Nicotinamidase activity is important for germination

Summary

It has been suggested that nicotinamide must be degraded during germination; however, the enzyme responsible and its physiological role have not been previously studied. We have identified an Arabidopsis gene, NIC2, that is expressed at relatively high levels in mature seed, and encodes a nicotinamidase enzyme with homology to yeast and bacterial nicotinamidases. Seed of a knockout mutant, nic2-1, had reduced nicotinamidase activity, retarded germination and impaired germination potential. nic2-1 germination was restored by after-ripening or moist chilling, but remained hypersensitive to application of nicotinamide or ABA. Nicotinamide is a known inhibitor of NAD-degrading poly(ADP-ribose) polymerases (PARP enzymes) that are implicated in DNA repair. We found reduced poly(ADP)ribosylation levels in nic2-1 seed, which were restored by moist chilling. Furthermore, nic2-1 seed had elevated levels of NAD, and germination was hypersensitive to methyl methanesulphonate (MMS), suggesting that PARP activity and DNA repair responses were impaired. We suggest that nicotinamide is normally metabolized by NIC2 during moist chilling or after-ripening, which relieves inhibition of PARP activity and allows DNA repair to occur prior to germination.

Introduction

Nicotinamide, the vitamin B3 precursor, has a diverse range of biological effects. In animals, vitamin B3 is essential for the release of energy from carbohydrates, fats and proteins, and deficiency causes pellagra. It is probable that intracellular nicotinamide levels are tightly regulated via both degradation and synthesis. Nicotinamide is a metabolite of NAD and is recycled via a salvage pathway back to NAD (Figure 1a; Gholson, 1966). Only one route is known for the degradation of nicotinamide in plants and yeast, the conversion of nicotinamide to nicotinic acid, catalysed by nicotinamidase (Figure 1a,b; Anderson et al., 2003). Animals have nicotinamide phosphoribosyltransferase activity but there is no evidence for this enzyme activity in plants (Zheng et al., 2005), suggesting that a pathway through nicotinamidase is the only way to degrade nicotinamide and salvage NAD. In yeast, nicotinamidase PNC1 activity regulates transcriptional silencing and longevity by preventing the accumulation of nicotinamide during times of stress (Gallo et al., 2004). Indeed, lifespan in yeast is dramatically reduced by the application of nicotinamide, and extended by low-intensity stresses such as heat and salt, factors that increase the levels of PNC1 (Anderson et al., 2003). In plants, application of nicotinamide has been shown to reduce H₂O₂-induced cell death (Amor et al., 1998), prevent development of tracheary elements (Sugiyama et al., 1995), and inhibit ABA-induced stomatal closure (Leckie et al., 1998). More recently, nicotinamidase activity has been detected in embryos of germinating mungbean, and application of nicotinamide has been shown to inhibit embryo growth (Zheng et al., 2005). In these experiments, application of radiolabelled nicotinamide resulted in incorporation of radioactivity into NAD(P), demonstrating that NAD is salvaged from nicotinamide during germination.

Figure 1.  NAD and nicotinamide metabolism in plants and yeast. (a) Potential enzymic sources and metabolites of nicotinamide in plant cells. Three potential pathways of NAD degradation are shown, catalysed by ribosyl cyclase, poly(ADP-ribose) polymerase (PARP) or sirtuin deacetylase. cADPR, cyclic ADP-ribose; OAc-ADPR, O-acetyl ADP-ribose; PNC1, yeast nicotinamidase; NAM, nicotinamide; NA, nicotinic acid. (b) Conversion of nicotinamide to nicotinic acid, the reaction catalysed by nicotinamidases. (c) Alignment of deduced sequences of yeast PNC1 (NP_011478) and bacterial PNCA (GI:81414154) with Arabidopsis homologues NIC2 and NIC3. Alignments were performed using Multalin (http://prodes.toulouse.inra.fr/multalin) and displayed using GeneDoc (http://www.flu.org.cn/en/download-47.html).

Nicotinamide levels are also affected by nicotinamide synthesis, and there is evidence for three enzymes in plants that produce nicotinamide by cleaving NAD: poly(ADP-ribose) polymerases (PARPs), ribosyl cyclases and sirtuins (Figure 1a; Hunt et al., 2004). In animals, each of these three enzymes are subject to feedback inhibition by their product, nicotinamide (Anderson et al., 2003).

We have investigated the role of nicotinamide metabolism in plant cells by identifying and studying nicotinamidases. We demonstrate here that proteins encoded by the adjacent Arabidopsis genes At5g23220 and At5g23230, with homology to yeast and bacterial nicotinamidases, have nicotinamidase activity in vitro. Our characterization of a nicotinamidase mutant (nic2-1) reveals a possible role for nicotinamidase activity in regulating seed germination potential.

