Morphological and cytological development and starch accumulation in hermaphrodite and staminate flowers of olive (Olea europaea L.)
- First Online:
- Cite this article as:
- Reale, L., Sgromo, C., Ederli, L. et al. Sex Plant Reprod (2009) 22: 109. doi:10.1007/s00497-009-0096-1
- 301 Views
In olive (Olea europaea L.), the formation of functionally staminate flowers rather than fully functional hermaphrodites is one of the major factors limiting fruit set, as flowers with aborted pistils are incapable of producing fruit. Studies conducted on various angiosperm species have shown a correlation between flower abortion and starch content. Thus, it is important to know if starch content plays a role in regulating pistil development in olive and if so, what mechanism regulates starch distribution. Cyto-histological observations of staminate and hermaphrodite olive flowers show that pistil development in staminate flowers is interrupted after the differentiation of the megaspore mother cell. At that stage, starch grains were only detected in the ovary, style and stigma of the hermaphrodite flowers. No starch was observed in the pistils of the staminate flowers. This finding suggests a tight correlation between starch content and pistil development. The secondary origin of starch within the flower is indicated by low chlorophyll content in the gynoecium, undetectable Rubisco activity in the pistils of these two kinds of flowers and by the ultrastructure of the plastids observed by transmission electron microscope analysis. The plastids have few thylakoid membranes and grana and in the staminate flowers appeared very similar to proplastids. Considering differences in starch content between staminate and hermaphrodite flowers and the secondary origin of the starch, differences in pistil development in the staminate and hermaphrodite flowers could be related to differences in the sink strength of these two types of flowers.
KeywordsOlea europaeaStaminate and hermaphrodite flowersPistil abortionPlastidStarch
Fruit set, more than flower differentiation, is the limiting factor for the yield in most angiosperms. In Citrus, the number of fruit harvested is <2% of the differentiated flowers (Monselise 1986) and in olive only 2% of the floral population sets fruit (Troncoso et al. 1978). Abscission is an important phenomenon that helps adjust fruit number to the metabolite supply of the tree. The production of an excess of female flowers provides the possibility for controlling the number and quality of fruit depending on environmental conditions and nutrient availability. The number of pollinated flowers, the total amount of resources, the distribution of the resources among fertile branches and the amount of damaged fruit seeds are unpredictable at anthesis, so the production of surplus flowers and fruit may provide plants with a choice of which offspring should mature (Stephenson 1981).
Flower and fruit abscission is influenced by environmental conditions, hormonal balance, particularly of ethylene and auxin, and nutrient supply. Developing flowers need carbohydrates from the floral initiation stage (Yu et al. 2000) to the complete differentiation of the floral organs (Clément et al. 1996; Rodrigo et al. 2000) and fruit setting (Iglesias et al. 2003). A nutrient deficiency in the inflorescences can be a putative cause of flower abscission such as demonstrated in grapevine (Lebon et al. 2004); in cv. Pinot noir, particularly, carbon metabolism is thought to be involved in the process of floral necrosis (Gu et al. 1996). Sugar supplied to the reproductive structures comes from reserves in the root and stem (Nishizawa and Shishido 1998) or from photosynthesis that occurs in mature leaves or in the flowers themselves. Reproductive structures may be partially independent of carbohydrate production and photosynthetic capacity is not the same in all species; however, photosynthetic activity of flowers is usually not sufficient to assure flower and fruit development. In Ambrosia trifida L., the net photosynthesis by reproductive structures amounts to 41 and 51%, respectively, of the carbon required for male and female inflorescences (Bazzaz et al. 1979), and Antlfinger and Wendel (1997) reported that in Spiranthes cernua Rich the reproductive structure produces only half of the carbon required for cell construction and maintenance. Studies conducted on pepper flowers by Aloni et al. (1996) suggest that photosynthesis is an important factor in determining pepper flower abortion since shading flowers enhanced floral abortion in several pepper cultivars; however, differences in photosynthetic rates alone do not explain the differences observed in cultivar susceptibility to shading. During flower and fruit differentiation a significant proportion of the total energy budget must be supplied by reserve storage structures and this depends on the sink strength of the flower or fruit, which may be defined as the capacity of the phloem in the sink region to import assimilates from other parts of the plants and release the imported substances into the sink apoplast.
