|Home | About | Journals | Submit | Contact Us | Français|
Pollination drops and nectars (floral nectars) are secretions related to plant reproduction. The pollination drop is the landing site for the majority of gymnosperm pollen, whereas nectar of angiosperm flowers represents a common nutritional resource for a large variety of pollinators. Extrafloral nectars also are known from all vascular plants, although among the gymnosperms they are restricted to the Gnetales. Extrafloral nectars are not generally involved in reproduction but serve as ‘reward’ for ants defending plants against herbivores (indirect defence).
Although very different in their task, nectars and pollination drops share some features, e.g. basic chemical composition and eventual consumption by animals. This has led some authors to call these secretions collectively nectar. Modern techniques that permit chemical analysis and protein characterization have very recently added important information about these sugary secretions that appear to be much more than a ‘reward’ for pollinating (floral nectar) and defending animals (extrafloral nectar) or a landing site for pollen (pollination drop).
Nectar and pollination drops contain sugars as the main components, but the total concentration and the relative proportions are different. They also contain amino acids, of which proline is frequently the most abundant. Proteomic studies have revealed the presence of common functional classes of proteins such as invertases and defence-related proteins in nectar (floral and extrafloral) and pollination drops. Invertases allow for dynamic rearrangement of sugar composition following secretion. Defence-related proteins provide protection from invasion by fungi and bacteria. Currently, only few species have been studied in any depth. The chemical composition of the pollination drop must be investigated in a larger number of species if eventual phylogenetic relationships are to be revealed. Much more information can be provided from further proteomic studies of both nectar and pollination drop that will contribute to the study of plant reproduction and evolution.
A large variety of animals rely on sugary secretions for their nutrition. This is why sugary secretions are largely involved in mutualistic interactions. Sugary secretions can be produced by animals, fungi and plants. Aphids (Hymenoptera: Aphididae) suck the sap of the plants and produce sugary droplets, called honeydew, eagerly collected by ants that defend the aphids from predators. Caterpillars of some species of blues (Lepidoptera: Lycaenidae) produce sugary droplets by a dorsal gland; these droplets also attract ants that protect the caterpillar from its enemies (Wäckers, 2002). Rust fungi (Basidiomycetes; Uredinales) produce sugar droplets among the maturing spores; insects (especially flies) feed on these droplets and disperses the spores (Wäckers, 2002). Sugary secretions are much more common in higher plants (pteridophytes, gymnosperms, angiosperms) where generally they maintain this double function: protection against predators by attracting ants (Hymenoptera: Formicidae) or attraction of animals that mediate the dispersal of spores or pollen. With the exception of the secretions by members of the Gnetales, gymnospermous sugary secretions are not generally involved in such mutualistic relationships. In most gymnosperms a sugary secretion, the so-called pollination drop, is the landing site for pollen grains. This secretion is produced by ovules, and it protrudes beyond the micropylar terminus (Gelbart and von Aderkas, 2002). After pollen lands, the pollination drop is withdrawn, thus transporting the pollen on to the surface of the nucellus (Tomlinson et al., 1997; Mugnaini et al., 2007, and references therein) where it germinates. According to Lloyd and Wells (1992), a major step in angiosperm carpel evolution was a change in the landing site for pollen from a pollination drop on the naked ovule to a wet stigma on the outside of a closed carpel (see also Heslop-Harrison and Shivanna, 1977). Another important reproductive feature that characterizes early angiosperms is the interaction with insects for pollen dispersal. Insects visiting flowers in search of a food reward unintentionally mediated the transfer of pollen to receptive surfaces, thereby favouring cross-pollination (Faegri and van der Pijl, 1979; Crepet, 1983; Crepet and Friis, 1987). Entomophily is considered a plesiomorphic character in angiosperms but it is apomorphic in gymnosperms, i.e. the gnetophytes (Lloyd and Wells, 1992). The majority of gymnosperms are anemophilous, entomophily having evolved in Cycadales and Gnetales (or Chlamydospermae) and a few extint lineages (Crepet, 1974; Labandeira et al., 2007). In each of these groups the types of insects that are attracted are different and how they carry out pollination also differs.
The reward to pollinating insects in early angiosperm flowers was most probably floral secretions (Endress, 1994). The pollination drop on the ovular micropyle and later the stigmatic secretion may have served as reward for pollinators, at first mainly flies (Diptera), micropterigid moths and beetles (Coleoptera) (Lloyd and Wells, 1992; Endress, 1994). Some authors (Lloyd and Wells, 1992; Endress, 1994) called these secretions ‘nectar’, focusing on their ecological meaning. They created some confusion. Real nectar, i.e. a sugary secretion produced by a defined organ that is the nectary, first appeared, according to the fossil record, in late Cretaceous flowers (Friis and Endress, 1990). Although the main function of the pollination drop and nectar is completely different, some similarities are reported in the literature, both in the chemical composition, as well as in the anatomy of secretory tissues (Owens et al., 1987; Bernardello, 2007).
The recent increase of information about the molecular biology and proteomics of the two secretion types (O'Leary et al., 2004; Poulis et al., 2005; Carter et al., 2007; Thornburg, 2007; Wagner et al., 2007) has added new insights into their function. New similarities have also been discovered. In this review, an updated comparison concerning the cytological, ecological and biochemical features of these two types of secretions is provided with the aim to stimulate further studies and discussion about their chemical, biochemical and functional complexity that still remains widely unknown.
The evolution of more specialized interactions between plants and animals frequently transformed their inter-relationship from predation to mutualism in which both ‘partners’ benefit: the animal is generally rewarded with food while the plant benefits in terms of both propagule dispersal (spores, pollens and/or seeds) and defence against herbivores (Pacini et al., 2008). Among the various tissues of higher plants, secretory tissues are most often involved in such interactions (Pacini et al., 2008). Many arthropods feed on carbohydrate-rich products elaborated directly by plants (most frequently nectar), or indirectly through insects sucking phloem sap (honeydew) (Wäckers, 2002).
