Introduction
After the Asteraceae, the Orchidaceae is the largest Angiosperm family, with 736 genera (Chase et al., Reference Chase, Cameron, Freudenstein, Pridgeon, Salazar, van den Berg and Schuiteman2015) and over 26,000 species, with around 500 new species described per year (Chase et al., Reference Chase, Cameron, Freudenstein, Pridgeon, Salazar, van den Berg and Schuiteman2015). There is evidence that orchids arose over 80 million years ago (Ramirez et al., Reference Ramirez, Gravendeel, Singer, Marshall and Pierce2007), and possibly up to 100 million years ago (Chase, Reference Chase, Pridgeon, Cribb, Chase and Rasmussen2001). Orchids are valued for their ornamental and medicinal uses, and are a prominent focus for plant conservation (Yam et al., Reference Yam, Arditti and Cameron2009). Over the past two centuries the interest in orchid species for their ornamental value has intensified and this, coupled with loss of habitat and changing climate, has increased the pressure on natural populations. Along with habitat preservation, ex situ conservation strategies such as seed banking are important for conserving threatened orchid species (Koopowitz, Reference Koopowitz2001; Machado Neto and Custódio, 2005; Seaton and Pritchard, Reference Seaton and Pritchard2008; Li and Pritchard, Reference Li and Pritchard2009; Seaton et al., Reference Seaton, Hu, Perner and Pritchard2010). Seed conservation offers a simple and fairly inexpensive means of conserving diverse genetic material in a relatively small space. Due to their tiny size many thousands of orchid seeds can fit into a single vial, while the space and resources required to maintain similar numbers of living plants would be prohibitive. Mature orchid seeds are very small and are classified as ‘dust seeds’ (Eriksson and Kainulainen, Reference Eriksson and Kainulainen2011). Seed sizes range from 0.28 to 10.09 mm in length, and 0.39 μg to 1.79 mg in weight (Yam et al., Reference Yam, Arditti and Cameron2009). Orchid seeds consist of rudimentary globular embryos of relatively few cells with no endosperm or cotyledonary storage reserves (Lee et al., Reference Lee, Yeung, Lee and Chung2008). Lipid and protein bodies within the embryonic cells form the main storage reserves, and orchid seeds tend not to accumulate starch. Instead, carbohydrates required during germination are provided by fungal symbionts to support early seedling growth (Peterson et al., Reference Peterson, Uetake and Zelmer1998; Rasmussen, Reference Rasmussen2002).
A thorough understanding of seed storage physiology is crucial for successful seed banking, and studies of orthodox seeds of agricultural species have aided understanding of how seed development contributes to seed longevity and germination (Ellis and Pieta Filho, Reference Ellis and Pieta Filho1992; Sanhewe and Ellis, Reference Sanhewe and Ellis1996a, Reference Sanhewe and Ellisb). Little is known about orchid seed reserve deposition and its consequences for the acquisition of desiccation tolerance, timing of harvest, storage behaviour, germination (Schwallier et al., Reference Schwallier, Bhoopalan and Blackman2011) and seedling vigour in comparison with well-studied agricultural species (Lee et al., Reference Lee, Yeung, Lee and Chung2008). Some studies have reported that seeds containing lipids as the major storage component are shorter-lived than seeds with mainly protein or carbohydrate reserves (Nagel and Börner, Reference Nagel and Börner2010), while others have found no association between seed storage reserves and longevity (Walters et al., Reference Walters, Wheeler and Grotenhuis2005; Probert et al., Reference Probert, Matthew and Hay2009). However, seed storability may be related to lipid composition rather than lipid content, particularly in relation to lipid stability (Ponquett et al., Reference Ponquett, Smith and Ross1992) and thermal properties (Crane et al., Reference Crane, Miller, van Roekel and Walters2003, Reference Crane, David Kovach, Gardner and Walters2006), including that of orchids (Pritchard and Seaton, Reference Pritchard and Seaton1993).