In Citrus, flower abortion within 30 days of flowering onset seems to be caused by a low sink strength, rather than limited carbohydrate supply because reserves are at a maximum in that period (Ruiz et al. 2001). In olive (Olea europaea L.), flower abscission is associated with the presence of staminate flowers, characterized by an androecium that is completely differentiated and an aborted gynoecium or one that is not completely differentiated; pistil abortion can occur at varying stages of gynoecium differentiation and the aborted pistil was smaller and had ovaries without ovules or ovules without embryo sacs in functionally staminate flowers (Reale et al. 2006). These staminate flowers abscise during or immediately after anthesis and do not contribute to fruit set. The number of staminate flowers is directly correlated with the total number of flowers for each inflorescence, since a higher total number corresponds to a higher number of staminate flowers (Reale et al. 2006). Staminate flowers are produced when an optimum resource investment in female function is reached; under those circumstances, secondary pedicels of the panicle where fruit set is unlikely are poorly nurtured and form staminate flowers in which the male function is unaffected. Uriu (1953) demonstrated that high leaf/flower ratios and nitrogen fertilization promote hermaphrodite flowers, which is an observation that is in agreement with other andromonoecious species in which favorable environmental conditions promoted the trend toward femaleness (Solomon 1985; Emms 1993). Observations on Prunusarmeniaca L. suggested that starch stored in the ovary at anthesis is involved in ovary development (Rodrigo et al. 2000). Similar research conducted on Lilium (Clément et al. 1994, 1996), Triticum aestivum L. (Dorion et al. 1996) and Oryza sativa L. (Sheroan and Saini 1996) showed that conspicuous starch reserves are associated with anther growth and pollen development. Due to this correlation between starch and flower development, it was of interest to study the development of staminate flowers in olive and particularly their starch accumulation in relation to the occurrence of hermaphrodite flowers. Differences in the metabolic activities of these two kinds of flowers could serve as the basis for future attempts to increase efficiency in flower fertilization and olive production. In this paper, the development of staminate and hermaphrodite flowers is described based on morphological and cyto-histochemical analyses, and flower development was correlated with variations in starch content in staminate and hermaphrodite flowers. A study was also made to determine if the starch detected was the product of photosynthetic activity of the floral organs or if it had been produced in other parts of the plant and stored in floral organs. Data obtained suggest that starch accumulation is not correlated with photosynthetic activity of the flower structures, so we hypothesize that differences in sink strength occur between hermaphrodite and staminate flowers.
Materials and methods
Plant materials and growth condition
Field observations and experiments were conducted during a 4-year period (2004–2007) in an olive grove of the “Istituto Sperimentale di Olivicoltura” of Spoleto located in Castel Ritaldi, near Spoleto in southern Umbria (42°49′N, 12°40′W) at 300 m a.s.l. Olive trees of the cultivars Leccino and Dolce Agogia were used.
Cyto-histological observations were carried out on flowers collected at different times of flower development (samples were collected weekly from 6 weeks before anthesis to anthesis). Observation with a stereomicroscope was used to distinguish between pistils of hermaphrodite and staminate flowers.
To obtain semi-thin and ultra-thin sections, flowers were fixed in 3% (w/v) glutaraldehyde in 0.075 M phosphate buffer, pH 7.2, for 5 h. The samples were then washed four times for 15 min each in 0.075 M phosphate buffer, pH 7.2, and post-fixed in 1% (w/v) OsO4 for 1.5 h. Samples were then dehydrated in increasing concentrations of ethanol and embedded in resin (Epon, 2-dodecenylsuccinic anhydride, and methylnadic anhydride mixture). Semi-thin sections (1–2 μm) were cut with an ultramicrotome (OmU2, Reichert, Heidelberg) equipped with a glass blade, stained with toluidine blue and observed using a light microscope (DMLB, Leica, Wetzlar, Germany). Ultrathin sections (~800 Å) were cut and stained with a saturated solution of uranyl acetate in ethanol (50%) and lead citrate; sections were observed under a transmission electron microscope (TEM 208, Philips).
