Morphological and histological characterization of the osmophores and nectaries of four species of Acianthera (Orchidaceae: Pleurothallidinae)
- First Online:
- Cite this article as:
- de Melo, M.C., Borba, E.L. & Paiva, E.A.S. Plant Syst Evol (2010) 286: 141. doi:10.1007/s00606-010-0294-1
- 486 Downloads
Nectar and floral odor are frequently associated with the presence and maintenance of specialized pollination systems in Orchidaceae. We studied flowers of four Acianthera species, a genus of myophilous orchids belonging to the largest fly-pollinated orchid group Pleurothallidinae, in order to characterize the secretory structures related to their pollination mechanism. Flowers at anthesis were sampled to detect volatile compounds and nectar; samples were fixed for light microscopy and scanning and transmission electron microscopy. The labellum presents epidermal cells and the first mesophyll layer involved with secretory processes. Cellular characteristics of these regions associated with the occurrence of sugars allowed us to recognize them as nectaries. Some portions of the sepals also shown to be involved with secretory processes and the presence of nitrogenated volatile compounds characterize them as osmophores. The production of nectar in these species makes the occurrence of sapromyophily questionable, even though these flowers present characteristics of this floral syndrome. The presence of osmophores on the sepals reinforces that this localization is common among the Pleurothallidinae, whilst they occur in a different region (labelum) in the other major fly-pollinated orchid group (Bulbophyllum).
KeywordsAcianthera Cell ultrastructure Myophily Nectaries Orchidaceae Osmophores
The subtribe Pleurothallidinae, with approximately 4,100 species, and the unrelated genus Bulbophyllum Thouars, with approximately 2,100 species are two of the largest fly-pollinated orchid groups. They exhibit some of the most interesting examples of floral convergence due to adaptation to the same group of pollinators (Dressler 1993). Although the flowers of these myophilous orchids share a number of morphological characters, floral biology studies have demonstrated that there is still a large diversity of pollination mechanisms associated with fly-pollination, involving both biotic and wind-assisted biotic mechanisms (van der Pijl and Dodson 1966; Borba and Semir 1998). Among them, different adaptations to attract these insects stand out; some of them are associated with the insects’ feeding and/or reproductive instincts (Borba and Semir 1998, 2001; Singer and Cocucci 1999). Evolution of these differences in many myophilous orchids seems to be related to the great specificity of the plant-pollinator relationship to a degree that was not initially expected.
Odor and nectar are important elements to maintain the plant-pollinator relationship. Floral odor is responsible for long distance attraction while nectar constitutes a principal floral reward (Proctor et al. 1996). Nectar is the principal reward to pollinators among the Orchidaceae (Dressler 1993) and is commonly present in orchids pollinated by anthophilous species of Diptera. The presence of nectar may result in high pollen transfer, as previously demonstrated in epidendroid orchids (Peter and Johnson 2009). Although nectar is not commonly found among sapromyophilic plants, as their pollinators are attracted by oviposition instinct (Faegri and van der Pijl 1979; Proctor et al. 1996), it seems to be an important element in the mechanism of deceit-pollination in some wind-assisted fly-pollinated Bulbophyllum species (Borba and Semir 1998) or in partially deceitful species of Acianthera (Borba and Semir 2001).
The genus Acianthera (Pleurothallidinae) comprises approximately 200 species distributed throughout tropical South America, especially southeastern Brazil (Pridgeon et al. 2005). As with other members of the subtribe, Acianthera in humid forests are predominantly epiphytes, or otherwise, lithophytes on rocky soils exposed to direct sunlight. Their flowers show features typical of myophilous flowers such as diurnal anthesis and, frequently, unpleasant odor. Borba and Semir (2001) studied the pollination biology of a group of Brazilian Acianthera species and found that species pollinated by Phoridae flies have nectar on the labellum, while those pollinated by Chloropidae flies are nectarless and use deceit-pollination. In all these species scent emission occurs in the sepals (E. L. Borba, UFMG, Minas Gerais, Brazil and J. R. Trigo, UNICAMP, São Paulo, Brazil, unpublished data).
