AtNAP, a NAC family transcription factor, has an important role in leaf senescence


*(fax: +1 607 255 0599; e-mail:


Leaf senescence is a unique developmental process that is characterized by massive programmed cell death and nutrient recycling. The underlying molecular regulatory mechanisms are not well understood. Here we report the functional analysis of AtNAP, a gene encoding a NAC family transcription factor. Expression of this gene is closely associated with the senescence process of Arabidopsis rosette leaves. Leaf senescence in two T-DNA insertion lines of this gene is significantly delayed. The T-DNA knockout plants are otherwise normal. The mutant phenotype can be restored to wild-type by the intact AtNAP, as well as by its homologs in rice and kidney bean plants that are also upregulated during leaf senescence. Furthermore, inducible overexpression of AtNAP causes precocious senescence. These data strongly suggest that AtNAP and its homologs play an important role in leaf senescence in Arabidopsis and possibly in other plant species.


The transition from a functional photosynthetic organ to an actively degenerating and nutrient-recycling tissue in a leaf's life history represents the onset of leaf senescence. This onset is a developmental switch that involves dramatic differential gene expression. Differential gene expression is believed to play an important role in leaf senescence. In a senescing leaf, many genes that are expressed in green leaves, including those genes involved in photosynthesis, are downregulated, while a subset of genes, generally referred to as senescence-associated genes (SAGs), are upregulated. Leaf senescence is under direct nuclear control, and SAG expression is required for senescence to proceed. Inhibitors of transcription or translation prevent leaves from senescing (Buchanan-Wollaston et al., 2003; Guo and Gan, 2005; Hadfield and Bennett, 1997; Lim and Nam, 2005; Smart, 1994). For the past decade, much effort has been made to isolate SAGs, and hundreds of SAGs have been cloned from various plant species including Arabidopsis, barley, Brassica, maize, cucumber, rice, tobacco, radish, asparagus and soybean (Buchanan-Wollaston et al., 2003; Gepstein et al., 2003; He and Gan, 2003). Recent application of genomics approaches has led to the identification of thousands of potential SAGs (Andersson et al., 2004; Bhalerao et al., 2003; Buchanan-Wollaston et al., 2003, 2005; Guo et al., 2004; Lin and Wu, 2004; Zentgraf et al., 2004). Analysis of a leaf senescence EST database (dbEST) indicated that approximately 10% (approximately 2500) of the Arabidopsis genes are expressed in senescent leaves (Guo et al., 2004). Microarray analysis of the global gene expression changes during developmental leaf senescence in Arabidopsis has led to the identification of >800 genes that show a reproducible increase in transcript abundance (Buchanan-Wollaston et al., 2005).

Changes in gene expression are often regulated by transcription factors that bind to specific cis elements of target gene promoters, resulting in the activation and/or suppression of the target genes. There are approximately 1500 transcription factor genes in the Arabidopsis genome that belong to >30 gene families based on their DNA-binding domains (Riechmann et al., 2000). Microarray analysis has identified 96 transcription factor genes with at least a threefold upregulation during leaf senescence (Buchanan-Wollaston et al., 2005), and analysis of the leaf senescence dbEST revealed 134 unique genes that encode transcription factors representing 20 different gene families (Guo et al., 2004). Among the largest transcription factor groups are NAC, WRKY, C2H2 type zinc finger, AP2/EREBP, and MYB proteins (Buchanan-Wollaston et al., 2005; Chen et al., 2002; Guo et al., 2004; Lin and Wu, 2004). Two WRKY transcription factor genes have been studied: WRKY53 plays an important role in controlling leaf senescence (Hinderhofer and Zentgraf, 2001; Miao et al., 2004; Robatzek and Somssich, 2002), while suppression of WRKY6 expression has little effect on either the onset or the progression of leaf senescence (Hinderhofer and Zentgraf, 2001; Miao et al., 2004; Robatzek and Somssich, 2002). The potential functions of the majority of the leaf senescence-associated transcription factors remain to be elucidated.

