• To investigate the proposed role for NAD metabolism in regulating seed dormancy, NAD metabolites and associated enzyme activities were analysed in seed of Arabidopsis thaliana ecotypes ranging from Col-0, which has low seed dormancy, to Cvi, which is highly dormant.
• Seed poly(ADP-ribosyl)ation levels did not correlate well with the depth of seed dormancy but did correlate with the sensitivity of germination to the DNA damaging agent MMS. Cvi seed had relatively high NAD and low NADP levels compared with the less dormant ecotypes and the NAD : NADP ratios correlated well with dormancy. The activity of NAD kinase was relatively low, and NADP phosphatase was relatively high in dormant Cvi seed, indicating that these enzymes may be involved in controlling the NAD : NADP ratio.
• Dormant fresh Cvi and nondormant after-ripened Cvi seeds were used to investigate further. Measurement of reduced and oxidised pyridine nucleotides indicated that breaking of dormancy was associated with a reduction in NAD levels but not with an increase in NADP levels.
• It is proposed that poly(ADP-ribose) polymerase is involved in protecting the seed from genotoxic stress, whereas the level of NAD affects the depth of dormancy, perhaps by enhancing abscisic acid (ABA) synthesis.
The pyridine nucleotides NAD(H) and NADP(H) are essential cofactors in many redox reactions, and recently many of the enzymes involved in plant NAD synthesis and salvage have been identified and characterized in Arabidopsis thaliana (Katoh et al., 2006; Wang & Pichersky, 2007; Hunt et al., 2007). In addition to its function as an enzyme cofactor, NAD is involved in eukaryotic cell signalling and gene regulation pathways, where NAD is cleaved by enzymes such as poly(ADP-ribose) polymerases (PARPs; Fig. 1) (Hunt et al., 2004; Noctor et al., 2006). The PARPs catalyse the transfer of ADP-ribose groups from NAD to nuclear protein acceptors, resulting in branched chains of ADP-ribose polymers. In the response to DNA damage, where PARP activity has been best studied, poly(ADP-ribosyl)ated proteins bind to specific protein targets and alter their function (Schreiber et al., 2006). In animal cells these NAD-cleaving enzymes, including PARPs, are inhibited by their product, nicotinamide.
We recently reported that Arabidopsis seeds with reduced nicotinamidase activity have increased dormancy, accompanied by increased NAD levels and reduced poly(ADP-ribosyl)ation levels (Hunt et al., 2007). We proposed that increased seed dormancy in the nicotinamidase mutant nic2-1 could result from inhibition of PARP activity through an increase in nicotinamide levels, or the significantly increased levels of NAD found in the mutant seed. The experiments presented here were designed to further investigate whether NAD and/or poly(ADP-ribose) (PAR) levels are involved in determining the depth of seed dormancy.
The control of seed germination is a complex process. Environmental factors and the hormone abscisic acid (ABA) are known to promote seed dormancy, and dormancy can be broken by a variety of treatments, including a period of after-ripening, moist-chilling or application of gibberellic acid (GA), nitrate and nitric oxide (Debeaujon & Koornneef, 2000; Finch-Savage & Leubner-Metzger, 2006). Most Arabidopsis ecotypes exhibit low seed dormancy but the ecotype from Cape Verde Islands (Cvi) is highly dormant and is a good model for studying the biochemical and genetic basis of seed dormancy (Ali-Rachedi et al., 2004). We have measured NAD, NADP and PAR levels, and associated enzyme activities, in the seed of three Arabidopsis ecotypes – Col-0, Ws and Cvi – which have low, intermediate and high levels of dormancy respectively, and have also examined the effect of genotoxic stress on their germination. Together our results suggest that NAD levels are a good indicator of the depth of seed dormancy, but that PARP activity is more likely to be involved in maintaining seed genomic integrity.