Acianthera hamosa (Barb Rodr.) Pridgeon & M. W. Chase, A. limae (Porto & Brade) Pridgeon & M. W. Chase, A. modestissima (Rchb. f. & Warm.) Pridgeon & M. W. Chase, and A. prolifera (Herb. ex Lindl.) Pridgeon & M. W. Chase constitute a group of species that have very similar overall flower morphology. However, they have subtle morphological differences that may be related to the attraction of distinct groups of pollinators (different families of Diptera) and, thus, contribute to their reproductive isolation (Melo 2008). These species have similar flower morphology to those studied by Borba and Semir (2001). Some areas of the sepals and labellum are shiny, and the base of the labellum is frequently moist, which may indicate that these regions can be involved in the release of secretions, like those from osmophores or nectaries (Vogel 1990; Pacini et al. 2003). This may suggest that the papillose regions of the labellum of those species pollinated by Phoridae flies (A. hamosa, A. limae, and A. modestissima) and Chloropidae flies (A. prolifera) are nectaries, even though such structures were believed to be absent in the latter (Melo 2008).
All four species seem to have osmophores located at the adaxial face of sepals. The micromorphology of osmophores in fly-pollinated orchids has been examined only in a few species of Bulbophyllum and Pleurothallidinae before (Pridgeon and Stern 1985; Vogel 1990; Teixeira et al. 2004). These studies have shown that osmophores are generally found in the labellum, in species of Bulbophyllum, or the sepals, in the Pleurothallidinae (Masdevallia, Pleurothallis, Restrepia, and Scaphosepalum) (Vogel 1990; Teixeira et al. 2004). If osmophores occur in different parts of the flowers in other species of the Pleurothallidinae, such as in Acianthera, is unknown. But if it does, this could reinforce the idea of divergence in anatomical and functional aspects of the flowers of Bulbophyllum and Pleurothallidinae species, despite the apparent morphological convergence of their flowers (van der Pijl and Dodson 1966).
In this study we investigated the micromorphological features of the labellum and sepals of four species of Acianthera pollinated by Chloropidae (A. prolifera) and Phoridae flies (A. hamosa, A. limae, A. modestissima) using light microscopy and transmission and scanning electron microscopy. We aim to (1) determine the occurrence of nectary glands and osmophores in flowers of Acianthera; (2) correlate these structures with the pollination mechanism observed in the group; and (3) correlate the morphology and location of these structures with that observed in the two largest myophilous groups in the family, Pleurothallidinae and Bulbophyllum.
Materials and methods
Flowers were collected from plants cultivated at the greenhouse of the Universidade Federal de Minas Gerais. These plants were collected from wild populations previously included in reproductive biology studies by our group (Melo 2008). In the wild these plants are usually found growing on rock outcrops (A. modestissima and A. prolifera) or on the forest floor of gallery forests (A. hamosa and A. limae) in areas of campos rupestres vegetation in Minas Gerais State, Brazil. Voucher specimens were deposited in the herbarium BHCB (A. hamosa, M.C. Melo 08; A. limae, M.C. Melo 07; A. modestissima, M.C. Melo 06; A. prolifera, M.C. Melo 04).
Flowers in full anthesis of all four species were fixed in Karnovsky solution (Karnovsky 1965) for 48 h, dehydrated in an ethanol series, and embedded in 2-hydroxyethyl-methacrylate resin (Leica). Transverse and longitudinal sections of 5 μm were made and stained with 0.05% toluidine blue at pH 4.3 (O’Brien et al. 1964). Cross-sections of fresh sepals and labellum were done by hand and used in histochemical tests using Lugol’s solution to detect starch (Johansen 1940) and Sudan Red B for lipids (Pearse 1980), with their respective controls. These tests were performed in flowers that have been in anthesis for up to 48 h. To detect starch dynamics, the Lugol test was repeated on flowers close to senescence (approximately 10 days in A. hamosa and A. modestissima and 20–25 days in A. limae and A. prolifera). Glycerin was applied to paradermal sections to observe stomatal movements (Jernstedt and Clark 1979). Descriptions of structural characteristics were made, and attention was especially given to cells and tissue in the sepals and labellum likely to be involved in secretory processes.
