Introduction
For the majority of vascular plants, flooding is deleterious to plant health and growth (McClean, Reference McClean2000) and may result in premature death (Bailey-Serres and Voesenek, Reference Bailey-Serres and Voesenek2008). During flooding, seed germination is limited in terrestrial, wetland and even some aquatic species (Gurnell et al., Reference Gurnell, Goodson, Thompson, Mountford and Clifford2007). Flooding restricts oxygen availability to the embryo, preventing germination or inducing dormancy (Hook, Reference Hook and Kozlowski1984; Kozlowski and Pallardy, Reference Kozlowski and Pallardy1997; Mollard et al., Reference Mollard, Insausti and Sánchez2007). However, some plants have evolved physiological adaptations to survive in ecosystems with regular flooding (Bailey-Serres and Voesenek, Reference Bailey-Serres and Voesenek2008; Parolin, Reference Parolin2009).
The environment of the Amazonian flooded forests is unique. Frequency and duration of flooding and the height of the flood-water level determine which species germinate, establish and reproduce along the flood-level gradient (Junk et al., Reference Junk, Bayley, Sparks and Dodge1989; Wittmann and Parolin, Reference Wittmann and Parolin1999; Piedade et al., Reference Piedade, Junk and Parolin2000; Parolin, Reference Parolin2009). For example, near Manaus, the water levels reach up to 11 m and last up to 130 days (Marinho et al., Reference Marinho, Lopes, Assis, Ramos, Gomes, Wittmann and Schöngart2013), with temperatures from 23 to 32 °C, and dissolved oxygen concentrations from 0.6 to 8.9 mg l–1 (Rai and Hill, Reference Rai and Hill1980). Seeds of some species have evolved adaptive strategies to survive, germinate and form seedlings, including Crateva benthami Eichler, Mora paraensis (Ducke) Ducke, Nectandra amazonum Nees, Vatairea guianensis Aubl. (Parolin, Reference Parolin2001) and Himatanthus sucuuba (Spruce ex Müll. Arg.) Woodson (Ferreira et al., Reference Ferreira, Piedade, Tiné, Rossatto, Parolin and Buckeridge2009). Others exhibit radicle protrusion but no shoot development underwater and the seeds of some species only swell and split the seed coat (Parolin, Reference Parolin2001). Submersion tolerance may be related to the major storage reserves. Starchy seeds of Eugenia inundata DC. and Pouteria glomerata (Miq.) Radlk. show enhanced seedling development underwater compared with oily seeds of Crataeva tapia L., Pseudobombax munguba (Mart.) Dugand and Simaba guianensis Aubl., with the exception of oily seeds of Laetia corymbulosa Spruce ex Benth. (Melo et al., Reference Melo, Franco, Silva, Piedade and Ferreira2015). We sought to determine some physiological, morphological and biochemical features that enable seeds of Eugenia stipitata ssp. sororia McVaugh (Myrtaceae) to survive and grow under laboratory conditions comparable to the flooded forests.
Eugenia stipitata occurs along rivers in Western Amazonia. Mature seeds have no endosperm (Flores and Rivera, Reference Flores and Rivera1989) and the large embryo does not show any apparent differentiation between the embryonic axis and the cotyledons (McVaugh, Reference McVaugh1958). The whole embryo has meristematic potential and has been termed totipotent, because of its capacity to re-sprout when cut into halves (Anjos and Ferraz, Reference Anjos and Ferraz1999; Mendes and Mendonça, Reference Mendes and Mendonça2012) or quarters (Calvi et al., Reference Calvi, Aud, Ferraz, Pritchard and Kranner2017), such that normal seedlings develop from all seed fragments. The seeds are capable of surviving and germinating subsequently during flooding (Mendes and Mendonça, Reference Mendes and Mendonça2012). Germination takes about 4 months (Anjos and Ferraz, Reference Anjos and Ferraz1999) and can be reduced to 2 months by removing the mechanical constraint of the fibrous seed coat (Gentil and Ferreira, Reference Gentil and Ferreira1999). The seeds are recalcitrant (i.e. desiccation-intolerant) and die when seed moisture content (MC) falls below 26% (Gentil and Ferreira, Reference Gentil and Ferreira1999). Based on this set of characteristics, E. stipitata was proposed as a recalcitrant seed model for the application of multiple measurements in single seeds (Calvi et al., Reference Calvi, Aud, Ferraz, Pritchard and Kranner2017).
