•Owing to its evolutional plasticity and adaptability, barley (Hordeum vulgare) is one of the most widespread crops in the world. Despite this evolutionary success, sexual reproduction of small grain cereals is poorly investigated, making discovery of novel genes and functions a challenging priority. Barley gene Jekyll appears to be a key player in grain development; however, its role in flowers has remained unknown.
•Here, we studied RNAi lines of barley, where Jekyll expression was repressed to different extents. The impact of Jekyll on flower development was evaluated based on differential gene expression analysis applied to anthers and gynoecia of wildtype and transgenic plants, as well as using isotope labeling experiments, hormone analysis, immunogold- and TUNEL-assays and in situ hybridization.
•Jekyll is expressed in nurse tissues mediating gametophyte–sporophyte interaction in anthers and gynoecia, where JEKYLL was found within the intracellular membranes. The repression of Jekyll impaired pollen maturation, anther dehiscence and induced a significant loss of fertility. The presence of JEKYLL on the pollen surface also hints at possible involvement in the fertilization process.
•We conclude that the role of Jekyll in cereal sexual reproduction is clearly much broader than has been hitherto realized.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Small grain cereals (wheat, barley and rye) form the foundation of most agricultural systems in temperate climates. Among these, barley (Hordeum vulgare) is one of the most widespread crops in the world (Morrell & Clegg, 2007). The evolutional plasticity and adaptability of barley stimulate interest to sexual reproduction in this cereal crop.
The plant life cycle alternates between the haploid and the diploid state. In barley and other angiosperms, the diploid sporophyte predominates, while the haploid gametophyte is typically reduced to a structure consisting of a few cells only. The gametophytes develop within specialized male (anther) and female (ovule) sporophytic organs.
The establishment of the same specialized cell types in the anther is rather conserved in monocotyledonous and dicotyledonous species (Scott et al., 2004; Wilson & Zhang, 2009) despite some considerable differences (Zhang et al., 2011). The gametophyte develops within a microsporangium derived from archesporial cells, which divide periclinally to form an inner primary sporogenous and an outer primary parietal cell (Feng & Dickinson, 2010). Meiotic division generates a tetrad of haploid microspores, which, after separation, undergo microsporogenesis, resulting in the production of mature pollen grains (Ma, 2005). The surrounding tissues, consisting of the tapetum, middle layer, endothecium and epidermis, are derived from the outer primary parietal cells (Goldberg et al., 1993; Feng & Dickinson, 2010). The essential role of the tapetum for microspore development and maturation has been demonstrated (Scott et al., 2004; Ma, 2005; Wilson & Zhang, 2009; Parish & Li, 2010; Chang et al., 2011). Control of tapetal cell identity, specification and tapetal programmed cell death (PCD) involves the interplay of many genes (Feng & Dickinson, 2007, 2010; Chang et al., 2011; Zhang et al., 2011). Tapetal degeneration is especially important for late male reproductive development and fertility in higher plants (Ma, 2005; Zhang et al., 2011). The rice PERSISTENT TAPETAL CELL 1 (PTC1) and its Arabidopsis ortholog MALE STERILITY1 (MS1) control programmed tapetal cell degradation (Li et al., 2011). APOPTOSIS INHIBITOR 5 (API5; Li et al., 2011) and TAPETUM DEGENERATION RETADRATION (TDR; Li et al., 2006) are other regulators of tapetal PCD. In cereals, degeneration of the tapetal layer contributes to the release of cell wall materials and other nutrients to the microspore via formation of Ubisch bodies or orbicules (Zhang et al., 2011). The endothecium is thought to be essential for anther dehiscence and pollen release (Ma, 2005; Yang et al., 2007). MYB26 plays a critical role in the regulation of endothecial maturation in a pathway distinct from that defined for the jasmonic acid (JA)-controlled dehiscence pathway (Yang et al., 2007). Mutations in basic leucine-zipper transcription factors TGA9 and TGA10 lead to male sterility by abnormal stability of the middle layer and lack of endothecium and epidermis degeneration, resulting in nondehiscent anthers (Murmu et al., 2010).
A few years ago we cloned and characterized the new gene Jekyll, encoding a small cysteine-rich protein which plays a key role in the sexual reproduction of barley (Radchuk et al., 2006). The gene is highly expressed in maternal seed tissues and is involve in sucrose allocation during early grain filling (Melkus et al., 2011). Jekyll transcripts are also present in flowers but at much lower abundances. However, the reduction of the grain set in transgenic lines with RNAi-mediated down-regulation of Jekyll expression was more severe than anticipated, raising questions as to whether and how JEKYLL might be involved in the development of functional floral organs or in fertilization.