Comparative longevity studies of orchid seeds have shown that seeds of 12 species are desiccation tolerant and have improved lifespan on desiccation (Pritchard et al., Reference Pritchard, Poynter and Seaton1999; Hay et al., Reference Hay, Merritt, Soanes and Dixon2010). These studies have shown that orchid seeds are relatively short-lived compared to non-orchid species. While storage at − 18°C may be acceptable for some species, seeds of other species might display temperature-specific cold sensitivity, which could be a factor limiting their longevity at some low temperatures (Pritchard et al., Reference Pritchard, Poynter and Seaton1999; Seaton et al., Reference Seaton, Kendon, Pritchard, Puspitaningtyas and Marks2013). Short longevity limits the effectiveness of conventional ex situ storage in seed banks as a conservation strategy, but alternative techniques such as cryopreservation show promise (Pritchard et al., Reference Pritchard, Poynter and Seaton1999; Merritt et al., Reference Merritt, Touchell, Senaratna, Dixon and Walters2005; Seaton et al., Reference Seaton, Kendon, Pritchard, Puspitaningtyas and Marks2013). However, overall data for orchid seed storage are limited, and questions remain concerning the most suitable environmental conditions for storage (Pritchard and Seaton Reference Pritchard and Seaton1993; Shoushtari et al. Reference Shoushtari, Heydari, Johnson and Arditti1994; Pritchard et al. Reference Pritchard, Poynter and Seaton1999; Machado-Neto and Custódio Reference Machado-Neto and Custódio2005; Hay et al., Reference Hay, Merritt, Soanes and Dixon2010; Hosomi et al., Reference Hosomi, Custódio, Seaton, Marks and Machado-Neto2012).
Despite the potential importance of the lipid reserves in relation to seed desiccation tolerance and longevity, no studies to date have attempted to characterize the fatty acid composition of orchid seeds. A search of the Seed Oil Fatty Acid (SOFA; http://sofa.mri.bund.de/) database for records of the Orchidaceae family yielded eight records with fatty acid composition. Six records related to the lipids of leaves, roots and flower spikes, rather than seeds, of Phalaenopsis and Cattleya hybrids. The remaining two records of fatty acid composition were for an Oncidium sp. and Dendrobium moniliforme, but it is not clear whether these records are for seeds. While orchids represent around 8% of all angiosperm species, < 0.1% of the 7000 plant species records on the SOFA database are for orchids. Similarly, Orchidaceae taxa represent < 1% of the 33,346 taxa on the Royal Botanic Gardens, Kew's Seed Information Database and, of the 289 Orchidaceae taxa, oil content data are available for just three species.
Biochemical studies of orchid seeds are limited by the microscopic nature of the seeds, which means that large quantities of seeds are needed for conventional assays. There are numerous methods for fatty acid analysis, most involve initial extraction in organic solvents such as hexane (Younis et al., Reference Younis, Ghirmay and Al-Shirby2000; Gören et al., Reference Gören, Kiliç, Dirmenci and Bilsel2006; Nedhi et al., Reference Nedhi, Zarrouk and Al-Resayes2011), petroleum ether (Matthaus and Özcan, Reference Matthaus and Özcan2011) or, most commonly, chloroform:methanol (2:1, v/v) (Bligh and Dyer, Reference Bligh and Dyer1959; Colville et al., Reference Colville, Bradley, Lloyd, Pritchard, Castle and Kranner2012). This is followed by transesterification to convert saponifiable lipids to fatty acid methyl esters for analysis by gas chromatography. However, these methods tend to use large amounts of tissue. A number of reports in the literature have shown that direct transesterification methods without prior lipid extraction can result in higher yields of fatty acids, are quicker and avoid the use of toxic chlorinated organic solvents (Lepage and Roy, Reference Lepage and Roy1984, Reference Lepage and Roy1986; Griffiths et al., Reference Griffiths, van Hille and Harrison2010; Cavonius et al., Reference Cavonius, Carlsson and Undeland2014). Smaller-scale direct transesterification methods have been developed for microalgae (Lee et al., Reference Lee, Yoo, Jun, Ahn and Oh2010) and for microheterotrophs (Lewis et al., Reference Lewis, Nichols and McMeekin2000). A sub-microscale assay for fatty acids using gas chromatography–mass spectrometry (GC–MS) has also been described (Bigelow et al., Reference Bigelow, Hardin, Barker, Ryken, Macrae and Cattolico2011). However, these small-scale assays have yet to be applied to orchid seeds.
