FAD2 and FAD7 desaturases are involved in cold acclimation of olive (Olea europaea) mesocarp. There is no research information available on cold acclimation of seeds during mesocarp cold acclimation or on differences in the cold response of the seed coat and embryo. How FAD2 and FAD7 affect seed coat and embryo cold responses is unknown. Osmotin positively affects cold acclimation in olive tree vegetative organs, but its role in the seeds requires investigation.
OeFAD2.1, OeFAD2.2, OeFAD7 and Oeosmotin were investigated before and after mesocarp acclimation by transcriptomic, lipidomic and immunolabelling analyses, and cytosolic calcium concentration ([Ca2+]cyt) signalling, F-actin changes and seed development were investigated by epifluorescence/histological analyses.
Transient [Ca2+]cyt rises and F-actin disassembly were found in cold-shocked protoplasts from the seed coat, but not from the embryo. The thickness of the outer endosperm cuticle increased during drupe exposure to lowering of temperature, whereas the embryo protoderm always lacked cuticle. OeFAD2 transcription increased in both the embryo and seed coat in the cold-acclimated drupe, but linoleic acid (i.e. the product of FAD2 activity) increased solely in the seed coat. Osmotin was immunodetected in the seed coat and endosperm of the cold-acclimated drupe, and not in the embryo.
The results show cold responsiveness in the seed coat and cold tolerance in the embryo. We propose a role for the seed coat in maintaining embryo cold tolerance by increasing endosperm cutinization through FAD2 and osmotin activities.
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.
Cold acclimation involves physical and biochemical restructuring of cell membranes, accumulation of cryoprotectants, and changes in F-actin (Örvar et al., 2000). It requires the acquisition of a persistent cold memory by exposure to gradual reductions in temperature (Levitt, 1980). Olea europaea is an evergreen species that has to endure cold winters. However, there are genotypes that are able to acclimate to cold (D'Angeli et al., 2003). A transient increase in cytosolic calcium concentrati1on ([Ca2+]cyt) is one of the early-sensing mechanisms for low temperatures (Kudla et al., 2010). In herbaceous plants, early cold-induced events include changes in the lipidome, leading to cell membrane rigidification and F-actin disassembly (Örvar et al., 2000). By contrast, when cold acclimation is acquired, cell-membrane fluidification, induced by an increased presence of unsaturated fatty acids (FAs) in membrane lipids, and F-actin stabilization occur concomitantly with the acquisition of [Ca2+]cyt resting concentrations (Martinière et al., 2011; Burgos et al., 2011). In accordance with this, noncold-acclimated mesophyll cells of olive trees show transient F-actin disassembly and increases in [Ca2+]cyt, whereas cold-acclimated cells show stable F-actin and [Ca 2+]cyt concentrations (D'Angeli et al., 2003; D'Angeli & Altamura, 2007).
The actin cytoskeleton requires a regulatory system to control its cycling, and the actin-binding proteins (ABPs) play a central role in this system (Papuga et al., 2010). In collaboration/competition with other ABPs, the actin-depolymerizing factors (ADFs) integrate transmembrane signals and are involved in lipid metabolism (Bamburg & Bernstein, 2010). Some ADFs regulate actin cycling by promoting actin-filament (AF) severing and/or facilitating pointed-end depolymerization, whereas other ADFs enhance the polymerization rate by stabilizing and crosslinking AFs (Tholl et al., 2011). In Arabidopsis, AtADF1 increases the AF-depolymerization rate, whereas AtADF9 enhances the AF-polymerization rate (Tholl et al., 2011).
The olive tree drupe is formed by an external epicarp, a middle mesocarp and an internal endocarp, which becomes totally lignified at end of the epi-mesocarp expansion growth. The endocarp encloses the seed, which is delimited by a seed coat, differentiating at the onset of endocarp lignification (Matteucci et al., 2011). The seed coat encloses the endosperm and embryo (Fig. 1). The drupes remain on the tree for a long time during autumn and winter, with winter chilling necessary to slow down of the ripening process (Palliotti & Bongi, 1996). By monitoring [Ca2+]cyt changes in response to cold shocks in drupes of genotypes differing in cold acclimation in the leaves, the mesocarp cells have been found to cold acclimate, but only in genotypes able to cold acclimate in the leaves (Matteucci et al., 2011). Moreover, cold acclimation is acquired by the mesocarp only at the completion of oleogenesis, that is, at maturation end (Matteucci et al., 2011). In the mesocarp, the expression of FA desaturation genes (FADs) changes during oleogenesis, with remarkable increases in FAD2.1 and FAD2.2 (oleate desaturase genes) and FAD7 (linoleate desaturase gene) transcription, and with further increases in FAD2.2 and FAD7 transcripts in response to cold shocks (Matteucci et al., 2011).
There is no information on the acquisition of cold acclimation by the olive tree seed, or on whether it can acclimate during the predispersal phase. It is possible that even when the seed is still enclosed in the drupe on the tree, all or part of seed compartments exhibit a different cold perception/acclimation in comparison with the mesocarp. Olive tree seeds produce their lipid reserves (Ross et al., 1993), with unsaturated FAs accumulating in the triacylglycerols (TAGs) of their oil bodies (OBs) (Hernández et al., 2005). Unsaturated FAs might be also involved in membrane fluidification in response to cold; however, how cold affects FAD2.1, FAD2.2 and FAD7 transcription, and the corresponding enzyme activities, within the seed compartments is unknown.