Materials and Methods
Wildtype and transgenic RNAi-mediated Jekyll down-regulated plants of barley (Hordeum vulgare L.) were grown under standard glasshouse conditions at 18°C with 16 h of light and a relative air humidity of 60%. Transgenic plants with weak (line N18), moderate (line N61) and strong decreases (lines N91) in Jekyll RNA levels were selected for the study (Radchuk et al., 2006). The determination of stages of anther development was performed by evaluation of microspores under a confocal microscope as described in the Supporting Information (Fig. S1). One of three anthers of a flower was used for examination of the developmental stage and fitness of microspores by confocal microscopy, while the other two anthers and a gynoecium were collected and frozen in liquid nitrogen. At least six flowers from six separate spikelets were collected for each developmental stage and used for quantitative reverse transcription polymerase chain reaction (RT-PCR). For cDNA arrays and quantitative RT-PCR of the transgenic plants, anthers and gynoecia from at least 50 different flowers of the same plant line were collected. For each microscopic analysis, 40–60 microspores were used. Probe was classified as ‘aberrant phenotype’ if at least 10 microspores were aberrant.
For the controlled crosses, the wildtype (WT) or transgenic flowers from the middle region of an isolated spike were emasculated while the other flowers were discarded. Emasculated and isolated flowers were hand-pollinated with transgenic or WT pollen and plants were further grown for grain maturation. Between five (for WT ×WT) and 20 spikes were tested for each crossing combination. Fertilization rates were calculated as the relative ratio of the number of developed grains to the number of pollinated flowers in a spike.
Histochemical techniques and microscopy
Histochemical techniques applied to flower organs as well as in situ hybridization, immunostaining and green fluorescent protein (GFP) visualization were performed as described previously (Radchuk et al., 2006). The 33P-labeled Jekyll DNA fragment derived by PCR with the gene-specific primers 5′-CGTGGATCCGATCTCCACAAGTGCTTCTG-3′ and 5′-GAAGAAGCTTAATTCTCGGCCTATACCG-3′ was used as a probe for in situ hybridization. Immunolocalization of the JEKYLL protein was performed by polyclonal anti-Jekyll antibodies, which were derived previously (Radchuk et al., 2006) and additionally tested on Escherichia coli cells expressing JEKYLL (Fig. 6i). TUNEL assay was performed essentially as described (Radchuk et al., 2011). For transient expression of JEKYLL-yellow fluorescent protein (JEKYLL-YFP) in Arabidopsis protoplasts, the coding region of Jekyll was fused in frame with the Yellow Fluorescence Protein using primers 5′-GCAACTCGAGATGGCGGCTCGCGGTGGGAA-3′ and 5′- GAAGCTCGAGGCGACATTGAACTCGCCGTG-3′ (the cloning site for Xho I is underlined), and cloned under the control of the CaMV 35S promoter. Arabidopsis protoplasts were isolated from suspension culture and transiently transformed with the JEKYLL-YFP construct as previously described (Ellerström et al., 2005). The expression of YFP alone driven under the same promoter was used as the control. YFP signals were measured in vivo with a Zeiss LSM510 META confocal laser scanning microscope. Preparation for electron microscopy and immunogold labeling was carried out according to Rutten et al. (2003), and scanning electron probes were examined in a Hitachi S4100 SEM (Hisco Europe, Ratingen, Germany) at 5 kV acceleration voltage.
For measurement of 14C-sucrose uptake, stems with spikelets were cut at the pollen developmental stage I and placed into 100 ml of solution containing 10 mM sucrose, 5 mM glutamine, 5 mM asparagine, 5 mM KH2PO4, 10 mM 2-(N-morpholino)ethanesulfonic acid (MES), pH 7.0, and 500 μl [U-14C] sucrose (7.4 MBq ml−1; Amersham-Buchler, Braunschweig, Germany). Incubation was performed in the light (400 μmol m−2 s−1) during stages I–III (according to Fig. S1). After incubation, the anthers and gynoecia of WT and transgenic flowers were collected separately and frozen in liquid nitrogen. Subsequently, the plant material was homogenized in 2 ml methanol (60%, v/v). Radioactivity was determined by a liquid scintillation counter type Wallac 1409 (Wallac, Germany) using Rotiszint eco plus (Roth, Karlsruhe, Germany). Counts were corrected for background and quenching by external standards.
Determination of phytohormones
Concentrations of phytohormones were measured as previously described (Luo et al., 2009). Plant material (500 mg) was extracted with 0.75 ml of methanol containing 10 ng of D6-JA, 30 ng of D5-oxo-phytodienoic acid (D5- oPDA), 10 ng of D4-JA-Leu each as internal standard (all three provided by Otto Miersch, Institute of Plant Biochemistry, Halle, Germany). Further procedures are described in detail in Methods S1.
cDNA array and data analysis
Total RNA was extracted from anthers and gynoecia of WT and Jekyll down-regulated lines N61 and N91 (Radchuk et al., 2006) using the Gentra RNA isolation kit (Biozyme, Oldendorf, Germany). The material was collected at stages III–IV (Fig. S1). The isolated RNA was treated with RNase-free DNase, purified using an RNeasy Plant Mini Kit (Qiagen) and used for the synthesis of 33P-dCTP (2′-deoxycytidine 5-triphosphate)-labeled probes. Probe preparation, hybridization and processing of 12K barley seed cDNA array was done as previously described (Sreenivasulu et al., 2006) and annotations were refined based on MapMan functional categories (Sreenivasulu et al., 2008). Images of hybridized nylon membranes were subjected to automatic spot detection using the MATLAB program and scored the signal intensities of 11 787 genes from the double spots, enabling us to assess two technical replications. Additionally, two biological repetitions were performed using RNA from independently grown plants to check the biological reproducibility. Quantile normalization was carried out on the complete data set (Bolstad et al., 2003). Fold changes between WT and transgenic probes were calculated from two technical and two biological replicates. P-values were calculated based on a moderated t-test to detect false positives. Only the twofold and higher expression differences with statistically significant regulated genes in anthers/gynoecia of transgenic lines were selected for further analyses. The detailed set of the normalized values, fold difference, and P-values of differentially expressed genes are provided in Tables S1 and S2 and the complete normalized data of all 12K genes are provided in Table S3.