Scanning electron microscopy
Tissue samples of the lateral sepals and labellum of A. limae, A. modestissima, and A. prolifera were prepared for observation under scanning electron microscopy. Due to low availability of flowers, A. hamosa was not included in this analysis. Tissue was taken from the flowers up to 48 h after the beginning of anthesis, fixed in 2.5% glutaraldehyde (0.1 M phosphate buffer, pH 7.2), dehydrated in an ethanol series, dried to critical point, and subsequently sputter-coated with ca. 10 nm of gold as described by Robards (1978). The samples were examined in a scanning electron microscope, model Quanta 200 (Fei Company, Hillsboro, OR, USA), and all images were processed digitally.
Transmission electron microscopy
Overall, the floral anatomy of all the Acianthera species included in this study is very similar. Therefore, only flowers of A. prolifera were used for the TEM study. This similarity was confirmed by observations of the sepals and labellum under light and scanning electron microscopy and the chemical nature of the substances they secrete. Tissue samples from the lateral sepals and labellum were obtained from flowers up to 48 h after beginning of the anthesis, fixed in Karnovsky solution (Karnovsky 1965) for 24 h, post-fixed in 1% osmium tetroxide (in 0.1 M phosphate buffer, pH 7.2), and processed using standard methodologies (Roland 1978) for observation under transmission electron microscope. Ultra-thin sections were stained with uranyl acetate and lead citrate and examined under a Philips CM 100 transmission microscope at 60 kV. Descriptions were focused on cells and tissues that appeared to be involved in secretory processes.
Preliminary analyses of nectar and volatile compounds
The presence of sugars was detected by thin layer chromatography (TLC) following Dafni (1992). Residual floral secretions on the adaxial surface of the labellum and on the lateral sepals of about ten flowers of each of the four species were soaked in situ with distilled water (2 μl droplet) for 10 min. These samples were then removed and applied to 10 × 10 cm silica gel TLC plates prepared with 0.02% sodium acetate. It is important to emphasize that samples from the labellum and sepals were analyzed separately. Standard marker solutions of fructose, glucose, and sucrose, as well as distilled water (control) were also prepared and applied to the same plates. These plates were run with a mobile phase of chloroform:methanol (6:4), dried at room temperature, sprayed with an orcinol-sulfuric acid solution, and then heated to 120°C for 5 min. Sugars stained as dark-purple bands against a yellow background (Stahl 1988).
The Whiff test (Amsel et al. 1983) was used to detect volatile amines, generally responsible for the unpleasant scent emitted by flowers. The dorsal sepal, lateral sepals and petals, and the labellum of ten flowers from each of the four species were immersed in a solution of 10% KOH, in a closed vessel for 1 min. As control, we used a flask with just the KOH solution and another flask with the same solution but with a piece of leaf from one of the species. Floral parts that gave off fishy odors were considered as sites where volatile compounds are probably emitted.
The labellum mesophyll is parenchymatous, and the superficial layer of the adaxial face is formed by cells with a denser cytoplasm than the internal cells; starch grains are also frequently observed here. There is considerably less starch in the mesophyll cells of flowers during the secretory phase and practically none in the mesophyll of presenescent flowers. Vascular bundles (one dorsal and two laterals) are present near the abaxial face, but no ramifications were observed to the adaxial face (Fig. 1a).
The epidermis of the sepals is uniseriate with juxtaposed cells and covered by cuticle. The epidermis on the adaxial face contains regions that are characterized by the presence of cells with dense cytoplasm and conspicuous nuclei (Fig. 1a, c, d). The epidermal cells on both sides of the sepal are projected slightly upwards, giving the sepals a subtle papillose appearance (Fig. 1c). The mesophyll, and especially the area next to the adaxial face, is formed by a two-to-three-cell layer with large numbers of starch grains (Fig. 1e).
The underlying parenchymatous mesophyll on the adaxial face of the sepals, towards the epidermal cells described above, includes cells with dense cytoplasm, conspicuous nuclei, and numerous plastids with starch grains (Fig. 1c–e). The quantity of starch grains decreases with the age of the flowers. Although starch grains can be observed throughout the mesophyll and epidermis, they are mostly concentrated in the cells containing dense cytoplasm.