Unsaturated FAs are very important in determining olive oil quality, with the best quality obtained with 55–83% of oleic acid (C18:1), 3.5–21% of linoleic acid (C18:2), and < 1% of α-linolenic acid (C18:3) (European Commission, 2003). The contribution of the seed to the oil of the drupe is c.10% (Connor & Fereres, 2005). However, if the seed were to become oil-rich in unsaturated FAs (e.g. C18:2 (ω6 FA)) in response to cold, the oil quality would be enhanced. This would render the oil more suitable for diets designed to reduce cardiovascular diseases, and for the treatment of some cancers and arthritis (Salas et al., 2000; Ursin, 2003). The benefits to human health might be echoed in agriculture, with improved orchard management and oil quality in regions with cold winters that have recently begun olive tree cultivation (e.g. China; Rugini et al., 2000).
Dormancy is normally induced during seed maturation on the mother plant, and the basic helix–loop–helix transcription factor SPATULA (SPT) is a marker of the dormant state in Arabidopsis seeds (Penfield et al., 2005). Dormancy is preceded by desiccation tolerance (Finch-Savage & Leubner-Metzger, 2006). RAB18 is a late embryogenesis abundant (LEA) gene coding a dehydrin involved in seed desiccation tolerance (Nylander et al., 2001), and also has protective functions during the osmotic stress caused by cold exposure (Catalá et al., 2003). Osmotin is another protein with osmo- and cryoprotectant roles, and it also has actin-binding activities (Zhu et al., 1995; Takemoto et al., 1997; Newton & Duman, 2000). It is a member of the pathogenesis-related (PR) type-5 protein family (Min et al., 2004). A role of osmotin as a cold-acclimating protein, involved in F-actin stabilization and in blocking [Ca2+]cyt transients, has been demonstrated in the vegetative organs of olive trees (D'Angeli & Altamura, 2007). Osmotin and osmotin-like mRNAs and proteins have been also found in seeds, for example, those of Benincasa hispida (Shih et al., 2001) and tobacco (Miele et al., 2011).
Arabidopsis thaliana NAC with Transmembrane Motif 1-Like (AtNTL6) protein integrates cold signals into defence responses, because it is cold-induced and its transcriptionally active form (NAC) induces the expression of PR-5 genes (Seo et al., 2010). This highlights the possibility that the homologue NAC sequence in O. europaea may induce osmotin transcription in the seed.
The research was designed to verify whether cold acclimation occurs in olive tree seeds before fruit dispersal, whether different seed compartments (seed coat vs embryo) differ in cold responsiveness/acclimation during drupe maturation and exposure to cold, and whether a crosstalk between different compartments may exist. To reach the goal, genes involved in F-actin cycling (ADF1 and ADF9), FA desaturation (FAD2.1, FAD2.2, and FAD7), cryoprotection (osmotin, and its inducer NAC), and dehydration (RAB18) and dormancy (SPATULA) were investigated through an integrated transcriptomic, lipidomic and cytohistological study.
The results show no cold acclimation in the seed coat, and cold tolerance in the embryo that is independent of the acquisition of cold acclimation by the mesocarp. Osmotin presence and increased C18:2 concentrations occur in the seed coat only, and only when the mesocarp is acclimated. The seed coat is involved in maintaining cold tolerance in the not-yet-dormant embryo by increasing endosperm cutinization through the activity of both the isoforms of FAD2 and osmotin.
Materials and Methods
Protoplast isolation and fluorescence measurements for [Ca2+]cyt detection
Protoplasts were obtained from the seed coats and embryos of 10 drupes randomly collected at a specific time after flowering (at 14 and 22 wk after flowering (WAF) when not alternatively specified), from five randomly selected 5-yr-old trees of Olea europaea L. cv Frantoio. The sampling WAF were chosen during the oleogenic period, when cold acclimation is acquired by the drupe mesocarp of the cold-acclimating cultivars (Matteucci et al., 2011). The trees were grown under the conditions adopted by Matteucci et al. (2011) in the Garden of the Environmental Biology Department (Rome) for three crop years (2008–2011). The temperature-lowering began at 14 WAF and its trend was similar in each of the three years (www.arsial.it/portalearsial/agrometeo).
Protoplasts were produced according to D'Angeli et al. (2003) and the [Ca2+]cyt changes after cold shocks at a cooling rate (ΔT/Δt) of 10°C over 60 s (10°C/60 s), starting from room temperature, were performed, evaluating Calcium-Crimson-AM fluorescence according to D'Angeli et al. (2003) and Matteucci et al. (2011). Heat shocks consisting of a 10°C rise over 60 s (10°C/60 s) starting from room temperature were also applied to embryo protoplasts at 14 and 22 WAF. The fluorescence variations were detected according to D'Angeli et al. (2003). The fluorochrome stability was verified in the protoplasts at room temperature for 400 s (Supporting Information, Fig. S1). The cell permeability of the fluorochrome was verified at room temperature using the calcium ionophore A23187 (Sigma), and showed the same results as in D'Angeli et al. (2003). Values were expressed in arbitrary units (AUs) and their deviation from the initial value was determined. A total of 30 protoplasts per replicate were examined. Experiments were carried out in triplicate with similar results (data from the second-year replicate are shown).
Fatty acid detection and lipoxygenase (13-LOX) activity
Samples of seed coats and embryos coming from batches of 30 seeds at both 14 and 22 WAF were freeze-dried in a Lio 5P-4K (Cinquepascal, Milan, Italy) for 3 d at −33(± 2)°C and 0.01 mbar. A quantitative extraction by a 3/2 v/v n-hexane/isopropyl alcohol mixture containing 0.0025% w/v BHT (Buthylhydroxytoluene) was performed in an orbital shaker (110 rpm) at 25°C for 30 min and the extraction was repeated twice. Nonadecanoic acid (19 : 0) was added as internal standard before extraction. In order to precipitate a pellet, a potassium sulphate aqueous solution (7% w/v) was added to the supernatant (v/v), and the organic phase was removed. The samples were then dried by nitrogen at room temperature, and transesterified by a boron trifluoride–methanol complex 14% (BF3; Sigma). The derivatization procedure was performed in accordance with Morrison & Smith (1964) with slight modifications (Reverberi et al., 2001). The resulting FA methyl esters were analysed by gas chromatography (GC-HP5890 series II) equipped with a flame ionization detector (FID) using a capillary column SPB™-PUFA 30 m × 0.25 mm i.d. with a 0.2 μm film thickness (Supelco Inc., Bellefonte, PA, USA). FAs were identified by comparison of retention times with authentic standards. Data were expressed as GC-mean areas mg–1 DW (± SD) of embryo and seed coat batches coming from the same batch of 30 seeds per WAF, and mean percentage (± SEM) distribution of C18:1 and C18:2 in the free FA (FFA), TAG and polar lipid (PL) eluted fractions determined by setting C18:1 and C18:2 values in the total FA fraction equal to 100%. Experiments were carried out in triplicate with similar results (data from the second-year replicate are shown).