To obtain paraffin sections, flowers were fixed in FAA medium (50 ml absolute ethanol, 5 ml acetic acid, 10 ml 37% formaldehyde and 35 ml distilled water), dehydrated in increasing concentrations of ethanol and embedded in liquid paraffin medium (Paramat, VWR International). Paraffin sections were stained with safranin and fast-green solutions (Johansen 1940).
Evaluation of starch content
To evaluate changes in starch deposition during flower development, 50 slides with paraffin sections and 25 slides with semi-thin sections of flowers at different developmental stages were observed. The following stages were considered: early pre-meiosis (from anther differentiation to the appearance of microspore mother cells); late pre-meiosis (microspore mother cells surrounded by callose wall); meiosis (from microspore mother cells undergoing meiosis to tetrad formation); post-meiosis (presence of young microspores) and anthesis (from anther dehiscence to petal drop). Paraffin was removed from paraffin sections using three washes of 10 min each in xylene. For semi-thin sections, resin was removed by immersion in sodium methoxide for 7 min. Sections were then rehydrated and treated with iodine–potassium iodide solution (Johansen 1940) and observed under a light microscope; the presence of starch grains was indicated by a blue-dark color.
The periodic acid Schiff’s (PAS) reaction (O’Brien and McCully 1981) was also used to check starch content and confirm data of iodine–potassium iodide reaction. Semi-thin sections were placed in 0.5% periodic acid for 30 min, rinsed with tap and demineralized water and placed in Schiff’s reagent for 15 min; the sections were then washed with tap water, SO2 water and finally with demineralized water. Slides were observed using a light microscope; the presence of starch was indicated by a red color.
Extraction and assay of Rubisco
Frozen leaves or ovaries (approximately 0.1 g fresh weight) collected just before anthesis were ground in a mortar and pestle and chilled by liquid nitrogen with 2 ml of 100 mM bicine pH 7.8, 20 mM MgCl2, 5 mM dithiothreitol (DTT), 1 mM ethylenediaminetetraacetic acid (EDTA), 10 mM NaHCO3, 2% (w/v) polyvinyl polypyrrolidone (PVPP) and 0.02% (w/v) bovine serum albumin (BSA). The extract was centrifuged at 13,000×g for 2 min at 4°C and the supernatant was immediately used to determine the activity of Rubisco. Spectrophotometric measurements were read in a cuvette containing 50 mM Bicine (pH 8.0), 1 mM EDTA, 15 mM MgCl2, 20 mM NaCl, 9.2 mM DTT, 9.2 mM NaHCO3, 0.5 mM ATP, 4.6 mM phosphocreatine, 1.3 U ml−1 creatine phosphokinase, 47 U ml−1 glyceraldehyde-3-phosphate dehydrogenase, 47 U ml−1 3-phosphoglycerate kinase, 0.1 mM NADH. The mixture assay was incubated for 3 min at 25°C and the reaction was started by adding 30 mM RuBP. The Rubisco activity was measured spectrophotometrically by determining the absorbance at 340 nm. For other details see Sharkey et al. (1991).
Liquid N2-frozen leaves or ovaries collected just before anthesis were ground in 80% acetone and the homogenate was centrifuged at 13,000×g for 5 min. The supernatant was used for chlorophyll quantification by measuring Abs 645 and 663 nm, considering that chla (mg l−1) = (12.7 × Abs 663) − (2.69 × Abs 645) and chlb (mg l−1) = (22.9 × Abs 645) − (4.68 × Abs 663) (Arnon 1949). The mean values were calculated from data collected from cvs. Dolce Agogia and Leccino; the mean values were analysed by Anova test at probability for no difference P ≤ 0.01.