Numerous mitochondria, ranging from globose to elongated and with well-developed cristae, were observed in the epidermal cells on the adaxial face of the labellum (Fig. 4b). Rough endoplasmic reticulum was associated with the smaller vacuoles, often surrounding them (Fig. 4b). The plastids were globe-shaped, had poorly developed membrane systems and dense stroma, and contained starch grains (Fig. 4c). Free ribosomes and dictyosomes were also observed here, although they were not frequent (Fig. 4b). The epidermal cells have ample periplasmatic spaces and present vesicles and flocculated material (Fig. 4b), which suggests secretory activity taking place here.
The parenchymatous cells of the mesophyll have a single large vacuole (Fig. 4d). Mitochondria and plastids are also seen in these cells, the latter containing starch grains, especially in young flowers.
Mesophyll cells of the adaxial faces of the sepals show ultrastructural characteristics similar to those seen in the epidermal layer (Fig. 5c, d). These cells are connected to each other and with the epidermis by plasmodesmata. Some plastids contained no starch reserves at this stage, and this was more evident in the cells of the internal layers of the mesophyll. Frequent fusion of plastids or fusion of plastids and vacuoles was also observed at this phase.
Analyses of nectar and volatile compounds
The TLC analyses indicated the presence of sugars on the surface of the labellum of the four species examined but were negative for the sepals. Samples of the labellum yielded bands with Rfs corresponding to sucrose (0.46) and to glucose and/or fructose (0.55). Two bands were observed for A. hamosa and A. modestissima, one intense band for sucrose and one weak band indicating a monosaccharide. Analysis of the samples of A. limae and A. prolifera, on the other hand, showed two bands with similar intensity.
The Whiff test was positive for all four species and strong fishy odor was perceived from the dorsal and lateral sepals. However, odor intensity was stronger in the lateral sepals. No odor was perceived from the labellum or any vegetative organ soaked in KOH for any of the species.
The structural and chemical analyses of the four species provide strong evidence to conclude that some regions of the labellum and the calyx are secretory. These are easily identified by their high cytoplasmic density and starch presence. Also, the ultrastructural analysis of A. prolifera showed an abundance of mitochondria, which are usually found in areas of elevated energetic demands due to secretory processes. The abundance of mitochondria in the tissue-forming part of osmophores and nectary glands in the Orchidaceae has been previously reported (Pridgeon and Stern 1983; Stpiczynska et al. 2005b).
The presence of sugars on the outer surface of the labellum of the four species of Acianthera allowed us to delimit the nectary gland. This was corroborated by the ultrastructural characteristics of the epidermal cells of the labellum in A. prolifera studied under TEM. Meanwhile, the presence of volatile nitrogenized compounds, high cytoplasmic density of the epidermal cells, and the presence of starch in the subepidermal layers allowed us to identify the location of osmophores in the papillose regions of the adaxial face of the sepals, which are larger in the lateral sepals than in the dorsal sepal. These histological and chemical characteristics have been observed before in other studies and seem to be typical of the osmophores in the Orchidaceae and other plant families (Pridgeon and Stern 1983, 1985; Vogel 1990; Garcia et al. 2007).
The overall morphology of the nectary glands examined here is similar to that previously described in the Orchidaceae, and it is formed by a secretory epidermis associated with layers of parenchymatic tissue (Stpiczynska and Matusiewicz 2001; Teixeira et al. 2004; Stpiczynska et al. 2005b; Vieira et al. 2007). The presence of papillae, like the ones observed in this study, or unicellular trichomes that increase the secretion surface has also been described in the nectary glands of other Orchidaceae species (Stpiczynska 1997; Stpiczynska and Matusiewicz 2001; Teixeira et al. 2004; Stpiczynska et al. 2005b).
Idioblasts containing raphides of calcium oxalate have been observed in the sepals and petals of other species of orchids (Stpiczynska et al. 2003, 2005a), and their presence near plant secretory structures has often been noted in diverse taxonomic groups (Schnell et al. 1963). According to Paiva and Machado (2005), the presence of these crystals may be related to the elimination of excess calcium from the cytosol.
The nectaries of the four species of Acianthera examined here are not vascularized, but they are associated with large starch reserves that can be used as energy sources for secretory cells. Starch accumulation in the presecretory phase is commonly seen in the nectaries of orchid species (Figueiredo and Pais 1992; Stpiczynska 1997; Davies et al. 2005; Stpiczynska et al. 2005a, b), as well as in other plants (Nepi et al. 1996; Paiva and Machado 2008). A significant reduction in the levels of starch, such as those observed here, in the tissue of the nectary gland and during the secretory phase, has also been reported in earlier studies (Nepi et al. 1996; Stpiczynska et al. 2005a, b; Vieira et al. 2007; Paiva and Machado 2008).