A total of 20 genes encoding NAC transcription factors are in the leaf senescence dbEST (Guo et al., 2004), representing almost one-fifth of all the predicted 109 members of the NAC superfamily in Arabidopsis (Riechmann et al., 2000). The NAC domain was originally defined by the highly conserved N-termini of the petunia NAM (NO APICAL MERISTEM) and Arabidopsis ATAF1 and CUC2 (CUP-SHAPED COTYLEDON2) genes. It exists widely in plants but not in other eukaryotes. Roles of the NAC family genes include embryo and shoot meristem development, lateral root formation, auxin signaling, defense and abiotic stress response (Olsen et al., 2005). Expression of the NAC family genes in senescing leaves has been reported by several groups (Andersson et al., 2004; Buchanan-Wollaston et al., 2005; Guo et al., 2004; John et al., 1997; Lin and Wu, 2004), but whether these genes play a part in leaf senescence is unknown.

Here we report the functional analysis of AtNAP, one of the NAC family transcription factor genes whose expression is associated with leaf senescence. Two T-DNA insertion lines of this gene display a significantly delayed leaf senescence phenotype, which can be complemented by introducing a wild-type copy of AtNAP. In contrast, inducible overexpression of AtNAP readily causes precocious senescence. Furthermore, homologs of AtNAP in kidney bean and rice are also upregulated during leaf senescence, and like AtNAP, the kidney bean and rice homologs are able to rescue the Arabidopsis atnap null mutant phenotype.


AtNAP is upregulated during leaf senescence in Arabidopsis

Digital expression profile analysis of the Arabidopsis leaf senescence dbEST and microarray analysis (Buchanan-Wollaston et al., 2005) revealed that AtNAP is one of the most abundantly transcribed transcription factor genes in senescing leaves (Guo et al., 2004). RNA gel blot analysis showed that the expression of AtNAP in rosette leaves of Arabidopsis is closely associated with the progression of leaf senescence (Figure 1). When leaves from the same phyllotactical position (leaf number 6 from the bottom of the plant) were studied, RNA messenger (mRNA) of AtNAP was detected only when the leaves start senescing (2–3 weeks after emergence; Figure 1a). Leaf senescence in Arabidopsis grown under non-stressful conditions is age-dependent and progresses sequentially from the oldest leaf (the first leaf at the bottom of a plant) to the top young leaves. As shown in Figure 1(b), AtNAP transcript was detected in the old, senescing leaves, but not in the young, green leaves. In a given leaf, senescence starts from the leaf tip and progresses towards the leaf base (petiole). The yellow tip showed stronger AtNAP expression than the proximal part of a leaf (Figure 1c).

Figure 1.

 RNA gel blot analysis of AtNAP expression during leaf senescence in Arabidopsis.
(a) AtNAP expression in leaves at different developmental stages. YL, a young leaf with half the size of a fully expanded leaf; NS, a fully expanded, non-senescent leaf; ES, an early senescent leaf, with <25% leaf area yellowing; LS, a late senescent leaf, with >50% leaf area yellowing.
(b) AtNAP expression in leaves 1–12 of a 30-day-old plant. Leaves are counted from the bottom of the rosette.
(c) AtNAP expression in different part of a senescing leaf. B, base; M, middle; T, tip.
The 18S rRNA autoradiographs (a, c) and ethidium bromide-stained gel (b) indicate the relative amount of total RNA loaded in respective lanes.

AtNAP is targeted to nuclei

Although AtNAP is predicted to be a nuclear protein by PredictNLS (Cokol et al., 2000) and PSORT (Nakai and Kanehisa, 1992), it does not have any obvious nuclear localization signal. To determine the subcellular localization of AtNAP, a chimeric gene containing a GFP–AtNAP construct driven by the 35S promoter was transiently expressed in onion (Allium cepa) epidermal cells using particle bombardment. Subcellular localization of the GFP fusion protein was visualized with a fluorescence microscope. DAPI (4′,6′-diamidino 2-phenylindole) staining of DNA revealed GFP–AtNAP protein localization in the nuclei, suggesting that AtNAP is a nuclear protein (Figure 2).

