Deterioration of willow seeds during storage

Willow (Salix spp.) seeds are able to tolerate desiccation, but differ from typical orthodox seeds in that they lose viability in a few days at room temperature, and in that the chloroplasts in embryo tissues do not dedifferentiate during maturation drying, thus retaining chlorophyll and maintaining intact their thylakoid membranes. In the present study, we investigated the damage generated in willow seeds during storage under appropriate conditions to exclude the eventual generation of reactive oxygen species by photooxidation. To this end, we measured different indicators of molecular damage, such as changes in the fatty acid profile, protein degradation, nuclease activities, and DNA damage, and evaluated normal germination and total germination in seeds stored for one, ten and sixteen years. We found: (i) a decrease in the fraction of unsaturated fatty acids; (ii) changes in the protein profile due to a decrease in protein solubility; (iii) activation of nucleases; and (iv) DNA fragmentation. Taken together, our findings identified programmed cell death as a key mechanism in seed deterioration during storage. We also found that, although the seeds maintained high percentages of total germination, the death program had already started in the seeds stored for ten years and was more advanced in those stored for sixteen years.

and reserve proteins and lipids in protein and lipid bodies respectively 17,18 . In addition, cells reserve minerals in the form of phytin and phytoferritin, which are included in protein bodies and chloroplasts, respectively. Unlike typical orthodox seeds, willow seeds tolerate very low water contents, and lose viability in a few weeks at room temperature [18][19][20][21] .
We have previously found that seeds of Salix nigra exposed to light are very susceptible to photooxidation 19 . In this respect, we proposed that lipid peroxidation is initiated in the granal membranes of the chloroplasts, but then spreads to other cellular membranes 19,20 .
Because in the present study we aimed to exclude the eventual generation of ROS by photooxidation, we investigated the damage generated during storage, by using willow seeds (specifically Salix nigra L.) that had been stored for one, ten and sixteen years under the following specific conditions: (i) manipulated in darkness and (ii) stored at −80 °C and 8% water content. We monitored different indicators of molecular damage, including changes in conductivity, changes in the content of photosynthetic pigments, changes in fatty acids, changes in profiles of proteins, activation of nucleases and RNases, and DNA damage, and evaluated total and normal germination.

Materials and Methods
Plant material. Seeds 21 and Roqueiro et al. 18 , under very low light intensity to reduce the eventual generation of ROS by photooxidation, and stored in sealed containers at −80 °C and around 8% water content expressed in dry weight basis (dwb).
Experiments were repeated with at least three independent biological replicates and, unless otherwise stated, the results were comparable across experiments. Also, due to the small size of the seeds, each replicate consisted of a large number of seeds (approximately 850 seeds contained in 100 mg of seeds).
Germination tests. Seeds were tested for germination following Maroder et al. 21 and Roqueiro et al. 19 . The germination was considered normal when (i) the seedling was erect; (ii) the cotyledons were undamaged and dark green instead of pale green; and (iii) the hypocotyl and the root were well-developed. Seedlings not meeting these criteria were classified as abnormal. Total germination was calculated as the sum of normal germination and abnormal germination. Fresh seeds germinated in 12-24 h (up to 72 h for low vigor seeds). Definitive counting was carried out 6 days after sowing in order to recognize abnormal seedlings with certainty.
Conductivity tests. Electrolyte leakage from seeds was determined by placing 150 mg of seeds in 6 mL distilled water at 20 °C as previously reported by Roqueiro et al. 19 . Conductivity was measured after 3 h of soaking, using an Altronix Model CTX-II conductivimeter (Brooklyn, NY, USA). Results are expressed as µSmg −1 of dry seeds (mean of three replicates ± s.d.).

Determination of pigments.
Pigment analysis was carried out as described by Arnon 22 and Lichtenthaler 23 . The contents of chlorophyll and total carotenes were analyzed simultaneously 24 . To quantify the content of chlorophyll a, chlorophyll b and carotenes, 17 mg (dry weight) of each sample was used; then, 1 mL of N, N-dimethylformamide was added and left in the dark overnight. Then, the samples were centrifuged at 12,000 g for 10 min and diluted 5-fold. The optical densities of the supernatant were measured at 664 nm, 647 nm, and 480 nm by means of a spectrophotometer at the same time and pigments were quantified using the equations elaborated by Lichtenthaler 23 .