Results

The Arabidopsis genome encodes two nicotinamidase proteins

BLAST analysis (http://www.ncbi.nlm.nih.gov/blast/) identified three Arabidopsis genes encoding products with homology to yeast nicotinamidase PNC1 (Figure 1c). At5g23230 (45% similarity, 25% identity to PNC1 over 124 amino acids, using the BLAST2 sequences program) and At5g23220 (44% similarity, 24% identity over 106 amino acids) are annotated as hypothetical nicotinamidases in Munich Information Center for protein sequences (MIPS, http://mips.gsf.de), and we have designated them as NIC2 and NIC3, respectively. A third gene, At3g16190 (42% similarity, 26% identity over 213 amino acids) is annotated as an isochorismatase. Isochorismatases are thought to be only found in bacteria, whereas nicotinamidase activity has been identified in bacteria (Purser et al., 2003), yeast (Anderson et al., 2003; Ghislain et al., 2002) and plants (Zheng et al., 2004). To investigate whether NIC2, NIC3 or At3g16190 encode proteins with nicotinamidase activity, we expressed His-tagged cDNA gene products in Escherichia coli cells. NIC2 and NIC3 (Figure 2a) were both able to degrade nicotinamide. No activity was detectable without added enzyme. NIC2 and NIC3 enzymes appeared to have specificity for nicotinamide as no activity was detected with nicotinic acid, NAD, NADP, glutamate or asparagine as substrate. The activity of both enzymes was stimulated by 5 mm MgCl₂ and 2 mm DTT, and inhibited by 5 mm ZnCl₂, 5 mm NiCl₂ and 5 mm CuCl₂ (Figure S1). No activity could be detected in similar assays with At3g161390 recombinant protein expressed with either a His tag or a GST tag (data not shown). After submission of this paper, we were informed of another gene (At2g22570) that encodes a functional nicotinamidase with homology to PNC1, which has been designated NIC1 (Wang and Pichersky, 2007).

Figure 2.  NIC2 protein degrades nicotinamide in vitro, and NIC2 gene expression is high during seed maturation. (a) HPLC traces demonstrating the nicotinamide-degrading activity of recombinant NIC2 and NIC3. Reaction products were separated by ion-exchange chromatography and detected by absorbance at 280 nm. Retention times were 4 and 9 min for nicotinamide and nicotinic acid standards, respectively. (b) Expression of NIC2 in 2-week-old Arabidopsis seedlings. Semi-quantitative RT-PCR analysis including RUB1 as control. (c–e) Histochemical staining of a plant expressing a P_NIC2::GUS fusion showing GUS activity in dissected testa/endosperm (c), embryo (d) and testa of intact seed (e). (f) Expression patterns of NAD metabolism genes during seed development. Transcript levels of NIC1, NIC2, Srt1, Srt2 and PARP3 are shown. AtGenexpress microarray data were extracted using Genevestigator (http://www.genevestigator.ethz.ch). The mean spot intensity for three replicates was plotted for silique-expressed genes from slides ATGE76–ATGE84. Stages 3–5 are whole siliques from 6-week-old plants; stages 6–10 are seeds from 7-week-old plants.

NIC2 is expressed during seed maturation

Semi-quantitative RT-PCR was used to examine the expression patterns of the NIC2 gene in seedlings. NIC2 mRNA was detected in both roots and shoots (Figure 2b). P_NIC2::GUS fusion experiments suggested that NIC2 is expressed weakly in the roots and vasculature in seedlings, but at higher levels in both the endosperm and embryo of mature seeds (Figure 2c–e). Analysis of available microarray data indicated that NIC2 is expressed at relatively high levels during seed maturation (http://www.genevestigator.ethz.ch; Zimmermann et al., 2004). Silique and seed mRNA levels reported for NIC2 and for genes encoding putative nicotinamide-generating PARP and sirtuin enzymes are plotted in Figure 2(f). NIC3 transcript was not detected in these experiments. There was a substantial increase in the expression level of NIC2 during seed development, reaching a peak at seed maturity (stages 8–10). A similar expression pattern was observed for a PARP homologue encoded by At5g22470 (50% similarity, 33% identity to AtPARP2; 44% similarity, 27% identity to human PARP-2), which we designate PARP3, that is expressed at particularly high levels in seed. NIC1 and the two Arabidopsis putative sirtuin genes, Srt1 and Srt2 (Pandey et al., 2002) were expressed at low constant levels throughout seed development. Further available microarray data suggests that NIC2 and PARP3 transcript levels return to background levels 24 h after imbibition (http://www.genevestigator.ethz.ch; Nakabayashi et al., 2005).