In the phylogeny of plants, nectar-like secretions appear for the first time in pteridophytes (Table 1). Extant genera of nectar-secreting pteridophytes include: Angiopteris, Cyathea, Drynaria, Hemitelia, Holostachyum, Merinthosorus, Photinopteris, Platycerium, Polypodium, Polybotra and Pteridium (see Koptur et al., 1982). Secretions are abundantly rich in sugar and insects can exploit them as food sources. The history of sugary exudates has resulted in nectar that is not produced by flowers being called ‘extrafloral nectar’, which in the case of pteridophytes represents an unfortunate terminological anachronism, as fern sugar secretions were the first occurrence among all vascular plants. According to Schmid (1988) it is better to regard the pteridophytes nectaries as foliar nectaries. There are conflicting reports in the literature about the function of these nectaries. Most probably they reward ants that protect the plant from herbivore attack, although this function was not always confirmed (Tempel, 1983; Heads and Lawton, 1985; Cooper-Driver, 1990, and references therein). Ants are not the only arthropod species that have been observed visiting these nectaries; parasitic wasps, coccinellid species, flies and spiders have also been observed (Tempel, 1983). Although nectaries on both developing and mature fronds attract and help maintain an ant-guard system, there may be other explanations. Foliar nectaries, especially those in close proximity to spore-producing structures such as sori, may also attract insects that will serve as spore dispersal agents (Koptur et al., 1982; Tryon, 1985; Walker, 1985).
Although gymnospermous nectaries are not a common phenomenon, currently being entirely restricted to the gnetophytes, sugar secretion is widespread, particularly in the form of pollination drops (Gelbart and von Aderkas, 2002) (Fig. 1 and Table 1). Pollination drops appeared very early in their phylogeny. According to Doyle (1945) and Rothwell (1977) they were common among pteridosperms (seed ferns) and early gymnosperms. This mechanism of pollen capture was widespread during the Carboniferous (Taylor and Millay, 1979) and has remained so to this day. Only one family – Araucariaceae, a family that has some primitive reproductive features such as recalcitrant pollen grains with 15 prothallial cells (Singh, 1978) – lacks a pollination drop system (Gelbart and von Aderkas, 2002).
The groups that are most controversial in gymnosperm phylogenetic studies – the three genera of Gnetales (Ephedra, Gnetum and Welwitschia) – all possess pollination drops. These genera are all dioecious with particular features. Some species have strobili in which sterile ovules are regularly associated with stamens in a morphologically bisexual, flower-like structures (Bino et al., 1984a; Lloyd and Wells, 1992, and references therein). More remarkably, in all three genera, sterile ovules of male plants produce pollination drops that resemble those produced by ovules of female plants (Lloyd and Wells, 1992, and references therein; Kato et al., 1995). Several insects (ants, bees, wasps, flies, moths and some Heteroptera) have been reported feeding on pollination drops of both male and female plants of all the three genera. Insect pollination has also been demonstrated in Gnetales (Lloyd and Wells, 1992; Kato et al. 1995; Wetsching and Depisch, 1999, and references therein). In addition to these ovular secretions, some Ephedra nectaries have been found not only on the bracts of male and female plants, but on integuments of the female plants (Bino et al., 1984b). This is the only case of the coexistence of both nectar and pollination drop secretions. Another interesting case is offered by Gnetum cuspidatum. Although male strobili lack sterile ovules in this species, they nevertheless secrete nectar between and on collars, which are vegetative structures associated with male strobili (Kato et al., 1995).
The new condition of enveloped ovules that characterizes angiosperms brings about a complication because the primary source of sugary exudates (i.e. the pollination drop) disappears. Insect mouthpart structures and plant reproductive features indicate that insect feeding on plant exudates far antedates the origin of angiosperms extending to the earlier Mesozoic and into the late Paleozoic (Meeuse, 1978; Crepet and Friis, 1987). Thus insects were already pre-adapted to feed on nectar-like liquids at the moment of angiosperm rise, and a strong selective pressure must probably have operated on provision of a sugary exudate (nectar) in the flowers of very early angiosperms. Another way to think about the problem of nectar tissue occurrence in early angiosperms is one in which flower petals and sepals are, from a developmental viewpoint, considered to be modified leaves. Convergent evolution of nectar production occurred in widely different species, whose leaves developed the ability to produce extrafloral nectar. Assembling modified leaves into a flower, it was only a matter of time before extrafloral nectaries ceased being extrafloral.
Nectar is present in a great number of unrelated angiosperm taxa (Fig. 2 and Table 1). It is believed that both floral and extrafloral nectaries independently arose a number of times (Brown, 1938; Meeuse, 1978; Friis and Endress, 1990). Unlike other floral structures whose relative positions are conserved throughout the angiosperms, the nectary is not located in the same position in all plants and can be easily lost or acquired within a lineage (Bernardello, 2007). The molecular basis for such great variability was recently discovered by Baum et al. (2001): the nectary is independent of the ABC floral homeotic genes that are responsible for floral organ identity specification according to their position. Thus the nectary is ‘free’ to move about the flower and plant during evolution in response to selection imposed by interactions with the environment and pollinators. The position of the nectar inside the flower has an ecological meaning and the place of nectar presentation can be different from the site of nectar production (secondary presentation) (Pacini et al., 2003). This finding reinforces the statement of Meeuse (1978) that‘. . . even the presence of nectaries in corresponding parts of the flower does not necessarily mean that they are homologous; in other words, the presence of nectariferous organs of seemingly the same origin does not necessarily imply that there is always a close taxonomic relationship’.
Extrafloral nectaries are also present in angiosperms where they maintain the function of ‘indirect defence’ (Heil, 2008a) by means of attraction of predatory animals (generally ants) that defend the plant from phytophagous insects (see section entitled ‘Functions’).
Structures that secrete water (e.g. hydathodes), proteins (e.g. digestive glands) or carbohydrates (nectaries) are considered part of a continuum of secretory structures (Fahn, 1979). A nectary, by definition, is a carbohydrate secretory structure, a more or less defined organ that is specialized for this function (Nepi, 2007). The diversity of anatomical organization and localization of the nectaries inside the flowers is incredibly broad, such that several morphological classifications were proposed (Fahn, 1979; Schmid, 1988; Smets and Cresens, 1988; Bernardello, 2007).