In the present study, E. stipitata seeds were kept for up to 1 year in running water at two water depths. In ‘shallow’ water, the seedlings could reach the water surface after germination, whereas in the ‘deep’ water condition they could not. With germination and seedling development the demand for oxygen increases, yet oxygen diffusion rates in water are around 104 times slower than in the air (Armstrong, Reference Armstrong and Woolhouse1979). To further elucidate the submersion tolerance of E. stipitata seeds, we studied concentrations and redox state of the antioxidant tripeptide glutathione (γ-glutamyl-cysteinyl-glycine; GSH), and changes in the GSSG (glutathione disulphide)/GSH redox couple, suggested as stress markers and modulators of programmed cell death (Kranner et al., Reference Kranner, Birtić, Anderson and Pritchard2006). We also assessed seed reserves, underwater germination and seedling morphology; and simulated mechanical damage, such as can happen due to attrition by fast-flowing water or consumption by fishes, by removing the emergent seedlings from the seeds, so as to assess the ability of seedlings to survive and produce a second shoot from the same seed. Furthermore, we were interested to know if the flooding tolerance of E. stipitata can facilitate the long-term preservation of this species underwater as an unconventional means of seed storage.
Materials and methods
Collection, seed processing and flooding experiments
Ripe fruits of E. stipitata were collected at the beginning (November; Lot 1) and towards the end of the rainy season (May; Lot 2) from an experimental plantation at the INPA Campus III in Manaus. Seeds were manually isolated from the fruits and cleaned by friction with moist sand, followed by rinsing under running tap water (Gentil and Ferreira, Reference Gentil and Ferreira2000). Only seeds with a diameter ≥10 mm were used for the studies, which began within 24 h after processing.
For the flooding experiments, the seeds were maintained submerged in the laboratory in running water with average minimum and maximum water temperatures of 21 ± 1.1 and 26 ± 1.8°C, respectively. Submersion under a 6 cm water column is referred to as ‘shallow water’ and that under a 26 cm water column as ‘deep water’. Throughout the test, water flow in the containers was approximately 0.01 litres s–1. In the period of submergence, the mean water pH was 5.7 ± 0.5 and contained 6.3 ± 0.3 mg l–1 dissolved oxygen (measured monthly).
About 1000 seeds of Lot 1 were placed in a white plastic basin (33 × 28 × 8 cm) perforated at a height of 6 cm to keep the water level constant (Fig. 1A). A perforated hose, connected to a tap and closed at the open end with a stopper, drained running water continuously. Once a week, the submersed seeds were gently stirred manually. At intervals for up to 12 months of flooding, 100 seeds each were taken randomly for germination testing. About 2000 seeds of Lot 2 were distributed into five round, transparent glass containers (26 × 13 cm, height × diameter) representing five repetitions (Fig. 1B). A hose connected to a tap was placed at the bottom of each container and drained water continuously throughout the experiment. The seeds were subjected to flooding for 15 days (first sample) and at intervals for up to 12 months. After each time interval, a sample of 45 seeds was taken from each container and used for germination testing, viability staining with tetrazolium and moisture content (MC) determination, as described below.
Schematic drawing of the submersion methods of Eugenia stipitata seeds used with Lot 1 (shallow water; A) and Lot 2 (deep water; B).
Germination tests
The seeds from both lots were submitted to a pre-germination treatment, where the fibrous seed coat was removed from the radicle protrusion region and at the opposite end of the seed (Calvi et al., Reference Calvi, Aud, Ferraz, Pritchard and Kranner2017). The seeds were sown in plastic trays (33 × 28 × 8 cm) at 2 cm depth in moistened vermiculite of medium granulation (Brasil Minérios®) and watered whenever necessary. Four repetitions of 25 seeds were used for Lot 1, and five repetitions of 20 seeds for Lot 2. As a control, the seeds were sown under the conditions described above, immediately after the pre-germination treatment. The trays were placed in a greenhouse with a transparent polyethylene roof. The diurnal temperature varied between a minimum of 25.6 ± 2.5°C and a maximum of 37.9 ± 2.6°C; mean relative humidity was 80%, and light intensity varied between 0.12 × 10 and 3.53 × 10 μmol m–2 s–1 using a digital luxmeter (SKP 200, Skye®). The emergence of the epicotyl on the substrate surface was assessed at least twice a week for 10 months, after which time germination of all viable seeds was completed. The results were expressed as percentage of total germination (TG) and mean germination time (MGT; Santana and Ranal, Reference Santana and Ranal2004). For the latter, only seeds germinated in plastic trays were considered.