The aim of this work was to develop a reliable and simple protocol for fatty acid profiling of orchid seeds using GC–MS. Comparison was made between two species: Dactylorhiza fuchsii (common spotted orchid), a temperate terrestrial species belonging to the Orchidoideae subfamily which is widespread across Europe; and Grammatophyllum speciosum (common names include giant orchid, tiger orchid and sugar cane orchid), a tropical epiphytic species belonging to the Epidendriodeae subfamily, and the world's largest orchid, native to Burma, Laos, Vietnam, Indonesia and Malaysia.
Material and methods
Seed material
Seeds of D. fuchsii were collected in the Wakehurst Place estate in Ardingly, UK in 2013 and G. speciosum seeds were obtained from the Prince of Songkla University, Thailand in 2011 and stored at 15% relative humidity (RH) and 5°C since their arrival in ripe capsules. Seeds of both species were cleaned and stored at 15% RH and 5°C in paper bags. Initial viability was determined using tetrazolium (TZ) staining as described by Hosomi et al. (Reference Hosomi, Santos, Custodio, Seaton, Marks and Machado-Neto2011). Germination was conducted in three Petri dishes, each with 5 mg seeds. Seeds were disinfected according to Hosomi et al. (Reference Hosomi, Custódio, Seaton, Marks and Machado-Neto2012) with 5 g l− 1 of dichloroisocyanuric acid sodium salt (DCCA; Sigma, Colorado, USA), rinsed three times and sown in Murashige and Skoog medium (Murashige and Skoog, Reference Murashige and Skoog1962) at half strength.
Fatty acid transesterification
Due to the ease of preparation and high fatty acid recovery (Cavonius et al., Reference Cavonius, Carlsson and Undeland2014) sulphuric acid in methanol was selected as the catalyst for direct transesterification in this study, following a method adapted from Christie (Reference Christie1989) and Colville et al. (Reference Colville, Bradley, Lloyd, Pritchard, Castle and Kranner2012).
Three replicates of each sample mass of 2.5, 5, 10, 15 and 20 mg were used to determine the minimum sample mass required for reliable determination of fatty acid composition. Seeds were placed into 20-ml glass vials with 1 ml of toluene containing 50 mg l− 1 butylated hydroxytoluene (BHT) as an antioxidant. As an internal standard, 10 μl of 10 mg ml− 1 heptadecanoic acid was added. The samples were incubated for 4 h at room temperature. Fatty acids were methylated with 2 ml of 1% (v/v) sulphuric acid in methanol overnight at 50°C with constant shaking at 150 rpm. The fatty acid methyl esters (FAMEs) were isolated by partitioning with 5 ml of hexane and 5 ml of NaCl (5%, w/v). The organic phase was transferred to a clean vial and the aqueous phase was washed with another 5 ml of hexane. Removal of the organic phase was undertaken carefully to avoid the aspiration of seeds into the Pasteur pipette. The organic phases were combined and dried at 45°C under a stream of nitrogen gas. The residue was dissolved in 1 ml of hexane and transferred to 2-ml autosampler vials for GC–MS analysis.
The effect of incubation time in toluene was investigated by extending the 4-h incubation to 16 h (overnight) before derivatizing the fatty acids as described above. In addition, comparison was made between the extraction efficiency from intact and crushed seeds using sample masses of 5 mg. Seeds were crushed between glass slides prior to overnight extraction in toluene as described above.
