Introduction

Desiccation tolerance (DT) is defined as the ability of seeds to resume regular metabolic activity following rehydration after almost total water loss in the protoplasm (Oliver et al. 2000; Hoekstra et al. 2001). DT allows seeds to endure adverse environmental conditions and ensures that the plants life cycle is continued (Alpert 2000; Oliver et al. 2000). Not all seeds are able to tolerate desiccation and storage; therefore, seeds are classified into three groups. Seeds that tolerate drying down to 7 % of the water content (fresh weight) as well as long-term storage at sub-zero temperatures are classified as orthodox. Seeds that do not tolerate a reduction in water content and that lose viability when the moisture level falls to approximately 20 % are classified as recalcitrant. Additionally, recalcitrant seeds do not tolerate long-term storage or low temperatures (Roberts 1973). Some seeds, classified as the intermediate seed group, exhibit intermediate behaviors between the orthodox and recalcitrant seed groups, tolerating reductions in water content to levels below those of recalcitrant seeds, but not to the same extent as those in orthodox seeds. Nor can the intermediate seeds be stored at sub-zero temperatures (Ellis et al. 1990).

When mature orthodox seeds are hydrated, DT progressively decreases until it is totally lost (Dasgupta et al. 1982; Buitink et al. 2003; Faria et al. 2005) because the seed resumes its metabolic activities towards radicle protrusion and subsequent seedling formation (Bewley 1997). Furthermore, different hydration conditions may result in different seed responses to drying and may cause divergent results in experiments aimed at understanding the extent of DT in seeds. It also has been reported that orthodox seeds subjected to imbibition followed by drying have a shorter storability (Hong and Ellis 1992).

Senna multijuga (Rich.) H. S. Irwin & Barn. is widely distributed in Brazil and grows in the Atlantic Forest and Brazilian savanna within different vegetation types such as rainforest, semideciduous forest, deciduous forest and gallery forest (Oliveira-Filho 2006). Senna multijuga is an important tree species and is mainly used for landscaping and ecological restoration. The seeds of S. multijuga are physically dormant (Rodrigues-Junior et al. 2014), with the ability to form soil seed-banks (Ferreira et al. 2004). This kind of dormancy prevents water uptake; consequently, the onset of germination is also prevented. There is limited information available on the seed biology of the S. multijuga species; thus, not much is known about DT or the effects of germination on DT or about the seed behavior of S. multijuga during storage. Such data may be useful to develop seed storage protocols for the species, which may be particularly helpful in ex situ conservation. The objectives of this study were to understand (1) the loss of DT in S. multijuga seeds during hydration under different conditions and (2) the storage behavior of both dry and germinating S. multijuga seeds after desiccation.

Materials and methods

Seed collection

Dry pods with mature seeds were collected in September 2011 from S. multijuga trees during the time of natural dispersal (dry season) on the campus of the Federal University of Lavras, in Lavras city, MG, Brazil (21°14′S; 45°0′W) at 900 m of elevation. The vegetation of the Federal University of Lavras campus is characterized as a transition between the Atlantic Forest and Brazilian savanna. After collection, seeds were manually processed, separated by floating in water and dried in a climate-controlled room [20 °C; 60 % relative humidity (RH)] for 2 days. Seeds were then stored in a sealed semi-permeable plastic bag in a cold room (5 °C; 40 % RH) until the beginning of the experiments one week later.

Water content determination and dormancy release

Seeds were randomly sampled from the seedlot for water content determination by weighing four replications of 10 seeds each, placing them in an oven at 103 °C for 17 h and then weighing the seeds again. The data were expressed as percentage of water on a fresh basis (ISTA 2004). To characterize water absorption during germination, 20 replications of one non-dormant seed each (made non-dormant by immersion in hot water, as described below) were kept individually under germination conditions (moistened paper at 25 °C under constant light of approx. 27 µmol m−2 s−1) and weighed several times until the protruded radicle reached 3 mm in length. Before each weighing, the superficial water was removed from the seeds using absorbent paper. After weighing, the seeds were returned to imbibition conditions and weighed regularly thereafter during this treatment. To release physical dormancy, seeds were immersed in hot water (initial temperature of 80 °C that reached room temperature within a few minutes) and kept for 12 h, as described by Rodrigues-Junior et al. (2014), and fresh mass was determined from the start of the hot water submersion.