Materials and Methods
Arabidopsis thaliana (L.) Heynh plants for seed collection were grown in compost with 16-h photoperiod, day temperature 25°C, night temperature 22°C, and a relative humidity (RH) of 60%. All seeds were freshly harvested from siliquae immediately after pod dehiscence. Fresh seeds were after-ripened by storing at 20°C for 6–9 months at ambient humidity. For determination of seed dormancy levels fresh seed was sown on moistened filter papers. Germination assays were carried out in a growth chamber with an 8 h photoperiod, with a day–night cycle of 22°C : 18°C. For other treatments freshly harvested seed was surface-sterilized (1 min 50% ethanol; 5 min 50% bleach) and sown on 0.6% water agar and either transferred immediately to a plant growth chamber or moist chilled for 4 d at 4°C in the dark.
Pyridine nucleotide extraction
Twenty milligrams of seed was homogenized with 1 ml 0.1 n HCl, extracts were neutralized for NAD and NADP and quantified as described (Bernofsky & Swan, 1973). The NADH and NADPH levels were measured by homogenization in 0.1 n NaOH and concentrations calculated using a control extract without alcohol dehydrogenase or glucose-6-dehydrogenase, respectively.
Seed protein extraction
Ten milligrams of seed was homogenized in a pestle and mortar with 500 µl of Cellytic solution (Sigma-Aldrich, Irvine, UK) with the addition of plant protease inhibitors. Samples were centrifuged at 13 000 g for 10 min and supernatant used for enzyme assays.
Levels of PAR in dry seeds were measured by dot blotting with polyclonal antibody to PAR as described in Hunt et al. (2007).
NAD kinase assay
Ten microlitres of seed extract was incubated at 20°C for 30 min in a final volume of 20 µl, containing 5 mm NAD, 5 mm ATP, 100 mm Tris HCl (pH 7.6), 2 mm MgCl2 and 2 mm CaCl2. Reactions were stopped by heating at 100°C for 5 min. The NADP levels were quantified using a cycling assay with glucose-6-phosphate dehydrogenase as previously described (Bernofsky & Swan, 1973).
NADP phosphatase assay
Ten microlitres of seed extract was incubated at 20°C for 30 min in a final volume of 20 µl, containing 2 mm NADP, 50 mm Tris HCl (pH 7.6), 2 mm MgCl2. Reactions were stopped by heating at 100°C for 5 min. The NAD levels were determined as previously described (Bernofsky & Swan, 1973), using 10 µl of reaction mixture and alcohol dehydrogenase in a cycling assay. Sample absorbance was measured at A570, after subtracting absorbance of control (equivalent seed extract heated at 100°C for 5 min before reaction and incubated for 30 min at 20°C).
Ten microlitres of seed extract was incubated in a final volume of 20 µl, containing final concentrations of 50 mm Tris (pH 8), 1 mm dithiothreitol (DTT), 1 mm MgCl2, 1 µCi of carbonyl-[14C]nicotinamide (American Radio Chemicals, St Louis, MO, USA) for 30 min at 22°C. 2 µl was added to microcrystalline cellulose thin-layer chromatography (TLC) plates F254 (Macherey-Nagel, Düren, Germany) and resolved using a solvent ratio of 60:20:20:1 butanol–methanol–water–ammonia (BMWA) (Zheng et al., 2004). Plates were dried and analysed by phosphoimaging (InstantImager; Canberra Packard, Pangbourne, UK).
[14C]Nicotinamide feeding experiments
Ten milligrams of freshly collected or after-ripened seed was added to a 1.5 ml tube with 75 µl H2O to which was added 150 kBq of [14C]nicotinamide. Seeds were incubated for 24 h, then washed three times in water and metabolites extracted in 80% methanol, 20 mm sodium dithiocarbamate (Zheng et al., 2004). Metabolites were separated by TLC as above using 60 : 20 : 20 : 1 butanol–methanol–water–ammonia as solvents or with isobutyric acid–ammonia–water in a 660 : 17 : 330 ratio. All standards were purchased from Sigma (Poole, UK) except 6-methylnicotinamide which was obtained from Fisher Scientific (Loughborough, UK). The radiolabelled spot with the highest Rf were putatively identified as methylnicotinic acid, in line with Wang & Pichersky (2007). We note that it cannot be ruled out that this may comigrate with 6-methylnicotinamide or 1N-methylnicotinamide, although neither of these has previously been identified in plants. Previous nicotinamide feeding experiments in plants revealed an unknown metabolite (Matsui et al., 2007; Wang & Pichersky, 2007) which our experiments suggest is the glycine conjugate of nicotinic acid, nicotinuric acid which has been previously identified in cold-treated Arabidopsis (Cook et al., 2004). Matsui et al. (2007) identified nicotinic acid glucoside as a metabolite in Arabidopsis leaves and suspension cells, but we were unable to obtain a standard for this compound.