The secretion mode of the nectaries of A. prolifera can be characterized as granulocrine, in which vesicles with secretion materials derived from the endoplasmic reticulum, the dictyosomes, or both, fuse to the plasma membrane. This process has been observed in the nectaries of other Orchidaceae, such as Hexisea imbricata (Stpiczynska et al. 2005a) and Platanthera bifolia (Stpiczynska 1997). We observed that the cuticle of A. prolifera does not constitute a barrier to nectar flow, a situation that has also been observed in other orchid species (Stpiczynska et al. 2003, 2005a).
The presence of starch reserves in the tissues underlying the glandular epidermis, as observed in the four species of Acianthera, is a common characteristic of the osmophores, including those in Orchidaceae (Vogel 1990). A number of studies have indicated that starch is utilized in these osmophores as a source of both energy and carbon for the synthesis of volatile substances (Vogel 1990). The morphological alterations observed in the plastids suggest the transition of these organelles to become vacuoles has been similar to that observed in the floral nectaries of other plants (Jiang et al. 2002; Gaffal et al. 2007; Paiva and Machado 2008).
The presence of rough endoplasmic reticulum, and the low frequency of dictyosomes in the secretory cells of the osmophores in the species of Acianthera examined here, is compatible with the nitrogen-rich nature of their secretions. Low frequency of dictyosomes was also reported for Restrepia, which likewise has a disagreeable odor associated with amines (Pridgeon and Stern 1983; Vogel 1990).
The secretion of volatile chemical compounds from the osmophores of A. prolifera seems to occur by ecrine processes, with a notable absence of vesicles in the secretory cells, as has been reported for other Pleurothallidinae such as Restrepia and Scaphosepalum (Pridgeon and Stern 1983, 1985). This is unusual because a granulocrine secretion mode seems to be common in other groups of orchids (Vogel 1990; Stpiczynska 1993, 2001).
Emission of volatile compounds by osmophores by cuticular diffusion processes has been observed in Orchidaceae before, such as in species of Scaphosepalum (Pridgeon and Stern 1985) and Stanhopea (Stern et al. 1987); or by cuticular pores in species of Restrepia and Restrepiella (Pridgeon and Stern 1983) and Gymnadenia conopsea (Stpiczynska 2001). The emission of volatile compounds in these species of Acianthera, however, seems to be associated with the presence of stomata. Stomatal pores were frequently observed on the surface of the nectaries that are involved in exogenous secretion, and Vogel (1990) suggested they could work as possible routes for volatile secretions. We found evidence here that the secretion products of species of Acianthera are liberated by the cells of the osmophores and accumulated in the periplasmatic and intercellular spaces. These compounds are probably volatilized by daylight temperatures and finally reach the outside environment through the stomatal pores. This hypothesis can be further supported by the liberation of these odors only during the hottest hours of the day. This has also been reported for other species of Acianthera (Borba and Semir 2001) and Bulbophyllum (Borba and Semir 1998).
The diversity of nectaries in Orchidaceae is intimately related to their adaptations to diverse groups of pollinators (Baker and Baker 1983; Pacini et al. 2003). Nectaries in most orchids are formed by protuberances (calli) or papillose regions located on the column, labellum, or between the two (Dressler 1993). The presence of exposed nectaries on the labellum, as in the species of Acianthera studied here, seems to be common in myophilous orchids, and it has been found in several species of Acianthera, Anathallis, Bulbophyllum, Octomeria, and Stelis (van der Pijl and Dodson 1966; Dressler 1993; Borba and Semir 1998, 2001; Singer and Cocucci 1999).