Loss of DT during hydration

Intact seeds were subjected to dormancy breaking treatment (immersion in water at 80 °C initial temperature for 12 h); thereafter, the non-dormant seeds were placed in moistened paper at 25 °C under constant light (germination conditions) or kept immersed in water (30 mL, replaced daily) at 25 °C to assess the loss of DT. Loss of DT was evaluated from the start of the dormancy breaking treatment. One hundred twenty seeds of each condition were removed—after different time intervals—and subjected to drying in silica gel. Seeds were kept in a Gerbox® plastic container over a net, above a layer of active silica gel which kept the relative humidity of the air to approximately 8 %. The Gerbox® containers were sealed with plastic film. A moisture test was performed to check the seed water content (ISTA 2004) during drying (four replications of five seeds per treatment). After reaching the initial water content (approximately 8 %, after 1 day of drying), seeds were kept for three more days under the same drying conditions as described by Buitink et al. (2003). In preliminary studies, the seeds returned to the initial water content after 1 day in silica gel, and a moisture test was performed to confirm this. After drying the seeds, four replications of 25 seeds per treatment were pre-humidified in a water-saturated atmosphere (approximately 100 % RH) at 25 °C for 24 h. Imbibition in humid air normally is performed to avoid imbibition damage. Following this, seeds were allowed to germinate on moistened paper at 25 °C under constant lighting conditions. Germination percentage, germination speed index (GSI) (Maguire 1962) and percentage of normal seedlings were each assessed for a period of 9 days. Seedlings were considered to be normal when all of their parts were well developed and healthy, which was observed following a maximum of 9 days. The criterion for assessing germination was a protruded primary root of greater than or equal to 1 mm. To assess DT, the criterion was the formation of normal seedlings. The seeds were considered dead if disruption of the seed coat due to cotyledon expansion without radicle protrusion occurred, followed by deterioration of the seed.

Storage after imbibition and drying

Seeds were subjected to dormancy release treatment and then dried in silica gel for 4 days either (1) immediately after treatment or (2) following imbibition for 10 h under germination conditions. The control treatment comprised dry seeds that were untreated by the aforementioned method to break dormancy. Fifteen samples containing 100 seeds for each treatment were placed in sealed plastic bags and stored in the dark at three temperatures: −18, 5 and 20 °C, for up to 240 days (0, 60, 90, 180 and 240 days). After storage, four replications of 25 seeds per treatment were pre-humidified (approximately 100 % RH) and placed at germination conditions, as described above. Dry seeds were subjected to the same breaking dormancy treatment described above and placed at germination conditions. Germination and formation of normal seedlings were both assessed for 9 days. The experiment was a full randomized factorial design with 45 treatment combinations (3 dormancy treatments × 3 storage temperatures × 5 storage durations).

Statistical analysis

The experimental design was completely randomized for all assays. A regression analysis was performed on the data of loss of DT. The model was tested at 5 % significance and evaluated by the coefficient of determination (R2) using the software SigmaPlot® v. 11.0. The storage data were analyzed using a generalized linear model (GLM) followed by a Tukey test at 5 % significance, using the software R for Windows® 2.12.0 (R Development Core Team 2011). The model included the effects of dormancy treatments, storage temperatures and storage durations as well as the interaction between these factors.