The activity of the 13-LOX enzyme was evaluated in mesocarp, seed coat and embryo samples by spectrophotometric assay according to Fabbri et al. (2000) using a Beckman DU 530 at 234 nm. Five determinations from three independent experiments yr–1 were carried out (data from the second-year replicate are shown).
F-actin (Olea europaea β-actin, Matteucci et al., 2011) variations in the seed coat and embryo protoplasts in response to cold shocks of ΔT/Δt = 10°C/60 s starting from room temperature were detected using Phallacidin-TRX (Molecular Probes, Eugene, OR, USA) according to D'Angeli & Altamura (2007). Values were expressed in AUs, and their deviation from the initial value determined. A total of 30 protoplasts per replicate were examined. Protoplast permeability to the fluorochrome was evaluated. Three minutes of incubation were enough for complete fluorochrome absorption, as in D'Angeli & Altamura (2007). The fluorochrome stability was verified in the protoplasts at room temperature for 400 s (Fig. S1). Experiments were carried out in triplicate with similar results (data from the second-year replicate are shown).
Ten drupes at 14, 22 and 26 WAF were randomly collected in each crop year. The seeds were separated from the endocarp, fixed in 70% ethanol, dehydrated, embedded in a resin (Technovitt 7100, Heraeus Kulzer, Wehrheim, Germany), sectioned at 5 μm thickness with a Microm Zeiss HM 350 SV microtome (Zeiss, Germany), and stained with toluidine blue for bright-field imaging. Oil bodies and suberin were detected by Sudan IV stain according to Capitani et al. (2005). The images were acquired with a Zeiss Axiolab microscope, equipped with a Leica DFC 320 (Leica, Germany) video camera.
Aniline blue staining was used to detect callose deposits in the cell walls of the seed tissues according to Falasca et al. (2000). The detection of cutin was performed using berberine-HCl 0.1% for 1 h (Cruz et al., 1992). The sections were rinsed in water, counterstained with calcofluor 0.1% for 25 min to detect cellulose (Cruz et al., 1992), and observed under epifluorescence with a 420–490 nm filter absorption and 515 nm filter emission. The fluorescence images were collected using a Leica DMRB microscope with a Leica DC500 camera.
Flow cytometry and Vybrant assay
Estimations of cell-cycle activity were performed on protoplasts from seed coats and embryos at 14 and 22 WAF for each crop year using a FaCSCalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA). Protoplasts were prepared and incubated in propidium iodide dye according to D'Angeli & Altamura (2007), and the values were estimated using Cell Quest software (Becton Dickinson). The analyses were performed on 100 000 cells per replicate (data from the second-year replicate are shown).
Protoplasts of seed coat, embryo and endosperm, prepared according to D'Angeli & Altamura (2007), were marked with the apoptosis detection kit (Vybrant apoptosis assay, Molecular Probe). The assay allows simultaneous imaging of viable protoplasts, dead protoplasts and protoplasts in programmed cell death (PCD) in the same sample through the evaluation of differences in cell-membrane permeability. The assay uses propidium iodide, which is excluded by viable protoplasts, but enters late-PCD and dead protoplasts; YO-PRO dye, which enters nuclei from early to late PCD; and Hoechst 33342 dye, which enters live or early-PCD nuclei (Darzynkiewicz et al., 1997; D'Angeli & Altamura, 2007).
Viable, dead and apoptotic protoplasts were counted at 14, 18, 22 and 26 WAF under the same epifluorescence microscope as described earlier. Mean percentages (± SEM) were calculated for a total of 2500 protoplasts per seed compartment and WAF, with similar results in the years (data from the second-year replicate shown).
RNA extraction, cDNA sequencing, and quantitative real-time reverse transcription-PCR (qRT-PCR)
Total RNA was isolated and cDNA synthesized using the protocol described in Matteucci et al. (2011), pooling the material from the seed coats and embryos of 10 drupes at 14 and 22 WAF, respectively, per crop year. The drupes came from the same trees used for the previous analyses.
The primers used in qRT-PCR on the total RNA of seed coats and embryos were designed by OligoExplorer 1.2 software (Genelink, Hawthorne, NY, USA) from the comparison of plant genes available in Genbank (www.ncbi.nih.nlm.gov/genbank) and olive tree genes available in the OleaEst database (http://184.108.40.206/oleaestdb/) (Table 1).
Table 1. Primer sequences used for quantitative real-time reverse transcription-PCR (qRT-PCR) amplification
All sequences designed on genes or expressed sequence tag of Olea europaea.
Three technical replicates in two independent experiments, using two independent RNAs for each sample, were carried out for qRT-PCR experiments and the means from six replicates were subjected to SEM calculation according to Matteucci et al. (2011) and expressed on a linear scale. Quantification was performed using standard dilution curves for each gene fragment, and the data were normalized for the quantity of Cryptochrome2 (Cry2) transcript (O. europaea.) (Livak & Schmittgen, 2001). Cryptochrome2 was selected as housekeeping gene after a qRT-PCR screening of various candidate genes (Table S1). Data from the second crop year are shown.