Gene expression analysis
RNA was extracted from dry seeds as described previously (Haslekas et al., 2003). A 2 µg sample of total RNA was reverse transcribed with Superscript II (Invitrogen, Carlsbad, CA, USA), diluted 100-fold and amplified with Taq polymerase using primers UBQ10, 5′-gatctttgccggaaacaattggaggatggt-3′, 5′-cgacttgtcattagaaagaaagagataacagg-3′; NADK1, 5′-tccgggttgatcacgaacct-3′, 5′-cacatgtgtgcaaggcgatg-3′; NADK2, 5′-tttggcaacgggaagttttca-3′, 5′-gggacggctcctttgaacaag-3′; NADK3, 5′-aaccgattgatccgtaccca-3′, 5′-cgtttaagactggcgcatcc-3′.
PAR levels correlate with resistance to genotoxic stress but do not correlate well with depth of seed dormancy
The percentage germination of freshly collected seed (germination potential) of the ecotypes used in this study ranged from high in Col-0 (100% germination) to low in Cvi (1% germination) (Fig. 2a). All three ecotypes reached 100% germination after a moist chilling pretreatment demonstrating that seed was viable (data not shown). We measured PAR levels in dry seed of Col-0, Ws and Cvi and found that these levels did not correlate well with the depth of seed dormancy. The PAR levels in Cvi were significantly higher than Ws (P < 0.006) but not Col-0 (P < 0.06) (Fig. 2b). As PARP activity is involved in the response to genotoxic stress the effect of the DNA damaging agent MMS on germination was also measured. Figure 2c shows that after moist chilling in the presence of 0.01% MMS Cvi germination was significantly (1.7-fold) less sensitive than Col-0, (P < 0.007) and eightfold less sensitive than Ws (P < 0.01). The level of MMS insensitivity correlated well with PARP activity.
NAD levels are relatively high and NADP levels are relatively low in highly dormant seed compared to nondormant seed
The concentrations of NAD and NADP were measured in freshly collected dry seed (Fig. 2d). The Cvi seed, with the highest dormancy, had the highest levels of NAD and the lowest levels of NADP. The NAD levels in Cvi seed were significantly higher than both Ws (P < 0.01) and Col-0 (P < 0.01), with levels approximately threefold higher, whereas NADP levels in Cvi seeds were significantly lower (twofold) than Col-0 (P < 0.03). Thus it appeared from our experiments that dormant seed have a high NAD : NADP ratio, and nondormant seed have a low NAD : NADP ratio. These results suggested an inverse relationship between seed germination potential and seed NAD concentration or NAD : NADP ratio.
Highly dormant seeds have relatively low NAD kinase and high NADP phosphatase activities
NAD kinases are able to catalyse the ATP-dependent formation of NADP from NAD, or NADPH from NADH (Fig. 1). They may therefore be important in regulating the NAD : NADP ratio in seed. It was found that Cvi seeds had significantly reduced NAD kinase activity compared with Ws (P < 0.002) or Col-0 (P < 0.0007) (Fig. 3a). NADP phosphatase catalyses the reverse reaction to NAD kinase, (i.e. the conversion of NADP to NAD; Fig. 1). NADP phosphatase activities in our seed extracts were approximately twofold higher in Cvi than Ws (P < 0.0002) or Col-0 (P < 0.003; Fig. 3b). These results support the suggestion that the NADP : NAD ratio may be regulated by these enzymes in seeds.
