Interspecific and Intersexual Differences in the Chemical Composition of Floral Scent in Glochidion Species (Phyllanthaceae) in South China

Plants of theGlochidion (Phyllanthaceae) genus are pollinated exclusively by host-specific Epicephala (Gracillariidae) moths. Floral scent has been thought to play key role in the obligate pollination mutualism between Glochidion plants and Epicephala moths, but few studies have been reported about chemical variation in floral volatiles of Glochidion species in China. Floral volatiles of male and female flowers of five Glochidion species in south China were collected by dynamic headspace absorption technique and then were chemically analyzed by using gas chromatography-mass spectrometry. A total of 69 compounds were identified from floral scents of five investigated species. Glochidion hirsutum and G. zeylanicum showed no qualitative differences in floral scent, whereas there were clear variations of floral scent among other species (G. eriocarpum, G. daltonii, and G. sphaerogynum) and also they distinctly differed from these two species. Male flowers emitted significantly more scent than female flowers.Glochidion plants exhibited qualitative and quantitative differences in floral scent between two sexes of flowers. The findings suggest that the volatile variation of floral scent among Glochidion species reflects adaptations to specific pollinators. Sexual dimorphism in floral scent has evolved to signal alternative rewards provided by each sex to Epicephalamoths.


Introduction
Mutualism is cooperative interactions in which each associated species benefits from its partner [1,2]. Obligate pollination mutualism between plants and their seed-parasitic pollinators is known as one of the most specialized plantinsect interactions [3][4][5][6], where insects usually exclusively pollinate for the host flowers in which they lay eggs; in turn the hosts offer seeds to feed pollinator larvae [7,8].
The classical examples of such nursery pollination systems are the fig-fig-wasp [4] and yucca-yucca-moth interactions [6]. Moreover, the obligate pollination mutualism between Epicephala moths (Gracillariidae) and tribe Phyllantheae plants (Phyllanthaceae) has been also documented, including the associations of Epicephala moths with Glochidion, 2 Journal of Chemistry of the seeds ensure that the hosts reproduce their offspring [5].
The most intriguing characteristics among Glochidion-Epicephala mutualism are extremely high taxonomic diversity and species specificity of both plants and moths [5,7,14]. Pollinators' encounter with host plants is the key step in these specialized pollination interactions. Floral scents are complex blends of volatile organic compounds, which have been thought to be crucial chemical signals guiding obligate nocturnal moths to locate host plants in Glochidion-Epicephala associations. Interspecific variations of floral scents among Glochidion plants are essential for maintaining species specificity of Glochidion-Epicephala mutualism [15][16][17], and Epicephala moths are also capable of discriminating their hosts using floral odor [16,18].
So far, studies on Glochidion-Epicephala mutualism have focused on pollination biology and insect taxonomy in China [19,20] and Japan [5,9,10]. However, little is reported about the chemical composition of floral scents from Glochidion plants in China. Recently, two pollination relationships between Glochidion and Epicephala were found in south China: (1) one Epicephala species exclusively pollinates one Glochidion species (G. eriocarpum, G. daltonii, and G. sphaerogynum) (unpublished data); (2) one Epicephala species jointly pollinates two Glochidion species (G. hirsutum and G. zeylanicum) [12]. For maintaining extremely high species specificity of Glochidion-Epicephala mutualism, floral scents may differ among Glochidion species pollinated by different Epicephala moths, while Glochidion species pollinated by the same Epicephala moth should produce similar floral signals. Therefore, it is speculated that variations in floral scent among different Glochidion species should reflect differences in Epicephala moth species with which the plants are associated.
The dioecious plants pollinated by animals often, but not always, transmit similar floral signals between the male and female flowers to attract the same pollinator to complete conspecific pollen transfer [21][22][23]. An implicit assumption in the fact is that pollinators use similar signals to seek the similar reward from flowers of both sexes (e.g., floral nectar) [21]. Conversely, if the plants offer different rewards to the specific pollinators on the male flowers (e.g., pollen) and female flowers (e.g., ovule), floral characters (e.g., shape, visual, and odor) may be selected to diverge between both sexes of the flowers to signal alternative rewards [23,24]. Interestingly, in Epicephala-pollinated Phyllanthaceae plants, there was sexual dimorphism of floral scents between male and female flowers, which is in contrast to scent from non-Epicephala-pollinated Phyllanthaceae [18,25]. Female Epicephala moths pollinate the host plants, involving distinctly different behaviors on male and female flowers: pollen collection from male flowers and then pollen deposition and oviposition on female flowers [5,[9][10][11]. Therefore, it was thought that sexual dimorphism in floral scent of Epicephala-pollinated Phyllanthaceae plants has evolved to signal alternative floral rewards provided by each sex to Epicephala moths [18]. In our field observations on Glochidion-Epicephala mutualism, Epicephala moths also exhibited similar high specific pollination behaviors on male and female flowers as found in previous studies [5,[9][10][11].
Based on the above discussion, we hypothesized that, in mutualism recently found in China between Glochidion and Epicephala, (1) variations in floral scent composition among five Glochidion species may reflect adaptations to species-specific pollinators with which the host plants are associated and (2) Glochidion plants should be selected to produce different fragrance blends between the male and female flowers to induce Epicephala moths to get the different rewards from flowers of each sex. Consequently, the floral volatiles of both male and female flowers in five Glochidion species were collected, respectively, by the headspace adsorption technique and chemically characterized using gas chromatography-mass spectrometry (GC-MS). Then, nonmetric multidimensional scaling (NMDS) was performed to investigate interspecific variations and intersexual differences in floral scent composition of five Glochidion species in south China.