After removing the endocarp, 10 seeds at 14 and 22 WAF for each year were fixed, dehydrated, embedded in resin and sectioned for in situ localization of osmotin following the procedure of D'Angeli & Altamura (2007).
Differences between the mean values (± SEM) were compared using Student's t-test.
Seed development during drupe maturation
The embryo showed the cotyledonary stage at 14 WAF (Fig. 1). No temperature decrease was experienced by the drupes from fertilization up to 14 WAF (i.e. half of the oleogenic period). They experienced a temperature decrease of 10 (± 2)°C between 14 and 22 WAF (i.e. the end of oleogenesis and maturation), and a further decrease of 4 (± 1)°C between 22 and 26 WAF (i.e. the advanced postmaturation phase).
The embryo increased in length between 14 and 22 WAF (from 8.5 ± 1 to 10.0 ± 1 mm, P < 0.01 difference, Fig. 2a,d). Since the mean cell area did not change significantly between 14 and 22 WAF, being 294 ± 10 and 300 ± 16 μm2, respectively, a contribution of cell expansion to growth was excluded. By contrast, embryo cells showed division activity between 14 and 22 WAF, as verified by flow cytometry analysis (Fig. 2b,e). Protoderm was already differentiated at 14 WAF. Neither cutinization nor modifications of any other type, such as callose and lignin deposition, occurred in the cell walls at later times (e.g. at 22 WAF; Fig. 3a). Dead and apoptotic cells were sporadic all over the investigation period, as verified by Vybrant assay. Oil bodies were already detected at 14 WAF, and their presence did not increase further (Fig. 3b, Table 2).
Table 2. Mean diameter (∅) and mean number per cell (± SEM) of oil bodies, detected by Sudan IV staining, in the seed coat, endosperm and embryo of olive tree (Olea europaea) seeds at 14 and 22 wk after flowering (WAF) during drupe maturation
∅ (μm) (seed coat)
Number/cell (seed coat)
∅ (μm) (endosperm)
∅ (μm) (embryo)
n =100 (data from the second replicate). **, P <0.01 in comparison with the value at 14 WAF within the same column. Values not followed by an asterisk within the same column are not significantly different.
At 14 and 22 WAF, OBs were also present in the endosperm cells, with diameters c. three- and sevenfold, respectively, higher than in the embryo (Fig. 3c, Table 2). Endosperm cells were parenchymatous with thin and cellulosic cell walls except in the outermost layer, showing cutinized external walls already at 14 WAF (Fig. 3d). Cutin deposition increased significantly (P < 0.01) during the following weeks, reaching a value at 26 WAF c. 1.5-fold higher than that at 14 WAF (Table 3, Fig. 3e), and continuing to increase thereafter (reaching 9.04 ± 0.14 μm at 28 WAF, i.e. at dispersal time). Callose, cellulose and suberin did not contribute to cell wall thickening. Few dead/apoptotic cells (20%) were present in the endosperm at 14 WAF, whereas the majority of cells were dead (c. 40%) or apoptotic (c. 20%) at 22 WAF.
Table 3. Cuticle thickness (mean value in μm ± SEM) in the external cell wall of the outer epidermis of the Olea europaea seed coat, and in the external cell wall of the outermost endosperm layer at different times of drupe maturation (14 and 22 WAF), and in postmaturation (26 WAF)
n =100 (data from the second replicate). WAF, wk after flowering.
**, P <0.01 in comparison with the value at 14 WAF in the same row. ***, P <0.01 in comparison with all the other values in the same row.Values not followed by an asterisk within the same row are not significantly different.
The seed coat (Fig. 1) was composed of c. 10 cell layers, excluding the vascular system (Fig. 3a,f–h). No increase in cell number occurred in these layers between 14 and 22 WAF, because the bulk of the nuclei were at G1 phase at both times (Fig. 2c,f). At 14 WAF, the cells of the outer epidermis already exhibited a cuticle on the external wall (Fig. 3d), which was thinner than in the outermost endosperm layer (Fig. 3d and Table 3). Internally to the outer epidermis, there was a parenchyma with chloroplasts (Fig. 3g). The seed vascular system was differentiated at 14 WAF (Fig. 3g,h). The parenchymatous layers located internally to the vascular system exhibited small-sized cells with rare chloroplasts. An inner epidermis with cellulosic cell walls completed the seed coat (Fig. 3i). The cells of all the seed coat layers already contained OBs at 14 WAF (Table 2), the number and diameter of which did not change at 22 WAF (Fig. 3j, Table 2). The cuticle thickness of the external wall in the outer epidermal cells did not change significantly up to 26 WAF (Table 3), and even up to 28 WAF (not shown). At 22 WAF the seed was not green in colour, because chloroplasts had turned into amyloplasts. Moreover, inclusions, appearing as very small spots and showing the same autofluorescence signal of cutin, possibly ‘cutinsomes’ according to Heredia-Guerrero et al. (2008), were present within the cells and in the walls of the internal epidermis at both 22 and 26 WAF (Fig. 3k–l). At the same time, the same spots occurred in the cuticle of the cell walls of the outermost endosperm layer (Fig. 3l–m). The percentage of dead cells in the seed coat was low at 14 WAF. Apoptotic cells appeared at 18 WAF (Table 4). Dead and apoptotic cells significantly (P <0.01 and P <0.05, respectively) increased at 22 WAF (Table 4), and, in accordance with the Vybrant assay results, a small class of nuclei with a DNA content lower than G1 was identified by flow cytometry (Fig. 2f). Mortality further increased at 26 WAF (Table 4). The de novo formed tracheids and fibres, which are dead cells at maturity, were not responsible for these increases, because no bundle branching occurred later than 18 WAF.
