ROS Accumulation as a Hallmark of Dehydration Stress in Primed and Overprimed Medicago truncatula Seeds

: Seed priming protocols implement incomplete imbibition phases, as well as physical, chemical or biological treatments, to activate pre-germinative metabolism and stress response, thus improving germination performances, seedling establishment and stress tolerance according to agricultural productivity requirements. The dehydration phase following priming treatments represents a critical variable, since an excessively prolonged imbibition (overpriming) impairs desiccation tolerance, compromising seed viability and seedling establishment. Priming protocols generally optimize imbibition-dehydration timing empirically to avoid overpriming. Hence, a better understanding of the dynamics underlying the loss of desiccation tolerance represents a promising route to test and develop efﬁcient and cost-effective priming techniques. In the present work, priming and overpriming conditions were deﬁned to explore the role of desiccation tolerance in seed priming efﬁciency in the model legume Medicago truncatula . The positive effects of hydropriming and kinetin-mediated hormopriming on germination parameters were screened in combination with conditions of short/prolonged priming and mild/severe overpriming. Biometric analyses highlighted contrasting responses in terms of germination performances and seedling development, while ROS (reactive oxygen species) levels measured during dehydration positively correlate with the loss of desiccation tolerance in early seedlings, suggesting possible applications to monitor priming progression and predict overpriming occurrence. combined effects of short and prolonged hydropriming and kinetin-mediated hormopriming were tested on a commercial M. truncatula seed lot. Germination tests conﬁrmed the effectiveness of short and prolonged


Introduction
Seed priming techniques are routinely employed to improve germination performances and stress tolerance of commercial seed lots [1,2]. Seed priming stimulates the activation of pre-germinative metabolism through an incomplete imbibition that prepares the seeds for an accelerated and coordinated germination and a more efficient stress response. Primed seeds are subsequently dehydrated (dry-back) and stored in view of sowing or commercialization [3,4].
Relevantly for priming protocols, orthodox seeds can withstand significant desiccation events through their endowment in osmo-protective proteins and sugars stabilizing cellular components, while antioxidant compounds protect lipids, proteins and nucleic during dehydration. Dehydration was carried out for 4 h (2 h-imbibed seeds) or 6 h (24 himbibed seeds, overprimed seeds).
As a proof-of-concept of the main conclusions of this study, Medicago sativa seeds (commercial genotype) underwent seed priming treatments as previously described for M. truncatula seeds and outlined in Figure 1. Overview of the experimental system aimed to compare the effects of short (2 h) and prolonged (24 h) hydropriming and kinetin-mediated hormopriming on M. truncatula and M. sativa seeds, including mild (radicle protrusion 1 mm) and severe (radicle protrusion 2 mm) overpriming conditions. Imbibition steps carried out in water are indicated in blue, imbibition steps carried out in presence of 0.5 mM kinetin are indicated in green, and dry-back steps are indicated in yellow. UP, water-imbibed unprimed seeds; KUP, kinetin-imbibed unprimed seeds; HP2h, 2 h-hydroprimed seeds; KP2h, 2 h-hormoprimed seeds; HP24h, 24 h-hydroprimed seeds; KP24h, 24 h-hor- Figure 1. Overview of the experimental system aimed to compare the effects of short (2 h) and prolonged (24 h) hydropriming and kinetin-mediated hormopriming on M. truncatula and M. sativa seeds, including mild (radicle protrusion 1 mm) and severe (radicle protrusion 2 mm) overpriming conditions. Imbibition steps carried out in water are indicated in blue, imbibition steps carried out in presence of 0.5 mM kinetin are indicated in green, and dry-back steps are indicated in yellow. UP, water-imbibed unprimed seeds; KUP, kinetin-imbibed unprimed seeds; HP2h, 2 h-hydroprimed seeds; KP2h, 2 h-hormoprimed seeds; HP24h, 24 h-hydroprimed seeds; KP24h, 24 h-hormoprimed seeds; HOP1mm, hydroprimed seeds with 1 mm radicle protrusion; KOP1mm, hormoprimed seeds with 1 mm radicle protrusion; HOP2mm, hydroprimed seeds with 2 mm radicle protrusion; KOP2mm, hormoprimed seeds with 2 mm radicle protrusion; DB4h, 4 h-dry-back; DB6h, 6 h-dry-back. As a proof-of-concept of the main conclusions of this study, Medicago sativa seeds (commercial genotype) underwent seed priming treatments as previously described for M. truncatula seeds and outlined in Figure 1.

