The results obtained in the present study allowed the schematization of which germination strategies are used by seeds produced in different climatic conditions (Fig. 3). As observed in the present study, the greater the number of HD cycles, the greater the germinability of P. catingicola subsp. salvadorensis produced in both rainy and dry seasons, regardless of the hydration time to which seeds were exposed between each cycle. Discontinuous hydration benefits in this cactus seeds had already been reported by Lima and Meiado (2017). In addition, the increase in water deficit severity reduced synchrony and germinability, but HD cycles increased the percentage of seeds germinated at more negative osmotic potentials, indicating an acquisition of tolerance provided by seeds discontinuous hydration, which corroborates the third hypothesis of this study. The benefit of HD cycles under water deficit conditions was observed in seeds produced in both seasons. However, seeds produced in the dry season showed greater acquisition of tolerance to the most severe conditions of water unavailability after the cycles (Fig. 3). These results corroborate our first hypothesis that seeds from the same population, produced, and dispersed in different seasons, express seed hydration memory differently during their germination process. On the other hand, the hypothesis that the hydration time between HD cycles influences seeds germinal response produced and dispersed in different seasons was not supported.
Patterns of water availability over time and space determine reproductive strategies in Seasonally Dry Tropical Forest species (Bullock 2009). The synchronization of leaf senescence, flowering, and plant growth with water availability in the soil throughout the year is not the only strategy for perpetuating species in these ecosystems that experience long periods of drought (Helbrook et al. 2009), as demonstrated by the present study. Unlike seeds of temperate forest species, which find it more difficult to germinate and establish themselves in face of water limitations (Retana et al. 1999), recent studies with forest species have shown that native species from Seasonally Dry Tropical Forests use this stressful condition in favor of optimizing the germination process (Lima and Meiado 2018b; Nascimento et al. 2021). Although previous works have demonstrated the influence of hydration time between HD cycles on seed hydration memory expression during germination of native seeds from semi-arid ecosystems (Lima and Meiado 2017; Lima et al. 2018), the present study demonstrated the opposite. Such response is related to optimization of germination process since each species requires specific combinations of abiotic factors to germinate. According to Schwienbacher et al. (2012), the difference in germination behavior between species established in the same community directly contributes to differentiation in germination niche, enabling coexistence and better use of limiting resources.
Whereas P. catingicola subsp. salvadorensis produces and disperses seeds throughout the year (Barbosa 2015) and inhabits an ecosystem that presents irregular precipitation with rainfall concentrated in a few months (Oliveira et al. 2014; Zappi and Taylor 2020), it is expected that different germination strategies have been selected throughout the evolutionary process so that seeds fulfill their ecological role of maintaining the species population dynamics. After being submitted to HD cycles, which simulate a condition of discontinuous hydration during germination process (Lima and Meiado 2017), this study results showed that seeds produced during the dry season and dispersed at the beginning of the rainy season present germination patterns different from seeds produced in the rainy season and dispersed at the beginning of the dry season, when submitted to water deficit conditions (Fig. 3). As suggested by Contreras-Queiroz et al. (2016), this results also show that seed hydration memory expression depends on the climate or microenvironment where the species is established.
Even in the rainy season, P. catingicola subsp. salvadorensis seeds are submitted to HD cycles, since they are dispersed in the soil surface layer, where water resource is spatially and temporally limited (Meiado et al. 2012) and seeds have an average germination time of 6.5 ± 0.4 before undergo HD cycles, when they reduce the time to 4.0 ± 0.2 (Lima and Meiado 2017). With this information available, it becomes possible to understand the germination strategy adopted by seeds that are produced in the dry season and dispersed at the beginning of the rainy season. As observed in the results, these seeds have a greater tolerance to water deficit after going through HD cycles. Therefore, after their dispersion, they are submitted to cycles during the rainy season, when they acquire tolerance to germinate and establish themselves in the following season, where more severe water deficit indices occur.