Table 4. Mean percentage (± SEM) of live, dead and apoptotic nuclei present in the olive tree (Olea europaea) seed coat protoplasts at different times of drupe maturation (14–22 WAF), and in postmaturation (26 WAF), evaluated by Vybrant assay (n =2500, data from the second replicate)
WAF, wk after flowering.
*, P <0.05 in comparison with the value at 18 WAF in the same row; **, P <0.01 in comparison with the values at 14 and 18 WAF for dead and live cells within the same row; ***, P <0.01 in comparison with all the other values in the same row. Values not followed by an asterisk within the same row are not significantly different.
Cold-induced [Ca2+]cyt signalling is active in the seed coat, but not in the embryo
The variations in [Ca2+]cyt were investigated in seed coat and embryo protoplasts belonging to drupes exposed to a cold shock of ΔT/Δt = 10°C/60 s at 14 and 22 WAF. The endosperm cells were excluded from this and the successive analyses because their high mortality at 22 WAF (i.e. c. 60% of dead and apoptotic cells).
The mesocarp cells were exposed to the same shock to verify whether the drupe was not cold-acclimated at 14 WAF and cold-acclimated at 22 WAF. The presence of a [Ca2+]cyt transient at 14 WAF and not at 22 WAF (Fig. S2) showed that the mesocarp was cold-acclimated at 22 WAF.
As shown in Fig. 4(a), there was no change in [Ca2+]cyt in the embryo cells at both times. The resting concentrations of [Ca2+]cyt revealed that the embryo was not responding to cold shocks (i.e. it was cold-tolerant) independently of the cold acclimation by the mesocarp. The seed coat showed a [Ca2+]cyt transient increase at both drupe stages (Fig. 4b), with mean percentage increases relative to the initial value of 23.6 ± 2.1 and 14.5 ± 2.4 at 14 and 22 WAF, respectively (P < 0.01, difference between 14 and 22 WAF). The transient increases in [Ca2+]cyt revealed that the seed coat was also not cold-acclimated when (22 WAF) the drupe was cold-acclimated.
To verify whether the embryo cells were also thermotolerant, a heat shock of ΔT/Δt = 10°C/60 s was applied at 14 and 22 WAF. Changes in [Ca2+]cyt at both WAFs, with mean percentage increases relative to the initial value of 43.3 ± 1.9 and 22.6 ± 1.3 at 14 and 22 WAF, respectively (P < 0.001, difference between 14 and 22 WAF), demonstrated that the embryo was capable of responding to increasing temperature, but not to reducing temperature (Fig. 4a,c, in comparison).
Cold induces changes in F-actin cycling in the seed coat, but not in the embryo
In olive tree mesophyll cells, actin cycling between filamentous (F-actin) and monomeric (G-actin) forms is affected by exposure to low temperature and cold acclimation, with a stable F-actin meshwork characterizing the cold-acclimated cells (D'Angeli & Altamura, 2007). Actin cycling was investigated according to D'Angeli & Altamura (2007). The fluorochrome used is a labelled phallotoxin with high affinity for F-actin and unable to bind G-actin (Wieland, 1986). After dye absorption, the changes in fluorescence intensity indirectly measure the stability/disassembly of F-actin (Ketelaar et al., 2003).
Protoplasts extracted from the embryo and the seed coat from drupes at 14 and 22 WAF were exposed to the same cold shock (i.e. ΔT/Δt = 10°C/60 s) by which the variations in [Ca2+]cyt had been investigated (Fig. 4a,b). In the embryo, no significant variation in F-actin meshwork occurred with the cold shock at both 14 and 22 WAF (Fig. 5a), indicating, together with the absence of [Ca2+]cyt transients already at 14 WAF (Fig. 4a), that the embryo was cold-tolerant before the acquisition of acclimation by the mesocarp. By contrast, the F-actin meshwork in the seed coat protoplasts showed a transient reduction in stability at both 14 and 22 WAF (Fig. 5b–c), albeit to a different extent, that is, a mean percentage decrease with respect the initial value of 22.16 ± 1.9 at 14 WAF and of 12.44 ± 0.3 at 22 WAF (P < 0.001 difference). The results of Figs 4(b) and 5(b–c) collectively indicate the absence of cold acclimation in the seed coat of the cold-acclimated drupes.
Expression levels of the olive tree actin gene (OeACT), and of OeADF1 and OeADF9, homologues of AtADF1 and AtADF9, coding for ABPs causing F-actin depolymerization and polymerization, respectively, in Arabidopsis (Tholl et al., 2011) were evaluated in both embryo and seed coat for an in-depth investigation of actin cycling in the two seed compartments.
In the embryo (Fig. 6a), actin mRNA was greatly increased at 22 WAF compared with 14 WAF, possibly to provide actin subunits to the cytoskeleton of the forming cells (Fig. 2e). ADF1 transcription was low at both 14 and 22 WAF (Fig. 6b), whereas ADF9 transcription was constantly higher (Fig. 6c), suggesting low and high activity, respectively, of the corresponding ABPs, resulting in a stable F-actin meshwork, in accordance with the epifluorescence results (Fig. 5a).
In the seed coat (Fig. 6a), the actin mRNA showed a significant (P < 0.001) reduction at 22 WAF compared with 14 WAF, coincidently with the cessation of cell division in the seed coat at 22 WAF (Fig. 2f). ADF1 levels were many-fold higher than in the embryo at the corresponding WAF. Moreover, the transcription abundance at 14 WAF was significantly (P < 0.001) higher than at 22 WAF (Fig. 6b), whereas the expression levels of ADF9 were similar at both times (Fig. 6c), and lower (P < 0.01) at 14 WAF, and higher (P < 0.01) at 22 WAF, than those of ADF1 (Fig. 6b,c, in comparison). The results support a possible activity of ADF1 in increasing F-actin depolymerization in the seed coat, giving monomers to be repolymerized to maintain the actin cycling necessary for cold response (Fig. 5b).