Germination Tests and Biometrical Analyses
For germination tests, M. truncatula unprimed seeds (UP, KUP), primed seeds (HP2h, KP2h, HP24h, KP24h) and overprimed seeds (HOP1mm, KOP1mm, HOP2mm, KOP2mm) were imbibed in parallel. Germination was assessed hourly for 48 h and the following germination parameters were calculated: G (germinability), MGT (mean germination time), MGR (mean germination rate), CVG (coefficient of velocity of germination), U (uncertainty index), Z (synchronization index), T 50 (time required to reach 50% of maximum germination) [17]. For each treatment, five independent replicates with 20 seeds per replicate were analyzed. Biometric data of seedling development were collected four days after the beginning of reimbibition (primed seeds) or imbibition (unprimed seeds). Radicle length was measured on millimetric paper (10 seedlings per condition), and fresh and dry weight data were retrieved using 5 replicates of 5 seedlings each. For dry weight measurements, seedlings were dried 24 h at 80 • C.

Relative Water Content (RWC) Measurement
RWC was measured for M. truncatula dry seeds and for primed and overprimed seeds along dry-back (0, 2, 4 and 6 h) and expressed as percentage over the seed fresh weight according to the following formula: RWC [%] = [(Fw − Dw)/Fw] × 100, where fresh weight (Fw) was measured at the indicated timepoints after removing the excess of superficial water from the seeds, when present, and dry weight (Dw) was measured after 24 h dehydration at 80 • C [34]. For these measurements, 5 replicates of 20 seeds each were used to calculate RWC.

ROS Detection by 2 ,7 -Dichlorofluorescin Diacetate (DCF-DA) Assay
ROS (reactive oxygen species) levels were quantified in dry seeds and in primed/ overprimed seeds before, during and at the endpoint of dehydration. ROS levels were assessed before (at 0 h), during (at 2 h) and at the endpoint of dry-back (4 h or 6 h). The assay was carried out using the fluorogenic dye 2 ,7 -dichlorofluorescin diacetate (DCF-DA; Sigma-Aldrich, Milan, Italy). The dye is converted to a non-fluorescent molecule following deacetylation mediated by cellular esterases, and it is subsequently oxidized by ROS into the fluorescent compound 2 ,7 -dichlorofluorescein (DCF), that can be detected by fluorescence spectroscopy with maximum excitation and emission spectra of 495 nm and 529 nm, respectively. The assay was carried out as described by Macovei et al. [35], with the following modifications. Seeds were collected at the indicated timepoints and dried on filter paper. Samples (5 replicates per timepoint, 5 seeds per replicate) were incubated for 15 min with 100 µL of 10 µM DCF-DA and subsequently fluorescence at 517 nm was determined using a Rotor-Gene 6000 PCR apparatus (Corbett Robotics, Brisbane, Australia), setting the program for one cycle of 30 s at 25 • C. A sample containing only DCF-DA was used as a control to subtract the baseline fluorescence. Relative fluorescence was expressed as relative fluorescence units (RFU).

H 2 O 2 Detection by 3,3 -Diaminobenzidine (DAB) Staining
The distribution of H 2 O 2 within M. truncatula dehydrated embryos was assessed through DAB staining performed at the endpoint of dry-back according to Kiran et al. [36]. Seeds were incubated for 30 min at room temperature in 1 mL of 1 mg/mL DAB (Sigma-Aldrich, Milan Italy) prepared in 10 mM disodium hydrogen phosphate (pH 7.4). For positive control treatments 100 µL of 1 M H 2 O 2 was also added. For negative controls 100 µL of 100 µM ascorbic acid was also added. The seed coat was manually removed and images of 5 seeds per treatments were acquired via stereomicroscope (Olympus SZX9; Olympus ® , Tokyo, Japan) and camera (Camedia C-5060 Wide Zoom; Olympus ® , Tokyo, Japan), supported by the relative software (Olympus DP-Soft 5.0).