Glutathione analysis, viability staining and determination of seed moisture content
After flooding, seeds of Lot 2 were decoated, and seed quarters were used for viability staining, seed MC determination, and analysis of GSH and GSSG in each individual seed (Calvi et al., Reference Calvi, Aud, Ferraz, Pritchard and Kranner2017). Two quarters were immediately frozen in liquid nitrogen and freeze-dried for 3 days to prevent redox reactions prior to analysis. One quarter was used for viability staining with 1% 2,3,5-triphenyl tetrazolium chloride (TZ; Vetec®, Brazil) at 30°C for 24 h in the dark and photographed. One quarter was used for the determination of embryo MC, determined gravimetrically after drying at 103 ± 2°C for 24 h and expressed on a percentage fresh weight basis (Brasil, 2009).
After TZ staining, the seeds of each replicate were sorted into viable seeds (stained) and dead seeds (unstained; Fig. 2). Germinated seeds were considered in a further category. Seeds of each group were freeze-dried and then ground to a fine powder with a ball mill (Fritsch, Pulverisette 7, Germany). The powder was stored at –80°C until it was sent by express mail to the Department of Botany, University of Innsbruck (Austria), where thiols and disulphides were measured.
Staining of viable and dead seeds of Eugenia stipitata seeds after treatment with 1% tetrazolium chloride (TZ) solution for 24 h at 30°C in the dark. Dark purple staining indicates high cell viability and no staining indicates dead tissue.
Thiol/disulphide redox couples were analysed as described by Kranner and Grill (Reference Kranner and Grill1996) and Bailly and Kranner (Reference Bailly, Kranner and Kermode2011). Low molecular weight thiols and disulphides were extracted on ice from the freeze-dried seed powder using 50 ± 2 mg seed material (exact weight recorded) by the addition of 1 ml of 0.1 M HCl, to minimize unwanted reactions. The extracts were vortexed for 20 s, then centrifuged at 4°C for 40 min. In one part of the extract, thiols + disulphides were measured (e.g. GSH + GSSG = total glutathione). For this, disulphides were reduced to thiols using dithiothreitol (DTT), followed by labelling of thiol groups with monobromobimane. Glutathione was separated from cysteine and the dipeptide thiols γ-glutamyl-cysteine and cysteinyl-glycine using an Agilent 1100 HPLC system (Agilent Technologies, Santa Clara, CA, USA) on a ChromBudget 120-5-C18 column (5.0 µm, Bischoff Analysentechnik u. -geräte GmbH, Leonberg, Germany) and fluorescence detection (excitation wavelength = 380 nm, emission wavelength = 480 nm). Disulphides were measured in the second part of the assay: thiol groups were first blocked with N-ethylmaleimide (NEM), then the excess NEM was removed by extracting five times with toluene and the remaining disulphides were reduced to thiols with DTT, labelled with monobromobimane and analysed as above. The half-cell reduction potential of the glutathione/glutathione disulphide redox couple (E GSSG/2GSH) was calculated according to the Nernst equation as described by Kranner et al. (Reference Kranner, Birtić, Anderson and Pritchard2006) and Birtić et al. (Reference Birtić, Colville, Pritchard, Pearce and Kranner2011), using the MC of the embryo (seed without seed coat) to estimate molar concentration of thiols and disulphides at an assumed cellular pH of 7.3.
