The increase of total class III peroxidase activity is due to only a few specific peroxidases
Class III peroxidase activity increased during development and senescence in different tissues of several plant species [18–21]. Class III peroxidases are known to be involved in several events taking place during development and maturation such as chrorophyll degradation, anthocyanin accumulation or lignification [22–24]. In the present study in accordance with the literature [25–27], anthocyanin and lignin content increased whereas chlorophyll content decreased while total peroxidase activity increased. Nevertheless a plethora of other functions have been described and will be certainly further linked in the future to peroxidase activity . However the specific genes involved in the general peroxidase activity increase as well as their specific function have not been identified yet in A. thaliana.
In this study, macroarrays allowed analysing temporal gene expression of the 73 class III peroxidase genes during flower and silique developments in A. thaliana. The expression profiles of peroxidase genes varied during flower and silique developments. Our data indicated a general lower gene expression level in the silique development stages but also an induction of different set of genes during the transition from flower to silique development. We identified a number of genes that are, within flowers, specifically or predominantly expressed in one development stage and are probable components of the gene networks involved in floral organ development. Indeed some of the events taking place in each of the development stages can be linked with peroxidase activity and localisation.
Class III peroxidases can regulate directly or indirectly the cell wall architecture through their catalytic and hydroxylic cycles. In growth stage F1, AtPrx33 and AtPrx43, already identified by a proteomic approach as having a role in cell elongation of hypocotyls , are highly expressed. At stage F2, the stigma is receptive, anthesis take place and fertilization occurs. Previously a stigma-specific class III peroxidase gene, SSP (stigma-specific peroxidase) was identified which is expressed exclusively in the stigmas of an Asteraceae, Senecio squalidus L. [29, 30]. Expression of SSP increases during flower development, to reach a maximum in newly opened flowers when stigmas are most receptive to pollen. So far its function is unknown but it may be related to regulation of H2O2 levels in stigmas. AtPrx58 appears in our study to have its highest level of expression in stage F2. In line with these data it has been identified as a stigma-specific gene and during pollen-pistil interaction by microarrays in two different studies [31, 32]. In stage F3 sepals, petals and stamens wither and fall from the fruit. AtPrx51 that is highly expressed in this stage according to our study, has also been identified during stamen abscission by microarrays . However, the floral abscission zone is also strongly stained in situ for peroxidase activity and suberin (unpublished data). Cross-linking of the phenolic monomers in the formation of suberin has been linked with peroxidase activity . Peroxidases could have a function in defense of the floral abscission zone against biotic attack, either through a cell wall cross-linking activity (formation of lignin, extensin cross-links, dityrosine bonds; ) or by creating a highly toxic environment by producing ROS [34, 35], which results in adverse growth conditions for microorganisms. A good candidate for such function could be AtPrx50 for example. Indeed this gene is highly expressed during F3 stage in our study and has been identified during various studies concerning stresses [36, 37] as well as stamen abscission . During stage S1, the fruit elongates to protect the seeds throughout their development. In addition, the valve margin and en b layer lignified. Growth and lignification are functions classically attributed to class III peroxidases . AtPrx02 has been shown to be involved in lignification, AtPrx34 in root growth, and AtPrx45 in cell elongation [28, 38, 39]. At the end of fruit development at stage S2 the fruit yellows. Chlorophyll breakdown has been linked to peroxidase activity . Nevertheless none of the genes expressed at this stage has been related yet to such a function.