Arabidopsis has three NAD kinases (Hunt et al., 2004): NADK1 activity is calcium independent, NADK2 requires calcium/calmodulin while NADK3 has a preference for NADH as a substrate over NAD and does not require calcium/calmodulin for activity (Turner et al., 2005a,b). To examine which NADK isoform genes are expressed in Arabidopsis seeds semiquantitative reverse transcriptase polymerase chain reaction (RT-PCR) was performed on freshly collected dry seed of each ecotype (Fig. 3c). All three isoforms were expressed in the three ecotypes, but Cvi seeds had lower levels of expression of NADK1 and NADK2 genes, which is in agreement with the lower NADK activity found in Cvi seed (Fig. 3a). Transcriptomics data available through http://www.genevestigator.ethz.ch also suggests expression of the NADK genes in seed, consistent with a potential role in the control of seed dormancy and germination. NADK3 is absent from the ATH1 Genechip.
Nicotinamidase activity does not correlate well with seed dormancy
The relatively high NAD levels found in Cvi seeds (Fig. 2b) may have been caused by a reduced level of nicotinamidase activity, as reported for nic2-1, which also has high seed dormancy and enhanced NAD levels (Hunt et al., 2007). Seed nicotinamidase activities were measured by the conversion of [14C]nicotinamide to [14C]nicotinic acid and were found to be twofold higher in Cvi than in Col-0 (P < 0.02) and threefold higher in Cvi than in Ws seeds (P < 0.006). Col-0 also had significantly higher nicotinamidase activity than Ws (P < 0.04) (Fig. 4a). These results indicated that the high NAD levels in Cvi seed were not caused by a reduction in nicotinamidase activity. Although the nicotinamidase activity did not correlate well with the depth of seed dormancy it did correlate with the PAR levels measured in dry seed of Col-0, Ws and Cvi (Fig. 2a). Thus it appears likely that nicotinamidase activity has an effect on PARP activity and resistance to genotoxic stress but does not directly affect seed dormancy.
We observed that germination of Cvi seeds, which had relatively high nicotinamidase activity, was particularly sensitive to exogenously applied nicotinamide. Figure 4b shows Col-0 germination to be significantly less sensitive to 5 mm nicotinamide in the germination medium than either Ws (P < 0.005) or Cvi (P < 0.0003). Ws was significantly more sensitive to 5 mm nicotinamide than Cvi (P < 0.02). There were no significant differences in germination sensitivities at 1 mm nicotinamide. Thus, unexpectedly, it appeared that seeds with higher nicotinamidase activities are also more sensitive to nicotinamide.
After-ripening alters pyridine nucleotide ratios
The differing NAD : NADP ratios found in the three ecotypes above led us to investigate whether changes in NAD metabolism were also associated with the reduction in seed dormancy that occurs during after-ripening. We carried out a more detailed analysis of all pyridine nucleotides in the same ecotype background using freshly harvested and after-ripened seeds of Cvi. After-ripened seed exhibited 100% germination frequency in contrast to fresh Cvi seed with approx. 1% germination (Fig. 2a) and was associated with a 30% drop in NAD levels during after-ripening of Cvi seeds (Fig. 5a; P < 0.045). However, NADP levels were unchanged by after-ripening (Fig. 5a). After-ripened seeds also showed a c. twofold increase in NADH levels (P < 0.015). These changes in pyridine nucleotide levels resulted in a significant drop in the NAD : NADP ratio (Fig. 5b; P < 0.001), and reductions in the NADH : NADPH (P < 0.009) and the NAD : NADH (P < 0.020) ratios in after-ripened seed (Fig. 5b). The NADP : NADPH ratio and the total level of pyridine nucleotides did not significantly change during after-ripening (data not shown). We also calculated the charge ratio, (the ratio of oxidized to reduced pyridine nucleotides) which reduced during after-ripening but this change was not statistically significant.
After-ripening alters nicotinamide and NAD metabolism
The metabolic fate of labelled nicotinamide in imbibed fresh or after-ripened seed was compared by TLC using unlabelled standards to identify as many compounds as possible (Fig. 6a). After-ripened seed had significantly increased label in nicotinamide (P < 0.04) and methylnicotinic acid (P < 0.009) and reduced incorporation into nicotinic acid (P < 0.002) compared with fresh seed, suggesting reduced nicotinamidase activity and increased methyltransferase activity (Fig. 6b). The unresolved compounds at the TLC origin (Fig. 6a) were separated using an alternative solvent system (Fig. 6c). This revealed significant decreases in radiolabel in after-ripened seed of NAD (P < 0.002) and NADP (P < 0.024) and increased flux into nicotinamide mononucleotide (NMN P < 0.004) (Fig. 6d). The conversion of NAD to NMN is catalysed by enzymes of the Nudix family in Arabidopsis (Ge et al., 2007) and our results suggest that this activity may be increased by after-ripening. Nicotinamide mononucleotide may then be dephosphorylated to nicotinamide ribose and converted to nicotinamide, although this activity has yet to be identified in plants.