Analysis of seed reserves
Seeds from Lot 1 were separated into embryo and seed coat (n = 3 biological replicates of 100 seeds) and analysed for starch, protein, fat, crude fibre, ash content and phenolic compounds, expressed on a dry weight basis. Phenolic compounds were solubilized in 50% methanol according to Goldstein and Swain (Reference Goldstein and Swain1963) and quantified according to Folin-Denis by estimating the absorbance at 760 nm compared with a standard curve with tannic acid (Shanderl, Reference Shanderl and Joslyn1970). Starch content was determined in 500 mg of ground tissue after extraction and hydrolysation with HClconc. (Instituto Adolfo Lutz, 1985) with standard dosage according to Somogyi-Nelson, described by Southgate (Reference Southgate1976), using glucose as a standard. Proteins were determined after digestion and distillation according to AOAC (1975), and a factor of 6.25 was used for the conversion of the total nitrogen into protein content. Total fat content was determined for samples of 1.5 g after fat removal by boiling in ether according to the methods used at Instituto Adolfo Lutz (1985). Embryo ash content was obtained by incineration of the fat-free material in a muffle furnace at 550°C, and the fibre content was calculated by the difference between the fat-free matter and ash. The presence of coumarin was detected with ultraviolet (UV) fluorescence. Three drops of alcoholic extract of ground embryo tissue were placed on filter paper alongside three drops of 1% KOH solution in alcohol. The overlapping zone of the two spots on the filter paper was analysed under a UV lamp (Matos, Reference Matos1988).
Capacity of a second sprouting after removal of the first seedling
After the pre-germination treatment, seeds were placed in three conditions: (1) moistened vermiculite in glass trays (25 × 15 × 5 cm) at 15°C; (2) as above, but at 25°C; and (3) flooded with running water under a 26 cm water column at 27°C (Fig. 1B). Two repetitions of 15 seeds were used for each of the three conditions. After 6 months, seedlings (root and shoot) were removed from the seeds by carefully cutting through the connection between seedling and seed (Fig. 3). The seedlings were transplanted into polyethylene pots containing a mixture of equal proportions of soil, vermiculite and coconut fibre, and placed in the aforementioned greenhouse. Seedling survival was assessed after 90 days. Then the seeds were placed again under the three conditions (as above) for an additional 9 months, when the development of roots (≥ 5 mm) and normal seedlings (development of the shoot with cataphylls until the second pair of leaves developed a leaf area ≥ 2 mm2, according to Calvi et al., Reference Calvi, Aud, Ferraz, Pritchard and Kranner2017) were assessed (Fig. 3). The formation of new roots and shoots is termed ‘resprouting’.
Illustration of the method used to test the capacity of a second germination event in Eugenia stipitata seeds after the removal of the first seedling. Germinated seeds had the seedling completely removed by carefully cutting through the connection between the seedling and the seed. Seedlings were transplanted and placed in a nursery, and seeds were placed again to germinate under the initial conditions (vermiculite at 15 or 25°C or flooded at 27°C) and a second germination event assessed.
Statistical analysis
The experimental design for the flooding experiments was randomized, with split-plot in time, where the plots were the flooding depths and the sub-plots were the evaluated periods. The effect of each flooding depth on percentage TG and MGT was evaluated by a variance analysis (ANOVA) and the means compared with Tukey's post-hoc test at P < 0.05. The same statistical analysis was performed for the GSH and GSSG concentrations, viability data and seed MC, which followed the same randomized experimental design. Survival of seedlings detached from the seeds and the ability for resprouting was tested for significance using the non-parametric test of Kruskal-Wallis at P < 0.05.
Results
Seed moisture content (MC) was 59% in fresh seeds (Lot 2) and did not increase significantly during underwater storage (P = 0.256), ranging from 58 to 66% over 12 months. Decoating reduced the seed MC (determined for Lot 1), as the fibrous seed coat had a slightly higher MC (54%) than the embryo (50%).
The seeds of E. stipitata survived underwater for up to 12 months, regardless of water depth (Fig. 4). Before flooding, TG was 97% in Lot 1 and was maintained throughout the flooding period. Germination was lower (63%) in Lot 2 before the flooding experiment, but increased to 96% after 2 months of flooding, similar to that observed in seeds of Lot 1 after 12 months of submersion. Tetrazolium staining confirmed that seeds of Lot 2 had viabilities between 80 and 100% with no statistical difference during the 12 months of underwater storage (P = 0.457; Fig. 4).
Total germination (TG) after 10 months under nursery conditions (bars) for Eugenia stipitata seeds of Lot 1 (A) and Lot 2 (B) after submersion in running water for different periods. Closed triangles (
) represent the percentage of seeds that germinated during submersion, and closed circles (●) denote the viability of seeds from Lot 2 after different periods of submersion as measured by tetrazolium staining. Means followed by different letters differ by Tukey's test at the 0.05 significance level; lower case letters are used for TG of each lot and upper case letters for the viability data. Error bars represent the SD.
