppcA1 does not contain any cis-regulatory elements that are essential for achieving high mesophyll-specific expression of a reporter gene. Accordingly, the specific binding of the FtHB proteins to this intron that was observed in vitro and in yeast onehybrid experiments [14,15] has no in planta relevance concerning the regulation of ppcA1 expression in C 4 leaves. However, our results do not necessarily indicate that the intron is completely dispensable for the regulation of ppcA1 gene expression. It is known that C 4 gene transcription is modulated by various metabolites such as sugar hexoses [25][26][27], and we cannot exclude that the first intron of the ppcA1 gene of F. trinervia might be involved in the metabolic control of gene expression.

Comparison of proximal ppcA promoter sequences from different Flaveria species
As reported above, cis-regulatory elements for leaf-specific enhanced transcription of the ppcA1 gene of F. trinervia could be allocated to the PR of the 5' flanking sequences, but their exact nature and localization was still unclear. To identify potential cis-regulatory enhancing elements, a sequence comparison between the PR of the ppcA1 gene of F. trinervia and equivalent promoter sequences from other Flaveria species was performed (Fig. 3). This approach was chosen because it was already known from northern analyses of ppcA transcript levels in different Flaveria species that ppcA RNA amounts in leaves increase gradually from C 3 to C 4 species [28]. This is consistent with the important function of PEPC during C 4 photosynthesis. The C 4 -like species F. brownii and F. vaginata exhibited ppcA RNA levels that were comparable to those of the C 4 plants F. bidentis and F. trinervia, and even in F. pubescens, a C 3 -C 4 intermediate with rather poorly developed C 4 -characteristic traits, ppcA transcript accumulation in the leaves was significantly higher than in the C 3 species F. cronquistii and F. pringlei [28].
Searching for known plant cis-regulatory DNA elements in the PLACE database [29] resulted in the identification of two distinct sequence motifs which might be involved in the regulation of ppcA expression levels (Fig. 3). Both of them, a putative MYB transcription factor binding site (GTTAGTT, [30]) and a CCAAT box [31], are present in all examined C 3 -C 4 , C 4 -like and C 4 species, but are missing in the two C 3 species (Fig. 3). Thus, these sequences are prime candidates for transcription-enhancing cis-regulatory elements. CCAAT boxes are common sequences that are found in the 5' untranslated regions of many eukaryotic genes [32]. They are able to regulate the initiation of transcription by an interaction of CCAAT-binding transcription factors with the basal transcription initiation complex [33]. There is no unifying expression pattern for plant genes containing putative CCAAT promoter elements, indicating that they may play a complex role in regulating plant gene transcription [32]. MYB proteins, on the other hand, comprise one of the largest families of transcription factors in plants, with almost 200 different MYB genes present in the Arabidopsis genome [34][35][36]. To test the physiological importance of the putative MYB and CCAAT binding sites (that are located within the PR of the ppcA1 promoter of F. trinervia) it will be crucial to inactivate these sequences in construct ppcA-PR Ft ∆Intron-DR(+) Ft by site-directed mutagenesis and to investigate whether this results in a decrease of reporter gene expression in the leaves of transgenic F. bidentis plants.
When searching for quantity elements in the PR of the ppcA1 promoter of F. trinervia, one should always keep in mind that high levels of reporter gene expression in the leaf mesophyll require the synergistic action of the distal and proximal promoter regions. The C 4 -PR alone exhibits very low transcriptional activity in all interior leaf cell types of transgenic F. bidentis [37], indicating that the cisregulatory elements for enhanced expression are only functional when the C 4 -PR is combined with the cognate C 4 -DR. One may speculate that a strong expression of the ppcA1 gene in the mesophyll cells of F. trinervia depends on the interaction of trans-acting factors which bind to cisregulatory elements within the PR with other transcription factors that are recruited to C 4 -specific cis-regulatory determinants in the DR. In the future, further dissection of the C 4 -PR of F. trinervia and expression analyses of additional DR-PR combinations from ppcA promoters of different Flaveria species in transgenic F. bidentis will be useful for uncovering the control of ppcA expression levels in C 4 leaves.

Conclusion
In this study, we have demonstrated that the proximal region (-570 to -1) of the ppcA1 promoter of F. trinervia (C 4 ) harbours cis-regulatory elements conferring high expression levels in leaf mesophyll cells of transgenic F. bidentis (C 4 ). It was further demonstrated that the deletion of an intron in the 5' untranslated leader region does not affect the C 4 -specific ppcA1 expression pattern and strength, indicating that the previously isolated zinc finger-homeobox transcription factors that specifically interact with this intron in vitro are not involved in regulating ppcA1 expression levels. Sequence comparisons resulted in the identification of potential cis-regulatory elements in the proximal part of the ppcA1 promoter that might play a role in controlling ppcA1 expression quantity. Genetic manipulation of these sequences and subsequent analyses in transgenic F. bidentis will clarify whether they are able to direct high ppcA1 expression levels in C 4 leaves.

Construction of chimeric promoters
DNA manipulations and cloning were performed according to Sambrook and Russell [38]. The construction of the promoter-GUS fusion ppcA-PR Ft -DR(+) Ft has been described in detail [13]. Plasmids ppcA-S-Fp [39] and ppcA-PR Ft -DR(+) Ft served as the basis for the production of ppcA-PR Fp -DR(+) Ft . The distal region (-2141 to -1566) of the ppcA1 promoter of F. trinervia was excised from ppcA-PR Ft -DR(+) Ft by digestion with XbaI. Insertion of this promoter fragment into XbaI-cut ppcA-S-Fp resulted in the generation of construct ppcA-PR Fp -DR(+) Ft .

Plant transformation
In all transformation experiments the Agrobacterium tumefaciens strain AGL1 was used [40]. The promoter-GUS constructs were introduced into AGL1 by electroporation. The transformation of Flaveria bidentis was performed as described by Chitty et al. [19]. The integration of the transgenes into the genome of regenerated F. bidentis plants was proved by PCR analyses.

Measurement of GUS activity and histochemical analysis
F. bidentis plants used for GUS analysis were 40 to 50 cm tall and before flower initiation. Fluorometrical quantification of GUS activity in the leaves was performed according to Jefferson et al. [41] and Kosugi et al. [42]. For histochemical analysis of GUS activity the leaves were cut manually with a razorblade and the sections were transferred to incubation buffer (100 mM Na 2 HPO 4 , pH 7.5, 10 mM EDTA, 50 mM K 4 [Fe(CN) 6 ], 50 mM K 3 [Fe(CN) 6 ], 0.1% (v/v) Triton X-100, 2 mM 5-bromo-4-chloro-3indolyl-β-D-glucuronid acid). After brief vacuum infiltration the sections were incubated at 37°C for 6 to 20 hrs. After incubation chlorophyll was removed from the tissue by treatment with 70% ethanol.