Plants.
Five studied Glochidion species, including G. eriocarpum, G. daltonii, G. sphaerogynum, G. hirsutum, and G. zeylanicum, generally inhabit forest margins of tropical and subtropical forests in south China [13]. Of these five species, G. hirsutum and G. zeylanicum grow together and occurred in sympatry at the study sites. In contrast, the other three species occurred in other different study sites and so their distributions did not overlap ( Figure 1).
Glochidion species investigated in this study are monoecious trees with inconspicuous green female flowers and yellow male flowers which are aggregated on the axils of the leaves ( Figure 2). Female flowers are sessile, more or less budlike, and appear to have specialized morphology: the style fused into a column at the trip with a small opening where the stigma is concealed in the inner surface, which makes it difficult to touch for flower visitors other than Epicephala moths. Male flowers are pedunculate with unfolded five or six sepals to expose the stamen. Glochidion flowers are small and are fragrant to the human nose only during the evening, and the floral scents of Glochidion species may be one of the key signals attracting nocturnal obligate pollinator Epicephala moths.
The five Glochidion species are associated with four Epicephala moth species in our study regions [20,26]. Among these five species, G. hirsutum and G. zeylanicum share similar morphology except that the trichomes on the epidermis are produced in the female flowers of G. zeylanicum rather than that of G. hirsutum (Figures 2(g) and 2(i)) and pollinated by the same Epicephala species, E. bipollenella. However, the other three species are, respectively, pollinated by other species-specific Epicephala moths.