GC–MS analysis of fatty acid methyl esters
FAMEs were separated by GC (Thermo Finnigan Trace GC Ultra) using a FAMEWAX column (30 m length, 0.25 mm internal diameter, 0.25 μm film thickness (df); Thames Restek UK Ltd, Saunderton, Buckinghamshire, UK) running a temperature program (initial temperature 70°C; 20°C min–1 until 195°C; 5°C min–1 until 240°C; 10 min at 240°C) with helium as the carrier gas (constant flow rate of 1 ml min–1). The compounds were detected by MS (Thermo Finnigan Trace DSQ; ionization energy 70 eV, scanning frequency m/z 10–500 at 0.3 s) and identified by comparison with the NIST mass spectral database (National Institute of Standards and Technology, Gaithersburg, Maryland, USA) and analytical standards (FAME Mix C4–C24, Supelco, Bellefonte, Pennsylvania, USA). FAME quantification was performed using quantitative standard curves (FAME Mix GLC-10, -30 and -50, Supelco).
Data analysis
Based on the molar percentage fatty acid composition, the double-bond index (DBI) and peroxidizability index (PI) were calculated according to Pamplona et al. (Reference Pamplona, Portero-Otín, Riba, Ruiz, Prat, Bellmunt and Barja1998). Analysis of variance and a post-hoc Tukey test (Genstat v.14; GenStat Committee, 2011) were used to determine statistical significance between sample masses and treatments. All differences quoted in the text were significant at P < 0.05.
Results and discussion
Seed viability
Germination and viability of the seed lots were assessed immediately prior to the fatty acid analyses. The total germination of D. fuchsii seeds was 89.7% and the viability according to the TZ test was 84.8%. G. speciosum seeds failed to germinate, possibly due to the use of inappropriate germination media or induction of dormancy, but TZ staining indicated that 87.6% of seeds were viable. The G. speciosum seeds had been stored at 15% RH and 5°C for 2 years since 2011, when the germination and viability of the fresh seeds were 57% and 99%, respectively. This indicates that the G. speciosum seeds had experienced some deterioration during storage, but that seed viability was comparable between the two species at the time of analysis.
Fatty acid yield and composition
The fatty acid composition of D. fuchsii and G. speciosum seeds was fairly similar in terms of the major fatty acid components, which were methyl linoleate, followed by methyl palmitate and methyl oleate (Table 1). These fatty acids are the most common polyunsaturated, saturated and monounsaturated fatty acids, respectively, found in plants. Methyl octadecenoate (unresolved double-bond position) and methyl stearate were present at low concentrations (Table 1). G. speciosum also contained methyl pentadecanoate, methyl linolenate and methyl arachidate as minor components. The total fatty acid concentration in D. fuchsii seeds was more than double that in G. speciosum seeds. There was no significant effect of sample mass on the concentration of methyl linoleate or methyl oleate in D. fuchsii seeds (Table 1), but for methyl palmitate, methyl stearate and methyl octadecenoate there were increases in concentration with sample mass. For G. speciosum seeds, where fatty acid concentrations were lower than in D. fuchsii seeds, there was an increase in all fatty acid concentrations as sample mass increased. At the lowest sample mass (2.5 mg) only four fatty acids were detected in G. speciosum seeds: methyl palmitate, methyl stearate, methyl oleate and methyl linoleate. This increased to six fatty acids when the sample mass was doubled to 5 mg, with the detection of methyl octadecenoate and methyl arachidate; and seven fatty acids when the sample mass was increased to 10 mg, with the detection of methyl pentadecanoate. Methyl linolenate was not detected at sample masses lower than 20 mg.
Table 1 Content of fatty acid methyl ester components in Dactylorhiza fuchsii and Grammatophyllum speciosum seeds obtained by direct derivatization of different seed masses following incubation in toluene for 4 h or 16 h. Values represent the mean (n=3) fatty acid methyl ester concentration. Lowercase letters indicate significant differences (P < 0.05) between the content of individual fatty acids in different sample masses, and uppercase letters indicate significant differences (P < 0.05) in the fatty acid content following 4 h or 16 h incubation in toluene
*Methyl octadecenoate – unresolved double-bond position.