The association of nectaries and pollination drops with reproductive structures shows strong differences. Although gymnosperm reproduction is simpler than angiosperm reproduction, and pollination drops are secreted from simple nucellar tissue, quite unlike the anatomically specialized nectaries of flowering plants, the biochemical composition of these drops is similarly complex and rich. A more comparable organ would be the angiosperm ovule. Although there is some evidence for ovular secretion in some angiosperms (Franssen-Verheijen and Willemse, 1993, and references therein), the nucellus does not play an active secretory role. Experimental and histological evidence shows that the gymnosperm pollination drop is produced by the nucellus (Tison, 1911; O'Leary et al., 2004). The nucellus of gymnosperms and that of angiosperms is quite different in some functions. A nucellus is a sporogenous tissue giving rise to the megagametophyte in the case of gymnosperms and the embryo sac in the case of angiosperms. By the time female tissues are receptive, the surrounding nucelli are quite different in size: gymnosperms have a thick nucellus but angiosperm nucellus can be only one or two cell layers thick, and therefore much reduced. The larger gymnosperm nucellus has many more functions that change during development. In some species, such as Picea spp., it may breakdown to form a pollination chamber (Singh, 1978). Whether it forms such a chamber or not, it is able to secrete a pollination drop in all genera except Araucaria, Abies and Thuja. After pollen has been captured, the nucellus may be long-lived, in the case of Pinus and at least 13 other genera in which there is a very long period (approx. 1 year) before the pollen tube finally penetrates the nucellus to enter the egg and bring about fertilization. After fertilization the role of the nucellus is to build up pectin and wax layers. By the time that the seed is mature and ready to be released, the nucellus is already a dead organ. In spite of a loss of vitality, it continues to play a physiologically important role. In many species, the nucellus is impermeable to water. In this way, the dead layer provides a barrier effectively forcing water movement along selective pathways into the germinating embryo (Tillmann-Sutela and Kauppi, 1995a, b, 1998; Tersikh et al., 2005). The angiosperm nucellus, by comparison, does not have a post-fertilization role similar to that described in gymnosperms. This adds some evidence that the nucellus performs many functions in gymnosperms that have been distributed among other tissues of the angiosperm ovule.
Gymnosperm nucellar tissue is polarized. Cytological studies have revealed that in some aspects the nucellar tip closely resembles certain types of nectaries in ontogeny, ultrastructure and secretory products. A group of small, starch-filled meristematic-like cells develop ultrastructural modifications associated with pollination drop secretion (Owens et al., 1987; Takaso and Owens, 1995). The dynamics of organelle development before and during the secretory stage – abundant dictyosomes in early stages, abundant endoplasmic reticula, the formation of many small vesicles and then large vacuoles – are similar to those found in some nectaries (Nepi, 2007). It is plausible that the secretory process in the nucellar tip is very similar at the cellular level to what occurs in the nectary. Accumulation of starch before secretion is a characteristic common to both nucellus and nectary. Starch hydrolysis provides the soluble sugars of the secretion (Carafa et al., 1992; Nepi, 2007). Cell degeneration at the tip of the nucellus may occur at the moment of secretion or soon thereafter (Owens et al., 1987; Carafa et al., 1992; Takaso and Owens, 1995), which is similar to holocrine secretion events in some nectaries (Vesprini et al., 1999; Horner et al., 2003; Vesprini et al., 2008). The merocrinal secretion, in which cells survive after releasing the secretory product, is much more common for nectar secretion (Pacini and Nepi, 2007). After secretion, programmed cell death has been shown to characterize degeneration in both nectary and nucellus (Li et al., 2003; Zhu and Hu, 2002). Another similarity is that the cuticle layer can show modifications, especially interruptions, at the moment of secretion. It becomes interrupted or stretched in the nectary (Nepi, 2007) and is almost absent from the nucellar apex (Takaso and Owens, 1995).
A difference between the two secreting structures concerns the vascular tissue. In general, the gymnosperm nucellus is not directly vascularized whereas the functional efficiency of the nectary relies on the presence of or proximity to xylem and/or phloem vessels (Nepi, 2007). In spite of such obvious differences, a recent systematic survey of floral nectaries by Bernardello (2007) considered the nucellar tip of Gnetales to be ovular, non-structural nectaries according to Schmid's broad topographical classification of nectaries (Schmid, 1988). However, it seems that the presence of such kinds of ovular nectaries and flower-like reproductive structures in the Gnetales does not mean necessarily a phylogenetic relationship with angiosperms, as previously stated, according to classic cladistic analysis based on morphological data (Crane, 1985; Loconte and Stevenson, 1990). More recent molecular work on the phylogeny of seed plants has grouped the Gnetales with the conifers, and placed the ancestor of the angiosperms among ancient gymnosperms (Chaw et al., 1997; Bowe et al., 2000; Donoghue and Doyle, 2000; Magallón and Sanderson, 2002; Willis and McElwaine, 2002.)