Dormancy is normally induced during seed maturation on the mother plant and is preceded by desiccation tolerance (Finch-Savage & Leubner-Metzger, 2006). In Arabidopsis seeds, RAB18 transcripts accumulate just before the dry seed stage (Nylander et al., 2001), and the transcription factor SPATULA (SPT) is a dormant-state marker (Penfield et al., 2005). To find whether the embryo cold tolerance was related to desiccation tolerance and/or dormancy, the transcription of OeRAB18 and of OeSPT (Table 1) was investigated at both WAFs. OeSPT transcription was quite undetectable at both WAFs (Fig. 7, inset), indicating that the embryo cold tolerance was unrelated to dormancy. By contrast, OeRAB18 transcription was already observed at 14 WAF, and increased many-fold (P < 0.001) at 22 WAF (Fig. 7), indicating that, at the latter WAF, its enhanced transcription was needed for the dehydration protection of the embryo. In the seed coat, OeSPT was undetectable, and OeRAB18 transcription did not change significantly between 14 and 22 WAF, and was many-fold lower than in the embryo at 22 WAF (P < 0.001; Fig. 7).
OeFAD2.1 and OeFAD2.2 transcription increase in the embryo of the cold-acclimated drupe, but linoleic acid decreases
The transcription of OeFAD2.1, OeFAD2.2 and OeFAD7 was investigated in the embryo before (14 WAF) and after (22 WAF) mesocarp cold acclimation.
FAD2.1 and FAD2.2 transcripts were present at 14 WAF at higher abundance than FAD7 transcripts (P < 0.0001) (Fig. 8a–c). The transcript abundances of the three FADs increased at 22 WAF, albeit to a different extent, that is, FAD2.1 and FAD2.2 transcripts were c. 400-fold and 130-fold, respectively, higher than at 14 WAF, and FAD7 transcripts were c. 40-fold higher. Even if increased, FAD7 transcript abundances were c. 1/80 and 1/17 of FAD2.1 and FAD2.2 abundances, respectively, and C18:3 (i.e. the product of FAD7 activity) was undetectable. By contrast, C18:2 (i.e. the product of the activity of the two FAD2s) was detected as early as 14 WAF (Fig. 9), suggesting that OeFAD2 transcription, and translation, had begun before this time point. C18:2 was c. 1/6 of C18:1, the main unsaturated FA present in the embryo (Fig. 9). At 14 WAF, about half of the content of the two FAs was present in the polar lipid (PL) fraction, including membrane phospholipids (Table 5). The enrichment in phospholipid unsaturation, which results in membrane fluidification (Martinière et al., 2011), was in accordance with the epifluorescence results (Figs 4a, 5a) in sustaining the embryo cold tolerance. At 22 WAF, the strong increases in OeFAD2 transcripts (Fig. 8a–b) did not couple with an increase in C18:2 content, which, by contrast, was significantly (P < 0.05) reduced (Fig. 9). Significant reductions occurred in all eluted fractions (i.e. free FA (FFA), P < 0.05; TAG, P < 0.01; and PL, P < 0.05), even if the PL fraction remained that with the highest percentage of C18:2 (P < 0.01, difference with the other fractions, Table 5). C18:1 also content decreased, and significantly (P < 0.01) in the PL fraction (Table 5).
Table 5. Comparison of unsaturated fatty acids (FAs), that is, oleic acid (C18:1) and C18:2 (linoleic acid), present in the lipid fractions (i.e. free FA (FFA), triacylglycerol (TAG), and polar lipid, (PL) fractions) of embryos and seed coats from drupes at 14 and 22 weeks after flowering (WAF)
Mean area ± SD
Mean% ± SEM
Mean area ± SD
Mean% ± SEM
Mean area ± SD
Mean% ± SEM
Mean area ± SD
Mean% ± SEM
Values from gas-chromatography (GC) analysis are expressed as mean (± SD) areas mg–1 DW. The percentage distribution (means ± SEM) of C18:1 and C18:2 in the eluted fractions is also shown per WAF and seed compartment, setting C18:1 and C18:2 values in the total FA fraction (Fig. 9) equal to 100%. Data from three independent determinations. Linolenic acid (C18:3) present only in traces. Statistical comparisons within the same seed compartment and WAF. *, P <0.05; **, P <0.01; differences with the other values within the same column. Values not followed by an asterisk are not significantly different.
Two 13-lipoxygenases (13-LOXs), both utilizing C18:2 as substrate, have been identified in the seeds of cv Arbequina and Picual, with the transcripts of Oe2-LOX2 showing differences related to the cultivar cold responsiveness (Padilla et al., 2009). To verify whether the C18:2 reduction in the embryo was the result of LOX activity for producing oxylipins as mediators in the stress response, as in other cases (Stelmach et al., 2001), the transcription of Oe2LOX2 and the activity of the coded enzyme were investigated. Oe2-LOX2 expression increased significantly (P < 0.05) at 22 WAF compared with 14 WAF (Fig. 10), but no enzyme activity was detected (Table S2). To exclude that the coded LOX was inactive in Frantoio fruit, it was also measured in the mesocarp and, in contrast with the embryo, an increase was observed in the presence of cold acclimation (Table S2).
OeFAD2.2 transcription increases in the seed coat of the cold-acclimated drupe, and linoleic acid content increases accordingly
As in the embryo, FAD2.2 levels rose 72-fold in the seed coat at 22 WAF in comparison with 14 WAF (Fig. 8b), whereas OeFAD2.1 of 1.5-fold (Fig. 8a), and OeFAD7 was always low (Fig. 8c). Low abundances of FAD7 transcripts were coincident with the absence of C18:3. At 22 WAF, the increased expression of FAD2.2 and the high transcription of FAD2.1 positively coupled with C18:2 production, which was significantly (P < 0.01) enhanced in comparison with 14 WAF (Fig. 9). The increases in C18:2 significantly occurred in the FFA (P < 0.01) and TAG fractions (P < 0.05), but the highest percentage of C18:2 was in the FFA fraction (P < 0.01 with the other fractions; Table 5).