Mutant nic2-1 seeds have delayed germination and reduced germination potential

An Arabidopsis mutant with a T-DNA insertion in the coding region of NIC2 was obtained and characterized (Figure 3a). No NIC2 transcript was detected in the nic2-1 mutant (Figure 3b), and Southern blotting indicated only one T-DNA insertion site in the nic2-1 genome (Figure S2). We presume therefore that this allele represents complete loss of NIC2 encoded function. The nic2-1 plants appeared to grow and develop normally. As NIC2 transcript levels were relatively high in mature seed, we investigated seed phenotype and observed a paler seed coat colour compared to Ws (parental background; Figure S3). Dissection of the seed showed that the embryo was the same colour in Ws and nic2-1, and confirmed that the nic2-1 testa had altered pigmentation. To confirm that nic2-1 seed has reduced nicotinamidase activity, we assayed for activity by following the conversion of [¹⁴C]-labelled nicotinamide to nicotinic acid. An approximately twofold decrease (P < 0.05) in nicotinamidase activity was seen in dry seed extracts of nic2-1 (Figure 4a). Although nicotinamidase activity was reduced, it was still detectable in nic2-1 extracts, suggesting that there may be other proteins capable of deaminating nicotinamide in Arabidopsis seeds. Germination experiments with moist-chilled (stratified) nic2-1 seeds showed that these seeds exhibited slightly retarded germination, but there was no difference in the final percentage germination (Figure 4b). However, in the absence of a moist chilling pre-treatment (non-stratified), germination was delayed and the germination potential of nic2-1 seeds was significantly lower than that of Ws (P < 0.05), indicating a hyperdormant phenotype (Figure 4b). Complementation of nic2-1 with a 4 kb genomic DNA fragment containing NIC2 was able to rescue normal germination (Figure 4c). Dormancy is usually removed by moist chilling or by after-ripening of seeds through dry storage (Bewley, 1997). To determine whether nic2-1 seeds show delayed after-ripening, the germination potential of stored seed was tested every 7 days, without moist chilling. After-ripening of nic2-1 seed was delayed in comparison to Ws seed, although after 21–28 days storage, the germination potential of nic2-1 reached that of Ws (Figure 4d). Seed maturation and germination are both ABA-dependent processes (Bewley, 1997), and germination of stratified nic2-1 seed was found to be hypersensitive to ABA (Figure 4e,f). This was unlikely to be due to altered permeability of the nic2-1 seed coat, as no differences in 2,3,5-triphenyltetrazolium chloride staining (Debeaujon et al., 2000) were seen between Ws and nic2-1 (data not shown). To investigate whether ABA hypersensitivity could be due to enhanced ABA levels, we estimated ABA levels in seed extracts using radioimmunoassay (Quarrie et al., 1988), and found no significant difference in ABA levels between Ws and nic2-1 seed. Both Ws and nic2-1 had an average of approximately 2 fg ABA per seed. We also found that nicotinamide is an inhibitor of Arabidopsis seed germination when added to the germination medium, and that nic2-1 seeds show hypersensitivity to nicotinamide (Figure 4g,h). At 5 mm, nic2-1 seeds were significantly (P < 0.02) more sensitive to nicotinamide than Ws. This result is in line with a previous observation that nicotinamide inhibits the growth of mungbean embryos during germination (Zheng et al., 2005).

Figure 3.  Characterization of the T-DNA insertion site in nic2-1. (a) nic2-1 contains a T-DNA insertion in the coding region of NIC2. The positions of primer sequences used for analysis (Lb2, F1, R1, R2 and R3) are marked with arrows. The white box represents the NIC2 coding region, hatched boxes represent non-coding regions, and lines represent intergenic regions. The triangle indicates the position of the T-DNA insertion. (b) PCR analysis of nic2-1. Amplification of genomic DNA confirmed the T-DNA insertion in nic2-1, and amplification of cDNA showed that no full-length or truncated NIC2 transcript was present in nic2-1. Primer positions are shown in (a).

Figure 4.  Characterization of nic2-1 seed germination. (a) Nicotinamidase activity in seed extracts. Conversion of [¹⁴C]-nicotinamide to [¹⁴C]-nicotinic acid was measured in extracts resolved by TLC. Mean relative spot intensities are shown for three independent extractions from seeds pooled from 6–8 plants. Activity is significantly reduced in nic2-1 seeds (P < 0.05). Bars are SE. (b) Germination of stratified (st) and non-stratified (nst) Ws and nic2-1 seeds. Each point is the mean of experiments on seed from pools of 6–8 plants (n = 3). The germination of non-stratified nic2-1 was significantly less than that of non-stratified Ws at 5–7 days after plating (P < 0.05). Bars are SE. (c) Complementation of nic2-1. Germination potential of Ws, nic2-1 and two independent lines of nic2-1 transformed with a 4 kb Arabidopsis genomic DNA fragment containing NIC2. Percentage germination of non-stratified seed after 7 days is shown. Each point is the mean of experiments on seed from pools of 6–8 plants (n = 3). Ws and complemented lines had significantly higher germination potential than nic2-1 (Ws, P < 0.04; line 1, P < 0.04; line 2, P < 0.0 006). Bars are SE. (d) Effect of after-ripening on the germination potential of Ws and nic2-1. The percentage germination of non-stratified seed was assessed 7 days after plating following a period of dry dark storage at 20°C as indicated. The results shown are representative of three independent experiments. (e) Effect of ABA on Ws and nic2-1 germination. Seeds were sown on medium supplemented with ABA or a methanol (0.02%) solvent control and stratified. Plates were then transferred to light for the time indicated. Points are the mean of three pools of 3–6 plants. Germination indices were calculated from six measurements over 3 days. Bars are SE. (f) Comparison of ABA sensitivities. Germination indices from (e) were expressed relative to indices for plants sown on medium without ABA. Germination of nic2-1 seed was significantly more sensitive to ABA at 1 μm (P < 0.03) and 2 μm (P < 0.04) compared with Ws. (g) Effect of nicotinamide on germination. Seeds from three pools of 6–8 plants were sown on medium containing nicotinamide and stratified. Germination indices were calculated from six measurements over 3 days. Bars are SE. (h) Comparison of nicotinamide sensitivities. Germination indices from (g) were expressed relative to indices for plants without nicotinamide. At 5 mm, nic2-1 seed was significantly more sensitive to nicotinamide than Ws (P < 0.02).