To evaluate our data we compared our results with other published studies. Another study using the same macroarrays has been done earlier on inflorescences . When comparing the two works, there was more convergence between highly expressed genes (8 out of 10) than lowly expressed genes (3 out of 10). A major difference was found for AtPrx16 that is reported as lowly expressed in Valério et al. (2004) study and is amongst the most expressed genes in our study. Other authors performed whole genome microarray analysis on various flower stages [40–43]. Nevertheless, none of these studies has been performed on mature and senescent siliques limiting comparisons with our work. Four of the genes (AtPrx03, AtPrx40, AtPrx42, and AtPrx63) that we found as highly expressed in flower buds F1 were also found as highly expressed in at least one of the other microarray studies. Only one gene (AtPrx07) that we found as expressed at low amounts was also found as lowly expressed in one other study . However, when comparing the 10 lowest and 10 highest expressed genes in flower buds only 2 and 7 genes coincided respectively between the two microarray analysis [40, 43]. This clearly suggested that the highest expression levels are more reliably monitored than the lowest expression levels with these array techniques, including macroarrays and microarrays. Other techniques (e.g. RNA-seq) need to be used to study low-expressed genes. However it has to be noted that class III peroxidases have homologies ranging from 28% to 98% at the nucleotide level . The advantage of our home-made macroarrays is that we used a set of primers exhibiting a maximum of only 70% homology with any other sequence of the A. thaliana genome . Such an approach was of course not possible with whole genome microarrays such as those used by other authors [41–43], since they examined thousands of genes. It can therefore not be excluded that these authors detect an unspecific peroxidase genes expression level due to some level of cross-hybridation . However, differences between our results and those of other scientists can be attributed to biological variation, growth conditions, experimental variations and use of different detection criteria. The use of a dedicated macroarray rather than a commercially available microarray also allowed analyzing data on genes not present on microarrays, such as AtPrx13. However, although these array studies are of great value, they provide information that need to be confirmed.
We therefore continued the study at the protein level. On IEF a modification of the pattern of peroxidase isoforms between different development stages was also observed. In addition, different expression patterns were observed between the different parts of the S1 siliques: the band of pI 8.74 was only present in the ovary. Nevertheless, only six major bands were visible on the IEF, indicating that from the 73 peroxidase genes only the more active are visible and/or also that one band-particularly the thicker ones-might be formed by several peroxidases of close pI. However, different specific isoforms were clearly induced or repressed in plant organs and probably as well in specific tissues and cell types over time. Our data indicated a general lower number of peroxidase isoforms in the silique S2 development stages. This result together with the macroarray analysis confirmed that the increase of total peroxidase activity observed with development in flower and siliques was due to only a few specific genes and not to an increase of the expression of all peroxidase genes. Senescence is known to be characterized by a progressive decrease of total protein content . Probably reduced synthesis and enhanced proteolysis are both responsible for protein loss observed during senescence. In this regard, synthesis of all thylakoid proteins is known to be severely curtailed in senescing bean leaves except for the D1 protein of photosystem 2 . The results presented in our study suggest that in the case of peroxidase a reduced synthesis might be predominant for the vast majority of the genes, therefore further supporting the existence of a functional specialization of peroxidases.
Peroxidases are involved in pod shatter and probably other cell separation processes
Presence of peroxidase activity has been previously reported in stigma, anthers and AZ from different plants [30, 46, 47], but to our knowledge it is the first time that peroxidases activity was localized in the en b and DZ. However, a common feature of several of these areas where peroxidases have been observed is their involvement in some kind of cell separation process including pod shatter, anthers dehiscence, and floral organ abscission [48–50].