Quantitative RT-PCR analyses
For quantitative RT-PCR analysis of the developing barley anthers, poly(A) RNA was isolated from anthers and gynoecia at stages I–IV (Fig. S1) of pollen development using the Dynabeads mRNA Direct Kit (Invitrogen). A quantity of 8 ng of mRNA was taken for linear amplification and cDNA synthesis using the ExpressArt mRNA Amplification Nano Kit (Amptec GmbH, Hamburg, Germany). For quantitative RT-PCR of the transgenic anthers, 5 μg of the total RNA isolated from the same stage as that used for cDNA arrays was used for reverse transcription by SuperScript III reverse transriptase (Invitrogen) with the oligo(dT) 20 primer. The resulting cDNAs were used as a template for quantitative RT-PCR analyses, which were performed as described previously (Radchuk et al., 2011). The efficiencies of PCRs were estimated using the LinRegPCR software (Ramakers et al., 2003). Primer sets for each gene are listed in Table S4. All samples were run in biological triplicates for each experiment. Dissociation curves confirmed the presence of a single amplicon in each PCR. The Ct of each gene of interest (GOI) from each sample was normalized against the endogenous reference gene actin (GenBank accession number AY145451) using the formula ΔCt = CtGOI– Ctactin and calculated as an arithmetic mean of the replicates.
For qRT-PCR of the developing barley anthers, ΔΔCt values were calculated in comparison to the lowest signal value observed in a tissue for each gene. Fold changes were calculated according to Livak & Schmittgen (2001). In order to point out the relative expression levels in transgenic vs WT anthers, the gene expression values are presented as (1 + E)−ΔCt according to Czechowski et al., 2005. All statistical data are presented as means ± SD and transformed to relative gene expression (%) values, where the highest gene expression is equal to 100%.
Down-regulation of Jekyll reduces grain set and alters carbon uptake but does not impair jasmonate metabolism
Development of transgenic barley caryopses expressing RNAi-mediated Jekyll down-regulation (e.g. lines N18, N61 and N91) is largely disturbed, resulting in reduction of yield (Radchuk et al., 2006). Reciprocal crosses of Jekyll-repressed plants with WT were performed here to investigate whether down-regulation of Jekyll expression might impair seed set. First, we performed reciprocal crosses between WT and Jekyll down-regulated plant line N61 (Radchuk et al., 2006). When flowers of the transgenic line N61 were pollinated with WT pollen, grain set was over 60% (Fig. 1a). When the male parent of crosses was the transgenic line N61, grain set fell to c. 20% independently of the female parent. This indicates that, although the down-regulation of Jekyll compromises the function of both gametophytes, its impact on the male gametophyte may be more severe.
Grain set can be negatively affected when either the supply of nutrients is inadequate or the hormonal balance of the gametophyte is disturbed (Schussler & Westgate, 1995; León & Sheen, 2003). The effect of Jekyll down-regulation on carbohydrate uptake by the spike was monitored by means of 14C-sucrose feeding experiments (Fig. 1b). The WT rate of isotope incorporation was about twofold that achieved by Jekyll down-regulated anthers and gynoecia. Thus, sucrose allocation is affected in both organs.
With respect to hormonal balance, it is known that JA is essential in flower development and plays a key role in anther dehiscence (Sanders et al., 2000). The concentrations of JA, 12-oPDA and dinor-oPDA, as well as key JA amino acid derivatives were measured in both the anther and the gynoecium of WT plants and two Jekyll down-regulated lines (Fig. 1c,d). JA and JA-Ile/Leu concentrations were higher in the gynoecia and anthers of both transgenic lines than in the WT. With respect to dinor-oPDA, concentrations were slightly elevated in the anthers, but not in the gynoecia of the Jekyll down-regulated lines. Overall, although the concentrations of both JA were elevated in both anther and gynoecia of Jekyll down-regulated lines at the time of pollination, there was little evidence for any disturbance in the JA biosynthetic pathway as a whole.
Jekyll is expressed in the sporophytic tissue surrounding both the male and the female gametophytes of barley
To explore Jekyll function in flowers we attempt to relate the Jekyll gene expression pattern to distinct developmental events in the WT flowers. Flower development was divided into four stages based on cytological examination of gametophytic development in anthers starting from free released microspores and going up to mature pollen grains (Fig. S1). Anthers and gynoecia at defined developmental stages were separated and used for quantitative RT-PCR. Jekyll mRNA was detectable in the anthers starting from the onset of microgametogenesis, and its expression peaked during pollen maturation (Fig. 2a). SERK1 was used to show the presence of RNA isolated from early developing anthers (Fig. 2a, lower panel). This gene has its expression peak during meiosis (Albrecht et al., 2005; Colcombet et al., 2005). Similarly, in the developing gynoecia, Jekyll mRNA was identified during macrogametogenesis, increasing its level toward anthesis (Fig. 2b).