The proportion of each fatty acid as a percentage of total fatty acid concentration showed a significant change with sample mass for D. fuchsii, and also for G. speciosum (Table 2). In both species the percentage of methyl linoleate declined, while the percentages of the fatty acids with lower abundance increased as sample mass increased. Due to their lower abundance, more saturated fatty acids were detected as sample mass was increased, which altered the ratio of unsaturated:saturated fatty acids, and also the double-bond index (DBI) and peroxidizability index (PI) (Table 2). Lewis et al. (Reference Lewis, Nichols and McMeekin2000) reported that variation in sample mass (1–20 mg) had no significant effects on the total amount or relative proportion of fatty acids from two microheterotrophs. Likewise, Li et al. (Reference Li, Beisson, Pollard and Ohlrogge2006) observed no difference in fatty acid composition of Arabidopsis seeds between direct transesterification of 10 mg of seeds compared to ~1 mg of seeds. It is likely that the minimum sample size for obtaining reproducible fatty acid composition data depends on the oil content of the sample and the limit of detection of the analytical method. In this study there were less significant changes in fatty acid concentration or relative abundance at sample masses ≥ 10 mg (Tables 1, 2), which indicates that a sample mass of 10 mg is sufficient to obtain reliable fatty acid composition data for D. fuchsii and G. speciosum seeds. However, optimization of sample mass may be required for other orchid species. Seed weights across the Orchidaceae family are reported to vary from 0.39 μg to 1.79 mg (Yam et al., Reference Yam, Arditti and Cameron2009), which means that the number of seeds required for a sample mass of 10 mg could range between 5 and 26,000. However, given that a single capsule of G. speciosum may contain up to 2 million seeds (Seaton et al., Reference Seaton, Kendon, Pritchard, Puspitaningtyas and Marks2013), while capsules of D. fuchsii may contain ~2000 seeds, each weighing ~1.9 μg (Marks et al., Reference Marks, Seaton and Pritchard2014), a sample mass of 10 mg is feasible for many studies.
Table 2 Fatty acid methyl ester composition of the seeds of two orchid species: Dactylorhiza fuchsii (a temperate species) and Grammatophyllum speciosum (a tropical species). Values represent the mean ± SD (n=3) abundance of each fatty acid as a percentage of the total fatty acid methyl ester (FAME) molar concentration. The unsaturated:saturated fatty acid ratio is shown along with the double-bond index (DBI) and peroxidizability index (PI), which were calculated as described in the text. Letters indicate significant differences (P < 0.05) between different sample masses
*Methyl octadecenoate – unresolved double-bond position.
To determine whether the efficiency of fatty acid extraction could be improved, the solubilization step in toluene prior to derivatization was extended from 4 to 16 h (i.e. overnight). The longer incubation time in toluene had no significant effect on the fatty acid yield from seeds of either species (Fig. 1), indicating that the 4-h solubilization step in toluene is sufficient for obtaining reproducible yields. Other direct transesterification methods, e.g. Lewis et al. (Reference Lewis, Nichols and McMeekin2000) and Li et al. (Reference Li, Beisson, Pollard and Ohlrogge2006), do not include a pre-incubation in toluene, but toluene is considered necessary for the solubilization of non-polar lipids such as triacylglycerols (Christie, Reference Christie1989), which form the bulk of seed-storage lipid reserves. Orchid seeds tend to have a hydrophobic, balloon-like testa, which causes the seeds to clump together and float on the surface of solvents. For this reason a pre-incubation in toluene was included to improve the access for organic solvents in the subsequent derivatization step.
Figure 1 The yield of fatty acid methyl esters from intact Dactylorhiza fuchsii and Grammatophyllum speciosum seeds (5 mg) following incubation in toluene for 4 h (light grey bars) or 16 h (white bars) compared to the yield from crushed seeds following incubation in toluene for 16 h (dark grey bars). Bars represent mean ± SD (n= 3). Letters indicate significant differences (P< 0.05) in the yield of individual fatty acids between different extraction protocols.