Hydropriming and Hormopriming Improve Germination Performances in M. truncatula
The optimal dehydration timing to be applied after short (2 h) priming protocols and prolonged (24 h) priming and overpriming protocols in presence/absence of kinetin was determined by assessing the decrease in RWC during dry-back in order to reach the RWC measured for dry seeds. As a result, 4 h dry-back was sufficient for 2 h-primed seeds, whereas 6 h dry-back was necessary for 24 h-primed and overprimed seeds. No significant differences were observed comparing hydroprimed and hormoprimed seeds ( Figure 2). and Tuckey-Kramer test, using the software developed by Assaad et al. [37] in order to compare priming conditions (UP, P2h, P24h, OP1mm, OP2mm) and priming treatment groups (hydropriming, hormopriming). The Pearson's correlation coefficient and the relative p-values were calculated using MetaboAnalyst 5.0 (Xia Lab, Ste. Anne de Bellevue, Quebec, Canada) [38] to assess the correlations between radicle length class (0 mm, 1 mm, 2 mm), ROS levels at 0 h, 2 h and 4/6 h of dry-back and seedling phenotype ('normal', 'aberrant', 'dead'). Statistically significant differences are indicated in the figures by the occurrence of different letters or by asterisks ('*' p < 0.05, '**' p < 0.01, '***' p < 0.001).

Hydropriming and Hormopriming Improve Germination Performances in M. truncatula
The optimal dehydration timing to be applied after short (2 h) priming protocols and prolonged (24 h) priming and overpriming protocols in presence/absence of kinetin was determined by assessing the decrease in RWC during dry-back in order to reach the RWC measured for dry seeds. As a result, 4 h dry-back was sufficient for 2 h-primed seeds, whereas 6 h dry-back was necessary for 24 h-primed and overprimed seeds. No significant differences were observed comparing hydroprimed and hormoprimed seeds ( Figure 2).
The effects of short and prolonged hydropriming and hormopriming on M. truncatula germination performances were assessed through multiple germination parameters to highlight the advantages and disadvantages of the tested priming protocols. Having already undergone germination during priming, overprimed seeds were excluded from post-priming germination tests. Germination speed (expressed as T50, MGT, MGR, CVG) increased in response to all priming conditions (HP2h, KP2h, HP24h, KP24h), as well as to exposure to kinetin without priming (KUP) compared to unprimed water-imbibed seeds (UP) ( Table 1). Prolonged hormopriming (KP24h) further increased germination speed and decreased germination uncertainty index (U) compared to prolonged hydropriming (HP24h), while prolonged hydropriming and hormopriming (HP24h and KP24h, respectively) impaired germination percentage (G) measured after 48 h imbibition (Table  1).
The effects of short and prolonged hydropriming and hormopriming on M. truncatula germination performances were assessed through multiple germination parameters to highlight the advantages and disadvantages of the tested priming protocols. Having already undergone germination during priming, overprimed seeds were excluded from post-priming germination tests. Germination speed (expressed as T 50 , MGT, MGR, CVG) increased in response to all priming conditions (HP2h, KP2h, HP24h, KP24h), as well as to exposure to kinetin without priming (KUP) compared to unprimed water-imbibed seeds (UP) ( Table 1). Prolonged hormopriming (KP24h) further increased germination speed and decreased germination uncertainty index (U) compared to prolonged hydropriming (HP24h), while prolonged hydropriming and hormopriming (HP24h and KP24h, respectively) impaired germination percentage (G) measured after 48 h imbibition (Table 1). Table 1. Germination parameters calculated on unprimed M. truncatula seeds in presence/absence of 0.5 mM kinetin, 2 h-hydropriming/hormopriming and 24 h-hydropriming/hormopriming, expressed as mean of 5 replicates (5 dishes each, 20 seeds/dish) ± standard deviation. G, germinability; T 50 , time required to reach 50% of G; PV, peak value; MGT, mean germination time; MGR, mean germination rate; CVG, coefficient of velocity of germination; U, uncertainty index; Z, synchronization index; UP, unprimed seeds; P2h, 2 h-primed seeds; P24h, 24 h-primed seeds; H, water imbibition and hydropriming protocols; K, kinetin imbibition and hormopriming protocols. Values that are followed by common letters are not significantly different (p < 0.05) as analyzed with two-way ANOVA and Tukey test. p-values are also provided referring to the difference observed between treatment comparison groups (H vs. K), priming comparison groups (UP vs. P2h vs. P24h) and their interaction.