We further studied the tissue-specific Jekyll localization in anthers (Fig. 3) and gynoecia (Fig. 4) by in situ hybridization and protein immunostaining at different developmental stages. In situ hybridization experiments detected Jekyll transcripts in the tissues immediately surrounding developing pollen grains but no labeling was detected in the outermost tissue layers of the anther (Fig. 3a,b). In anthers of transgenic lines expressing GFP under control of the Jekyll promoter, fluorescence distribution revealed a similar pattern (Fig. 3c). Immunoassaying of JEKYLL using polyclonal antibodies at stage II was not able to detect deposition of protein, either in pollen grain or in anthers (not shown). Initially, at stage III, the immunoassaying highlighted traces of JEKYLL protein on the surface of developing pollen grains and in the tissues surrounding the locule (Fig. 3d,e). The epidermis remained unlabeled. Shortly before pollen maturation (stage IV), the protein spread within all the tissues of anther walls, including the epidermis and labeling of the pollen surface, became pronounced (Fig. 3g,h). It thus seems that Jekyll expression is initiated with the onset of microgametogenesis at stage II (Figs S1, 2a) and results in protein deposition spreading from the innermost tissues enveloping pollen grains toward peripheral layers of the anther (Fig. 3).
In the gynoecium (schematic representation, Fig. 4), both in situ hybridization and expression of GFP fused to the Jekyll promoter confirmed that Jekyll expression at stage II was very much restricted to the central region identified as the nucellus (Fig. 4a–c). Immunodetection further showed that the gene product was deposited in the nucellar tissue before anthesis (stage IV, Fig 4d,e,g,h) and after anthesis (Fig 4f,i). Labeling was strong within the fully expanded nucellar cells facing the embryo sac vacuole, but not in the antipodal cells (Fig. 4e,h), the egg apparatus (Fig. 4f,i) or the developing embryo after anthesis (Fig. 4g,j). The deposition of JEKYLL in the nucellus coincided with cell vacuolization and further disintegration of fully expanded cells, as demonstrated by the TUNEL assay (Fig. 5a,b) and electron microscopy (Fig. 5c–e). JEKYLL protein is not observed in the outer regions of nucellar projection (Fig. 4h,i), where the tissues show no signs of PCD (Fig. 5a,b). To conclude, Jekyll is transiently expressed in the innermost tissues of sporophytes, which surround the gametophytes, but not in the gametophytes themselves.
JEKYLL is predominantly localized to the intracellular membranes
In nucellar cells labeled with JEKYLL-specific antibodies, the protein was present in the cytoplasmic space, rather than in the nucleus, plastids, vacuole or cell wall (Fig. 6a). This localization was confirmed using Arabidopsis thaliana mesophyll protoplasts transiently transformed with a plasmid containing Jekyll fused to YFP (Fig. 6b). The signal appeared in the form of fiber-like structures, which may well reflect the association of the protein with intracellular membranes (Fig. 6c). In control cells, expression of YFP without Jekyll resulted in a uniform labeling of the complete cytoplasm and nucleus (Fig. 6d). An ultrastructural study on the distribution of JEKYLL by immunogold labeling confirmed a localization to endoplasmic reticulum (Fig. 6e,g), but also Golgi elements (Fig. 6f) and small vesicles (Fig. 6h) were specifically labeled. All other organelles and the cell wall were free of label.
Anther maturation is retarded by down-regulation of Jekyll
The anthers of Jekyll down-regulated plants developed more slowly than those in the WT. They seldom appeared desiccated, their anther lobes were closed and thus pollen release was hampered (Fig. 7a,b). This phenotype well correlated with the down-regulation of Jekyll expression (Fig. 8). Transcript abundance of MYB26, which plays an important role in the secondary thickening of the anther endothecium and in anther dehiscence (Yang et al., 2007), was also reduced in transgenic anthers (Fig. 8). In the mature transgenic anther, the majority of pollen was deformed, often displaying large vacuoles and substantially less starch depositions, while at the same time the tapetal layer was still present (Fig. 7c–f). The characteristic features of tapetal secretion shown by the WT were not present in transgenic lines. In the WT, numerous spheroid electron-dense structures, called orbicles or Ubisch bodies, and associated with the secretory tapetum (Parish & Li, 2010), surrounded the microspore (Fig. 9a,c). Vesicle secretion continued until the tapetum had completely disintegrated. In the transgenic lines, however, the degeneration of the tapetal cells was delayed (Fig. 9b,d). The cytoplasm of these cells showed no signs of PCD, the nuclei were still spherical (arrowed, Fig. 9b), and there was also no indication of cell wall degradation (not shown). Although the number of vesicles surrounding individual microspores was much lower (up to 30–40%) than in the WT, there was a very striking accumulation of ellipsoid-shaped electron-dense structures accumulating in that part of the tapetum cell facing the microspores (arrowed, Fig. 9d).