Floral Scent Collections.
To analyze floral scent profiles of both male and female flowers of five Glochidion species, odor collections were performed using the dynamic headspace adsorption technique [27]. The details of the study sites, dates, and sample sizes are given in Figure 1 and Table 1. Before scent collection, specimens of each studied species were gathered   to save in our lab as a voucher. Female and male flowers of each species were removed from plants and put separately into the polyvinyl acetate bags for floral scent collection (Toppits; Melitta Scandinavia AB, Klippan, Sweden). Air was purified by active charcoal (Sigma-Aldrich, USA) and then introduced into the bag through Teflon tubes (VICI Jour Ltd., Sweden) with a mini air-pump. Air-containing floral volatiles were sucked through the small glass tube (outer diameter 6 mm, length 75 mm) filled with 100 and 50 mg of Tenax-TA (mesh 60/80; Sigma-Aldrich) to adsorb floral scent. Meanwhile, air samples were also collected from empty bags as controls to identify possible background contamination during volatile collections. The flow rate of air in the inlet and outlet of the sample bag was set to 400 mL min −1 . Odor collection was done at ambient temperature (25-28 ∘ C) indoors and continued for 3 h during the night. After collection, the filters were sealed with Parafilm and then wrapped with silver paper and stored at −20 ∘ C until used for analysis. Before analysis, the collected floral scents in Tenax-TA filters were eluted with 2 mL of hexane into brown glass vials, and then 10 L of n-Eicosane (1 mg mL −1 ) was added as the internal standard to calculate release amount of individual scent compounds. The eluate containing the floral volatiles was carefully concentrated down to 100 L under N 2 flow.

Chemical Analysis of Floral Scents.
To analyze chemical composition of floral scents, the headspace samples of Glochidion species were subjected to GC-MS, using an HP 6890 gas chromatograph (Agilent Technologies Inc., USA), and coupled with an HP 5975 mass spectrometer (Agilent Technologies Inc.). The GC was equipped with HP-5MS capillary column (30 m × 0.25 mm × 250 m film thickness; Agilent Technologies Inc.). Helium was used as the carrier gas at a flow rate of 1.0 mL min −1 . For each sample, 2 L was injected in a split mode (the split ratio was 1 : 1) for 1 min with an injector temperature of 250 ∘ C. After injection, the oven temperature was programmed to be kept at 40 ∘ C for 5 min, followed by an increase of 3 ∘ C min −1 to 150 ∘ C and 10 ∘ C min −1 to 200 ∘ C, where it was held for 5 min. Mass spectra were recorded at 70 eV in the electron impact ionization mode, and the ionic source temperature was 230 ∘ C.
Compound identification was carried out using the NIST 08 mass spectral libraries and further confirmed by comparison of retention indices with those reported in the literature [28,29], the NIST Chemistry WebBook (http://webbook.nist.gov/chemistry/) [30], and The Pherobase (http://www.pherobase.com/) [31]. Compounds in the samples presented at higher or similar percentages in controls were considered as contaminants and excluded from the analysis. For the quantities of compounds, the amount of each volatile compound was calculated as its peak area relative to the total peak area on gas chromatograms.

Chemical Composition of Floral
The relative contents of four chemical classes of compounds from male and female flowers among the respective five species are shown in Figure 3. For male flowers, the floral scents emitted by the five species could be divided into four groups: G. eriocarpum (86.65%) and G. daltonii (94.76%) dominated by sesquiterpenoids; G. sphaerogynum (67.87%) dominated by monoterpenoids; aliphatics with monoterpenoids in G. hirsutum (56.46 and 23.33%, resp.); aliphatics with aromatics in G. zeylanicum (38.56 and 29.85%, resp.). Likewise, the floral scents produced by female flowers of the five species were also divided into four groups: G. eriocarpum (51.73 and 32.68%, resp.) dominated by monoterpenoids with aromatics; G. sphaerogynum (67.95%) dominated by sesquiterpenoids; G. hirsutum (73.71%) and G. zeylanicum (78.02%) dominated by monoterpenoids. The findings indicated degrees of difference in chemical composition of floral scents among the five species.
In addition, there were clear differences in scent composition pattern of four different chemical classes between the male and female flowers of the five species. For example, in G. daltonii, the major components of male flowers were sesquiterpenoids, whereas those of the female flowers were monoterpenoids.            the results showed evident quantitative differences in floral scents between male and female flowers of the Glochidion species.

Sexual Dimorphism of Floral Scents.
The NMDS analysis of floral scents in male and female flowers of each Glochidion species is shown in Figure 6. In each investigated Glochidion species, the headspace samples of male flowers were clearly separated from those of female flowers (G. daltonii: 2 = 6, = 0.034; G. eriocarpum: 2 = 9, = 0.03; G. sphaerogynum: 2 = 5.33, = 0.021; G. hirsutum and G. zeylanicum: 2 = 3.857, = 0.05). The results indicated a qualitative difference between floral scents produced by male and female Glochidion flowers, which is so-called sexual dimorphism of floral scents.