Figure 2.

 Nuclear localization of GFP–AtNAP fusion proteins.
(a, b) Fluorescent images of the GFP–AtNAP fusion proteins expressed in living onion epidermal cells.
(c, d) DAPI (4′,6′-diamidino 2-phenylindole) staining of the same images to show the positions of the nuclei (indicated by arrows).

The AtNAP expression is knocked out in one T-DNA line and knocked down in another line

The AtNAP gene consists of three exons and encodes a protein with 268 amino acids (Figure 3a). We obtained two Salk T-DNA lines (Columbia background) from the Arabidopsis Biological Resource Center (ABRC) at Ohio State University (Alonso et al., 2003). Line 1 (SALK_005010) has a T-DNA insertion in the second exon, and line 2 (SALK_004077) has a T-DNA insertion in the promoter region (at -227 from the translation start site; Figure 3a). RNA gel blot analysis showed that the AtNAP transcript in senescent leaves of the homozygous line 1 was not detectable, while AtNAP transcript levels in senescent leaves of line 2 plants were reduced to 5% of that in age-matched wild-type leaves (Figure 3b). This suggests that line 1 is an atnap null mutant while line 2 is a knockdown line.

Figure 3.

 Expression of AtNAP in two T-DNA insertion lines.
(a) Gene structure of AtNAP and locations of T-DNA inserts.
(b) RNA gel blot analysis of AtNAP expression in senescing leaves (approximately 50% yellowing) of wild-type (WT), line 1 and line 2 plants.

Leaf senescence is significantly delayed in the atnap null mutant plants

To compare any phenotypic changes in growth and development among line 1 (the atnap null mutant), line 2 (the atnap knockdown mutant) and wild-type (Columbia accession), these plants were grown side by side in an Arabidopsis growth chamber. There were no visible differences in growth and development, except for the significantly delayed leaf senescence phenotype in the atnap null plants, and the less significantly retarded leaf senescence phenotype in the atnap knockdown plants (Figure 4).

Figure 4.

 Delayed leaf senescence phenotype of the T-DNA insertion lines compared with that of age-matched wild-type (WT) plants.
(a) Early stages of plant development in the null mutant (line 1) and WT plants.
(b, c) Senescence in the mutant lines and WT plants (note: the null plants are otherwise developmentally normal).
(d) Leaves excised from age-matched plants in (c). Leaves were numbered from bottom to top. Under our growth conditions, an adult Arabidopsis (accession Columbia) plant typically produces 12 rosette leaves.

We further characterized the atnap null mutant plants. As shown in the mortality curves in Figure 5(a), leaves of the atnap null mutant plants senesced later than those of wild-type plants. Consistent with a delayed visible yellowing phenotype (Figure 4), chlorophyll levels in individual rosette leaves of the null line were generally higher than in counterpart leaves of the age-matched wild-type plants (Figure 5b). The Fv/Fm ratios in individual leaves of the null line were also higher than in counterpart leaves of the age-matched wild-type plants (Figure 5c). The Fv/Fm ratio reflects the photochemical quantum efficiency of photosystem II, as well as the photoreduction efficiency of the primary electron-accepting plastoquinone of photosystem II. In contrast, ion leakage in individual leaves of the null plants was less than that in wild-type plants (Figure 5d). Ion leakage is an indicator of the intactness of the plasma membrane. The plasma membrane of a senescing cell becomes fragile and leaky.

Figure 5.