nic2-1 seeds have increased NAD levels

As NIC2 is likely to be involved in recycling NAD (Figure 1a), we measured pyridine nucleotide levels in nic2-1. NADP levels were not significantly altered in nic2-1 seed, but NAD levels showed a significant 1.75-fold increase (P < 0.05; Figure 5a). We also used an in vivo assay to investigate whether nic2-1 seeds have enhanced levels of NAD. The enzyme alcohol dehydrogenase converts NAD and ethanol to NADH and acetaldehyde, the latter of which is toxic and inhibits germination (Holm, 1972). Increased levels of NAD are believed to increase the rate of ethanol conversion to acetaldehyde. We found that application of ethanol had a much greater inhibitory effect on the germination of stratified nic2-1 seeds compared to Ws (Figure 5b,c). Ws seed showed a small increase in germination rates at 0.1 and 0.25% ethanol, but a slight inhibition at 0.5%. At 0.1% and 0.25% ethanol, there was a significant difference in germination (P < 0.05). At 0.25% ethanol, the germination of Ws was 110% of that for untreated controls (no ethanol), compared to 58% of untreated nic2-1. These results indicate that NAD levels are increased in nic2-1. The effect of the nic2-1 mutation on germination is unlikely to be due to a lack of NAD through reduced recycling as this might be expected to reduce NAD levels.

Figure 5.  nic2-1 seeds have elevated NAD levels and germination is sensitive to NAD. (a) NAD and NADP levels in dry Ws and nic2-1 seed. Nucleotides were extracted from freshly harvested seeds of Ws and nic2-1. Bars are means of three independent extractions from three pools of 3–6 plants. Error bars are SE. (b) Effect of ethanol on germination of stratified Ws and nic2-1 seeds. Values are expressed as germination index 5 days after transfer to 22°C. Bars are SE. (c) Comparison of ethanol sensitivities. Germination indices from (b) were expressed relative to indices without ethanol. nic2-1 seeds were more sensitive to ethanol at all concentrations (0.1%, P < 0.05; 0.25%, P < 0.02; 0.5%, P < 0.03). (d) Effect of reduced and oxidized pyridine nucleotides on Ws and nic2-1 germination. Seeds were sown as in (b) but with final concentrations of 1 mm NAD, NADP, NADPH or NADH added to the medium. Germination indices were measured 5 days after transfer to 22°C. Points are the mean of three independent experiments from pools of six plants. Bars are SE.

Seed germination is inhibited by NAD but not NADH

As nicotinamide inhibits germination (Figure 4g,h), and is enzymatically derived from NAD, we examined the effect of adding NAD or NADH to the germination medium. NAD (1 mm) and NADP (1 mm) significantly (P < 0.02 and P < 0.05, respectively) inhibited germination of moist-chilled Ws seed by approximately 40%, and a similar inhibition was seen for nic2-1 (Figure 5d). Seedlings that did germinate on NAD had inhibited primary root growth. Treatment with 1 mm NADH promoted a small but significant increase in germination (P < 0.05) of nic2-1 seeds (Figure 5d).

nic2-1 seeds have reduced poly(ADP)ribose (PAR) levels

The decrease in nicotinamidase activity in nic2-1 might be expected to increase nicotinamide levels, which in turn could inhibit the activity of the nicotinamide-producing enzymes PARP, ribosyl cyclase and sirtuin. However, as NAD is a substrate for these enzymes and we found nic2-1 plants to have enhanced NAD levels, it is also possible that activity might be stimulated. We investigated whether altered activity of these nicotinamide-producing enzymes might be affecting the germination potential in nic2-1.

As NIC2 and PARP3 showed similar expression patterns with increased levels in maturing seed (Figure 2d), we investigated whether they might act in the same pathway to regulate germination. Dot blotting with an anti-PAR antibody indicated that freshly harvested nic2-1 seeds had a decreased level of poly(ADP)ribosylation compared to Ws seed (P < 0.01), although there were no significant differences in PAR levels after moist chilling of fresh seeds (Figure 6a). PARP activity is induced by oxidative and genotoxic stress (Amor et al., 1998; Doucet-Chabeaud et al., 2001), so we investigated the effect on these factors on the germination potential of nic2-1. Germination in the presence of methyl methanesulphonate (MMS), a DNA-damaging agent showed germination of nic2-1 seeds to be hypersensitive to MMS. Only 25% of nic2-1 seed germinated on 0.005% MMS compared to 95% of Ws seed, relative to those that were untreated (Figure 6b,c). These results indicate that PARP activity and DNA repair responses are impaired in the nic2-1 seed, and that this may reduce their germination potential.