Analysis of fatty acids and malondialdehyde (MDA) content. The Bligh and Dyer method 25 with minor modifications was used to extract the lipid fraction from willow seeds for fatty acid analysis 26 .
Following Burrieza et al. 26 , fatty acid composition was determined by gas chromatography after derivation of extracted oils to fatty acid methyl esters (FAME) according to the AOAC Official Method 19 th EDITION, 2012. The content of each fatty acid was expressed as mean ± standard deviation.
MDA was determined using the thiobarbituric acid (TBA) reaction (Kai and Feng, 2011). Briefly, seed samples (100 mg) were homogenized in 20% trichloroacetic acid. The homogenate was centrifuged at 18,500 g for 20 min and the supernatant was used to evaluate the MDA content. A 0.6-mL sample of 0.5% TBA was added to 0.6 mL of the supernatant, and the mixture was heated at 95 °C for 30 min and then cooled in an ice bath. Then, the mixture was centrifuged at 18,500 g for 15 min. The absorbance of the supernatant was measured at wavelengths of 440, 532 and 600 nm, separately.
Protein analysis. Seeds (50 mg) were ground to a powder with liquid nitrogen. The flour was diluted in 1 mL miliQ water. The homogenate was incubated for 1 h at 4 °C. As described by Burrieza et al. 26 , the protein analysis was based on water (Fraction 1) and saline (Fraction 2) solubility. In addition, a final fraction was obtained using a strong denaturing and reducing buffer containing SDS, beta-mercaptoethanol and urea (Fraction 3). The gels were stained with Coomassie brilliant blue R-250. A rabbit polyclonal anti-RuBisCO antiserum raised against RuBisCO's large subunit (kindly provided by J. J. Guiamet, Universidad Nacional de La Plata, Argentina) was used to carry out a western blot analysis. Western blot analysis was performed following the protocol used by López Fernández et al. 27 . Biorad Precision plus (Bio-Rad #161-0373) was used as molecular mass standard. Protein glycation was determined by the dot blot assay, following the protocol described by Wehr and Levine 28 . Briefly, extracts derivatized with 2,4 DNPH were applied to Dry Immobilon-FL PVDF membranes. Derivatized proteins were detected with anti-DNP antibody. Dots corresponding to oxidized proteins were visualized using secondary antibodies conjugated with 3,3′-diaminobenzidine (DAB) from Roche (Merck, Germany) as substrate. For glycation quantitation, a circle to encompass the largest spot on the blot was drawn with the ImageJ software (http://rsbweb.nih.gov/ij/).
In-gel nuclease activity assay. Nuclease  . Semi-thin sections were obtained with an ultramicrotome and mounted on glass slides. DNA fragments were detected in situ using the terminal deoxyribonucleotidyl transferse (TdT)-mediated biotin-16-dUTP nick-end labeling (TUNEL assay), using the "In situ Cell Death Detection Kit fluorescein" (Roche, Darmstadt, Germany). TUNEL was conducted according to the manufacturer's instructions. A negative control was included by omitting the Terminal deoxynucleotide Transferase (TdT) enzyme from the reaction mixture. As a positive control, sections were incubated with DNase I (1 U/µl, Thermo Scientific ™ ) for 15 min before the labeling reaction. Control treatments were prepared for each set of slides. The sections were counterstained with 0.2 µg/mL 4′,6-diamidino-2-phenylindole (DAPI) (Sigma). Images were obtained as previously reported in López-Fernández and Maldonado 29 . The percentage of TUNEL-positive nuclei was calculated from 150 randomly selected nuclei, for each section. At least three whole-mounts of seeds from each harvest were observed.
Statistical analysis. When applicable, data were analyzed by one-way ANOVA (GraphPad Prism version 6.0), and differences between treatments were determined following Tukey HSD post-hoc test, at P ≤ 0.05.

Results
Total germination, as evaluated by the standard germination test, varied between 77.5% and 100% (Fig. 1A). In contrast, normal germination decreased progressively in seeds collected in 2016, 2006 and 2000 (84, 75 and 33% , respectively) (Fig. 1A). The damage produced in the seeds was reflected as abnormal and taken into account in the evaluation of normal germination.