Comparison was also made between the yield of fatty acids from intact seeds and seeds that had been crushed prior to 16 h incubation in toluene. Crushing the seeds prior to direct transesterification tended to reduce the fatty acid yield, but the reduction was only significant for methyl palmitate (P= 0.012) in D. fuchsii seeds, and methyl stearate (P= 0.035), methyl octadecenoate (P= 0.007) and methyl arachidate (P= 0.007) in G. speciosum seeds. This suggests that crushing the seeds led to loss of a portion of the sample, possibly due to incomplete recovery of material from the surface of the glass slides. Direct transesterification of intact Arabidopsis seeds was also reported to result in similar oil yield and fatty acid composition as the derivatization of ground seed samples (Li et al., Reference Li, Beisson, Pollard and Ohlrogge2006), indicating that for small seeds with thin seed coats grinding is unnecessary, and in some cases may lead to sample loss.
Fatty acid unsaturation indices and relationships with seed longevity
The data for the 10 mg sample mass was used to compare the fatty acid unsaturation indices between the two species. The ratio of unsaturated to saturated fatty acids was lower in D. fuchsii compared to G. speciosum (P= 0.001; Table 2), indicating a greater degree of lipid saturation. D. fuchsii had a DBI of 154.7 ± 0.18 and a peroxidizability index (PI) of 70.8 ± 0.15, which were both significantly lower (P < 0.001) than those of G. speciosum (DBI = 164.7 ± 1.12; PI = 79.2 ± 0.87). This suggests that D. fuchsii seed lipids are less susceptible to oxidative damage than G. speciosum seed lipids. This could have implications for seed storage because lipid peroxidation is considered to play a role in seed deterioration during storage (Sung, Reference Sung1996; Goel and Sheoran, Reference Goel and Sheoran2003; Ratajczak and Pukacka, Reference Ratajczak and Pukacka2005). Preliminary storage experiments indicate that G. speciosum seeds may be shorter-lived than D. fuchsii seeds. Germination of G. speciosum seeds declined from 50% to 0% during 24 months of storage at 15% RH and 20°C, compared to germination of D. fuchsii which declined from 80% to 62% under the same conditions (T.R. Marks, unpublished data). Both the DBI and PI of the mitochondrial membrane were found to be inversely correlated with maximum life span in mammals (Pamplona et al., Reference Pamplona, Portero-Otín, Riba, Ruiz, Prat, Bellmunt and Barja1998). Similarly, a negative relationship between DBI and longevity was reported for desiccated anhydrobiotic plant systems, e.g. seeds, pollen and spores (Hoekstra, Reference Hoekstra2005), which supports the possibility that fatty acid unsaturation could be a factor contributing to seed longevity. Although Ponquett et al. (Reference Ponquett, Smith and Ross1992) did not observe a significant relationship between total lipid unsaturation and longevity for seeds of eight species, relations between seed life span and indices of fatty acid unsaturation require more detailed evaluation.
Conclusions
A straightforward method for profiling the fatty acid composition of orchid seeds using GC–MS is presented. The method can provide reliable composition data using sample masses as low as 10 mg. Direct transesterification of fatty acids requires no pre-processing of seed samples and eliminates time-consuming extraction procedures and reduces the potential for sample loss. The fatty acid profiles of the seeds of two orchid species (D. fuchsii and G. speciosum) were determined using this method, and are presented for the first time. Characterization of orchid seed lipids will provide a valuable insight into storage-reserve deposition and the role of lipids in seed storability and longevity.
Financial support
This work was supported by Science without Borders fellowships (C.C.C. grant number 245783/2012-1, and N.B.M.N. grant number 245777/2012-1; both from Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brazil); Royal Botanic Gardens, Kew receives grant-in-aid from Defra.
Conflicts of interest
None.