Figure 6.  nic2-1 seed have reduced levels of poly(ADP)ribosylation, and germination is hypersensitive to MMS. (a) Analysis of poly(ADP)ribose (PAR) levels in dry and moist-chilled Ws and nic2-1 seeds. Nuclear extracts (1 μg) from dry and stratified seed were analysed by dot blot analysis with anti-PAR antibody. Mean spot intensities from four independent extractions from seed pooled from three plants are expressed relative to dry nic2-1, which had the lowest average levels of PAR. Bars are SE. The lower panel shows a representative dot blot. (b) Effect of MMS on germination. Seed of Ws and nic2-1 were sown on medium containing 0, 0.005 or 0.01% v/v MMS, moist-chilled, then transferred to the light at 22°C. Points are the mean of three independent experiments from pools of three plants. Germination indices were calculated from six measurements over 3 days. Bars are SE. (c) Comparison of MMS sensitivities. Germination indices from (b) were expressed relative to indices without MMS.

Ribosyl cyclase activity does not appear to regulate germination

As no ribosyl cyclase gene has been identified in plants, it was not possible to investigate whether this gene is expressed during germination. Treatment of seed with either the substrate or product of this reaction, ADP-ribose or cyclic ADP-ribose (cADPR), added to the germination medium had no effect on the germination potential of Ws or nic2-1 (Figure S4).

Discussion

A null mutation in a seed-expressed nicotinamidase causes reduced germination potential and increased ABA sensitivity for germination

We have identified two Arabidopsis proteins with in vitro nicotinamidase activity, NIC2 and NIC3, and characterized nic2-1 plants with a mutation in the NIC2 gene. Another Arabidopsis protein with nicotinamidase activity, NIC1, has recently been reported (Wang and Pichersky, 2007). Although nicotinamide has been shown to inhibit stomatal closure and tracheary cell development (Leckie et al., 1998; Sugiyama et al., 1995), we observed no difference in the ABA-induced stomatal closure response or in xylem vessel formation in the nic2-1 mutant plants (data not shown). This is perhaps because the NIC2 gene is predominantly expressed in seed, during late embryogenesis, and it is likely that other proteins with nicotinamidase activity are active in other tissues. Mutant nic2-1 seeds showed reduced nicotinamidase activity and a seed germination-related phenotype, with delayed germination, reduced germination potential, delayed after-ripening and ABA hypersensitivity. The effect of the nic2-1 mutation may be a result of impaired function during embryo maturation or during seed imbibition. Indeed, the expression pattern of NIC2 in testa and endosperm (Figure 2c,e) and the altered colour of the nic2-1 testa (Figure S3) suggest that the impaired germination potential of nic2-1 maybe due, at least in part, to a maternal effect. However, other pathways regulating germination still operate normally in nic2-1, as germination was largely restored by moist chilling or post-harvest storage. These results suggest that the metabolism of nicotinamide is required in the breaking of seed dormancy and in enhancing the effectiveness of after-ripening.

As nic2-1 seed exhibited hypersensitivity to ABA, mechanisms for the inhibition of germination in nic2-1 could include an increase in ABA synthesis during imbibition or prevention of ABA catabolism, both of which would decrease germination potential (Okamoto et al., 2006). However, we found no difference in ABA levels in nic2-1 seed, suggesting that nicotinamidase activity does not affect ABA metabolism. Several other loci have been identified that influence the ABA sensitivity of germination potential in a similar way to NIC2, by reducing the effectiveness of ABA repression of germination. These encode proteins with a variety of different functions (e.g. ABH1, nuclear mRNA cap binding protein, Hugouvieux et al., 2001; SAD1, Sm-like snRNP, Xiong et al., 2001; AtLPP2, lipid phosphate phosphatase, Katagiri et al., 2005), suggesting that ABA sensitivity is regulated by a number of different cellular processes during germination. Our results add another process to this list, implicating nicotinamide metabolism as an important function in ABA-related repression of germination. The reduced germination potential of freshly harvested nic2-1 seeds may be the result of enhanced sensitivity to endogenous ABA, which could also account for nic2-1 seeds taking longer to after-ripen. However, the eventual loss of the nic2-1 phenotype from stored after-ripened seeds suggests that removal of nicotinamide is associated with the repression of ABA effects in imbibed seeds, rather than with post-ABA-affected germination pathways. This is also evident through the observation that moist chilling of fresh seeds (which has been shown to induce gibberellin synthesis, Yamauchi et al., 2004) removed the block to nic2-1 germination in the absence of exogenous ABA.