Embryo MC, determined for one decoated seed quarter, was maintained during the 12 months of underwater storage (Table 1). Glutathione increased constantly and reached significantly higher values after 12 months. At this time, it was possible to distinguish non-germinated and germinated seeds, the latter showing significantly lower GSH values (Table 1). Glutathione disulphide, expressed as a percentage of total glutathione (GSH + GSSG), increased (P = 0.002) in the first 15 days of underwater storage and recovered slightly after 1 month (not significant). After 3 months, %GSSG significantly increased again, independent of whether the seeds had started to germinate or not. After 12 months, %GSSG decreased to control levels in non-germinated seeds but remained high in germinated seeds. Accordingly, E GSSG/2GSH was significantly more positive (i.e. more oxidizing) in germinated than in non-germinated seeds after 12 months of underwater storage (Table 1).
Effects of flooding of Eugenia stipitata seeds for different periods on moisture content (MC), glutathione (GSH) concentration and glutathione disulphide (GSSG) concentration, % GSSG (expressed as a percentage of total glutathione) and the glutathione half-cell reduction potential (EGSSG/2GSH) analysed in single seeds. Asterisks denote germinated seeds. Data represent mean ± SD. Least significant difference (LSD). Coefficient of variation (CV)
In columns, means marked by different letters differ significantly at P < 0.05 (one-way ANOVA with post-hoc comparison of means using the Tukey's test).
Regardless of water depth, the seeds of both lots started germinating underwater after 2 months. At the end of month 12, all viable seeds (87% TG in Lot 1 and 96% in Lot 2) showed radicle protrusion, and many of them had developed a shoot (Fig. 4). Non-germinated seeds were dead according to the cut test, which revealed blackened tissue.
Although water depth did not affect TG, it altered seedling morphology. Under a water column of 6 cm, the seedlings reached the water surface, expanded their leaves and showed normal growth, as did seedlings grown in the nursery (Fig. 5A). However, seedlings that developed under a water column of 26 cm were distinctly different, with short internodes and no expansion of the leaf blades (Fig. 5B).
Normal seedling of Eugenia stipitata in the nursery (A) and the morphology of a seedling that developed below the water surface (B), with details of the short internodes and eophylls without leaf expansion (inset).
Flooding slightly increased MGT in both seed lots within about 30 days (Fig. 6), coinciding with elevated %GSSG. Thereafter, MGT decreased to control values and even lower in Lot 2. This may indicate that an initially stressful period was overcome, i.e. seeds had acclimated to flooding and germination proceeded underwater. It was only possible to assess MGT in the nursery during the first 5 months of the flooding experiments (Fig. 6), because after this period more than 50% of the seeds had germinated underwater (Fig. 4), which made MGT calculations unfeasible.
Mean germination time (MGT) of seeds of Eugenia stipitata Lot 1 (circles) and Lot 2 (triangles) after submersion in water for different periods. Means followed by different letters differ significantly at P < 0.05 (one-way ANOVA with post-hoc comparison of means using Tukey's test); lower case letters for Lot 1 and upper case letters for Lot 2. Error bars represent the SD.
Starch was the major seed reserve with 76.7% of dry weight (DW), equivalent to 37.6% of fresh weight. Other compounds found were protein (5.9%), fibre (2.0%), lipids (1.9%) and ash (1.4%). Phenolic compounds were present in high concentrations in the embryo (3.48% DW) and the seed coat (1.64% DW). These values may be even higher, because with methanol (50%) only the most common phenols were extracted. No coumarin was detected in the embryo with the qualitative test.
After removal from the seedling, the seeds had a capacity to ‘germinate’ a second time. When ‘re-sown’ in moistened vermiculite and maintained for 9 months at 15°C, or at 25°C, or kept underwater at 27°C, the seeds produced new roots and developed a ‘second’ seedling (Table 2). No significant effects of the three conditions on regeneration of new roots (P = 0.333) and the development of seedlings (P = 0.067) were found. The survival level of the first seedlings transferred to the nursery was high (>85%) and independent of previous environmental conditions (P = 0.800; Table 2).