Hermaphrodite and staminate flower development
Previous studies (Reale et al. 2006) have shown that, in olive, micro- and megagametogenesis in a flower are paired and, if various cultivars are grown under the same environmental conditions, the time interval between archesporial cell identification and micro- and megagametophyte appearance was the same in all cultivars examined. The major difference was the time of flowering onset, which was cultivar-dependent. In this study two olive cultivars, Leccino and Dolce Agogia, were characterized with respect to different aptitudes for staminate flower differentiation; as previously demonstrated (Reale et al. 2006) the percentage of staminate flowers is higher in Dolce Agogia than in Leccino. Histological studies did not show any cultivar-based differences concerning the development of hermaphrodite and staminate flowers. Staminate flowers resulted from pistil abortion at varying stages of gynoecium differentiation (Reale et al. 2006); in this research we considered only the staminate flowers with ovules differentiated but without embryo sacs.
Differences between hermaphrodite and staminate flowers start to appear during megaspore mother cell differentiation. Four or five weeks before anthesis, inflorescences are less tightly packed, branches are more evident, and bracts have started to abscise (Fig. 1d). The exothecium, transition layers, endothecium and tapetum differentiate in the anthers of both kinds of flowers, and the microspore mother cells differentiate but are still very close to each other (Fig. 1e). At this stage, carpels are not yet completely closed to form the style and stigma; the nucellus and integument are evident in the ovules (Fig. 1e).
As the microspore mother cells become surrounded by the callose special wall, the carpels become enclosed within the hermaphrodite flowers (Fig. 1f). Style and stigma are readily identifiable and the megaspore mother cells with callose (arrowhead) are evident in the ovules. In the staminate flowers, pistil differentiation becomes arrested at this stage, with an identifiable, but poorly formed stigma (Fig. 1g).
Before anthesis, differences between hermaphrodite and staminate flowers become macroscopically evident and observations of hermaphrodite flowers show that the pistil was green, turgid and completely differentiated (Fig. 1h). In contrast, the staminate flowers show a sparse, necrotic style near anthesis, and a sketchy stigma with a small, yellow, withered ovary (Fig. 1i). Cyto-histological observations of the ovules of hermaphrodite flowers at this stage show the presence of a completely differentiated bisporic (Allium type) embryo sac (Johri et al. 1992) (Fig. 1j) with hooked synergids (Fig. 1k). At the same time in staminate flowers (Fig. 1l), the megaspore mother cell had not undergone meiosis. In both kinds of flowers, however, microgametogenesis in the anthers proceeded without obstacles to form tricolporate pollen grains.
The starch content in hermaphrodite and staminate flowers was examined at different stages of development by staining with iodine–potassium iodide solution and PAS reaction with similar results on starch abundance and distribution. According to micro- and megagametogenesis, five stages of flower development were considered, in which no differences were observed between cvs Dolce Agogia and Leccino; for each stage, 15 slides corresponding to 15 different flowers were observed.
Just after the differentiation of the tightly packed microspore mother cells, starch grains were also evident in the exothecium (Fig. 2c), which was formed by cells with large vacuoles and a thick cuticle along the external periclinal wall.
After callose deposition around the microspore mother cells, starch grains were evident in all the layers of the anther wall (Fig. 2d) and in the petal and sepal, particularly on the adaxial side of apical and basal portions. In the pistil of the hermaphrodite flowers, starch grains were observed in the stigma, style, ovary wall and ovule integument (Fig. 2e), while in the pistil of the staminate flowers, no starch was detected (Fig. 2f).
At this stage, starch grains were always present in the peduncle/receptacle, sepal and petal (basal and apical portions). In the anthers, starch was observed in the exothecium, endothecium and transition layers, but not in the tapetum, dyad or tetrad (Fig. 2g). In hermaphrodite flowers, starch was detected in the style, stigma, ovary wall and ovule integument, but not in the nucellus (Fig. 2h). No starch grains were observed in the pistil of staminate flowers.
The starch distribution was the same as that observed during the meiotic stage. Starch was also present in the anther, where microsporogenesis was interrupted and the microspore mother cells or young microspores were degenerated.