nic2-1 seeds have increased levels of NAD

The reduced levels of nicotinamidase in nic2-1 were shown to increase the NAD concentration in seed, thereby potentially changing the NAD/NADH ratio in dry seeds and perhaps disrupting the biochemical processes required to break dormancy. In our experiments, application of NAD or nicotinamide inhibited germination, whereas application of NADH did not (Figure 5c). The strong effect of exogenous NAD that we observed on germination of stratified seed suggests an inverse relationship between NAD levels and germination potential (Figure 5c). A report that NADP phosphatase activity, which converts NADP to NAD, is higher in dormant than non-dormant barley seed supports this hypothesis (Gallais et al., 2000). There are two possible explanations for the elevated NAD levels in nic2-1 seeds. The block in conversion of nicotinamide to nicotinic acid in nic2-1 seeds could stimulate de novo NAD synthesis (Katoh et al., 2006), or alternatively cause feedback inhibition of nicotinamide-generating enzymes, thereby inhibiting NAD breakdown. nic2-1 seeds, which have reduced ability to recycle nicotinamide, would therefore be more sensitive to applied NAD or nicotinamide, as we observed (Figures 4f and 5c).

The reduced germination potential of nic2-1 seeds is associated with reduced PARP activity

There is evidence for three types of NAD-degrading and nicotinamide-producing enzymes in plants that could be inhibited by nicotinamide or stimulated by NAD, and we therefore investigated two of these enzymes to determine whether they are involved in the regulation of seed germination potential. PARPs attach branched chain polymers of ADP-ribose to proteins and are believed to be involved in the plant DNA repair response (Amor et al., 1998; Doucet-Chabeaud et al., 2001;De Block et al., 2005). Ribosyl cyclases catalyse production of the second messenger cADPR, which in plant affects ABA-induced alterations in stomatal aperture (Leckie et al., 1998) and induction of ABA-inducible genes (Wu et al., 1997). Sirtuins, or Sir2-like proteins, catalyse an NAD-dependent histone deacetylase (Blander and Guarente, 2004) or mono-ADP-ribosyltransferase reaction (Haigis et al., 2006; Liszt et al., 2005), although their role in plants has not been well studied and was beyond the scope of this study. Our results did not provide any evidence of a role for ribosyl cyclase in regulating germination potential, but did suggest PARP as a potential target for regulation by nicotinamide.

The correlation between NIC2 and PARP3 expression patterns during seed maturation (Figure 2d) and the reduced PAR levels in dry nic2-1 seeds (Figure 6a) suggest that NIC2 and PARP3 could act in the same pathway to regulate germination potential. In the model presented in Figure 7, we illustrate how impaired NIC2 activity may affect seed germination potential via regulation of PARP3 activity and increasing NAD levels. In animals, PARPs are involved in DNA repair and cell-cycle checkpoints (Yu et al., 2003), and can interfere with G₁/S progress (Augustin et al., 2003). The delay in germination (in addition to reduced germination potential) in nic2-1 seeds (Figure 4b,d,e) may potentially be due to a less efficient DNA damage checkpoint capability. This is supported by the germination delay that we observed in the nic2-1 seed in the presence of the DNA-damaging agent MMS (Figure 6b,c). It is well known that imbibed seeds undergo a period of DNA repair during the early stages of phase 2 of germination in both dormant and after-ripened seeds, and this is a prerequisite to cell division that occurs in association with germination (Bewley, 1997). Although cell division may not be required for germination, seeds take several hours after imbibition to move out of G₁, and this is most likely to be a period of DNA repair (Vazquez-Ramos and Sanchez, 2003) before progress into S phase, which occurs before radicle emergence (Barroco et al., 2005). Contrary to previous suggestions that cell-cycle resumption occurs after radicle emergence, mitosis has recently been detected before radicle emergence, and mutations in D-type cyclins have been shown to delay germination in Arabidopsis (Masubelele et al., 2005). We suggest that PARP activity may be required for DNA repair to be accomplished before progress into S phase can proceed, and that increased levels of nicotinamide may reduce PARP activity and delay germination as seen in nic2-1 seed.

Figure 7.  Proposed model to account for the increased levels of NAD and decreased levels of poly(ADP)ribosylation observed in the nic2-1 mutant seed. Putative pathway for nicotinamide synthesis and degradation in wild-type (a) and nic2-1 (b) seed. The bar represents the proposed inhibition of PARP, and the double bar indicates the block in the pathway at NIC2. Bold text and arrows represent increased and decreased levels of NAD and PAR, respectively. NAM, nicotinamide; NA, nicotinic acid.

In summary, we propose that NAD metabolism is important during germination, and that nicotinamidase activity in seed may normally promote germination by a number of routes including altering NAD levels, reducing ABA sensitivity, and releasing inhibition of PARP activity. Further work will be required to define how NIC2 activity influences the enhancement of germination potential, and to define how this enhancement is related to ABA repression of germination.