Plants not only secrete sugary exudates, they can also reabsorb them. The ability of the plant to reabsorb sugary secretions previously produced may allow the plant to recoup part of the energy involved in the secretion process. This energy can be used elsewhere in the plant for other functions. Nectar resorption has been demonstrated in several but not all the studied species (Pacini and Nepi, 2007, and references therein; Nepi and Stpiczyńska, 2008). Two principal functions can be recognized: the recovery of resources invested in nectar production and a homeostatic mechanism during nectar secretion and presentation (Nepi and Stpiczyńska, 2008). The first function can be strongly involved in the regulation of the energy balance of the plant. In fact, nectar secretion is an expensive process; Southwick (1984) estimated that from 4 % to 37 % of daily photosynthate assimilated by Asclepias syriaca during blooming was secreted as nectar sugar. He also reported that the energy invested in nectar production by Medicago sativa is twice the energy invested in seeds. Recovery of resources is therefore an important advantage for plants that reuse this source of carbohydrates left over as uncollected nectar. The homeostatic mechanism has an important ecological function, as it is utilized to maintain nectar characteristics, such as volume and concentration, in a range suitable for the foraging pollinators. The resorbed material can be transferred from the nectary tissue to other floral or vegetative parts (Pedersen et al., 1958; Nepi and Stpiczyńska, 2007). Developing ovaries have been recognized as the main sink of such resorbed substances, at least in Platanthera chlorantha (Orchidaceae) and Cucurbita pepo (Cucurbitaceae) (Nepi and Stpiczyńska, 2008). The epidermis and the parenchyma of the nectary are mainly involved in the resorption process, although other floral parts (petals and ovary) can also play a role. To date, a mechanism for nectar resorption has not been elucidated. The mechanism most probably involves changes in cell turgor that, in turn, respond rapidly to changes in osmolality (Castellanos et al., 2002; Nepi et al., 2007).
Even less is known about pollination drop withdrawal. In the Podocarpaceae, the process is merely physical, attributable to passive evaporation independent of the presence of pollen. However, in most gymnosperms, such as Pinus, Callitris, Chamaecyparis, Cryptomeria, Thuja and Juniperus, withdrawal seems to be a metabolic process that is triggered by the presence of pollen that also stops drop secretion (Tomlinson et al., 1997, and references therein; Gelbart and von Aderkas, 2002; Mugnaini et al., 2007). When triggered by pollen, drop withdrawal is generally very quick, so that a resorption process can be hypothesized. The resorption mechanism is not understood, nor is it clear which tissue is performing the resorption. It is for these reasons that the term drop withdrawal is generally used instead of drop resorption. The actual stimulus for withdrawal–resorption is obscure. Recently, it was pointed out that it may depend on the size of the deposited particles. In addition, an unidentified biological stimulus carried by viable pollen plays a role in the response (Mugnaini et al., 2007). It is interesting to note that, in some cases of particular angiosperm species, nectar resorption is also stimulated by pollination (Koopowitz and Marchant, 1998; Luyt and Johnson, 2002). Regardless of the mechanism responsible for the drop withdrawal, the major function of the drop is to transport pollen grains to the nucellus surface. Pollination drop withdrawal is essential in the pollination mechanism of gymnosperms, whereas nectar resorption appears to be a facultative phenomenon (Table 1) (Búrquez and Corbet, 1991; Nepi and Stpiczyńska, 2008).
The volume of the secretion per nectary or per pollination drop is widely variable but, generally speaking, the volume of nectar is higher (especially flower nectar) than that of the pollination drop. Pollination drop volume is 60 nL on average (Seridi-Benkaddour and Chesnoy, 1988), reaching 250 nL in the single-ovuled cone of Taxus baccata (Fig. 1B). For nectar, the range of variability is from about 50 nL, as in single florets of Asteraceae (Wist and Davis, 2006) to 9·4 mL in Ochroma lagopus (Bombacaceae, bat pollinated) (Opler, 1983). Variability of volume in extrafloral nectar is in the lower range of floral nectar. The variability of floral nectar volume was analysed at the intraspecific level and can be correlated with the age of the flower, the position of the flower in the plant or in the inflorescence, environmental parameters (temperature and relative humidity) and type and body size of pollinator (Opler, 1983; Pacini and Nepi, 2007). More uncertain and less studied were the reasons for the variability of pollination drop volume. Putatively, volume variability can be related to the number of ovules per cone, environmental parameters during the period of drop production, the mean number and dimension of the pollen arriving on the drop, but currently no detailed study has been carried out.
Nectar and pollination drops have a similar basic chemistry: the main components are sugars, amino acids and proteins (Table 1). Nectar is a complex mixture of a surprising variety of substances: sugars, inorganic ions, amino acids, proteins, lipids, organic acids, phenolics, alkaloids, terpenoids, etc. (Nicolson and Thornburg, 2007). Generally sweet, its high sugar content was recognized early. The first quoted analysis of individual sugars appeared in the second half of the nineteenth century (see Baker and Baker, 1983a). It was not until the middle of the twentieth century that it was realized that many other substances besides sugars can be present in nectar. Sugars are the best studied component because there is a relationship between the relative proportions of the three main sugars found in nectar (sucrose, glucose and fructose) and the type of animal that is attracted (Baker and Baker, 1983b). Nectar amino acids also have similar ecological importance (Baker and Baker, 1986). Sugar and/or amino acid composition of nectar has been reported for numerous species.
Ovular secretions in conifers were first described in the literature by Vaucher in (1841). However, the ephemeral nature and small quantity of the secretions has significantly limited the ability to study the constituents of the drop. As a result, there have been very few reports on composition. Early studies by Fujii (1903) reported the presence of glucose, amino acids, malic acid and calcium in Taxus baccata. Similar results were found in Cupressus funebris by Tison (1911): glucose, calcium and malic acid. Since then, fructose, sucrose, mixed carbohydrate polymers, organic acids, amino acids and proteins have been identified in the ovular secretions of a handful of conifer species (Table 2;Gelbart and von Aderkas, 2002). More recently, plant-defence proteins have been found in the ovular secretions of a number of gymnosperms (O'Leary et al., 2007; Wagner et al., 2007; Table 3).
In both nectar and pollination drop it was demonstrated that their chemical composition may reflect phylogenetic relationships (Nicolson and Thornburg, 2007, and references therein; Wagner et al., 2007), although nectar chemical composition can also be driven by adaptation to specific pollinators (Nicolson and Thornburg, 2007).
Fructose, glucose and sucrose are the main sugars present in both the pollination drops and nectar (Table 1). Total sugar concentration of the pollination drop ranges from 1–2 % in anemophilous conifers (e.g. Pinus; McWilliam, 1958) to 10–80 % in entomophilious Gnetales (Bino et al., 1984a, b). Sugar content of nectar varies more or less in the same range (8–80 %) (Nicolson and Thornburg, 2007), but a high variability of nectar concentration was found in individual angiosperms and is related mainly to flower age and microclimatic effects (Nicolson and Nepi, 2005; Pacini and Nepi, 2007).