Radicle Emergence Impairs Seed Desiccation Tolerance and Post-Priming Seedling Development
The effects of the different priming conditions on seedling development, in terms of seedling morphology, radicle growth, fresh and dry biomass, were assessed 4 days after reimbibition. The 4-day-old seedlings displayed three main contrastive phenotypes: (i) 'normal', (ii) 'aberrant', (iii) 'dead'. 'Normal' seedlings displayed the same morphology as the seedlings developing from unprimed controls (UP or KUP), 'aberrant' seedlings were characterized by a substantially impaired radicle growth and by green cotyledons, whereas 'dead' seedlings did not display radicle growth after reimbibition nor cotyledon greening (Figure 3b). The occurrence of the three phenotypes varied according to the priming condition. The seedlings developing from unprimed seeds (UP, KUP), short priming (HP2h, KP2h) and prolonged priming without radicle protrusion (HP24h, KP24h) displayed 100% 'normal' morphology, whereas the percentage of 'normal' seedlings was reduced to 62-68% in response to mild overpriming (HOP1mm, KOP1mm) and to 5-8% in response to severe overpriming (HOP2mm, KOP2mm) (Figure 3a). Agronomy 2022, 12, x FOR PEER REVIEW 7 in 'normal' 4-day-old seedlings ( Figure 4a). Fresh and dry biomass of 'normal' seedli did not appear to be influenced by exposure to kinetin or priming condition, and fr biomass appeared to be consistently reduced (20-30% lower) only in 'aberrant' seedli compared to 'normal' seedlings ( Figure 4b).  Hormopriming treatments (KP2h, KP24h, KOP1mm, KOP2mm), as well as prolonged exposure to kinetin (KUP), consistently led to a 26-51% reduction in radicle length in 'normal' 4-day-old seedlings ( Figure 4a). Fresh and dry biomass of 'normal' seedlings did not appear to be influenced by exposure to kinetin or priming condition, and fresh biomass appeared to be consistently reduced (20-30% lower) only in 'aberrant' seedlings, compared to 'normal' seedlings ( Figure 4b).

Overpriming Results into ROS Accumulation after Dry-Back
Given the critical importance of dehydration phases in seed priming protocols and the risks connected to the loss of desiccation tolerance on seedling establishment, this work focused on the identification of stress signatures during dry-back. The fluorogenic dye DCF-DA was used to quantify ROS levels in dry unprimed seeds and in primed/overprimed seeds before dry-back (DB0h, i.e., at the end of the imbibition), after 2 h dry-back (DB2h) and at the end of dry-back. Based on the dehydration curves (Figure 2), the endpoint of dry-back was established at 4 h (DB4h) for 2 h-primed seeds (HP2h, KP2h), and at 6h for prolonged priming/overpriming conditions (HP24h, HOP1mm, HOP2mm, KP24h, KOP1mm, KOP2mm), when the RWC of DS was restored. The results of DCF-DA assay are shown in Figure 5a. For all priming/overpriming conditions, ROS emission before dry-back was not significantly different from the ROS emission of dry seeds and did not display a significant increase at 2 h dry-back. Short priming treatments (HP2h, KP2h) did not result in significant changes in ROS levels during and after desiccation, despite a general increasing trend. The highest ROS emission was detected at 6 h dehydration of severely