A decreasing tendency in the contents of chlorophyll a and b and carotenoids was detected throughout the storage period, but differences were not significant (Fig. 1B).
The electrolyte leakage increased with storage time (Fig. 1C). In addition, permeability was higher in seeds of the older harvests (Fig. 1C) and differences were significant (P ≤ 0.05) between the different harvests.
Changes in the fatty acid profile and MDA production during storage time. Changes in fatty acid composition provide an indirect measure of the extent of lipid oxidation. Table 1 shows the fatty acid composition and percentage of saturated, monounsaturated and polyunsaturated fatty acids of stored seeds. Samples from different harvests, i.e. stored for different times, presented differences in their fatty acid profile ( Table 1). Changes in the proportions of saturated fatty acids (SFA) and unsaturated fatty acids were mainly attributed to the oxidative damage occurred during storage. The SFA fraction was slightly affected, which could be explained considering the absence of double bonds in their structure. By contrast, changes in the unsaturated fatty acid fraction were more noticeable. The increase in C18   MDA, a product of lipid peroxidation, showed a gradual increase during storage time. MDA was detected in the seeds collected in 2016 (Fig. 2) and was progressively higher in the older collections. Differences were significant between successive collections (Fig. 2).
Protein solubility decreases during storage. The different protein fractions were sequentially extracted from the three collections studied (Fig. 3A). The main results were: (i) the proteins resolved into distinct bands that spanned a range of molecular weights from 210 to about 10 kDa; (ii) the band of 53 kDa would correspond to the RuBisCo large subunit 33 . The identity of this band was corroborated by western blot analysis (Fig. 3B); during storage, the large subunit of RuBisCo decreased progressively in fractions 1 and 2 (Fig. 3B); (iii) fraction 2, which should be rich in reserve proteins of the globulin type; here showed a very low number of bands, especially in lanes 5 and 6, which correspond to the seeds collected in 2006 and 2002, respectively (Fig. 3A); (iv) the denaturing and reducing buffer used to obtain fraction 3 showed to have been efficient in the extraction of the remaining proteins, which could correspond to reserve proteins; in this sense, for four species of Salicaceae, i.e. three species of Populus (P. grandidentata, P. balsamifera and P. deltoids) and one of Salix (S. microstachya), Beardmore et al. 34 reported polypeptides of various sizes (60, 58, 36, 32, 22, 18, and 14 kDa), which would correspond, approximately, to those here observed in fraction 3 (62, 53, 32, 25, 20, 16, and 14 kDa); and (v) as determined by the dot blot method (Fig. 3C) during the storage time, some proteins were modified by non-enzymatic glycation. In fact, glycation in seeds collected in 2000 was significantly different from that in seeds collected in 2006 and 2016, and glycation in seeds collected in 2006 was significantly different from that in seeds collected in 2016 ( Fig. 3C; Supplementary Fig. 1).

Some nucleases and RNases are activated during storage, whereas others decrease. SDS-
PAGE was used to identify the activities of the nucleases expressed in the embryo tissues during willow seed storage. These enzymes represent different several classes based on their pH and ion dependence and are associated with nuclear DNA and RNA fragmentation.
When double-stranded DNA (ds DNA) was used as substrate, seeds stored of the three collections showed the following differences: (i) in the presence of Ca 2+ at pH 8.0, DNase activity gel staining revealed three bands with masses of 20, 30 and 37 kDa; n20 was clearly active in the younger seed collection, and n30 and n37 were active in the oldest one; (ii) in the presence of Zn 2+ at pH 5.0, DNase activity gel staining revealed four bands corresponding to n20, n30, n37, and n44; (iii) n44 was detected in the older samples, using Zn 2+ as cofactor; (iv) a similar pattern was detected when single-stranded DNA was used as substrate, although n30 showed a clear preference for single-stranded DNA; (v) no nuclease activity was revealed in the absence of cofactors (Fig. 4A).
When RNA was used as substrate, three bands (n15, n30, and n47) were identified at pH 8 using Ca 2+ as cofactor. In the presence of Zn 2+ at pH 5, the n15 band increased its activity in the older samples, suggesting that it is associated with the PCD process. In contrast, the other two bands decreased (Fig. 4A). Figure 4B shows that proteins decreased during storage and that such decrease was progressive and significantly different between the three collections. The DNA showed no significant differences between seeds collected in 2016 and 2006 but differences were significant between 2006 and 2000 samples.