Experimental procedures

Plant material

Arabidopsis nic2-1 seed was obtained from INRA Versailles, France (seed stock FLAG415C12), and Ws-4 seed was obtained from the Nottingham Arabidopsis Stock Centre (NASC), UK. Plants for seed collection were grown in compost with a 16 h photoperiod, day temperature 25°C, night temperature 22°C, relative humidity 60%.

Heterologous expression of nicotinamidases in E. coli

Arabidopsis cDNAs in vector pUni51 (Yamada et al., 2003) were obtained from the Arabidopsis Biological Resource Centre (ABRC) and recombined into PrSETE (Invitrogen; http://www.invitrogen.com/) for NIC2 and NIC3, or PHB3-His6 (ABRC stock CD3-595) for At3g16190 using Glutathione S-transferase-Cre recombinase (Liu et al., 1998). Plasmids were transformed into E. coli BL21 plysS grown to an OD₆₀₀ of 0.6, and induced using 0.5 mm IPTG. Cells were harvested after 20 h at 20°C. Fusion proteins were purified by nickel-based affinity chromatography.

Nicotinamidase assays

Purified protein (10 μg) was incubated overnight at 30°C in the presence of 10 mm nicotinamide, 50 mm Tris, pH 8, 1 mm MgCl₂ and 1 mm DTT, separated by an HPLC AKTA purifier (Amersham-Pharmacia; http://www5.amershambiosciences.com/) by ion exchange (SAX 5 mm, 250 mm, Alltech, http://www.discoverysciences.com/gb) and reaction products were detected at 280 nm. Elution was performed in buffer A (25 mm NaH₂PO₄/Na₂HPO₄/acetic acid, pH 4.5) for 8 min, followed by a 20 min gradient to 70% buffer B (25 mm sodium phosphate, pH 7.4, 2 m NaCl) at a flow rate of 0.5 ml min⁻¹. Retention times, used to identify reaction products, were 4.7 and 10.5 min for nicotinamide and nicotinic acid standards, respectively.

To measure activity in seed extracts, 20 mg of dry or imbibed seeds were ground in liquid nitrogen and allowed to thaw in 400 μl Cellytic (Sigma; http://www.sigmaaldrich.com/). Samples were centrifuged at 13 000 g to pellet insoluble matter, and 10 μl of supernatant was incubated a final volume of 20 μl, containing final concentrations of 50 mm Tris, pH 8, 1 mm DTT, 1 mm MgCl₂ and 1 μCi of carbonyl-¹⁴C-nicotinamide (American Radiolabelled Chemicals, http://www.arcincusa.com/arc) for 30 min at 22°C. A 2 μl aliquot was added to microcellulose TLC plates F₂₅₄ (Sigma), and resolved using a solvent ratio of 60:20:20:1 butanol:methanol:water:ammonia (Zheng and Ashihara, 2004). Plates were dried and analysed by phosphoimaging (InstantImager, Canberra-Packard Ltd., Berks, UK). Spots were quantified by densitometry using scion image (http://www.scioncorp.com).

To measure the effect of metal ions on activity, an ammonia release assay was used. Samples contained, in a final volume of 200 μl, 50 mm Tris, pH 8, 10 mm NADH, 10 mm ketoglutarate and 10 mm nicotinamide. The A₃₄₀ was measured, and the reaction was started by the addition of purified protein (10 μg). After 30 min incubation at 30°C, 2 units of glutamate dehydrogenase (Roche, http://www.IbuyBiochem.co.uk) were added, and the decrease in A₃₄₀ was measured after 5 min.

Germination assays

Seeds were freshly harvested and removed from siliques immediately after pod dehiscence. For all germination assays, freshly collected seeds were surface-sterilized and sown on 0.6% Phytagar (Invitrogen) dissolved in deionized water, in rows of 30–100. Stock solutions of inhibitors added to the agar solution were prepared in deionized water except for ABA, which was dissolved in methanol. Moist chilling (stratification) was carried out at 4°C for 4 days in the dark. Germination assays were carried out in a growth chamber with an 8 h photoperiod, with day/night temperatures of 22/18°C. For experiments measuring both dose dependency and time dependency of germination, values were expressed as a germination index (GI; Reddy et al., 1985), which gives greater weight to seeds germinating early in the time course. This was calculated using the formula GI = (72 × n₁ + 48 × n₂… 12 × n₅…) × 100)/(total hours in time course × total number of seeds), where n₁, n₂, …n₅ are the numbers of seeds that germinated at the first, second and subsequent measurements multiplied by the weighting given to the number of seeds germinated at the first, second and subsequent measurements, respectively. Statistical significance was calculated using t-tests for pairwise comparisons, or anova for experiments where the number of biological samples was not identical for Ws and nic2-1.