Overpriming Results into ROS Accumulation after Dry-Back
Given the critical importance of dehydration phases in seed priming protocols a the risks connected to the loss of desiccation tolerance on seedling establishment, t work focused on the identification of stress signatures during dry-back. The fluoroge dye DCF-DA was used to quantify ROS levels in dry unprimed seeds and in primed/ov primed seeds before dry-back (DB0h, i.e., at the end of the imbibition), after 2 h dry-ba (DB2h) and at the end of dry-back. Based on the dehydration curves (Figure 2), the en point of dry-back was established at 4 h (DB4h) for 2 h-primed seeds (HP2h, KP2h), a at 6h for prolonged priming/overpriming conditions (HP24h, HOP1mm, HOP2m KP24h, KOP1mm, KOP2mm), when the RWC of DS was restored. The results of DCF-D assay are shown in Figure 5a. For all priming/overpriming conditions, ROS emission fore dry-back was not significantly different from the ROS emission of dry seeds and d not display a significant increase at 2 h dry-back. Short priming treatments (HP2h, KP did not result in significant changes in ROS levels during and after desiccation, despit general increasing trend. The highest ROS emission was detected at 6 h dehydration severely overprimed seeds (HOP2mm and KOP2mm). Considering the effects of kinet Since the DCF-DA assay evidenced more contrastive patterns of ROS accumulation at the endpoint of dry-back, this timepoint was selected for DAB staining to qualitatively assess the distribution of H 2 O 2 within M. truncatula dehydrated embryos. Although no significant differences were observed in response to hormopriming compared to hydropriming, the DAB-positive coloration was observed only in the radicle tips of all mildly (OP1mm) and severely (OP2mm) overprimed seeds (Figure 5b), consistent with the higher ROS levels detected through DCF-DA assay at the endpoint of dry-back in overprimed seeds (Figure 5a).

The Loss of Desiccation Tolerance Correlates with Radicle Protrusion Length and with ROS Levels during Dry-Back
The analysis of ROS levels during dry-back (Figure 5a) provided useful hints to monitor oxidative stress in primed and overprimed seeds and suggested possible correlations with the occurrence of 'aberrant' seedling phenotypes in overprimed seeds (Figure 3a). To statistically assess this hypothesis, a correlation analysis was performed between the length of radicle protrusion before dry-back (0 mm, 1 mm or 2 mm), seedling phenotype (expressed as percentage of 'normal', 'aberrant' and 'dead' seedlings), and ROS levels Since the DCF-DA assay evidenced more contrastive patterns of ROS accumulation at the endpoint of dry-back, this timepoint was selected for DAB staining to qualitatively assess the distribution of H2O2 within M. truncatula dehydrated embryos. Although no significant differences were observed in response to hormopriming compared to hydropriming, the DAB-positive coloration was observed only in the radicle tips of all mildly (OP1mm) and severely (OP2mm) overprimed seeds (Figure 5b), consistent with the higher ROS levels detected through DCF-DA assay at the endpoint of dry-back in overprimed seeds (Figure 5a).

The Loss of Desiccation Tolerance Correlates with Radicle Protrusion Length and with ROS Levels during Dry-Back
The analysis of ROS levels during dry-back (Figure 5a) provided useful hints to monitor oxidative stress in primed and overprimed seeds and suggested possible correlations with the occurrence of 'aberrant' seedling phenotypes in overprimed seeds (Figure 3a). To statistically assess this hypothesis, a correlation analysis was performed between the length of radicle protrusion before dry-back (0 mm, 1 mm or 2 mm), seedling phenotype 'aberrant' and 'dead' phenotypes (0.796 and 0.629, respectively) and a strong negative correlation (−0.804) with the percentage of 'normal' seedlings. Weaker significant correlations were also highlighted. For example, ROS levels at 2 h dry-back (DB2h) correlated with radicle protrusion length (0.503), ROS levels at 4/6 h (0.469), percentage of 'aberrant' seedlings (0.537), percentage of 'dead' seedlings (0.330) and percentage of 'normal' seedlings (−0.527). Figure 6. Correlation analysis carried out between the radicle protrusion length before dry-back, the ROS emission during dry-back and seedling morphology. DB0h, ROS levels before dry-back; DB2h, ROS levels after 2 h dry-back; DB4/6h, ROS levels at the endpoint of dry-back. The Pearson's correlation coefficients are indicated. The significance of the Pearson's correlation is shown by asterisks ('*' p < 0.05, '**' p < 0.01, '***' p < 0.001). NA, not applicable.