DNA electrophoresis revealed that DNA internucleosomal fragmentation increases during storage time. DNA fragmentation is a typical outcome during PCD in many systems. To evaluate DNA integrity, DNA was isolated from seeds of the three collections and analyzed by DNA gel electrophoresis (Fig. 4C). We found that: (i) in seeds collected in 2016 (lane 1), the extracted DNA produced one well-bound band of high molecular weight that allowed identifying the non-fragmented DNA; however a minor smear was detected; (ii) in the DNA extracted from the seeds collected in 2006 and 2000 (lanes 2 and 3, respectively), the band of high molecular weight was clearly weaker; and (iii) a DNA smearing lower than 600 bp was detected in lanes 2 and 3 (red arrows). This clearly demonstrated an advanced point of nuclear DNA degradation, which is associated with the progression of PCD. The TUNEL assay detected progressive DNA damage during storage. Under blue excitation, the TUNEL assay detected nuclear DNA damage as a green fluorescent labeling in the nuclei of cells. Figure 5A shows representative TUNEL assay images of embryo tissues from the three collections. In seeds harvested in 2000, images of the root apical meristem (RAM) exhibited a large number of positively labeled nuclei (Fig. 5). The number of nuclei positively labeled was reduced in seeds stored from 2006 and inexistent in seeds stored from 2016 (except for some nuclei of procambial cells) (Fig. 5). In samples from 2016, only few nuclei of the procambial tissue of the cotyledons presented TUNEL-positive labeling, whereas the rest of the nuclei were TUNEL-negative (Fig. 5). In the cotyledons, the first events of DNA fragmentation were detected in the seeds collected in 2006 and were progressively higher in the older collection (Fig. 5). The percentage of nuclei labeled in the RAM was 0, 7, and 20% in 2016, 2006, 2000 respectively. Analysis of DAPI-stained nuclei by fluorescence microscopy exhibited progressive changes in nuclear morphology (Fig. 6): initially large and round, then nuclei were smaller and became fusiform; also, a progressive increase in chromatin condensation was observed [35][36][37] . Positive and negative control treatments were conducted for each set of slides ( Supplementary Figs 2 and 3).

Discussion
Even under optimal storage conditions, seeds tend to lose viability 38 , which has been associated with the accumulation of cellular damage to macromolecules, including lipids, proteins, and DNA 39 . In this study, we monitored different indicators of molecular damage, such as lipid peroxidation, changes in protein profiles, protein glycation, activity of nucleases, and DNA fragmentation, in seeds of Salix nigra collected in 2000, 2006 and 2016 and stored according to the protocol previously reported [18][19][20] , which exclude the eventual generation of ROS by photooxidation. We aimed to evaluate their state of conservation considering that: (i) seeds tolerate desiccation but lose viability in a few weeks at room temperature; and (ii) the chloroplasts of the embryo tissues conserve their chlorophyll and granal membranes [18][19][20][21] .
It is extensively agreed that aging in orthodox seeds involves the gradual loss of the integrity of membranes, which are barriers that play an essential role in biochemical or physiological events during germination. The deterioration of membranes involves lipid peroxidation, and this is associated with the leakage of ions, amino acids and sugars 3,11 . In the present study, we found that, although the willow seeds studied maintained very high percentages of total germination, normal germination decreased significantly during storage. Likewise, electrolyte leakage occurred not only in the seeds that had been stored for more than ten years but also in those stored for less than one year, i.e., in a short time following harvest and storage.
It has long been suggested that lipid peroxidation by autoxidation is one of the main causes of seed deterioration. According to Stewart and Bewley 13 , lipid peroxidation and free radical formation are the main causes of the deterioration of oil seeds in storage; during storage, fatty acids are subjected to slow but consistent attack by oxygen, which leads to the formation of hydrogen peroxides, other oxygenated fatty acids and free radicals. The free radicals are unstable and may react and damage nearby molecules. In the absence of enzyme activity, due to the dry state of orthodox seeds during storage, the embryo tissues accumulate oxygenated fatty acids, damaging cellular components and leading to seed deterioration 13 . In this study, we showed that the fatty acid profile of Salix nigra seeds was modified during storage. The decrease in the more sensitive unsaturated fatty acid fraction was higher in the samples stored for a longer time, indicating the prevalence of lipid peroxidation at the storage conditions studied.