Sucrose is the preferred compound for carbon transfer in plants (Akazawa and Okamoto, 1980) and both developing nectary and ovules act as sinks for sucrose. Sucrose is very frequently found in angiosperm floral nectar, although it is not ubiquitous. In an extensive study (765 species) of nectar sugar composition, Baker and Baker (1983b) found that sucrose was present in the 89 % of the species. Almost all studies of carbohydrate composition of gymnosperm pollination drops found that sucrose is not the dominant sugar, if it is present at all (see Gelbart and von Aderkas, 2002). There is only one exception: the pollination drop of Ephedra helvetica has high sucrose concentrations (approx. 25 %; Ziegler, 1959). It is interesting to note that extrafloral nectars from several Polypodium species (Pteridophyta; Koptur et al., 1982) as well as angiosperm extrafloral nectars are generally sucrose-poor (Koptur, 1994).
Pollination drops are a poor food source for insects for the following general reasons: they possess low concentration of sugars, of which some, such as xylose, are unattractive to pollinators such as hymenopterans, and those sugars that are attractive, such as sucrose are present in low concentrations, if they are present at all (Baker and Baker, 1983b). Xylose was recently found in the pollination drop of Larix and Pseudotsuga (M. Nepi et al., unpubl. res.). Presence of xylose in angiosperm floral nectars, though rare, was considered to be a deterrent to insects and birds (Jackson and Nicolson, 2002).
The dominant sugar that was usually found in the pollination drop was fructose (Table 2). In vitro experiments with Pinus mugo pollen indicated that during germination pollen takes up fructose preferentially over other sugars (Nygaard, 1977). This implies that the chemical composition of the pollination drop is likely to be more suited for pollen nutrition rather than for insect consumption.
Uronic and galacturonic acids as well as xylose can be found in the pollination drops of some species (Table 2). The presence of these substances is due to breakdown of the nucellus tip cells, as these molecules form complex polymeric structural components of the plant cell wall and can be dispersed in the secretion when the cell wall ruptures.
Amino acids have been found in foliar nectar of pteridophytes, gymnosperm pollination drops, and extrafloral and floral nectar of angiosperms. Serine, glycine, alanine, arginine and proline are frequently and abundantly detected in all these secretions (Keeler, 1977; Koptur et al., 1982; Baker and Baker, 1983a; Pate et al., 1985) (for pollination drop see Table 2). Of these, proline seems to have a special importance for insects for a variety of reasons: it contributes a taste preferred by insects (Alm et al., 1990), it stimulates the insect's salt cell, a labellar chemosensory receptor of insects, resulting in increased feeding behaviour (Hansen et al., 1998), and it can be metabolized very rapidly. In honeybees, proline is the most abundant amino acid in the haemolymph and is required for egg laying (Nicolson and Thornburg, 2007, and references therein). Oxidative proline degradation is a very efficient, short-term burst of energy that can be utilized during the initial ‘lift’ phase of insect flight, whereas glucose is used for extended flight (Carter et al., 2005).
The accumulation of proline in plant tissue is generally taken as a sign of stress. Drought, salinity and freezing increase biosynthesis of proline (Carter et al., 2005). Proline dominates the free amino acid composition – it can be up to 70 % – of angiosperm pollen free amino acids; it is much less abundant in gymnosperm pollen (Stanley and Linskens, 1974; Zhang et al., 1982). Proline is used by the germinating pollen grains as a substrate or in the synthesis of hydroxyproline-rich wall proteins (Shivanna, 2003). Only a small proportion of the proline pool is used in this way; the major pool is required for other functions (Shivanna, 2003). Uptake experiments demonstrated that mature as well germinating pollen rapidly take up proline by means of a specific transporter (Schwacke et al., 1999). This adds further support to the suggestion that abundance of proline in the pollination drop appears to be correlated with its interaction with pollen grains, although this function requires experimental verification.
The existence of proteins in floral nectar was reported long ago (Beutler, 1935), but only recently was this shown in gymnospermous pollination drops (Poulis et al., 2005; O'Leary et al., 2007; Wagner et al., 2007). Although different proteins have been identified in nectar and pollination drop (Tables 3 and and4),4), similar putative functions can be ascribed to some proteins. Two general functional classes of proteins that are found in both nectar and pollination drops include carbohydrate-metabolizing enzymes and defence proteins.
A group of proteins comprising mainly invertases, transfructosidase and transglucosidase are involved in simple sugar metabolism. Among them invertases have been detected in gymnosperm pollination drops (Poulis et al., 2005), and in floral and extrafloral nectar (Pate et al., 1985; Heil et al., 2005; Nicolson and Thornburg, 2007). Invertases catalyse cleavage of sucrose into fructose and glucose. This activity can rapidly change the basic sugar composition of the secretions as well as their osmolarity. It has been shown in extrafloral nectar of Vigna unguiculata inflorescence stalks that invertase activity was osmotically regulated (Pate et al., 1985).
Another group of proteins is involved in defence against fungi and bacteria. Sugary solutions should be excellent media in which to grow fungi and bacteria, especially as these solutions are openly exposed to the environment. Plants must defend their sugary secretions, protecting these solutions for their reproductive purposes, while preventing their use as a base for microbes to further attack the plant's reproductive systems. Glucan 1,3-β-d-glucosidases, chitinases and thaumatin-like proteins are recognized as plant pathogenesis-related proteins. They have been found recently in pollination drops of some gymnosperm species (Table 3) (Wagner et al., 2007). Defence proteins are also known from floral nectar (Table 4). In Allium nectar, both alliinase and a mannose-binding lectin have antibiotic properties (Peumans et al., 1997).
A nectar redox cycle involving five enzymes first discovered in ornamental tobacco has been found to be more widespread (Carter et al., 1999, 2007; Carter and Thornburg, 2000, 2004). The nectar redox cycle maintains high levels of hydrogen peroxide in nectar 40 times the level produced by human neutrophils in response to microbial attack. The nectar redox cycle defends the metabolic-rich nectar from invading micro-organisms carried to the flower by wind or by non-sterile pollinators (Thornburg et al., 2003).