The Correlation between ROS Accumulation and Loss of Desiccation Tolerance Is Reproducible in Medicago sativa
The strong significant correlations between ROS levels at the endpoint of dry-back (DB4/6h) and the occurrence of 'aberrant' seedling phenotype observed in M. truncatula seeds ( Figure 6) indicated a possible route to predict the loss of desiccation tolerance in early seedlings. In order to test this hypothesis on a related species, a proof-of-concept was carried out on a commercial genotype of Medicago sativa, by measuring ROS levels through DCF-DA assay at the endpoint of dry-back and the occurrence of 'normal', 'aberrant' and 'dead' seedling morphology after 4 days of germination. The endpoint of dryback (DB4/6h) was selected because it was the timepoint that consistently displayed the highest level of ROS accumulation for all the priming conditions tested on M. truncatula seeds. The same priming and overpriming conditions and the same morphological criteria previously described for M. truncatula were applied to M. sativa. The results are shown in Figure 7a. The pattern of ROS accumulation in M. sativa seeds at the endpoint of dry-back was comparable to the results obtained in M. truncatula. Specifically, the highest ROS emission was detected in severely overprimed seeds (HOP2mm and KOP2mm) in association with the highest percentage of 'aberrant' seedlings. 'Dead' seedling morphology was present in response to all mild and severe overpriming conditions (HOP1mm, KOP1mm,  HOP2mm, KOP2mm). Considering the effects of kinetin, ROS emission appeared to be higher in response to kinetin-mediated severe overpriming (KOP2mm) compared to the Figure 6. Correlation analysis carried out between the radicle protrusion length before dry-back, the ROS emission during dry-back and seedling morphology. DB0h, ROS levels before dry-back; DB2h, ROS levels after 2 h dry-back; DB4/6h, ROS levels at the endpoint of dry-back. The Pearson's correlation coefficients are indicated. The significance of the Pearson's correlation is shown by asterisks ('*' p < 0.05, '**' p < 0.01, '***' p < 0.001). NA, not applicable.

The Correlation between ROS Accumulation and Loss of Desiccation Tolerance Is Reproducible in Medicago sativa
The strong significant correlations between ROS levels at the endpoint of dry-back (DB4/6h) and the occurrence of 'aberrant' seedling phenotype observed in M. truncatula seeds ( Figure 6) indicated a possible route to predict the loss of desiccation tolerance in early seedlings. In order to test this hypothesis on a related species, a proof-of-concept was carried out on a commercial genotype of Medicago sativa, by measuring ROS levels through DCF-DA assay at the endpoint of dry-back and the occurrence of 'normal', 'aberrant' and 'dead' seedling morphology after 4 days of germination. The endpoint of dry-back (DB4/6h) was selected because it was the timepoint that consistently displayed the highest level of ROS accumulation for all the priming conditions tested on M. truncatula seeds. The same priming and overpriming conditions and the same morphological criteria previously described for M. truncatula were applied to M. sativa. The results are shown in Figure 7a. The pattern of ROS accumulation in M. sativa seeds at the endpoint of dry-back was comparable to the results obtained in M. truncatula. Specifically, the highest ROS emission was detected in severely overprimed seeds (HOP2mm and KOP2mm) in association with the highest percentage of 'aberrant' seedlings. 'Dead' seedling morphology was present in response to all mild and severe overpriming conditions (HOP1mm, KOP1mm, HOP2mm,  KOP2mm). Considering the effects of kinetin, ROS emission appeared to be higher in response to kinetin-mediated severe overpriming (KOP2mm) compared to the respective hydropriming treatment (HOP2mm). To statistically assess the correlation between ROS levels after dry-back and the loss of desiccation tolerance in M. sativa seedlings, a Pearson's correlation analysis was carried out (Figure 7b). Significant positive correlations were obtained between ROS levels and the percentage of 'aberrant' and 'dead' seedlings (0.827 and 0.472, respectively), whereas the correlation with the percentage of 'normal' seedlings was significantly negative (−0.797). These correlation trends are comparable to those to those obtained for M. truncatula seeds ( Figure 6). tween ROS levels after dry-back and the loss of desiccation tolerance in M. sativa seedlings, a Pearson's correlation analysis was carried out (Figure 7b). Significant positive correlations were obtained between ROS levels and the percentage of 'aberrant' and 'dead' seedlings (0.827 and 0.472, respectively), whereas the correlation with the percentage of 'normal' seedlings was significantly negative (−0.797). These correlation trends are comparable to those to those obtained for M. truncatula seeds ( Figure 6).