At anthesis, observations of the anther in both kinds of flowers showed thickening of the radial wall of the endothecium cells and the degeneration of the protoplast in exothecium cells and transition layers. Bicellular pollen grains at the dispersal stage were rich in starch grains, but no starch was detected in the anther wall (Fig. 2i). In the pistil of hermaphrodite flowers, starch grains were evident in the cells of the stigma, style, ovary wall and ovule integument, and particularly in the chalazal and micropylar portions of the ovules (Fig. 2j). No starch grains were observed in the nucellus, which at this stage was greatly reduced. In staminate flowers, the starch distribution in the stamens was the same as that observed in hermaphrodite flowers, but no starch grains were found in any of the pistil portions (Fig. 2k). The ovules of staminate flowers did not differentiate embryo sacs. Starch was observed in the petal and receptacle cells.
In the staminate flowers, the plastids observed in the peduncle/receptacles could not be distinguished from those observed in the hermaphrodite flowers; no starch grains were observed in the cells of ovary wall or in the style or of the pistils and only a few, very disorganized thylakoid membranes were observed (Fig. 3c). Proplastids were sometimes observed in staminate flower pistils (Fig. 3d).
Chlorophyll and Rubisco quantification
The differentiation of staminate flowers in olive is one of the limiting factors of fruit set because these flowers abort precociously and do not produce fruit. Staminate flowers are characterized by a well differentiated androecium but poorly and incompletely differentiated gynoecium. Cyto-histological observations showed that during the first stages of flower development, until microspore and megaspore mother cell differentiation, staminate and hermaphrodite flowers are indistinguishable. The first differences in pistil development were observed as the microspore mother cells were enveloped by a callose wall. At this time, it was possible to distinguish identify a pistil with an ovary, style and stigma in the hermaphrodite flower. In contrast, in the staminate flowers, pistil differentiation had stopped, and although the ovary and the style were present, the stigmatic portion was not well defined. Later, mature embryo sacs were formed in hermaphrodite flowers and fertilization was possible; however, the gynoecial structures of staminate flowers maintained similar characteristics to those observed before microsporogenesis. That failure of megaspore mother cell differentiation and the successive meiosis seems to mark a divergence point in flower development, commitment to meiotic division may mark the decision point. This has already been observed in Vitis, where the first hallmarks of flower abscission are the abnormalities in ovary development during megasporogenesis (Fougere-Rifot et al. 1993). A strong correlation has often been noted in woody plants between the successful development of fertile sexual organs and the amount of starch available in the flower at various stages of development (Rodrigo et al. 2000; Jean and Lapointe 2001; Ruiz et al. 2001). Histochemical observations to study the pattern of starch synthesis and storage in hermaphrodite and staminate olive flowers showed that starch appeared early in the sepal and petal of both kinds of flowers. Successively, starch grains were observed in the anther and just before meiosis in the pistil of the hermaphrodite flowers; whereas no starch grains were detected in the pistil of the staminate flowers. The differences observed in starch content between hermaphrodite and staminate flowers in the late pre-meiosis stage closely correlates with the arrest of pistil development in staminate flowers. These observations emphasize the important role of megasporogenesis in pistil differentiation in olive flowers and suggest that starch accumulation is a key indicator of success in pistil differentiation and confirm that the absence of starch in flower abscission is an indicator of abortion, as suggested by research conducted with pepper (Aloni et al. 1996; Marcelis et al. 2004), apricot (Rodrigo et al. 2000) and grapes (Lebon et al. 2004). In apricot, some flowers in both pollinated and non-pollinated populations continued to grow, whereas the development of others was arrested (Rodrigo et al. 2000). This behavior is correlated with starch content. The ovary reserves of different flowers are not all exhausted at the same time, and the depletion of these reserves prior to availability of new photoassimilates could be a critical factor determining fruiting. Aloni et al. (1996) showed that pepper cultivars with differing susceptibilities to flower abscission may differ in their capacity to produce sucrose and accumulate starch during the light periods until pollination takes place. An accumulation of reserves in the flowers of these cultivars would be required for the dark period during which starch is remobilized (Preiss 1982). If starch quantities are too low, as in susceptible pepper cultivars, night respiration could deplete a large part of the flower carbohydrate reserves and, thereby, cause abscission.