Defence proteins have also been found in extrafloral nectar (Gonzalez-Teuber et al., 2009). We can state that these two functional groups of proteins (carbohydrate metabolism-related proteins and defence proteins) are likely to be common to all plant sugar secretions.
A third group of proteins restricted typically to gymnosperm pollination drops are probably involved in pollen development and germination. These enzymes influence plasticity of the cell wall, cell expansion, and use of specific substrates (Table 3). This group of proteins includes 1,3-β-d-glucosidases (a multifunction protein), glycosyl hydrolases, xylosidases and subtilisin-like proteinases (Wagner et al., 2007).
The protein profiles recently revealed in Juniperus communis, Juniperus oxycedrus, Chamaecyparis lawsoniana and Welwitschia mirabilis seem to reflect phylogenetic relationships. In each species of Juniperus nearly all peptide sequences and identified proteins had an identical match in the other species. There was an overall similarity in protein profile for all the species in the Cupressaceae (J. communis, J. oxycedrus and C. lawsoniana), but to a lesser degree than species within Juniperus. There were no obvious similarities between W. mirabilis and all the other species studied (Wagner et al., 2007).
Pollination drops contain calcium (Fujii, 1903; Tison, 1911). Atomic absorption spectroscopy of pollination drops has shown calcium concentration differences between conifer species (K. Gill and P. von Aderkas, unpubl. res.), implying that drops vary in this element, one that is critical for pollen germination (Brewbaker and Kwack, 1963). Such differences between species and genera may confer a germination advantage to homospecific pollen over heterospecific pollen landing on the same pollination drop.
Very few studies have looked at ion concentrations of nectar. The most abundant ions detected were K+ and Na+, while Ca2+ and Mg2+ were present in lower concentration (Hiebert and Calder, 1983; Heinrich, 1989; Kronestedt-Robards et al., 1989). Findings by Hiebert and Calder (1989) suggested a phylogenetic component to patterns of nectar ion composition. On the other hand, Barclay (2002) demonstrated that nectar ion composition may reflect adaptation to specific pollinators.
Nectar is considered to be the most common floral food resource that is exploitable by the largest variety of animals, in particular animals that can fly. The sugar content is considered to be a ready-to use energy resource that can be used by feeding animals generally to power their flight (Nicolson, 2007). Nectar amino acids, besides contributing to the taste of nectar, are also an important source of nutrients for animals, especially for those that are exclusively dependent on nectar for their nutrition, such as butterflies. Insect preferences for single sugars or for particular mixtures of sugars and amino acids are known (Baker and Baker, 1983b; Blüthgen and Fiedler, 2006). Although floral nectar production represents a high cost for the plant (Southwick, 1984; Pyke, 1991), it ensures a higher possibility of seed set and thus higher reproductive success and fitness (Neiland and Wilcock, 1998).
Extrafloral nectar predates floral nectar by many hundreds of millions of years. Extrafloral nectaries are known from >700 genera of plants (about 4000 species), including pteridophytes, gymnosperms and angiosperms (see http://www.biosci.unl.edu/emeriti/keeler/extrafloral/worldlistfamilies.htm). In many cases, extrafloral nectaries are a mutualistic response of plants to ants. Ants are the most common small-sized predator in the animal kingdom; it would appear that to rid themselves of herbivores, plants have recruited carnivores. Extrafloral nectaries are not the only reward – indeed they supply only a biased diet. Myrmecophytes also produce Beltian bodies and domatia to accommodate the ants, who, in return, provide the plant with a rapid defence (Heil, 2008a). In this way, sugary exudates are part of a tritrophic innate defence, which may be composed of a number of elements, including one or more of the following: (a) herbivore feeding induces greater formation of Beltian bodies (food bodies), domatia and/or extrafloral nectaries, which, in turn, attract herbivore predators; and (b) herbivore feeding induces release – by the plant – of volatile organic chemicals, which attract predators. The plant therefore plays a significant role in communicating to carnivores that herbivores are present. For the predators, there is a rich protein diet of herbivores supplemented with either fat-rich Beltian bodies or sugar-rich extrafloral nectary secretions. Although studies have clearly shown the advantage that plants reap in their fitness when they attract ants as protectors, other such relationships between insect trophic levels and plants require experimental dissection in the context of fitness.
In the fossil record, extrafloral nectaries appeared long before ants (early Cretaceous; Ward, 2007). Thus the extrafloral nectar initially had a function not involving ants. According to De la Barrera and Nobel (2004) nectar may originally have developed independently of any interaction with animals. Because extrafloral nectaries are frequently associated with developing organs (stems or leaves), the ‘leaky phloem’ hypothesis of the origin of nectar (De la Barrera and Nobel, 2004) can be applied to extrafloral nectaries. According to this hypothesis, nectar secretion may thus have originated as leakage of phloem solution, resulting from the structural weakness of developing tissues exposed to high pressure in the phloem. Interaction of extrafloral nectaries with ants may have evolved after the latter's appearance in the early Cretaceous (Ward, 2007). Since a symbiotic relationship with ants increased plant fitness, acting as an indirect defence against herbivores, it may have been established and then reinforced by differentiation of other ant-attracting structures, such as domatia and food bodies.
It is now clear that the classical definition of angiosperm floral nectar as simply a food reward for pollinating or defending animals is outdated. As Pedersen et al. (1958) stated some decades ago ‘nectar is not a static product remaining outside the plant once produced but is in close contact with the plant system’ and the substances present in the nectar can be dynamically exchanged with the nectary itself in response to ecological and physiological constraints. This exchange maintains a relatively constant nectar concentration to ensure insect visits (nectar homeostasis) and reallocates resources, especially during development of the ovules and ovary after fertilization (Nepi and Stpiczyńska, 2008). Although nectar is much more than a simple reward for visiting insects, the presence of nectaries, in either reproductive (floral nectar) or vegetative parts of a plant (extrafloral nectar), represents a basic interaction between plants and animals. This is not applicable to the pollination drop of most gymnosperms, with the notable exception of the Gnetales.