Stressful environmental conditions produce extensive changes in the regulation of gene expression, gene activation or suppression, and biochemical modulation of signal transduction pathways that can promote plant survival during this period (Amara et al. 2020). Saux et al. (2020) demonstrated that tolerance to water deficit in the germination phase is conferred by parental genetic conditions during seed production and that it is conserved at different stages of the plant life cycle. In addition, mechanisms of oxidative stress control and abscisic acid hormone (ABA) biosynthesis capacity are responsible for water deficit tolerance (Gonçalves et al. 2020). In turn, HD cycles, in addition to significantly increasing the germination percentage of several species (Lima et al. 2018; Becerra‑Vázquez et al. 2020; Ferreira et al. 2021; Nascimento et al. 2021), act in the upregulation of genes involved in DNA damage repair and antioxidant defense (Forti et al. 2020). This explains the condition of seeds produced during the dry season, which are more tolerant to water deficit. According to Matzrafi et al. (2020), parental plants under water deficit conditions can produce more drought tolerant seeds. In addition, Lima and Meiado (2018a) have already shown that, after passing through HD cycles, seeds of P. catingicola subsp. salvadorensis originated seedlings with greater root and stem lengths, which allows greater exploitation of available resources during initial development and, consequently, greater chances of recruitment. Thus, we can affirm that seeds produced in the dry season and dispersed in the rainy season, which are more tolerant to water deficit after HD cycles, play a fundamental ecological role in maintaining seedling recruitment and population maintenance during the following dry season, corroborating the hypothesis that seeds produced during the dry season show greater acquisition of tolerance to water deficit after passing through HD cycles.
During the dry season, seeds that were produced in the rainy season are also submitted to HD cycles, since, according to Oliveira et al. (2014), precipitation during this period is not zero in the study area. Certainly, the hydration time between cycles during the rainy and dry seasons are different, however, as seen in the present study, there is no influence due to this factor on the species germination. Yi et al. (2019) highlighted that the variation in the germination behavior of seeds submitted to different environmental conditions can increase the long-term reproductive success rate within the population, temporarily spreading risks and maximizing its fitness. In this way, the cycles to which seeds produced in the rainy season are submitted during the dry season also promote an increase in the germination percentage, increasing the chances of seedling establishment in the following rainy season, since the seeds produced in this period are less tolerant to more severe water deficit conditions, as observed in the results.
According to Souza et al. (2020), with the gradual reduction of water availability in the soil, there is an increase in the concentration of abscisic acid (ABA) and reduction of gibberellin (GA) in the seeds, resulting in a reduction in the synthesis of key enzymes that encode mRNA for endosperm degradation and/or embryonic cells expansion. Thus, these seeds do not germinate in the soil, even without some type of dormancy, but maintain the pool of mRNA-encoding enzymes produced in the previous rainy season. In the following rainy season, the concentration of GA increases and the enzymes are reactivated, promoting germination. This process explains the ability of P. catingicola subsp. salvadorensis seeds dispersed during the dry season to increase the percentage of germination in the following rainy season, as they are submitted to a greater number of HD cycles. Since they are dispersed during the dry season, reduced soil water availability can trigger an increase in ABA concentration and a reduction in GA concentration in seeds. As demonstrated by Dubrovsky (1996), seed hydration memory is the ability to maintain biochemical changes induced by imbibition before the next dehydration cycle. Therefore, in the next event of water availability where the seeds succeed to germinate, seed hydration memory expression occurs, as in the case of the studied species, where HD cycles promoted a higher percentage of germination and acquisition of tolerance to water deficit.
The differentiated germinal response between seeds produced in the dry and rainy seasons may also be related to epigenetic changes induced by abiotic stress which, according to Chinnusamy and Zhu (2009), have adaptive advantages for conservation of species in situ. It is worth noting that Seasonally Dry Tropical Forests are among the most threatened ecosystems and can become increasingly arid areas due to increasing droughts and global temperatures (Baccini et al. 2017; Rivas et al. 2021). In future climate scenarios, with reduced rainfall, strategies that allow the population dynamics of species in these ecosystems may become inefficient to overcome such drought conditions and, consequently, limit germination, as pointed out by Dantas et al. (2018).