Consequent with delayed degeneration of the tapetum, the expression levels of TDR and GAMYB genes were decreased (Fig. 8). TDR and GAMYB were shown to play a key role in tapetal PCD and pollen exine formation (Li et al., 2006; Li & Zhang, 2010). At the same time, expression of other genes known to be involved in the control of the tapetal PCD, such as API5, PTC1 and MADS3, was not affected (Fig. 8).
In summary, the Jekyll down-regulated tapetal cells did not follow the WT developmental pathway, showing reduced secretion and delayed cell degeneration, features which could negatively affect pollen development.
The late stages of pollen development are impaired by the down-regulation of Jekyll
In both the WT and transgenic microspores, an asymmetric cell division resulted in the formation of a vegetative and a generative cell. The nucleus of the vegetative cell eventually migrated to the pollen pore. In Jekyll down-regulated plants, however, the vegetative and generative nuclei tended to remain in close proximity, and some pollen grains acquired a star-like pattern (Fig. 10a–d) indicative of stress and/or the acquisition of embryogenic competence (Maraschin et al., 2005). In the WT, the generative cell underwent a mitotic division to form two sperm cells, but in the transgenic plants, some pollen grains retained a single generative cell rather than producing two sperms (Fig. 10e). In Jekyll down-regulated anthers, the expression of the barley homolog of the DUO3 transcription factor, known to be a regulator of generative cell division and sperm cell specification in A. thaliana (Brownfield et al., 2009), was reduced (Fig. 8). Mature WT pollen grains contained starch granules and oil bodies (Fig. 9f,h), but these were sparse in the transgenic pollen (Fig. 9g,i) with the starch granules being smaller and seemingly deformed (Fig. 10g). Furthermore, transgenic pollen contained more vacuoles (Fig. 10i) and Golgi bodies (not shown), which are characteristic of a less advanced stage of development. As a result, the transgenic pollen was not adapted to the desiccation process taking place in the context of anthesis.
Transcript profiling of the Jekyll-deficient anthers and gynoecia
A 12K handmade cDNA array (Sreenivasulu et al., 2006) was used to characterize the transcriptome in the anthers of WT and the two transgenic lines N61 and N91, while only the line N91 was used to compare the expression profiles of the gynoecium of WT and transgenic lines (Tables 1, 2, S1 and S2). Anthers and gynoecia from stages III–IV (which correspond to the maximum of Jekyll expression) were collected and two biological replicates with two technical replicates each were performed. Transcript abundances are given as ratios between WT and corresponding transgenic tissues. Of 12 000 cDNA clones present on the cDNA array filter, 199 were down-regulated and 91 were up-regulated (Tables S1, S2). The largest group of up-regulated genes with known function belongs to the transcription and translation groups (37 genes) followed by sugar conversion and starch biosynthesis (11 genes), as well amino acid metabolism (eight genes) and energy production (eight genes) (Fig. S1). Down-regulated genes were predominantly enriched by PCD-related genes (14 genes), cell wall synthesis and modification (11 genes), lipid synthesis (nine genes), and others (Table S2).
Table 1. Partial list of up-regulated genes from selected categories in transgenic anthers and gynoecia
Gene identification [species]
Fold up-regulation in N91
Fold up-regulation in N61
Fold up-regulation in N91
*Clone ID is taken from http://pgrc.ipk-gatersleben.de/est/index.php. A full list of up-regulated genes is given in Table S1. 2,0 and more fold differences in transgenic anthers and 1.5 and more differences in transgenic gynoecia are in bold.
Table 2. Partial list of down-regulated genes from selected categories in transgenic anthers and gynoecia
Gene identification [species]
Fold down-regulation in N91
Fold down-regulation in N61
Fold down-regulation in N91
*Clone ID is taken from http://pgrc.ipk-gatersleben.de/est/index.php. A full list of up-regulated genes is given in Table S2. 2,0 and more fold differences in transgenic anthers and 1.5 and more differences in transgenic gynoecia are in bold.
Protein degradation and programmed cell death related
Genes regulating transcription and translation The largest group of up-regulated transcripts in both transgenic anthers and gynoecia was formed by genes involved in general transcription and translation (Table 1). This group includes genes encoding H4 histones, RNA helicase and splicing factors, α- and β-importins, translation initiation factors, various ribosomal proteins and a putative nuclear transport protein. By contrast, DNA methyltransferase and methionine synthase genes were down-regulated (Table 2). Most of the methionine produced by methionine synthase is used for DNA and histone transmethylation (Ravanel et al., 1998), suggesting that general patterns of gene expression were modified in the Jekyll down-regulated organs, a conclusion supported by the down-regulation of JmjC domain-containing histone lysine demethylase and histone acetyl transferase. In rice, histone lysine demethylase is important for the development of the floral organs (Sun & Zhou, 2008).