Eugenia stipitata seedling survival in the nursery 90 days after separation from the seeds (%). Development of new normal seedlings after resowing in vermiculite at 15 and 25°C or submersion in running water after nine months (second germination) is also shown. Data represent mean ± SD
Means followed by different letters in the columns differ by Kruskal-Wallis test at 0.05 significance level.
Discussion
Earlier submergence studies on seeds of this species under standing water indicated decomposition after 40 days (Pinedo et al., Reference Pinedo, Ramírez and Blasco1981) or only rare germination events after 90 days underwater (Mendes and Mendonça, Reference Mendes and Mendonça2012). We show exceptional resilience of the seeds to 12 months submersion in partially oxygenated running water, and with high TG underwater (Fig. 4). Compared with nursery plants, these seedlings are shorter, due to reduced internodes, and do not elongate their leaf blades (Fig. 5), but they continually produce nodes with rudimentary leaves. This could be an ecological strategy of staying small and waiting for better conditions, which would require slow consumption of seed reserves.
Metabolic response to reduced O2 levels is governed by the availability and mobilization of carbohydrates, because anaerobic metabolism is costly in terms of carbohydrate consumption compared with normal aerobic respiration (Parolin, Reference Parolin2009). Storage of carbohydrates in underground organs before the wet season has been associated with flooding tolerance (Crawford, Reference Crawford1992; Scarano et al., Reference Scarano, Cattânio and Crawford1994). We found carbohydrates, mostly starch, to be the main seed storage compound (77%) in E. stipitata, in agreement with earlier findings (Melo et al., Reference Melo, Gonçalvez, Mazzafera and Santos2009; Mendes, Reference Mendes2011). Slow consumption of the seed reserves during germination has been reported for E. stipitata (Melo et al., Reference Melo, Gonçalvez, Mazzafera and Santos2009; Mendes, Reference Mendes2011), and a low level of metabolism is presumed to be sufficient to support the retention of viability over many months of flooding. In rice, the amount of stored carbohydrates in plant organs is positively correlated with the level of submergence tolerance (Jackson and Ram, Reference Jackson and Ram2003). Although non-structural carbohydrates are recognized as primary energy sources for rapid consumption, starch in E. stipitata seeds seems to play an additional role, i.e. sustaining the slow developmental switch of the relatively undifferentiated embryo into a germinating seed underwater. It seems that the seeds are dispersed before maturity (Anjos and Ferraz, Reference Anjos and Ferraz1999; Mendes and Mendonça, Reference Mendes and Mendonça2012); final germination and germination rate increased significantly after the first few months of submergence (Figs 4 and 6), and only after 12 months of underwater storage did the seeds reach the same germination level as seeds in vermiculite in the nursery after 2 months. In this regard, the slow germination (long MGT) of E. stipitata seeds is in contrast to most other recalcitrant seeds, many of which germinate quickly (Daws et al., Reference Daws, Garwood and Pritchard2005).
Seed stress response has been described as a tri-phasic process involving an initial ‘alarm phase’, in which a stress factor is recognized, a ‘resistance phase’ during which seeds can acclimate (or, in the course of evolution, adapt) and an ‘exhaustion phase’ characterized by the loss of vigour and viability (Kranner et al., Reference Kranner, Minibayeva, Beckett and Seal2010). Eugenia stipitata seeds kept submersed showed a 1.3- to 1.5-fold increase in the antioxidant GSH between 3 and 12 months, compared with controls (Table 1), consistent with ‘resistance’ and acclimation to the anoxic conditions. The half-cell reduction potentials of the glutathione/glutathione disulphide redox couple (E GSSG/2GSH) has been suggested to be a useful seed viability marker; in aged seed populations, a shift towards more positive E GSSG/2GSH values (indicative of more oxidizing intracellular conditions) correlated with decreasing seed viability, and values between –180 and –160 mV are typically found in seed populations with 50% viability (Kranner et al., Reference Kranner, Birtić, Anderson and Pritchard2006). E GSSG/2GSH values measured in the present study were more negative, between 206 and 224 mV, even after 12 months of submersion, in agreement with high levels of germination (Fig. 4). Seeds submerged for different time intervals showed significant, albeit subtle, variations in E GSSG/2GSH (Table 1). This variation could indicate differences in the developmental stage: in callus cultures of the orchard grass, Dactylis glomerata, a shift towards more reducing conditions was associated with cell division, whereas more oxidizing conditions accompanied differentiation processes associated with somatic embryo formation (Zagorchev et al., Reference Zagorchev, Seal, Kranner and Odjakova2012). In E. stipitata, the most positive (i.e. most oxidising) E GSSG/2GSH values found after 2 weeks of submergence could have been related to the initial stress sensing. The second oxidative shift found after 12 months in germinated seeds, which showed significantly more positive values of E GSSG/2GSH than non-germinated seeds (Table 1), could mean that flooding stress was more intense for seedlings than for seeds, but could also be indicative of the requirement for more oxidizing conditions upon root and shoot development.
