Discussion

In recent years sequencing technologies made possible the accessibility to more complete genomes and more genetic and genomic resources. This allowed the discovery of particular gene arrangements and promoters in the genome that can be used in the study of plants and crop improvement, to drive specific transgenes. In the genome sequence, different genes arrangements can be found, two neighboring genes can be distributed in a divergent, convergent or parallel configuration [Yang, 2011a]. In the divergent or bidirectional distribution, genes are located in opposite strands of the genome and with head-to-head orientations. The pair of genes arranged in this way is called bidirectional genes, while the intergenic region between them is defined as a bidirectional promoter. This divergent structure has been demonstrated to be quite common in the A. thaliana genome, about 13.3 % of the genes are arranged in this distribution [Wang, 2009]. The most probable function of these arrangements is the co-regulation of two adjacent genes by only one promoter. In recent years, bidirectional promoters have received considerable attention because they can initiate transcription in two directions and thus express simultaneously two downstream genes [Wei, 2011; Banerjee, 2013; Wang, 2016].

In this work, the functional activity of four putative bidirectional promoters from Arabidopsis was critically analyzed. The study was not only centered on the characteristics of the intergenic regions but also on the neighbor divergent genes. An initial search for genes coding for ACD proteins which have a divergent orientation with other genes in the A. thaliana genome was performed and the pairs At4g25650-At4g25660, At5g51430-At5g51440, At2g35490-At2g35500, At1g06460-At1g06470 were selected. In previews genome-wide studies, it has been proposed that bidirectional genes are usually correlated in expression and are involved in similar functions [Wang, 2009]. However, a broad range of functions and expression patterns were observed in the genes considered in this work. The genes in the first pair (At4g25650-At4g25660) include a protein that may be implicated in translocation processes in chloroplasts (ACD1-like) and a peptidase located in the cytoplasm (PPPDE). The genes At5g51430 and At5g51440 are both located in organelles (Golgi complex and mitochondria) while the pair At2g35490-At2g35500 codes for chloroplastic proteins, a fibrillin protein and a kinase. Finally, the genes of the fourth pair At1g06460-At1g06470 code for an ACD protein present in peroxisomes and for a transporter protein located in the tonoplast of cells. The expression of the genes in the four pairs was demonstrated by q-PCR to be higher in one direction.

Transcripts of At4g256660, At5g51430, At2g35490, and At1g06460 strongly accumulated compared to the level of the correspondent adjacent gene in Col-0 plants. In some pairs, genes displayed similar or opposite responses to the tested abiotic stresses what indicates a common but complex regulation of the genes expression. In Col-0 plants, At4g25650 had reduced expression in salinity, osmotic and heat conditions while the adjacent At4g25660 gene showed

exactly the opposite response to the same treatments. On the contrary, the expression of the genes in the pair At2g35490 and At2g35500 exhibited the same pattern in low and high temperature, they both accumulated in these conditions. Exclusive expression pattern was also observed for At1g06460 and At1g06470. These genes showed the same expression response to salinity and the opposite to heat stress conditions. To better understand the complexity in the regulation of bi-directional genes, the intergenic regions between them were in silico and in vivo analyzed. The in vivo activity of the four putative BDP was evaluated by successfully cloning these regions in opposite orientation to the β-glucuronidase (GUS) reporter gene and by exposing the transgenic Arabidopsis plants to various stress situations.

The four intergenic regions displayed GUS activity in both directions in all the tested lines proving that they are all indeed bidirectional promoters. That is the same genomic sequence coordinates and initiates the expression of the two divergent genes. Promoter activities were, nevertheless especially different in the positive and in the negative directions, as well as in the tested stresses. In only one BDP, located between At1g06460 and At1g06470, the GUS staining was similar in the 5 independent lines of both contructs and thus the activity could be considered to be divergently the same. In addition to this, the promoter activities tested by GUS staining did not always reflect what was found in the gene expression experiments by q-PCR in Col-0 plants (Table 4.10). The activity of the BDP in At4g25650-At4g25660 pair was stronger in the PPPDE direction, as it could be seen from GUS assays and q-PCR experiments. Same patterns were also found for this promoter under abiotic stress conditions.