Discussion
In total, 69 volatiles are detected from the floral scents of the five Glochidion species. Most of these reported volatiles are quite common in floral odors of many flowering plants [33,34]. In particular, 26 compounds also occur in floral fragrances of other Phyllantheae species, close relatives of the investigated Glochidion species, which are also exclusively pollinated by Epicephala moths [16,18,35]. For instance, (E)--ocimene (compound 25) and linalool (compound 28) occurred together in the five studied Glochidion species and are also produced by the flowers of four other reported Glochidion species [16,18]. In addition, these two compounds can attract a wide range of pollinators of flowering plants, such as moths [35,36], butterflies [37], and bees [38]. Therefore, our results suggest that these interactions between Glochidion species and Epicephala moths may be mediated by conventional floral compounds. This is similar to findings in most obligate insect-plant mutualism, where identified floral volatiles are also regarded as general floral volatiles [16,25,39]. However, Chen et al. [40] showed that an unusual compound, 4-methylanisole, mediated the species specificity in Ficus semicordata and Ceratosolen gravelyi wasp mutualism through a "private channel." In highly specific plant-insect interactions, the floral scent profile of host plants is sometimes dominated by one or few compounds [16,39,41,42], and these dominant compounds may account for variation in floral scents among species [16]. Similarly, our GC-MS showed the floral odor of each of the five species was also dominated by one compound at least, such as, for male flowers, -elemene (compound 35,28 results revealed that both male and female flowers of the five species could be divided into four distinct groups: G. zeylanicum and G. hirsutum clustered together and the other species were spaced with each other. Thus, floral odors of investigated Glochidion species could be clearly discriminated mainly due to the dominant compounds. This is in contrast to findings for the odor profile of five other Glochidion species that were distinguished by minor compounds [16]. These results suggest that the presence of dominant compounds in floral scents of Glochidion species may be crucial for attracting specific Epicephala pollinators. However, electroantennographic detection analysis and behavioral tests with authentic chemicals are needed for further screening of specific compounds responsible for this host specificity.
Highly interspecific variation in floral scent has been observed in many plant species [16,39,[43][44][45][46][47][48][49][50]. It is likely that variation in floral odor composition between species is due to adaptation to host-specific pollinators [16,50]. The species specificity in the Glochidion-Epicephala mutualism is extremely high, where each host plant is usually exclusively pollinated by specific pollinator species. Thus, to facilitate host specificity in these interactions, the floral scent should be selected to be strongly divergent among species. Interspecific variation in floral scent profile of Glochidion species observed in this study could contribute to the maintenance of the species specificity of the mutualism and may reflect adaptation to specific Epicephala moths. Pollinators can also find and locate their hosts by discriminating the difference in floral odor among species [16,40]. Our NMDS analysis showed a similar floral scent profile between G. hirsutum and G. zeylanicum. In Glochidion-Epicephala mutualism, E. bipollenella moths jointly pollinated G. hirsutum and G. zeylanicum [20]. Similarly, G. obovatum and G. rubrum, two parapatric species pollinated jointly by the same Epicephala species, produced similar floral scents, more similar than those emitted by other cooccurring Glochidion species pollinated by other Epicephala species [16]. In addition, Ficus natalensis and F. burkei jointly sharing the same pollinator wasp species emitted similar floral scent profiles, which are different from those produced by other African Ficus species [51]. When different plant species are pollinated by the same group of pollinators and especially by closely related pollinator taxa, they may be expected to have floral volatiles of similar chemical composition [43]. This implies that the similarity in odor of different Glochidion species in this study may suggest evolutionary convergence to attract the same Epicephala pollinators [51].