Pollination drops have a major function in regulating pollen behaviour, which is not the case with nectar. Here lies the most significant functional difference. Pollen requires water from the female counterpart to germinate. It also requires calcium and carbohydrates, as has been shown in many in vitro studies (Stanley and Linskens, 1977; Shivanna, 2003). The indisputable function of the pollination drop is that of a landing site and germination medium for pollen grains (Table 1). The drop's resorption serves to transport pollen grains inside the ovule, putting them within close proximity to the megagametophyte and its gametes. A number of compositional features of the pollination drop point to its role in nourishing pollen. These include the following: its sugar content, which is too low a concentration to benefit insects (with the exception of Gnetales pollination drops); the presence of calcium (Fujii, 1903), proline (see Table 2) and specific proteins related to pollen germination and tube growth (Wagner et al., 2007; Table 3). Although phylogenetic relationships can be recognized, the basic chemical composition of pollination drops is different from species to species. These differences may represent a direct method of homospecific pollen selection. For example, pollen that enters the ovule of its own species will encounter a favourable environment, both osmotically and chemically. This environment is very different in heterospecific ovules (Gelbart and von Aderkas, 2002).
The sucrose composition of pollination drops of gymnosperms may regulate pollen behaviour. Gnetales have high sucrose concentrations, whereas that of cycads (Tang, 1987) and conifers (McWilliam, 1958) is low. Sucrose concentration can affect germination as has been shown in vitro studies of many conifers (Dumont-BéBoux et al., 1999, 2000). Furthermore, in some conifers, proteomic profiling has revealed the presence of secreted invertases, which are thought to be responsible for the general absence of sucrose and the presence of fructose and glucose in the drop (Poulis et al., 2005). It is postulated that invertases, at least not functional ones, are unlikely to be found in the Gnetales.
Interestingly, differences in sucrose concentration between Gnetales (high sucrose) and cycads (low sucrose) may be related to differences in the foraging behaviour of pollinators: Gnetales are pollinated by insects that use the pollination drops as a reward, while cycads are pollinated by insects that do not feed on pollination drops. High levels of sucrose in sugary secretions may be maintained in gymnosperms and angiosperms under a selection regime of insect pollination.
Did pollination drops provide a source of nutrition to insects present or of the past? In a sweeping review of pollination drops and possible insect pollination in gymnosperms during the Mesozoic era that immediately preceded angiosperm radiation, Labandeira et al. (2007) make a compelling case that mecopteroids and brachyceran dipterans had siphonate proboscides type mouthparts adapted to uptake of nectar and sugar-rich pollination drops. The authors contend that the most plausible food sources would have been the hidden micropylar secretions of gymnosperms. Two extinct plant species – Frenelopsis alata and Cycadeoidea dacotensis – were used to illustrate paleobotanical and paleoentomological evidence for insect consumption of pollen and pollination drops. The evidence for Mesozoic insect feeding on the former is much stronger than the latter, which is mainly indirect. With the radiation of angiosperms and their associated insects, novel groups such as lepidopterans and new types of hymenopterans appeared, and thus, the types of insects that feed on gymnosperm pollination drops have changed.
Despite a significant body of literature about sugar composition of nectar, there is a notable lack of studies characterizing the presence of proteins. Although the presence of proteins in nectar has been long reported (Beutler, 1935), only three species [Allium porrum (Liliaceae), Nicotiana tabacum (Solanaceae) and Jacaranda mimosifolia (Bignoniaceae)] have been studied closely (Peumans, 1997; Carter et al., 1999, 2000, 2004, 2007; Kram et al., 2008). Nectar protein characterization is in its infancy and nectar protein profiling can be very complex. As an example, >30 proteins have been separated by 2D electrophoresis in Cucurbita pepo (Nepi et al., 1996). Several studies have been attempted to demonstrate whether nectar sugar composition is shaped by phylogenetic or ecological constraints (Nicolson, 2007, and references therein). Although results are not always consistent, it seems that phylogenetic history is the primary determinant of nectar chemistry, with pollinators having a secondary effect (Nicolson, 2007). It appears also that phylogenetic constraints can be more or less relaxed in different taxonomical groups (Krömer et al., 2008).
On the other hand, there is no research that has taken into consideration the phylogenetic relationships of nectar proteins.
Our knowledge of pollination drop composition is still relatively incomplete for both sugars and proteins. A limited number of species have had their sugar composition characterized (Table 2) and even fewer (seven) species have had their proteins studied (Table 3).
Proteomic analysis of pollination drops from four species of gymnosperms recently revealed a phylogenetic pattern (Wagner et al., 2007). Results of protein and sugar analyses are generally difficult to compare because of the limited number of species sampled, and the lack of uniform methods of analysis. These prevent us from drawing any further general conclusion about phylogenetic relationships in the composition of the pollination drops. We can only hypothesize that since pollination drops are not generally utilized by animals, the composition must be shaped predominantly by phylogenetic constraints and less by ecological ones.
Defence proteins look to be widespread in carbohydrate secretions of gymnosperms and angiosperms. Chitinases, as a functional class of proteins, seem to be conservative among the gymnosperms (Wagner et al., 2007) but nothing is known about their presence in angiosperm nectar as well as in foliar nectar of pteridophytes. Are some defence proteins conservative throughout the carbohydrate secretions of vascular plants?
When leaves of plants are attacked by fungi, they may respond with apoplastic pH signalling, as has been found in barley leaves (Felle et al., 2004). Such apoplastic responses have yet to be studied in nectar of angiosperms or in pollination drops of gymnosperms, as we are unaware of any studies on the response of these drops to microbial attack. Is the function of proteins of pollination drops and nectar to cleave oligosaccharides from microbes or fungi, and that these, in turn, act as signals to up-regulate defence genes in the nucellus or in the nectary? Little is known at present.