Part of the armoury of plant defence that enables seed persistence in natural environments is the accumulation of chemicals that protect against predator and pathogen attack (Long et al., Reference Long, Gorecki, Renton, Scott, Colville, Goggin, Commander, Westcott, Cherry and Finch-Savage2015), for example phenols (Weidner et al., Reference Weidner, Chrzanowski, Karamać, Król, Badowiec, Mostek and Amarowicz2014). We found phenols in both the fibrous seed coat (1.65% DW) as well as in the embryo (3.48% DW) of E. stipitata. The seeds also contain tannin and terpenoids in the seed coat and in idioblasts in the embryo (Mendes, Reference Mendes2011). These compounds are also potential germination retardants (Mendes, Reference Mendes2011) that could be leached faster during submergence in running water, thereby enabling seed germination over time (Fig. 4).
Predation on large seeds may be partial, and other mechanical damage to submerged seeds could be imposed by physical buffeting in water currents, resulting in damage to the embryo and emerging roots and shoots. To cope with these risks, E. stipitata has evolved to produce seeds that are totipotent, i.e. they can regrow from cut surfaces (Anjos and Ferraz, Reference Anjos and Ferraz1999; Mendes and Mendonça, Reference Mendes and Mendonça2012; Calvi et al., Reference Calvi, Aud, Ferraz, Pritchard and Kranner2017). It is a trait found in other species in the same genus: E. involucrate DC., E. uniflora L. and E. brasiliensis Lam. (Silva et al., Reference Silva, Bilia and Barbedo2005); E. cerasiflora Miq., E. umbelliflora O. Berg and E. pruinosa D. Legrand (Delgado et al., Reference Delgado, Mello and Barbedo2010); and E. pyriformis Cambess. (Silva et al., Reference Silva, Bilia, Maluf and Barbedo2003). Two genera from the Clusiaceae family, Garcinia and Allanblackia, also have cut-tolerant seeds (Ha et al., Reference Ha, Sands, Soepadmo and Jong1988; Malik et al., Reference Malik, Chaudhury and Kalia2005; Joshi et al., Reference Joshi, Kumar, Gowda and Srinivasa2006; Asomaning et al., Reference Asomaning, Olympio and Sacande2011; Ofori et al., Reference Ofori, Asomaning, Peprah, Agyeman, Anjarwalla, Tchoundjeu, Mowo and Jamnadass2015). However, the possibility of using the regenerative ability of the monoembryonic seed to initiate more than one germinative event has not previously been assessed.
The exceptional flooding tolerance and the regenerative ability in E. stipitata seeds offer the possibility for non-conventional seed storage underwater for several months, where even rough handling of seeds and seedlings can be tolerated. It also reinforces the potential importance of this species as a model for physiological and ecological studies.
Acknowledgements
The authors thank Raylton dos Santos Pereira for his assistance with germination assessments and Marianne Magauer for glutathione analysis.
Financial support
This work was supported by the National Council of Technological and Scientific Development of Brazil – CNPq (CT-Amazonia Project number 575889/2008-0) and was part of G.P. Calvi's PhD thesis in the Tropical Forest Science Program (PPG-CFT) of INPA, supported by a CAPES fellowship (28 months) and A.M.G. Anjos's Master's thesis in the Botany Program (PPG-BOT) of INPA by a CNPq fellowship. I.D.K. Ferraz was supported by a CNPq research fellowship. The Royal Botanic Gardens, Kew, received a grant-in-aid from the Department of Environment, Food and Rural Affairs (Defra), UK.