C18:1 was also increased (P < 0.01) at 22 WAF in comparison with 14 WAF (Fig. 9), and in the FFA and PL fractions, in particular (Table 5). The latter increase (P < 0.05 in comparison with 14 WAF) suggests a partial fluidification of seed coat membranes caused by an enhancement in C18:1-phospholipids, possibly resulting in a reduction in the capacity to respond to cold shocks. In accordance, both the peaks in [Ca2+]cyt and the F-actin disassembly were reduced at 22 WAF in comparison with 14 WAF (Figs 4b, 5b). Because the highest C18:2 content occurred in the FFA fraction at 22 WAF (Table 5), the possibility that C18:2 could be the 13-LOX substrate was investigated. The transcription of Oe2-LOX2 did not change between 14 and 22 WAF (Fig. 10), and no LOX activity was detected (Table S2), excluding the use of C18:2 for a 13-LOX-mediated oxylipin formation. Also C18:1 significantly (P < 0.01) increased in the FFA fraction at 22 WAF (Table 5). Both FAs are components of seed cutin (Järvinen et al., 2010). They were possibly produced for the building up of ‘cutinsomes’ for the endosperm, in accordance with the histological results (Fig. 3k–m).
Osmotin protein is active in the noncold-acclimated seed coat
In the embryo, the transcripts of the O. europaea Osmotin gene (OeOsm) and OeNTL6 (NAC domain) (i.e. the possible OeOsm inducer (see 'Introduction')), were already present at 14 WAF, and their abundance did not change at 22 WAF (Fig. 11a–b). The immunolabelling technique revealed the absence of osmotin protein in the embryo at both stages of drupe maturation (Fig. 12a), suggesting an OeOsm down-regulation after transcription (Figs 11a–b, 12a, in comparison).
In the seed coat, the transcripts of OeOsm and NAC were already present at 14 WAF, as in the embryo, and at a similarly low abundance (P < 0.05 difference with the embryo). However, in contrast to the embryo, both the transcripts increased c. eight- and fivefold, respectively, at 22 WAF (Fig. 11a–b). The immunolabelling technique showed the absence of osmotin in the seed coat and in the endosperm at 14 WAF (Fig. 12b), but presence at 22 WAF (Fig. 12c–g). In the seed coat, the signal was randomly present in all layers (Fig. 12c), and, in particular, in the lumen and cell walls of the inner epidermal cells (Fig. 12d, large arrows, and e). In the outer endosperm, the osmotin signal was localized in the lumen of scattered groups of cells (Fig. 12e–f), and in the cell walls of the outermost layer cells in particular (Fig. 12d, small arrows, and e). In the inner endosperm, the immunolabelling was present in the cell lumen (Fig. 12g). The results suggested that NAC was positively related to OeOsm transcription, that a post-transcriptional control of osmotin was present at 14 WAF, as for the embryo, and that this control was inactivated, giving a functional protein, at 22 WAF, in contrast with the embryo (Fig. 12a, c–g, in comparison).
The results show absence of cold acclimation in the seed coat, and presence of cold tolerance in the embryo, as indicated by the presence of cold-induced [Ca2+]cyt signalling and F-actin cycling in the former and not in the latter. OeFAD2 transcripts increase in both the seed coat and embryo with exposure of the drupe to temperature-lowering, but the enzyme activity results in increases in C18:2 in the seed coat only. The production of C18:2 and of C18:1 in the seed coat seems related to cutin synthesis. Osmotin occurs in the endosperm and the seed coat, and is absent in the embryo. Its localization couples with cutinsome extrusion from the seed coat towards the endosperm.
Embryo cold tolerance is independent of mesocarp acclimation
The results show that the embryo was cold-tolerant from 14 WAF, that is, many weeks before the completion of drupe maturation and cold acclimation (Matteucci et al., 2011; and present results). An early acquisition of cold acclimation by the embryo is excluded, because it experienced no temperature-lowering before WAF 14. An early acquisition of dormancy is also excluded, because dividing cells were present at 14 WAF, whereas dormancy implies a growth arrest (Finch-Savage & Leubner-Metzger, 2006). Moreover, the embryo cells were not thermotolerant, showing transient increases in [Ca2+]cyt induced by heat shocks at both 14 and 22 WAF. Had dormancy been acquired, unresponsiveness to heat shock would be observed. In addition, the transcription of SPATULA, a marker of seed dormancy (Penfield et al., 2005), was undetectable at both 14 and 22 WAF.
Dormancy is preceded by desiccation tolerance (Finch-Savage & Leubner-Metzger, 2006), and, in Arabidopsis, RAB18 transcripts accumulate just before desiccation (Nylander et al., 2001). Desiccation tolerance reduces membrane fluidification (Leprince et al., 1993). The present results show that OeRAB18 transcription increased many-fold at 22 WAF compared with 14 WAF, and, by coincidence, C18:1 and C18:2 decreased in the PLs, including membrane phospholipids, supporting the idea that the embryo had also become desiccation-tolerant at 22 WAF.