Results

Water content determination and imbibition curve

Seeds having an initial water content of 8.4 % showed a triphasic pattern of imbibition. There was a rapid increase in the seed fresh mass that extended for approximately 18 h, characterizing phase I of imbibition. From this period onwards, seed fresh mass was maintained (phase II) and the water absorption remained stable for approximately 36 h. After this period, the radicle began to protrude and a further increase in the seed fresh mass occurred, characterizing phase III of imbibition. Radicle protrusion (visible germination) started after 32 h of imbibition. After 36 h of imbibition, 50 % of the seeds exhibited protruded radicles.

Loss of DT during hydration

Germination (radicle protrusion) of the seeds that were imbibed in paper and dried was high until 32 h of imbibition and exhibited reduction after this imbibition period (data not shown). Regarding the formation of normal seedlings (the criterion used to assess loss of DT), tolerance was maintained until 22 h of imbibition with a resultant normal seedlings amount of 86 %. After the 22 h imbibition period, the resultant normal seedlings amount dropped reaching 62 % after 27 h of imbibition and 13 % after 32 h. Normal seedlings did not form after 36 h of imbibition and desiccation (Fig. 1). The loss of DT was characterized by the high percentage of seedlings with damaged radicles (Fig. 2b). These seedlings showed necrosis in the radicle, which prevented radicle development. After 36 h of imbibition (and radicle emergence beginning after 32 h), seeds completely lost DT. The GSI ranged from 23.03 for dry seeds to 1.6 for seeds undergoing imbibition for 36 h (Fig. 2a).

Fig. 1
figure 1

Percentage of normal seedlings for Senna multijuga seeds in response to desiccation treatment following imbibition under normal germination (filled circle; R2 = 0.92) or continuous soaking (open circle; R2 = 0.94) conditions (mean ± standard deviation)

Full size image
Fig. 2
figure 2

Germination speed index (GSI) (a) and percentage of normal seedling formation (b) for Senna multijuga seeds in response to desiccation treatment following imbibition under normal germination (filled circle; R2 = 0.91, a; R2 = 0.98, b) or continuous soaking (open circle; R2 = 0.90, a) conditions (mean ± standard deviation)

Full size image

Seeds that were soaked in water kept their DT for a longer time. Germination followed the same trend of the formation of normal seedlings with the percentage remaining above 70 % after 72 h of soaking. Soaking beyond the 72 h period (i.e., for 96 h) caused a decrease in normal seedling formation to 48 %, with no seedling survival occurring after 168 h of soaking (Figs. 1, 2b). Cotyledon expansion without radicle protrusion occurred in some cases and the germination speed also decreased as the soaking time advanced (Fig. 2a). No trend was observed between increasing formation of abnormal seedlings and longer soaking times (P > 0.05) (Fig. 2b). The reduction in the percentage of germination and formation of normal seedlings was thought to be due to seed mortality.

Storage after imbibition and drying

Dormancy treatment, storage temperature and storage duration treatments significantly (all P ≤ 0.05) affected seed germination and the resulting proportion of normal seedlings. There were interactions between dormancy treatment and storage duration as well as between storage temperature and storage duration on seed germination. Interactions were also observed between storage temperature and dormancy treatment on normal seedling formation (all P ≤ 0.05). Combining all storage temperatures, the germination of stored dry seeds and of those that were stored after dormancy breaking treatment and drying did not decrease during the 240 days of storage. However, seeds that were stored after a dormancy breaking treatment followed by 10 h of imbibition and drying showed decreased germination at 240 d of storage (Fig. 3A). Combining all dormancy treatments, no reduction in germination of seeds stored for 240 days at 5 and 20 °C was observed; however, seeds stored at −18 °C did exhibit reduced germination (Fig. 3b). The percentage of normal seedlings remained the same throughout the length of the storage of dry seeds; however, germinating seeds stored after drying exhibited an increase in the formation of normal seedlings (Fig. 3c). When all dormancy treatment results were combined, it was found that among the temperatures used for storage, −18 °C produced the lowest percentage of normal seedlings (Fig. 3d). Storage temperatures had no effect on the formation of normal seedlings of the seeds stored in the dry state; however, storage at −18 °C decreased the formation of normal seedlings of the germinating seeds stored after drying. In addition, the storage of dry seeds led to a high percentage of normal seedlings, irrespective of the storage temperature (Fig. 4).