We previously showed a link between seed dormancy, NAD and PAR levels, and resistance to genotoxic stress, using a null mutant in a seed expressed nicotinamidase (Hunt et al., 2007). To further dissect these relationships we measured the levels of enzymes and metabolites involved in NAD signalling in three Arabidopsis ecotypes which have a range of dormancy levels. This study represents the most comprehensive analysis of pyridine nucleotides in Arabidopsis seed to date. However, we measured only the levels of free nucleotides and it is possible that a significant amount of nucleotide is bound to proteins (Kasimova et al., 2006). It is also likely that within seed the nucleotide levels vary between the tissues and subcellular compartments. For example PARP is an exclusively nuclear enzyme, whereas NAD kinases are known to be both chloroplastic and cytosolic.
The PAR levels correlate with seed resistance to genotoxic stress but not dormancy
The level of poly(ADP-ribosyl)ation varied between seeds of Cvi, Ws and Col-0 ecotypes and correlated well with their nicotinamidase activities, but did not correlate well with the depth of seed dormancy. This leads us to the conclusion that dormancy strength is not determined by PARP activity, which is supported by a comparison with our previous findings. In our previous study nic2-1 seeds, which exhibit increased dormancy were found to have reduced PAR levels (Hunt et al., 2007) whereas in this study the more dormant Cvi seed had relatively high PAR levels. An explanation for this apparent discrepancy in PAR level measurements could be that PARP activity is inhibited in nic2-1 via increased nicotinamide levels. Our results suggest that PARP activity is regulated by nicotinamidase activity in seed, and we observed a correlation between these two enzyme activities in the three Arabidopsis ecotypes tested here. The mechanism by which NAD levels affect dormancy in the three ecotypes tested here appears to be independent of PARP activity.
Poly(ADP-ribose) polymerase activity is believed to be involved in the response to genotoxic stresses such as those caused by oxidative stress or ionizing radiation (Amor et al., 1998; Doucet-Chabeaud et al., 2001; De Block et al., 2005). We tested the sensitivity of seeds of the three Arabidopsis ecotypes to the DNA-damaging agent MMS and found that resistance to MMS correlated well with both seed PAR levels and nicotinamidase activities (but did not correlate well with dormancy or NAD levels). These results suggest that although PARP does not have a direct role in controlling seed dormancy, it may be involved in the response to genotoxic stress during dormancy and the early stages of germination but we restricted analysis of PAR levels to dry seed. The more dormant ecotypes of Arabidopsis such as Cvi and Shakdara are both better able to survive after controlled deterioration tests (CDTs; Bentsink et al., 2000). This may reflect a greater ability to withstand genotoxic stresses, which in Cvi at least could be related to high PARP activity. It is probable that this allows these seeds to survive longer in a quiescent state and germinate successfully with their genomes undamaged. Mature seeds are likely to be subjected to high levels of stresses in their desiccated state and show particularly high levels of expression of genes involved in ROS detoxification, such as catalases and superoxide dismutases (Nakabayashi et al., 2005). It has been suggested that during storage and upon imbibition, seeds accumulate free radicals, which must be detoxified during dormancy after-ripening and germination to prevent damage to proteins, membranes and DNA (McDonald, 1999: Kibinza et al., 2006).