Osmotin is a cold-acclimating protein in olive tree vegetative organs (D'Angeli & Altamura, 2007); however, we show that it did not function in the embryo, even if its gene was expressed, together with its inducer OeNTL6 (NAC domain). A down-regulation after transcription is possible, as in tobacco seeds (LaRosa et al., 1992). In olive trees, even cold-exposed meristems show an absence of osmotin (D'Angeli & Altamura, 2007). Embryo cells share with meristem cells gene expression related to stem specification (Savona et al., 2012). Osmotin is known to induce PCD, for example, in olive tree leaves (D'Angeli & Altamura, 2007). PCD is not compatible with the stem condition, because genes (e.g. SERKs), involved in stem specification/maintenance in embryos and meristems are negative regulators of PCD (He et al., 2007; Savona et al., 2012). Taken together, OeOsm may be post-transcriptionally controlled in both olive tree embryos and meristems for avoiding PCD induction by the protein.
A positive relationship between [Ca2+]cyt signalling and OeFAD2.1, OeFAD2.2 and OeFAD7 transcription in noncold-acclimated mesocarp cells has been found using the calcium ionophore A23187 (M. Matteucci et al., unpublished), indicating that cold-induced [Ca2+]cyt changes trigger the expression of FAD genes. By contrast, we show that OeFAD2.1, OeFAD2.2, and OeFAD7 transcripts were present in the embryo independently of [Ca2+]cyt signalling. ABA might have induced FAD gene expression, because it regulates the expression of a stearoyl-acyl carrier protein FAD, catalysing the stearic to C18:1 conversion, in cv Koroneiki embryos (Haralampidis et al., 1998). FAD2.1/FAD2.2 and the stearoyl-acyl carrier FAD might be important for the membrane fluidification necessary to embryo cold tolerance. In fact, at 14 WAF, very high concentrations of their products (i.e. C18:1 and C18:2) were observed, and in the PLs, in particular. By contrast, C18:3 was undetectable, possibly by a post-translational control on OeFAD7, as occurs for GmFAD7 in soybean seed (Andreu et al., 2010).
C18:1 and C18:2 reduction at 22 WAF, possibly caused by the acquisition of desiccation tolerance, highlights the possibility that the embryo needed the cooperation of the other seed compartments for maintaining cold tolerance in time.
The seed coat produces cutin for the endosperm as a way to protect the embryo
The continuous presence of [Ca2+]cyt transients and F-actin cycling show that, differently from the embryo, the olive tree seed coat maintains the capacity to respond to cold, that is, it does not become cold-acclimated.
The inner epidermis of the seed coat showed interesting features of secretory epithelium, able to extrude cutin-like materials. It has been proposed that the cutin monomers, that is, complex mixtures of lipid polyesters, might be assembled through an intracellular process, with larger oligomers requiring exocytosis of OBs (Pollard et al., 2008). These aggregates were named cutinsomes (Heredia-Guerrero et al., 2008), and their presence as the building units of seed cutin has been demonstrated in numerous berries (Järvinen et al., 2010). We show that OBs were present in the seed coat. Considering that starch deposition also occurred in the seed coat contributing to its reserves, it is possible that at least part of the OBs were not involved in storing. In grapefruit inner seed coat, the C18 family of monomers constitutes the major portion of cutin (Espelie et al., 1980), and in Arabidopsis seed coat inner integument, ω-hydroxy-FAs, derived from C18:1 and C18:2, are observed in the cutin (Molina et al., 2008). The essential role of FAD2 in providing precursors for epidermal polyester synthesis is sustained by the Arabidopsis fad2.2 mutant, exhibiting altered leaf cutin composition (Bonaventure et al., 2004). Taken together, the increased production of free C18:1 and, primarily, C18:2 in the seed coat at 22 WAF might be functional to cutin build-up. Following this hypothesis, the seed coat needs to remain cold-responding for the maintenance of active [Ca2+]cyt signalling-induced transcription of OeFAD2.1 and OeFAD2.2, with this expression necessary to sustain FAD2 activity for cutin production in time. Since there was no cuticle in the inner seed coat epidermis, but that cuticle was present in the outermost endosperm layer, and its thickness increased in time, we suppose that the FA-derived cutin polymers were needed for endosperm cuticle, in accordance with epifluorescence imaging.
Osmotin affects cold-induced PCD. The protein was present in the seed coat and endosperm at 22 WAF, as in the cold-acclimated mesocarp (Fig. S3). Since PCD occurred in the seed coat and in the endosperm from 22 WAF, as in the mesocarp (Matteucci et al., 2011), but not in the embryo, the protein might occur in tissues in which PCD must function. Its preferential localization in the seed coat internal epidermis, that is, the layer involved in cutinsome formation and extrusion towards the endosperm, might suggest a secondary role for osmotin in lipid trafficking. Osmotin belongs to the PR-5-proteins called permatins because are thought to create transmembrane pores, and it is known to cause an increase in membrane permeability leading to cell leakiness (Abad et al., 1996). The possible role of osmotin as a lipid-transfer protein might be supported by the presence of osmotin-like proteins in tomato berry cuticle (Yeats et al., 2010). In the Arabidospsis line silenced in the DSO/ABCG11 transporter, which is necessary for endosperm cuticle formation, the expression of a large number of genes involved in the export of cuticular lipids from the plasma-membrane is suppressed, but osmotin and NAC are overexpressed (Panikashvili et al., 2010). Thus, osmotin and NAC overexpression might be an alternative to DSO/ABCG11 silencing, and, accordingly, no dramatic change in mutant flower cutin was observed (Panikashvili et al., 2010).
A model summarizing events activated by cold in the seed coat of olive seeds contained in cold-acclimated drupes is proposed in Fig. 13.
In conclusion, a crosstalk between seed coat and embryo, involving the outer endosperm, seems to exist and to be aimed at maintaining embryo cold tolerance.
We thank S. Errico and G. Perrotta of UTT Trisaia ENEA (Italy) for bioinformatics support, and C. Fanelli, A. A. Fabbri, M. Reverberi, A. Ricelli, and C. Bello of Sapienza University (Italy) for helpful discussions and support in the LOX and FA analyses.