Generally, but not always, the animal-pollinated dioecious plants produce similar floral signals between male and female flowers. This is because attraction of the same pollinator to flowers of both sexes is essential for successful conspecific pollen transfer; thus the hosts should be selected to transmit such similar floral scents in the flowers of both sexes [21][22][23]. Underlying the fact above is that pollinators use similar cues to find the same floral rewards (e.g., floral nectar) from male and female flowers [21,24]. Selection for similarity in signals from male and female flowers is thought to be particularly strong when one sex (often the female) does not produce any reward and effectively mimics flowers of the other sex, so deceptive pollination occurs in the female [52]. For instance, chemical mimicry in floral scent occurs in fig-fig-wasp [40,52,53]. In addition, chemical mimicry between sexes of the flowers has also been documented in cycad species [50,[54][55][56][57]. Hemborg and Bond [24] proposed that different rewards offered by male and female flowers of some plant species to pollinators could promote the evolution of sexually dimorphic floral signals of flowers. Svensson et al. [25] and Okamoto et al. [18] found that Epicephala-pollinated Phyllanthaceae plants consistently showed clear differences in scent between male and female flowers. Similarly, our NMDS results also revealed in Glochidion-Epicephala a remarkable sexual dimorphism in floral scent between sexes of the five investigated Glochidion species. In Epicephala-pollinated Phyllanthaceae, when Epicephala moths pollinate for the host, male flowers can offer pollen to pollinators as food, whereas female flowers can provide the ovule for Epicephala adults as egg-laying sites and seeds for Epicephala larvae as foods [5,[9][10][11][12], and it was thought that Epicephala moths could get access the different and essential rewards from the flowers of both sexes of hosts; thus sexual dimorphism in floral scent of two sexes in Epicephala-pollinated Phyllanthaceae is selected to signal an alternative reward provided by each sex to Epicephala moths. Epicephala can also distinguish male and female flowers by odor and is crucial for successful reproduction of the moths and plants [18]. Our results also provide strong support that different rewards in male and female flowers result in the sexual dimorphism in floral scent of flowers of Glochidion species in this study [24].
Our results showed quantitative differences in floral scent emission between the sexes in each studied Glochidion species. In addition, there was a similar finding for B. vitisidaea, a close relative of Glochidion, which is also obligately pollinated by species-specific Epicephala moths [25]. Okamoto et al. [18] showed that Epicephala moths preferred the scent of male flowers over that of female flowers in G. zeylanicum. According to sexual selection theory, the males are limited in their reproductive success by access to mates, while females are mostly limited by available resources [58]. In animal-pollinated plants, attraction of pollinators and successful pollination are crucial for reproductive success; thus males in monoecious plants are expected to be under selection to increase their attractiveness to pollinators by investing more than females in floral traits that enhance pollinator visitation. Generally, male plants should produce more floral scent than the females to increase pollinator attraction. Waelti et al. [59] suggested that, in the Silene latifolia-Hadena bicruris nursery pollination system, male flowers emitted significantly more fragrance, especially of the pollinator-attracting compounds, than female flowers, and pollinator-moths preferred male over female flowers in behavioral tests, implying that evolution of the divergence in the floral signal of S. latifolia is shaped by sexual selection. Ashman et al. [60] also showed that the smaller hermaphroditic flowers in the gynodioecious strawberry Fragaria virginiana produced larger amounts of foliage volatiles, which resulted in attraction of more pollinators compared to female flowers. Therefore, more probably, the observed quantitative difference in floral scent of flowers between the sexes in the five species of Glochidion may be an outcome of sexual selection. It is possible that the stronger odor emission of male flowers in Glochidion species resulted in increasing attractiveness for Epicephala moths compared with female flowers.

Conclusions
Our findings suggest that the volatile variation of floral scent among Glochidion species reflects adaptations to specific pollinators. Sexual dimorphism in floral scent has evolved to signal alternative rewards provided by each sex to Epicephala moths. It is concluded that floral scent has been thought to play key role in the obligate pollination mutualism between Glochidion plants and Epicephala moths.