Genes involved in secretion and PCD Subtilisin-like serine protease encoding genes were down-regulated in the transgenic anthers but not in the gynoecia (Table 2). Subtilisin-like proteases are probably involved in the PCD response (Coffeen & Wolpert, 2004). The down-regulation of genes encoding mitochondrial Lon protease, cytosolic peptidase A1, aspartic proteinase and aminopeptidase, together with those encoding peptidase inhibitors, was a further indication that protein degradation was delayed in the Jekyll-deficient anthers. Several genes encoding components of the ubiquitin–proteasome complex were also down-regulated, as well as others encoding proteins used for transport vesicle formation and other transport processes (COPII, coat protein Sec23/Sec24, clathrin and SNARE SED5). The down-regulation of a Na+/H+antiporter and aquaporin shows that other cellular transport processes were affected by compromised Jekyll expression. Given that the molecular basis of PCD and secretion in plant cells is still incompletely understood (Hwang & Robinson, 2009), it is possible that many of the genes involved in these processes presently are classified as being of unknown functions. Approx. 43% (84 genes) of the down-regulated genes fall into this class, whereas only 12% (10 genes) of the up-regulated transcripts had unknown functions (Tables S1, S2). Both PCD and secretion are key components of the developmental program of nurse tissues (Scott et al., 2004; Parish & Li, 2010). Since the repression of Jekyll altered the expression of a number of genes controlling these functions in both tapetum and nucellus, it is implied that Jekyll is a component of the establishment of the bordering tissues. Since the gametophyte-flanking nurse tissues are established but do not reach maturity and, thus, full functionality, it seems more likely that JEKYLL is a component of the maturation machinery active in the cells of the tapetum and nucellus.
Genes involved in energy production and biosynthetic processes The up-regulation in both transgenic anther and gynoecia of genes encoding energy-producing enzymes (triosephosphate isomerase, glyceraldehyde-3-phosphate dehydrogenase and phosphoenolpyruvate carboxykinase), coupled with the down-regulation of those encoding sedoheptulose-bisphosphatase and fructose-1,6-bisphosphatase, suggested the premature activation of the late stages of glycolysis. Similarly, the pentose phosphate pathway genes encoding glucose-6-phosphate dehydrogenase, phosphoglucomutase and transaldolase were up-regulated. The entire set of genes (except for the one encoding fructan 6-fructosyltransferase) involved in starch synthesis was specifically up-regulated in transgenic anthers but not in the gynoecia. Another prominent class of up-regulated genes was those involved in amino acid metabolism, specifically S-adenosyl-L-homocysteine hydrolase, succinic semialdehyde dehydrogenase, proline transport protein and δ-1-pyrroline-5-carboxylate synthetase, together with a 24 kDa seed maturation protein, suggesting an elevated amount of protein synthesis in the transgenic anthers. To conclude, a shift in energy and storage metabolism in JEKYLL-deficient anthers and gynoecia was apparent with the more severe alterations in the anthers.
Genes related to pollen wall formation and desiccation The predominant pollen wall polymer sporopollenin is lipid- and phenylpropanoid-based. At least eight genes (including ω-3 fatty acid desaturase, β-ketoacyl reductase and lipoxygenase) involved in lipid synthesis were down-regulated in the transgenic anthers. The down-regulation of a lipid transfer protein gene in transgenic anthers may also affect sporopollenin synthesis or wax formation (Piffanelli et al., 1998; Rowland et al., 2007). Two peroxidase-encoding genes were down-regulated in the transgenic anthers; this enzyme participates in the late stages of phenylpropanoid metabolism, as well as in other metabolic pathways. Genes encoding galacturonosyltransferase, galactinol synthase and two β-galactosidases, all of which participate in cell wall lipopolysaccharide synthesis, were also down-regulated in the transgenic anthers, as were the genes encoding UDP-D-glucuronate decarboxylase, xyloglucan endo-1,4-β-D-glucanase, KORRIGAN, α- and β-expansin, which are involved in cell wall extension (Cosgrove, 2005), and two cellulose synthases. KORRIGAN, a membrane-bound endo-1,4-β-D-glucanase, is required for cellulose formation (Zuo et al., 2000), while UDP-D-glucuronate decarboxylase, xyloglucan endo-1,4-β-D-glucanase and cellulose synthase-like enzymes are probably involved in the synthesis of matrix polysaccharides, pectins and hemicelluloses (Cosgrove, 2005). The amounts of glycine-rich protein (GRP) encoding genes were also decreased. GRPs are of unknown functions, but many are associated with the pollen surface (Mayfield et al., 2001). The expression of almost all these genes was unaffected in the Jekyll down-regulated gynoecium, where desiccation and pollen coat formation do not take place.
In the angiosperms, the sporophytic tissues are responsible for the nutrition and development of the miniature gametophyte. Here, we have shown that in barley the JEKYLL protein is present in sporophytic tissues surrounding both male and female gametophytes, and that RNAi-mediated knockdown of Jekyll affects the development of these tissues. Our results not only suggest that JEKYLL is involved in the sporophytic control of the barley gametophytic development, but also indicate that the development and function of both female and male bordering tissues may be under the control of similar regulatory mechanisms.