Nectar is usually offered as a reward to anthophilous species of Diptera (Singer and Cocucci 1999; Albores-Ortiz and Sosa 2006) or in mechanisms of partial deceit-pollination by saprophytic flies (Borba and Semir 1998, 2001). The species in the present study have purple flowers with disagreeable odors that are generally associated with sapromyophily (Proctor et al. 1996), but the presence of nectar makes the occurrence of strict deceit-pollination doubtful (van der Pijl and Dodson 1966; Faegri and van der Pijl 1979). Additionally, no oviposition events were observed in the species pollinated by Chloropidae (A. prolifera) or by Phoridae (A. hamosa, A. limae, and A. modestissima) in this group (Melo 2008). Some authors, however, argue that the absence of nectar is a feature of sapromyophily s.s., since the insects would be attracted by oviposition instinct rather than food instinct (van der Pijl and Dodson 1966; Faegri and van der Pijl 1979). At least in Orchidaceae, absence of nectar in sapromyophilous systems can be effective especially in trap-like mechanisms, as observed in some Bulbophyllum species (Teixeira et al. 2004). However, in systems where it is necessary to keep the insect in the flower for a longer period of time, the presence of nectar may be crucial for pollination to occur successfully, even if the animal has been attracted by oviposition instinct (Borba and Semir 1998). Thus myophily s.s. and sapromyophily s.s. may represent only the extremes of a continuum in which the pollination mechanisms of these species cannot be precisely placed.
The four orchids studied here show pollination mechanisms similar to those reported by Borba and Semir (2001) in another group of Acianthera. In the study by Borba and Semir (2001), the two Phoridae-pollinated species offered nectar while the three species pollinated by female flies of Chloropidae were deceitful. Female flies were only observed laying their eggs in the flowers of two of them. Micromorphological analyses of the flowers of these latter species, similar to the study we have reported here, could help to clarify how these two groups of species, with apparently considerable macromorphological convergence, exploit a similar pollinator group but possess different pollination mechanisms.
Osmophores may be located in different parts of the perianth in orchids (Pridgeon and Stern 1983, 1985; Stern et al. 1987). In the Pleurothallidinae, including the species studied here, the osmophores seem to be only in the sepals or on appendages that originate from them, while in Bulbophyllum osmophores are commonly found on the labellum. These two groups have flowers with very similar external morphologies and share many adaptations to myophily (van der Pijl and Dodson 1966; Dressler 1993), but there are notable differences in floral micromorphology associated with each group, as mentioned above. At least initially, these differences may not play an important role in attracting specific groups of pollinators or determining different pollination mechanisms, as Bulbophyllum and Pleurothallidinae present a wide variation in both parameters (Borba and Semir 1998; Singer and Cocucci 1999; Borba and Semir 2001; Blanco and Barboza 2005; Albores-Ortiz and Sosa 2006). Rather, they may represent only phylogenetic constraints in both groups. However both studies on floral biology and morphology are still very limited for this to be determined in such large groups.
Flowers pollinated by saprophilous flies generally emit odors similar to proteinaceous material in decomposition, mainly nitrogen-containing compounds such as amines, ammonia, and indols (Proctor et al. 1996). These substances are associated with the odors of various typically sapromyophilous species belonging to Pleurothallidinae and Bulbophyllum (Vogel 1990). However, other studies in species of these groups have not always confirmed their presence (Kaiser 1993). The latter could be due to the sampling method, which used polymers associated with the technique of head-space and used solvents for the polymer elution (Kaiser 1993). This technique may be inappropriate when dealing with nitrogenous substances of low molecular weight, such as ethylamine and dimethylamine, which are apparently common in Acianthera and other fly-pollinated species (E. L. Borba, UFMG, Minas Gerais, Brazil, and J. R. Trigo, UNICAMP, São Paulo, Brazil, unpublished data), where the technique of solid injection seems more appropriate (Silva et al. 1999). Also, due to the use of very small amounts of plant tissue in the solid injection technique, identifying the precise location of the osmophores in the flower is necessary. Thus, the Whiff test may be an important indicator as to which technique is best suited to be employed in these myophilous groups, with the solid injection being indicated when nitrogenous substances are present, such as the Neotropical Pleurothallidinae, in contrast to the Neotropical Bulbophyllum species, where these substances are apparently rarer.
We thank the technical staff of the Centro de Microscopia Eletrônica, Instituto de Biociências, UNESP Botucatu, for their help in preparing the samples, and an anonymous reviewer for improvements to the manuscript. This work was supported by grants from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and from Pró-Reitoria de Pesquisa/UFMG to E. Borba. M. Melo received a fellowship from Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG). E. Borba and E. Paiva are supported by a productivity grant (PQ2) from CNPq.