Discussion
In the experimental system considered in this work, the combined effects of short and prolonged hydropriming and kinetin-mediated hormopriming were tested on a commercial M. truncatula seed lot. Germination tests confirmed the effectiveness of short and prolonged hydropriming and hormopriming in improving germination performances in M. truncatula seeds as previously observed [12,39], with further improvements in germination speed in response to hormopriming and an increase in germination consistency (expressed as a decrease in the uncertainty index and as an increase in the synchronization index) in response to prolonged hormopriming.
Although increases in germination speed and consistency represents an added value in agricultural contexts [40], the requirements of prolonged priming protocols to improve germination parameters may lead them to apply imbibition-dehydration timing patterns that operate at the limits of desiccation tolerance windows. Avoiding the occurrence of overpriming is crucial to optimize seed priming protocols but needs to account for the variability of different subpopulations within a seed lot in their response to priming and dry-back. In the experimental system considered in this work, mild and severe overpriming conditions were also defined for M. truncatula primed seeds.
A significant increase in the proportion of 'aberrant' seedlings was observed with increasing length of the radicle protrusions before dry-back. This observation was corroborated by a positive correlation index and is in agreement with previous literature, where it is explained as the result of the higher sensitivity of radicle protrusion to desiccation stress compared to the other embryonal tissues [21,25]. The exposure to osmotic agents has been proven effective to improve desiccation tolerance of primed seeds and early seedlings [6,21]. Conversely, in the present work, kinetin-mediated hormopriming per se does not appear to alter overpriming occurrence compared to the same physiological stage (1 mm and 2 mm radicle protrusion) of hydroprimed seeds. Nonetheless, treatments that are specifically able to synchronize germination may represent an alternative route to decrease the exposure to desiccation stress within the defined frame of a priming protocol, whereas the underlying variability of a seed population would represent an added value under less controllable natural environments [41,42]. This might be an applicative outcome of kinetin as a priming agent.
Although kinetin treatments did not alter the incidence of overpriming in the present work, they were effective in accelerating and synchronizing germination. The positive effects of cytokinins, including kinetin, on dormancy release and germination performances have been observed in A. thaliana, T. aestivum, M. truncatula and other crop and model species. These properties have been attributed to the capacity of cytokinins to promote cell division and antagonize the signaling pathways of auxins and abscisic acid, which are the key hormones in dormancy induction and maintenance [39,[43][44][45]. On the other hand, as evidenced in the present work, prolonged exposure to kinetin results in reduced radicle growth. This observation agrees with previous studies [45], where the inhibitory effects of kinetin on radicle development in M. truncatula have been correlated with global metabolomic depletion and accumulation of DNA double strand breaks. Subsequently, prolonged exposure to kinetin to accelerate and synchronize germination needs to be considered along with its side effects.
Considering the observed relation between radicle protrusion length before dry-back and the occurrence of 'aberrant' seedling phenotypes, M. truncatula represents an unambiguous model system for the study of overpriming conditions in species in which such correlations are less explicit and predictable [20]. In the context of the tested priming protocols, the occurrence of overpriming has provided a model to identify possible hallmarks of desiccation stress in M. truncatula, specifically in terms of accumulation and distribution of ROS along the dry-back phase.