Seed NAD(H) levels may regulate dormancy
The levels of NAD and NADP were analysed in fresh seed of the Arabidopsis ecotypes Cvi, Col-0 and Ws. Our experiments revealed that NAD levels were proportional to seed dormancy, whereas NADP levels were inversely proportional to seed dormancy (Fig. 2a,b). It therefore appeared that the NAD : NADP ratio may be involved in determining or controlling the depth of seed dormancy. To further investigate this we measured changes in nucleotide levels in fresh and after-ripened seeds of the most dormant ecotype, Cvi. While NADP levels did not change, NAD levels reduced significantly during after-ripening. Our measurements of seed enzyme activities indicated that the seed NAD levels may be determined by differential regulation of the enzymes which catalyse the interconversion of NAD and NADP. The activity of NAD kinase was significantly lower in the more dormant Cvi seed than either Ws or Col-0 ecotypes, and NADK1 and NADK2 genes were found to be expressed at reduced levels in Cvi seed. By contrast, NADP phosphatase activity was significantly higher in Cvi seed. The nature of the protein responsible for NADP phosphatase activity is not known and so patterns of gene expression could not be investigated for this enzyme This activity has been shown to be higher in dormant than nondormant Avena seeds (Gallais et al., 1998). On the basis of these results it appears that the depth of seed dormancy may be related to the NAD level or the NAD : NADP or NAD : NADH ratios. This suggestion and our results are in accord with our previous finding that exogenously applied NAD is an inhibitor of seed germination (Hunt et al., 2007).
nic2-1 and Cvi seed both have a high level of seed dormancy but unlike nic2-1, which has reduced seed nicotinamidase activity, we found that Cvi had relatively high nicotinamidase activity which appeared to be reduced by after-ripening (Fig. 6a,b). This suggests that nicotinamidase activity is not directly correlated with the depth of seed dormancy. In our experiments seed of ecotypes with high dormancy had high nicotinamidase activity, but germination was more sensitive to added nicotinamide, indicating that a metabolite of nicotinamide, rather than nicotinamide itself, may act to increase seed dormancy. Nicotinamide is recycled to NAD via a salvage pathway involving nicotinamidase activity (Fig. 1). Thus the relatively high nicotinamidase activity in dormant seeds, such as Cvi, may contribute to an increased level of NAD and an increased NAD : NADP or NAD : NADH ratio which, as we discuss above, are associated with high seed dormancy.
Flux through the NAD salvage pathway may be reduced by after-ripening
Our experiments with radiolabelled nicotinamide suggested that after-ripening reduces nicotinamidase activity and increases the activity of methyltransferase and Nudix enzymes. Nudix hydrolases have a wide variety of substrates and here we did not attempt to distinguish between NAD/NADH or NMN/NMNH by TLC. These flux alterations may contribute to the observed reduction in NAD levels during after-ripening by reducing the flux through the NAD salvage pathway and increasing the conversion of NAD to nicotinamide mononucleotide. This may represent a change from quiescence into preparation for germination by changing the ratio of NAD to NADH as metabolism resumes. A Nudix hydrolase (ATNUDT3) was shown to vary between dormant and nondormant seeds (Chibani et al., 2006).
Metabolism of NAD may regulate dormancy breaking enzymes
The capacity to generate sufficient energy for biosynthetic reactions must be important to ensure successful germination. The oxidative pentose phosphate pathway (OPPP) generates NADPH from NADP in nonphotosynthetic cells to provide reducing power for biosynthesis (Kruger & von Schaewen, 2003). However, we found no evidence to support the hypotheses that NADPH levels limit germination of dormant seed by either OPPP or the nitric oxide synthesis pathway as we observed no significant changes in NADP or NADPH levels following after-ripening of Cvi seed.
The differences in dormancy levels between ecotypes are likely to involve alterations in ABA biosynthesis and catabolism. NAD is required for ABA synthesis by xanthoxin dehydrogenase (Gonzalez-Guzman et al., 2002); while NADPH is required to degrade ABA by ABA hydroxylase (Kushiro et al., 2004). Thus, the high NAD levels that we report in fresh Cvi seed may be expected to have a positive effect on ABA levels, which in turn would strengthen seed dormancy.
In conclusion, we observed a correlation between seed NAD(H) levels and the depth of dormancy. We propose that seed dormancy may be regulated by the relative levels of pyridine nucleotides, which could affect the activity of enzymes that regulate the metabolism of compounds involved in dormancy. By contrast, poly(ADP-ribosyl)ation levels did not correlate well with seed dormancy, and we propose that PARP activity is involved in protecting the embryo from genotoxic stress during the period of seed dormancy.