Gene expression analysis

Seed were surface-sterilized and grown on 0.5 × MS, 15 g l⁻¹ sucrose, 0.6% phytagar plates. RNA was extracted using an RNeasy kit (Qiagen; http://www.qiagen.com/) and DNase-treated (Ambion, http://www.ambion.com). Dry seed RNA was extracted as described by Haslekas et al. (2003). cDNAs were synthesized using Superscript II (Invitrogen), and semi-quantitative RT-PCR was performed in an ABI2400 thermal cycler (Applied Biosystems; http://www.appliedbiosystems.com/) . For each primer pair, the number of temperature cycles to achieve non-limiting reaction conditions was determined by sampling every five cycles. Primers for RT-PCR had annealing temperatures of 60°C unless stated: NIC2 F1 5′-TACGAGACACGAAAGCGAGA-3′, R1 5′-ACGAGATAAGCGAAGCCAAA-3′ (67°C); RUB1 (At4g36800; Seki et al., 2002) 5′-GCGAACTTCGTCTTCACAAA-3′, 5′-GGAAAAAGGTCTGACCGACA-3′.

A 2 kb upstream DNA fragment of the NIC2 promoter was amplified from Ws genomic DNA using the primers 5′-CACCATGATCCTCAGCCTTAAA-3′ and 5′-CGGTGGAGATATATTTGGGAA-3′) and Pfu polymerase (Promega; http://www.promega.com/). This was recombined into pDONR/D/TOPO (Invitrogen), and then recombined upstream of the β-glucoronidase gene in pGWB3 binary vector (kindly provided by Dr T. Nakagawa, Shimane University, Matsue-Shi, Japan) using LR clonase plus (Invitrogen), transformed into the Agrobacterium C58 strain, and then into Arabidopsis (Col-0) by floral dipping (Clough and Bent, 1998). Transformants were selected with 50 mg l⁻¹ kanamycin, 50 mg l⁻¹ hygromycin. Seeds of primary transformants were imbibed for 3 h and dissected into embryo and endosperm before vacuum infiltration and staining with 100 mm potassium phosphate, 0.5 mm potassium ferrocyanide, 0.5 mm potassium ferricyanide, 0.1% Triton X-100, 2 mm 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid, 10 mm EDTA, overnight at 37°C. Several independent lines showed similar staining patterns.

Characterization of nic2-1

Confirmation of the presence of a T-DNA insertion within the NIC2 gene was performed using left border primer (5′-CAACCCTCAACTGAAACGGGCCGGA-3′) and gene-specific primer (Lb2 5’-GGTGGAACGGCGATCTAATCC-3′). Homozygous lines were detected using the insertion spanning primers F1 (as above) and R2 5′-TGTCGCGATGTAATCGGTCTTT-3′. RT-PCR with insertion-spanning primers confirmed the absence of a full-length transcript. A reverse primer before the insertion site (R3 5′-CGATCTTGTCCAGCTTCTCC-3′) was designed to detect any truncated transcript. Insertion position was verified by DNA sequencing. To genetically complement nic2-1, a 4 kb genomic fragment containing NIC2 was amplified from Ws genomic DNA using primers.

5′-AAACGAACCAATTCCAAGTGA-3′ and 5′-CGTTCTCGATGATGAAGGATCCATT-3′ and Pfu polymerase. This was then adenylated using Taq polymerase, and ligated into pGemT Easy (Promega). The plasmid was digested with EcoRI, and partially digested with SalI, then ligated into the EcoRI/XhoI sites of pCAMBIA1300 and transformed by the floral dip method (Clough and Bent). Transformants were selected by germination on 0.5 × MS, 15 g l⁻¹ sucrose plates containing 15 mg l⁻¹ hygromycin.

Pyridine nucleotide extraction

Seeds (20 mg) were homogenized with 1 ml 0.1 n HCl for NAD/NADP extraction, and extracts were neutralized and quantified as described previously (Bernofsky and Swan, 1973). Seed mass was estimated by counting 200–400 seeds, and the mean of three different counts was calculated.

Protein extraction

Seedlings were ground in liquid nitrogen and extracted in 1 ml per 500 mg tissue of 50 mm Tris, 250 mm sucrose, plus protease inhibitor cocktail (Roche). Extracts were centrifuged at 3000 g, and the supernatant was removed to a fresh tube. Nuclear proteins were isolated as described previously (Waterborg et al., 1987). The pellet was resuspended in 200 μl 1X SDS–PAGE loading buffer, heated at 95°C for 5 min, and then centrifuged at 13 000 g for 5 min. Protein concentrations were determined by the TCA-Lowry (Peterson, 1977) method using BSA as a standard.

PAR immunoassay

Aliquots (0.5 μl) of 2 mg ml⁻¹ seed nuclear protein extracts were applied to nitrocellulose, allowed to dry, blocked with 5% milk and probed with a 1:1000 dilution of anti-PAR (Affiniti-Res, http://www.affiniti-res.com), washed with Tris-buffered saline, 0.01% v/v Tween-20, probed with a 1:5000 solution of alkaline phosphatase-linked secondary antibody for 1 h, washed with 100 mm Tris-HCl, pH 9.5, 5 mm, MgCl₂, 100 mm NaCl, and developed through the addition of NBT/BCIP. Filters were photographed, and densitometry was performed using scion image software.