It worth to note that there was no statistically significant difference in electrolyte leakage of willow seeds collected in both 2006 and 2000, which was unexpected with the results of germination test and MDA measurement. At present, we do not find an explanation to this result. This issue deserves to be deepened.
According to Walters et al. 40 , protein oxidation either leads to the formation of smaller molecules with reactive carbonyl or nitrogen groups that easily diffuse through cells or generates an additional reaction between carbohydrates, proteins and nucleic acids, causing intermolecular cross-linking and further degradation into advanced glycation end-products. In Arabidopsis, Nguyen et al. 14 confirmed that protein oxidation (carbonylation) is involved in seed deterioration and identified two major storage proteins, crucifernins (12 S globulins) and napins (2 S albumins), which protect seeds from oxidative stress. Here, we analyzed the protein profiles of Salix nigra seeds during storage and compared them with those reported in seeds of various species of Salicaceae by Beardmore et al. 34 . The main and novel differences were: (i) by western blot, we identified that the band of 53 kDa corresponds to the RuBisCo large subunit and that, during storage, this band decreased progressively; (ii) by using saline buffer, we separated a very low amount of polypeptides; and (iii) by using a strong denaturing and reducing buffer, we managed to separate distinct various-sized polypeptides, mostly globulins; such polypeptides were similar in size to those resolved by Beardmore et al. 34 using a saline buffer. We infer that this is caused by a decrease in the solubility of the proteins, which could be attributed to a cross-linking between proteins or between proteins or nucleic acids and carbohydrates; in fact, we found a gradual increase in the content of glycated proteins in the different seed collections, as determined by the dot blot method. This particular issue will be addressed in further studies.
As mentioned above, deterioration in seed quality has been associated with the accumulation of cellular damage to macromolecules, activation of nucleolytic enzymes 41 and changes in nuclear morphology [35][36][37] . Progressive deterioration of 18S rRNA has been observed in dry non-viable rye embryos and wheat embryos 42 and lesions have been found to become amplified with progressive imbibition, either as a result of the intrinsic properties of the rRNA or due to an increase in the activity of other components such as nucleases 43,44 . Progressive rRNA deterioration has also been recorded in embryos of aging seeds of Triticum durum 45 . PCD, as revealed by DNA fragmentation and TUNEL assay, has been detected in aged seeds of Phaseolus vulgaris and Helianthus annuus, coincident with a loss of viability 46,47 . Fragmentation has also been observed to accompany the loss of viability in aged Secale cereal seeds 43 . Also, cell death in Zea mays (maize) seeds has been associated with a dramatic loss in cellular organization 48,49 . According to Rogers 50 , a significant decrease in DNA content often occurs during plant PCD without forming the characteristic DNA ladder. The occurrence of PCD is more precisely determined when structural analyses are associated with the TUNEL assay 51 . In the present study, DNA degradation, which increased with storage time, was associated with changes in the activity of RNases and different nuclease types (acidic, basic, or salt-stimulated). We inferred that the PCD program was advanced in seeds stored for sixteen years but that the program had already begun in the seeds harvested in 2006. However, no clear internucleosomal DNA fragmentation was observed, but rather a smear indicating DNA degradation, which increased throughout storage time. The lack of detection of a clear laddering could be interpreted as a result of the DNA analyzed from samples composed of approximately 85 seeds with probably different timing of the onset of PCD. We concluded that, during storage, willow seeds deteriorated by autoxidation, and that such deterioration was progressive, affecting all biologically relevant macromolecules, i.e., lipids, proteins and nucleic acids. Taken together, our findings allowed identifying several indicative markers for PCD and suggest that this mechanism is involved in the deterioration that slowly occurs in Salix nigra seeds during storage. Figure 6. Details of nuclei from the three different harvests following DAPI staining. Scale bar: 12 μm. In each case, the figure is a representative result of the observation of at least three whole-mounts of seeds from each harvest. Scale bar: 12 μm.