Although starch accumulation was consistently observed in the anthers of hermaphrodite and staminate flowers and patterns of accumulation in both flowers proceeds through consistent patters of stamen development in which starch grains first appear in the exothecium, just after its differentiation, and successively in the all layers of the anther wall until anthesis, with exception of the exothecium and in the endothecium. No starch grains were ever observed in the tapetal cells, such as the nucellus cells, in either kind of flower; these tissues have a strong respiratory activity (Felker et al. 1983) and consequently do not store starch. In olive, starch accumulation in the anthers is not strictly correlated with success in the microgametogenesis process. In anthers where microgametogenesis was arrested due to the degeneration of the microspore mother cells or young microspores, starch grains were still observed in the anther wall. No differences were observed in wheat (Dorion et al. 1996) or in male-sterile soybean (Smith et al. 2002) between anthers that contained sterile pollen grains and those containing fertile pollen grains. In these plants, differences in the pollen grains were only observed after the starch disappeared from the anther wall. In the olive flower anthers, where microgametogenesis was not completed and pollen grains were not formed, it was not possible to detect any differences in starch content in the different layers of the anther wall.
Starch stored in flower organs is usually synthesized from carbohydrates supplied by photosynthesis, which occurs in the green vegetative organs of the plant (Tirosh and Mayak 1988; McConchie et al. 1991). These assimilates reach the flower bud and are distributed among the various floral organs. In some plants, however, floral photosynthesis supplies a significant portion of the carbon requirement of reproduction; Bazzaz et al. (1979) showed that the amount of carbon supplied from in situ photosynthesis with respect to total carbon required for the production of mature seed varied from 64.5% for Acer platanoides to 2.3% in Quercus macrocarpa Endl. The low chlorophyll content and undetectable Rubisco activity which was observed in pistils of hermaphrodite and staminate flowers of olive suggest that the pistil does not have any detectable photosynthetic activity and that the stored starch was of another origin. Plastid ultrastructure, observed by TEM investigations, confirmed this hypothesis. The pistil of the hermaphrodite flowers contained plastids that were characterized by a large quantity of starch grains and did not have well organized thylakoid membranes. In contrast, the plastids in the staminate flower never had any starch grains and had a few, very disorganized thylakoid membranes; proplastids were occasionally observed. Similar plastids were observed in mature Dendrobium flowers (Khoo et al. 1999), in which a decrease in photosynthetic activity during flower development was observed in concomitance with the deterioration of chloroplasts into chromoplasts. These results suggest that in olive differences in starch content between hermaphrodite and staminate flowers and consequent pistil abortion were not correlated with the different photosynthetic activities, rather with different sink strengths. In pepper, and in maize, it is suggested that the inflow of sucrose into the flower increases the flower sink activity, inhibits abscission and induces metabolic changes, that ensure fruit set (Zinselmeier et al. 1995; Aloni et al. 1996; Marcelis et al. 2004). The sucrose inflow is influenced by sucrose synthase and invertase activity; these enzymes cleaved sucrose and maintained a favorable gradient for continued delivery of sucrose from the phloem into the developing flower organs that ensures continuous growth. Sucrose synthase is the most important enzyme for establishing and maintaining reproductive sink strength, in pepper flower and tomato fruit, while in maize, invertase plays the key role.
In olive, further investigation of the activity of these two enzymes and their eventual involvement in pistil abortion is needed to confirm the correlation between sink strength and flower development. Different activities of invertase and sucrose synthase could explain different sink strengths in hermaphrodite and staminate flowers and consequently explain the different starch contents.
This research was supported by a grant from MIUR (National Project, COFIN 2006). The authors thank Prof. Nicoletta Rascio (Department of Biology, University of Padova) for the critical review of this manuscript and Prof. Scott Russell for his useful suggestions.