Fig. 3
figure 3

Percentage of seed germination (a, b) and normal seedlings (c, d). Interaction effect between dormancy treatment and storage duration (a, c) and between storage temperature and storage duration (b, d). Means with the same letter are not significantly different (P ≤ 0.05) among (lower case letters) and within each storage duration (upper case letters)

Full size image
Fig. 4
figure 4

Interaction effect between dormancy treatment and storage temperature on proportion of normal seedlings. Means with the same letter are not significantly different (P ≤ 0.05) among (lower case letters) and within each storage temperature (upper case letters)

Full size image

Discussion

S. multijuga seedlings that originated from seeds that were imbibed, dried and rehydrated had damaged radicles, as necrosis and expansion of the cotyledons without elongation of the hypocotyl-radicle axis, suggesting that there are different DT responses among the embryo parts, with the radicle being the most sensitive part (Reisdorph and Koster 1999). In the Tabebuia impetiginosa seedlings, necrosis and subsequent death of radicles after treatment with polyethylene glycol and desiccation did not prevent seedling growth because adventitious roots that emerged from the hypocotyls allowed seedling survival (Vieira et al. 2010). However, adventitious roots were not observed in S. multijuga seedlings throughout the evaluation period. In addition to the visible damage (described above), there was a reduction in germination speed following longer imbibition, and this may be related to damage at the cellular and/or molecular level that must be repaired for germination to begin.

The hydration conditions assessed herein resulted in a changed seed response to desiccation, and this may be related to differences in metabolism. Conditions limiting metabolic activity in seeds may reduce the damage caused by desiccation (Leprince et al. 1995). The reduction of oxygen availability, which may have occurred during soaking of the seeds, reduces metabolic activity of plant tissues (Geigenberger 2003). The latter may then prolong the ability of cells to tolerate desiccation. Seed tolerance to hypoxic conditions varies among species, ranging from a few hours to days (Perata and Alpi 1993). A decrease in metabolism may have prevented cellular damage to S. multijuga seeds, allowing them to withstand drying after longer soaking times in comparison to seeds imbibed under germination conditions. During the soaking process, seeds do not reach phase III of imbibition; instead, the seeds remain at phase II. Thus, the response of seeds to hydration conditions may differ, and this should be taken into consideration in studies performed to assess the DT.

Orthodox seeds with low water content are able to withstand storage over long periods (Roberts 1973). These seeds stored at −18 °C with a water content below 10 % are considered to be optimal for germplasm conservation (Genebank Standards 1994; Pritchard and Dickie 2003). Storing dry seeds (<10 % water content) of S. multijuga at −18 °C for up to 240 days had no harmful effects but caused damage to germinating seeds after drying. Interestingly, these seeds lost the characteristics that enabled them to tolerate extremely low temperatures, which decreased their storability at sub-zero temperatures. This finding supports previous studies showing a change in storage behavior of orthodox seeds after imbibition and drying by Hong and Ellis (1992). A similar behavior during drying and storage was found in Coffea arabica, Carica papaya and Elaeis guineensis seeds, which are classified as intermediate (Ellis et al. 1990, 1991a, b). Thus, it important that the seeds of S. multijuga should be stored at the correct water content level, but storability (at same seed water content level) will also be affected if the seeds have been previously hydrated.

It is concluded that the conditions in which seeds remain during hydration can affect the physiological response to DT as it relates to the seed metabolic state. S. multijuga seeds can be dried down to 10 % water content and stored at −18 °C for up to 240 days; therefore, seeds of this species can be considered to have orthodox seed characteristics. However, germinating seeds are harmed by sub-zero temperatures after they have been dried.