ROS accumulation, either as a byproduct of the reactivation of pre-germinative metabolism or as a consequence of biotic/abiotic stress, causes oxidative damage to lipids, proteins and nucleic acids, potentially compromising cellular structures and seed viability [3,6,46]. Nonetheless, within controlled ranges, ROS also have physiological functions in the context of germinative metabolism, promoting cell wall plasticity, endosperm weakening, stress response and reservoir mobilization, either directly or by modulating hormonal signaling [3,46,47]. As a consequence of their dual role, ROS levels in seeds are controlled by an endowment of antioxidant compound accumulated during seed maturation and by the reactivation of antioxidant response pathways during imbibition [48]. The DCF-DA assay carried out in this work highlighted contrastive responses in terms of ROS accumulation during the dehydration phase following short/prolonged imbibition. Direct correlations were evidenced between radicle protrusion length before dehydration, ROS levels at the endpoint of dehydration and the occurrence of 'aberrant' seedling morphology. DAB staining allowed the assessment of the distribution of H 2 O 2 within M. truncatula embryos at the endpoint of dehydration, indicating a localized accumulation in the tips of the radicle protrusions of overprimed seeds. The correlation between radicle elongation and the progressive decrease of desiccation tolerance has been reported in Pisum sativum, Fagus sylvatica and other model systems, and it has been linked to the progressive depletion of compounds, including LEA (Late Embryogenesis Abundant) proteins, stabilizing proteins and cellular structures. In these contexts, the accumulation of H 2 O 2 observed in early seedlings exposed to dehydration stress was explained in terms of ROS leakage through damaged membranes [32,[49][50][51].
Following this direct route of ROS accumulation under drought stress, ROS can subsequently modulate hormonal signaling (specifically ABA) to promote the expression of genes involved in desiccation tolerance, antioxidant response and DNA repair to restore redox homeostasis and preserve seed viability [33,52,53]. The maintenance of desiccation tolerance in germinating seeds implies transcriptomic and metabolomic changes that partially overlap those involved in the acquisition of desiccation tolerance during seed maturation. These include ROS detoxification and synthesis of osmo-protectants, but also lipid, starch and oligosaccharide reservoir mobilization, as highlighted comparing desiccation-sensitive and desiccation-tolerant M. truncatula early seedlings [54]. From a molecular standpoint, these dynamics likely underlie the transition from a prolonged and reversible priming into a condition of irreversible overpriming, as observed in the present work.
To support molecular interpretations, ROS accumulation is measured through multiple detection techniques, including nitroblue tetrazolium chloride staining, fluorescent dyes, confocal microscopy, and electronic spin resonance spectroscopy, with different degrees of sensitivity and specificity [33,55], also allowing the use of ROS as reliable sensors to monitor the occurrence of stress conditions in a variety of experimental systems [56]. In the present work, DCF-DA assay and DAB staining were used to quantify and localize, respectively, the emission of ROS in M. truncatula embryos during post-priming dry-back. The results obtained from DCF-DA assay confirmed ROS accumulation with dry-back progression and increasing radicle length, displaying significant correlations with the loss of desiccation tolerance in both M. truncatula and M. sativa.

Conclusions
This work reported the beneficial effects of hydropriming and kinetin-mediated hormopriming on germination performances in the model legume Medicago truncatula. The contrastive effects of short/prolonged priming protocols and mild/severe overpriming conditions supported the concept of 'overpriming' in terms of loss of desiccation tolerance during the radicle protrusion phase, resulting in impaired seedling development. ROS accumulation during dry-back in response to overpriming suggests the potentialities of ROS as hallmarks to monitor seed priming progression and assess dehydration stress with reliable and resource-effective approaches. These conclusions were confirmed with a proof-of-concept on a commercial variety of Medicago sativa, defining a reproducible experimental system to study and test seed priming protocols in model and crop species.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.