To further identify the genes potentially expressed in en b and DZ, we analyzed class III peroxidase gene expression profiles in mutants related to pod-shattering and compared to their expression in WT. AtPrx13 and AtPrx30 were significantly down-regulated, on the contrary AtPrx55 was significantly up-regulated in the three loss-of-function shp mutant lines, suggesting that these three genes are expressed in en b. The three genes were mainly expressed in flower or siliques. In addition, several regulatory sequences identified in the promoter regions of these three genes further supported their role in flower development. The expression of AtPrx13, Atprx30 and AtPrx55 were also monitored in a panel of other loss-of-function mutants related to en b development (ag, ful, rpl, ind and alc). FUL, IND, ALC, SHP1/2 appear to be expressed together in the en b layer, although the nature of the interaction of these antagonistic factors is unclear and their precise role in en b development has remained elusive [16, 17]. However the loss of en b lignification is only seen in the ful ind alc shp1 shp2 quintuple mutant, indicating that all genes are involved in en b lignification. In our model (Figure 9) it can be observed that AtPrx55 expression level is affected in all these mutant lines, indicating a good candidate for lignification of siliques except that the gene was down-regulated by SHP1 and SHP2. Indeed, we would expect the opposite regulation for a gene involved in lignification of en b. However, class III peroxidases can generate highly reactive ROS which can possess an intrinsic activity, or can act as part of signal transduction pathways [51, 52]. Maybe AtPrx55 is one of the peroxidases with such activity. Unfortunately, AtPrx55 is not documented in the literature preventing further hypothesis on its putative function. AtPrx30 on the other hand is up regulated by the SHP transcription factors and has also been identified by microarrays in stamen AZ and in monolignol polymerization suggesting its putative involvement in both silique DZ as well as stamen AZ lignification [33, 38]. All cell separation process involves the differentiation of specialized cell types and a tight co-ordination of molecular and biochemical events . There is accumulating evidences that common mechanisms exist between the different cell separation process [48, 49, 54]. For instance RDPG1, an endo-polygalacturonase involved in cell wall breakdown during silique opening of oilseed rape (Brassica napus), has been found also in dehiscence zones of anthers and floral abscission zones and stylar tissues during pollen tube growth in Arabidopsis and Brassica . Moreover transcription factors (ALC, AG) known to be involved in regulation of pod shattering have also been identified in a microarray study concerning stamen AZ . In addition, several peroxidase genes (AtPrx03, AtPrx17, AtPrx21, AtPrx31, AtPrx33, AtPrx34, AtPrx42, AtPrx45, AtPrx50, AtPrx51, AtPrx52, AtPrx53, AtPrx67, AtPrx71) have been identified in the study concerning stamen AZ, illustrating the redundancy of this protein family and the complexity in assigning a function to a class III peroxidase genes. Six of these genes (AtPrx21, AtPrx31, AtPrx33, AtPrx34, AtPrx53, AtPrx71) were identified in our study on shp loss-of-function mutants, further supporting the existence of common mechanisms in the various cell separation processes. Several of these genes have also been reported in studies concerning responses to abiotic or biotic stresses (AtPrx03, AtPrx21, AtPrx33, AtPrx34, AtPrx45, AtPrx50, AtPrx52, AtPrx67, AtPrx71) or lignin synthesis (AtPrx17, AtPrx53) giving some additional indication on their putative role [7, 35–38, 55–61]. AtPrx13 on the opposite is not documented in the literature and in the microarray databases. However, the precise role of each peroxidase gene in cell separation processes needs to be elucidated.
In A. thaliana, the DZ and the en b are composed by highly specialized cells essentially involved in the pod shatter mechanism. A lignification of the en b layer happens at stage 17 of silique development, and is necessary for a proper shatter mechanism [14, 16]. A well known function of class III peroxidases is lignification . In our study, the observation of peroxidase activity in these specialized tissues further supported a possible involvement through lignification of various peroxidase genes in cell separation processes and particularly in the pod shattering mechanism. In the present study we showed that plants treated with peroxidase inhibitor produced siliques with lower lignin content resulting in a delay of pod shattering. A good spatio-temporal correlation was found between abscission zone weakening and increased peroxidase activity in Phaseolus, cherry fruit, cotton, and Citrus [47, 62–65]. Nevertheless several mechanisms that can be related to peroxidase activity have been observed in the abscission zone. Expression data associated with the leaf abscission in Citrus indicated the occurrence of a double defensive strategy mediated by the activation of a biochemical program including ROS scavenging, defense and PR genes, and a physical response mostly based on lignin/suberin deposition . ROS scavenging, defense and lignin deposition are linked to peroxidases and may be related to the considerable increase in peroxidase activity [47, 66]. Lignin has been proposed to have a double role as a physical barrier of the protective layer developed on the part of the organ that remains attached to the plant and also as component of the fracture line favoring cell separation . On the other hand class III peroxidases can also produce ROS [34, 35], which results in adverse growth conditions for microorganisms. During cell separation processes, peroxidase expression could also be triggered as a preventive defense mechanism against pathogen attacks .