JEKYLL is required for performing nutrition function of sporophytes
Nutrient delivery from sporophyte to gametophyte in anthers is mediated by tapetum, which releases nutrients into the anther locule via secretion (Wang et al., 2003; Zhang et al., 2011). The similar function in gynoecium is adopted by nucellus (Rangan & Rangaswamy, 1999; Luigi et al., 2006). The membrane-rich structures of the tapetal cells are responsible for the formation of transport vesicles characteristic for cereal tapetum and called orbicules or Ubisch bodies (Denecke, 2007; Zhang et al., 2011). JEKYLL was shown to be present in the endoplasmic reticulum, the Golgi apparatus and the small vesicles (Fig. 6), all of which are components of the intracellular transport system (Hwang & Robinson, 2009). The repression of Jekyll led to the down-regulation of several genes encoding proteins involved in cellular traffic in the anther (Table 2). In contrast to the rice api5 mutant anthers, where delayed degeneration of the tapetum does not affect orbicule formation (Li et al., 2011), the extrusion of orbicules is strongly suppressed in Jekyll down-regulated barley anthers, reflecting a decrease in secretion capacity (Fig. 9; Table 2). As a result, the nurse function of tapetum in these plants was compromised. Secretion normally continues until the whole cellular content of the tapetum has been supplied to the gametophyte and coupled with tapetal PCD (Vizcay-Barrena & Wilson, 2006; Parish & Li, 2010). The relationship between secretion and cell death is not clear, but it is certainly possible that a functional secretory pathway is required for the spatiotemporal orchestration of developmental PCD (Guo & Ho, 2008; Cacas, 2010). This suggestion is reinforced by the fact that, in the Jekyll down-regulated anthers, the expression of genes related to PCD (Table 2) and degeneration of the tapetal and subsequent cell layers is significantly suppressed (Figs 7, 9). Suppression of PCD and disturbed nutrient delivery are observed in the nucellus of Jekyll down-regulated caryopses (Radchuk et al., 2006). While PCD reached maximal amount in caryopsis (Radchuk et al., 2011), it is initiated in gynoecia of flowers (the innermost layer of the nucellus). Correspondingly, down-regulation of PCD-related genes was already apparent in the gynoecia of transgenic plants, where Jekyll expression is repressed (Table 2). The reduced sink strength of the transgenic male gametophyte probably reflects a disruption of the nutrient flux through the tapetum. The tapetal cells of transgenic plants seemed unable to mediate supply of the necessary amount of nutrients, and this triggered a deviant up-regulation of the genes responsible for nutrient storage in the anther (Table 1).
Our hypothesis, therefore, is that Jekyll is required to ensure a sufficient nutrition flux from the sporophyte to the gametophyte by programmed deterioration of nurse tissues. Thus Jekyll is important for the establishment of a balanced sink–source relationship during gametogenesis. While direct monitoring of 13C sucrose movement from sporophytic tissues to the gametophyte is challenging as a result of the small size of barley flowers, using caryopses of the same transgenic lines, it was shown that repression of Jekyll resulted in deceleration of sucrose allocation via nucellar tissues (Melkus et al., 2011). Collectively these results imply that Jekyll has an impact on nutrient translocation between two generations.
JEKYLL is involved in the sporophytic control of pollen maturation
Jekyll repression did not induce aberrations in the way that MS1 (Vizcay-Barrena & Wilson, 2006) and EMS1/EXS1 do (Canales et al., 2002; Zhao et al., 2002; Feng & Dickinson, 2010). This is consistent with the absence of Jekyll expression at these developmental stages. However, we observed multiple aberrations during the late microgametogenesis, when massive nutrient translocation, storage product accumulation and the formation of pollen wall occur (stages II–IV, Fig. S1). An aberrant star-like pattern was common in the pollen grains, and the second pollen mitosis of the generative cell often failed to occur (Fig. 10). In addition, there was a pronounced down-regulation of DUO3 (Fig. 8), a gene required for the expression of certain germline genes (Durbarry et al., 2005; Rotman et al., 2005). The loss of either DUO1 or DUO3 function compromises male germline specification in A. thaliana (Brownfield et al., 2009). The pollen produced by the Jekyll down-regulated transgenic anthers accumulated less starch (Figs 7, 10). Decrease in storage was also demonstrated in barley caryopses, when sucrose delivery to the endosperm was compromised as a result of repression of Jekyll in nucellar tissue (Radchuk et al., 2006; Melkus et al., 2011). Genes putatively involved in lipid and phenylpropanoid synthesis, as well as those encoding enzymes necessary for cell wall synthesis were also down-regulated in the transgenic anthers (Table 2). The formation of the outer coat of pollen, the exine, also relies on the deposition of sporopollenin, raftins and pectocellulose from the sporophytic tissues (Piffanelli et al., 1998; Wang et al., 2003; Blackmore et al., 2007); is controlled by GAMYB and TDR transcription factors (Zhang et al., 2011); and is vital for the pollination (Kawanabe et al., 2006; Borg et al., 2009). The nucellar cell remnants also contribute to the exine formation (Piffanelli et al., 1998; Blackmore et al., 2007). Both GAMYB and TDR genes are down-regulated in the transgenic anthers (Fig. 8), indicating tapetal PCD delay in transgenic anthers which might compromise incorporation of cell remnants into the exine. Thus, JEKYLL function in sporophytic anther tissues is essential for the microgametophyte development, including pollen formation, accumulation of storage products and the synthesis of pollen coat, which were all impaired by the repression of Jekyll.