On the contrary, the higher expression of At5g51430 compared to At5g51440 obtained by q-PCR could not be reproduced when the activity of this BDP fused to the GUS gene was tested, GUS staining was considerably stronger for the sHSP23.5 strand in all the 5 isolated lines.

Both, gene expression of At5g51440 and GUS assays in this direction accumulated under high temperature. The same discordance between q-PCR and GUS experiments was observed for the third BDP, between At2g35490 and At2g35500 genes. The results of genes expression and promoters activities were not always comparable in the tested conditions. The behavior of the BDP::GUS in At1g06460-At1g06470 pair showed no notable changes in response to the abiotic stress treatments, making it difficult the comparison to the q-PCR data. These particular expression patterns found in the genes and promoters suggest that more intricated mechanisms are involved in the bidirectional transcriptional regulation, and further experiments will be necessary to a better understanding of this complexity.

DNase I hypersensitive sites (DHSs) were found in all the four BDPs indicating that this sites of open chromatin may be cis-regulatory enriched and transcriptionally active regions. To identify putative cis-elements involved in the co-regulatory function of the BDPs, the four intergenic regions were investigated using the PlantCare and PLACE databases. This analysis revealed the presence of CAAT and TATA boxes in both strands of the four BDP which have

an important role in the transcription initiation. Furthermore, numerous stress-related elements were located in the identified BDPs. The low responsive LTREATLTI78 and LTRECOREATCOR15 elements were predicted in the minus strand of BDP At4g25650-At4g25660. In this strand, the promoter GUS activity was stronger under cold treatments. Also, in this strand, one cis- salt related element was identified (GT1GMSCAM4), which could be involved in the salt stress response of At4g25650 (a reduction in the expression was observed after 3 h of 150 mM ClNa). Special interest has the BDP of At5g51430 and At5g51440 since it regulates the expression of a mitochondrial small heat shock protein. As it was formerly postulated, the expression of sHSPs is transcriptionally regulated by highly conserved cis-elements named Heat Shock Element (HSE) [Sun, 2002]. In this BDP, 8 HSE sites were localized in the positive strand, in which the GUS staining and q-PCR experiments showed strong heat response. The intergenic region between At2g35490 and At2g35500 has the LTR cis-element and the LTRE1HVBLT49 binding site in the minus direction. These regulatory motifs may be responsible for the induction of At2g35490 in response to low temperatures. In the opposite strand, heat and salt stress related motifs were allocated to this BDP and could explain the response of At2g35500 to the mentioned stresses. Asides from salt-related cis-elements, the sequence of the BDP located between At1g06460 and At1g06470 includes low and high temperatures responsive motifs probably controlling the expression of the divergent genes in this abiotic stress situations. It is worth mentioning the identification of several light regulated elements in the 4 BDP that were analyzed. Additional studies need to be performed in plants exposed to high or low luminosity in order to gain insight into the role of these cis-regulatory elements.

The conservation of a bi-directional arrangement in other species suggests that this kind of structure may be beneficial for plants and indicates a possible co-evolution of bi-directional genes and BDP.

Here, 4 different bidirectional promoters in the A. thaliana genome were successfully identified and characterized. Among them, the promoter of At1g06460-At1g06470 showed comparable high activity in both directions what represents a useful tool to be used in genetic engineering.

The rest of the promoters showed greater strength in one side and can be considered as asymmetric bidirectional promoters. The data suggest that the BDP in the pair At5g51430-At5g51440 is strongly heat induced in one direction but not considerably stress induced in the other. These promoters are extremely useful because they can drive the simultaneous expression of two genes and at the same time, they have the potential to induce specifically one of them in a certain condition (high temperature) when it is required.

CHAPTER II

4.2 Functional characterization of mitochondrial small heat