Another field open to research is the interaction between proteins of pollen grains and proteins in the pollination drops and the nucellus of gymnosperms. Particularly important in this respect is the report of arabinogalactan proteins in the pollination drop and in the micropylar region of the nucellus in Taxus × media (O'Leary et al., 2004). Arabinogalactan proteins are proteins expressed at the cell surface and in the intercellular spaces and are present throughout the plant kingdom. Arabinogalactan proteins were found abundantly in pistil exudates of angiosperm species where they have been recognized as adhesive, nutritional, protective and chemotropic agents during pollen–pistil interactions (Showalter, 2001). Their detection in gymnosperm pollination drops and the nucellus reveals biochemical adaptation during pre-zygotic events in common with angiosperms. How taxonomically widespread arabinogalactan proteins are in pollination drops and nucellus of gymnosperm is unknown at present.
It is clear that both nectar and pollination drop, although being simple liquid secretions, hide a biochemical and physiological complexity that requires further study for clarification. Results from these studies may reveal new information about the evolution of vascular plants, their phylogenetic relationships, their reproduction and animal associates.
Pollination drops of gymnosperms and the nectar of angiosperms are analogous but not homologous. They are analogous in several aspects: their basic chemical composition (sugars, amino acids, proteins) exhibits similarity, the secretory process involves similar cell biological processes, and both types of secretion are known to be reabsorbed in some species. They are not homologous since they are produced by ontogenetically unrelated organs. Pollination drops are secreted by the nucellus, providing evidence that reproductive organs are able to secrete a wide range of proteins, from carbohydrate-modifying invertases to pathogenesis-related proteins and proteins interacting with pollen development. In angiosperms, this ability to produce proteins during reproduction did not disappear with the evolutionary loss of the pollination drop. From the genetic analysis of Arabidopsis, it is clear that proteins are secreted from the filiform apparatus, e.g. MYB98 (a transcription factor) that regulates gene expression in cells of the embryo sac, in particular, synergid cells (Punwani et al., 2007). This triggers gene networks responsible for guiding pollen tubes toward the egg (Franssen-Verhaijen and Willemse, 1993; Fortescue and Turner, 2005). Due to the presence of the gynoecium, these secretions are essentially internal.
Nectar and pollination drops share a number of classes of compounds. These classes show qualitative and quantitative differences in their composition. Sugar composition varies and protein profiles differ. This reflects the different functions of the two types of secretions. Pollination drops act as landing sites for gymnosperm pollen. Afterwards, the drop provides water for pollen hydration, carbohydrates and amino acids during germination. In contrast, nectar is a food source for pollinating or defending animals. In addition, pollination drops have cations, mainly Ca2+, as well as specific proteins able to promote pollen development and germination (1,3-β-d-glucosidases, glycosyl hydrolases, xylosidases, subtilisin-like proteinases).
The differences between nectar and pollination drop function are reflected in the different ranges of sugar concentrations. Pollination drop sugar concentrations range from 5 % to 10 %, whereas nectar sugar concentration is almost without exception much higher. Gymnosperm pollen cultivated in vitro absorbs fructose preferentially; fructose was also found to be the more abundant sugar of pollination drops in several gymnosperm species (Table 2). The total sugar concentration is higher in nectar. Of the three most common sugars – glucose, fructose and sucrose – sucrose is the most common form found in nectar, but least common in pollination drops. The first appearance in the fossil record of advanced pollinator groups, including many Hymenoptera and glossate Lepidoptera, dates approximately to 140 million years ago, which is very close to the first appearance of angiosperms (Willis and McElwain, 2002). The food preferences of these two groups of insects – especially Lepidoptera – for sucrose-rich nectars (Baker and Baker, 1983b) may have driven the selection of individual plants producing increased carbohydrate rewards.
Some chemical components are functionally very similar in both nectar and pollination drops. Defence proteins show convergent evolution: they are necessary for defence of exposed sugary secretions from fungi and microbial attack. Sugar solutions are perfect media for growth of fungal and bacterial spores. Microbes can cause profound changes in the physico-chemical properties of nectar and pollination drops, to the detriment of pollinators and pollen, respectively.
Other components common to these various secretions are enzymes involved in sugar metabolism. Among these, invertase is a key enzyme in sugar metabolism. Invertase hydrolyses sucrose to form glucose plus fructose. Its presence in nectars (both floral and extrafloral) and pollination drops indicates that these secretions may be able to alter their composition and osmolarity after they are produced. Previously thought to be static, these secretions may prove to be more dynamic. The presence of these enzymes may also offer a way in which osmolarity may be regulated in response to variable environmental parameters such as temperature and relative humidity allowing a certain degree of homeostasis.
There is fossil evidence that some groups of insects able to feed on sugary secretions, such as pollination drops, already existed in the Jurassic period (Labandeira et al., 2007; Heil, 2008b). These associations are specialized relationships that pre-date those occurring later in angiosperms (Labandeira et al., 2007). In the context of these gymnosperm–insect associations, the presence of proteins in pollination drops may be an exaptation (Gould and Vrba, 1982) for co-opting insects into feeding on these secretions and providing them with a source of carbon (sugars) and nitrogen (proteins).
To return to the question presented in the title, we can say that functionally and ontogenetically the nectar and pollination drops are very different, but chemically they show strong similarities with significant differences. Sugars, although present in different concentration and composition, are strictly necessary for visiting animals' diet and for pollen germination. Defence proteins and carbohydrate metabolism-related proteins protect from changes in composition and osmolarity imposed by factors both biotic (fungi and bacteria) or abiotic (changes in environmental parameters such as temperature and relative humidity).
We are grateful to Martin Heil (Department of Genetic Engineering, CINVENSTAV—Irapuato, Mexico) and Malgorzata Stpiczyńska (Department of Botany, Lublin Agricultural University, Lublin, Poland) for providing Fig. 2A and D, respectively. We thank two anonymous referees for the helpful review of the manuscript. Research related to this review was supported by PAR (University of Siena) and by PRIN (MIUR, Italian Ministry for University and Research to authors M.N., S.M. and E.P.) as well as the National Research Council of Canada (to authors P.v.A., A.C. and R.W.).