JEKYLL is important for the coordinated anther dehiscence
The controlled dehiscence of the anther, which is critical for effective pollen shed, is brought about by PCD in the sporophytic tissues (Beals & Goldberg, 1997; Sanders et al., 2000). The initiation of PCD takes place in the tapetum and normally extends radially to the endothecium and epidermal cells (Varnier et al., 2005). The factors triggering this process have remained elusive (Wilson & Zhang, 2009). Here, we have shown that the pollen release from Jekyll-repressed anthers was disturbed (Fig. 7b). The accumulation of JEKYLL starts from the innermost tapetal layer of the anther, then spreading to the outermost tissues of the sporophyte (Fig. 3), coinciding with degeneration of these tissues. JEKYLL is able to switch the cell fate to death (Radchuk et al., 2006). The repression of Jekyll might decelerate cell death in sporophytic tissues of the anther during anther dehiscence (Wetzel & Jensen, 1992; Varnier et al., 2005). The much reduced expression of PCD-related genes (Table 2) and the activation of those involved in biosynthesis and energy metabolism (Table 1) provided further evidence for the loss of control over dehiscence in the transgenic anthers.
The dehiscence process is under hormonal control (Wilson & Zhang, 2009), specifically involving JA in its early stages (Sanders et al., 2000; Stintzi & Browse, 2000; Ishiguro et al., 2001; Nagpal et al., 2005), and auxin in its later ones (Cecchetti et al., 2008). In Arabidopsis, loss-of-function double mutants for the auxin response factors ARF6 and ARF8 produce indehiscent anthers and short filaments (Nagpal et al., 2005), and the barley ARF6 homolog was strongly down-regulated in both Jekyll RNAi lines (Table 2). The MYB26 factor was also down-regulated in both RNAi lines (Fig. 8). In A. thaliana, MYB26 is involved in the regulation of anther dehiscence and, along with MYB24, is a component of the JA response (Mandaokar & Browse, 2009) during the late stages of anther development. The JA and JA-Ile-Leu (the active form of JA) content of the transgenic anthers during the period when Jekyll expression was at its peak was higher than in the WT anthers (Fig. 1). JA is known to be able to activate its own synthesis via a positive feedback loop (Farmer, 2007; Wasternack, 2007). However, despite the high concentration of JA present in the transgenic anthers, they still suffered from delayed dehiscence. The data clearly indicate that JEKYLL is involved in the regulation of pollen release, but further study focused to a greater extent on JEKYLL interplay with other hormones governing the late stages of anther development (e.g. GA, auxin) is required to understand the mechanism of JEKYLL in anther dehiscence.
The impact of JEKYLL on fertility
Defects in the generative nuclei, reductions in the amount of storage products available and imperfect pollen walls all impair the success of pollination. All these aberrations occurred in the pollen produced by Jekyll down-regulated plants, and so can explain the major reduction in seed set experienced when their pollen was used for fertilization (Fig. 1a). A more direct role for JEKYLL is also imaginable. The protein composition of the pollen wall is important for adhesion, signaling and sexual compatibility (Edlund et al., 2004). JEKYLL protein was detectable on the pollen grain surface (Fig. 3g). The down-regulation of Jekyll would therefore likely reduce the amount of JEKYLL present on or within the pollen wall. During the course of the interaction between the germinating pollen grain and the stigma, PCD is triggered as a means of incompatibility (Bosch & Franklin-Tong, 2008), but it is also required for the normal growth of the pollen tube (Wu & Cheun, 2000; Crawford & Yanofsky, 2008). JEKYLL has structural similarity (see Radchuk et al., 2006) to pollen wall GRPs in the Brassicaceae (Fiebig et al., 2004). Mutation of GRP17 in Arabidopsis has been shown to delay pollen hydration and to decrease the rate of effective pollination (Mayfield & Preuss, 2000).
Genes involved in the reproductive system are prone to rapid evolutionary change, which provides a means of curtailing fertilization by nonselfs (Fiebig et al., 2004). The Jekyll sequence is present as a single copy in barley, with homologous sequences known in wheat and rye but not in more distant species (Radchuk et al., 2006). In Arabidopsis and some related species, a family of pollen-specific GRPs appears to have diverged markedly through a series of duplications and deletions, as well as through mutations in their C-terminal repetitive sequences (Fiebig et al., 2004). A comparison of Jekyll homologs has also identified divergence in the central and C terminal repetitive regions, while the N termini have remained rather well conserved (Radchuk et al., 2006). Such divergence in genes involved in sexual reproduction can promote speciation by preventing wide crossing and/or by reducing the extent of self-incompatibility expressed (Edlund et al., 2004; Wheeler et al., 2009). While the evolutionary significance of JEKYLL remains to be explored, its role in the sexual reproduction of barley is clearly much broader than has been hitherto realized.
We thank U. Wobus, T. Altmann, J. Thiel and D. Weier for valuable discussions. Special thanks to M. Strickert and S. Friedel for support in quantitative normalization and statistics. We greatly appreciate U. Siebert, A. Stegmann, G. Einert, and E. Fessel for excellent technical assistance. This work was supported in part by the German Ministry of Education and Research within the German Plant Genome Initiative (GABI-sysSEED, FKZ 0315044A and GABI-GRAIN, FKZ0315041A).