 Physiological and molecular analyses of atnap null mutant plants (line 1).
(a) Leaf survival curves (combination of leaves 9 and 10) of wild-type (WT, n = 27) and line 1 (n = 22).
(b–d) Chlorophyll content (b), Fv/Fm ratio (c) and ion leakage (d) in individual rosette leaves of age-matched WT and line 1 plants.
(e) RNA gel blot analysis of SAG12 and RBCS in the 12 rosette leaves of age-matched WT and line 1 plants.

We also monitored the expression of SAG12 and the Rubisco small subunit gene (RBCS). SAG12 is a highly senescence-specific gene in Arabidopsis and has been widely used as a molecular marker for leaf senescence, while RBCS is a typical senescence downregulated gene. As shown in Figure 5(e), the expression of SAG12 was readily detectable in leaf number 7 of a 30-day-old wild-type plant, but it was barely detectable in leaf number 4 of an age-matched null plant.

All the data described above indicate that the leaf senescence process was dramatically delayed in the atnap mutant plants (approximately 10 days).

AtNAP restores the atnap null mutant plants to wild type

To confirm that the T-DNA insertional null mutation in AtNAP is responsible for the delayed senescence phenotype, we performed a complementation test experiment. The wild-type copy of AtNAP, including the 2 kb promoter region, was introduced into the atnap null mutant plants. The introduced AtNAP was expressed in senescing leaves (see ‘nap + AtNAP’ panel in Figure 6a). We characterized the senescence phenotype in leaves that are either detached or in planta. The leaves detached from the AtNAP complemented lines senesced in the same manner as the leaves from wild-type did, both phenotypically (Figure 6b) and in terms of the Fv/Fm ratio (Figure 6c). In planta leaves of the complemented plants also senesced in the same manner as wild-type leaves (Figure 6d). These data confirmed that loss of AtNAP expression in the atnap null mutant was the only cause of the delayed senescence phenotype.

Figure 6.

 Complementation of Arabidopsis atnap null plants with AtNAP, OsNAP (rice) and PvNAP (kidney bean).
(a) RT-PCR analysis of expression of AtNAP (left lanes), OsNAP (middle lanes) and PvNAP (right lanes) in wild-type (WT), atnap null mutant, and atnap null mutant transformed with AtNAP, OsNAP or PvNAP. 18s rRNA serves as an internal standard of equal loading.
(b) Phenotype of detached leaves of WT, null mutant, and various complementation lines. The leaves were kept in darkness for 4 days.
(c) Fv/Fm ratios of leaves shown in (b).
(d) Leaf senescence in intact plants of WT, null mutant and various complementation lines. The plants were grown side by side in an Arabidopsis growth chamber.

Inducible overexpression of AtNAP causes precocious senescence

We further investigated the role of AtNAP in leaf senescence by performing gain-of-function analysis. Considering the fact that constitutive expression of this gene might be lethal, the chemical inducible gene expression system (Aoyama and Chua, 1997) was used. First, we generated transgenic lines that harbored either the pTA7001 or pGL1167 constructs (Figure 7a). pTA7001 contains the chimeric transcription factor GVG consisting of a DNA-binding domain of the yeast transcription factor GAL4, a transactivation domain of the herpes simplex virus transcriptional regulatory protein VP16 and a glucocorticoid receptor domain (Figure 7a), while pGL1167 is a construct in which AtNAP is driven by a promoter containing six tandem copies of the GAL4 upstream activation sequence (Figure 7a). Treatment with dexamethasone (DEX, a synthetic glucocorticoid) caused precocious leaf yellowing (Figure 7b) and a significant reduction of the Fv/Fm ratio (Figure 7c) in F1 plants (pGL1167 homozygous plants × pTA7001 homozygous plants) but not in controls (wild-type, plants containing pGL1167 or pTA7001 only). RNA blot analysis showed that AtNAP expression was strongly induced in the F1 plants but not in the controls (Figure 7d). The precocious leaf yellowing was a senescence process because SAG12 and SAG13 were both expressed (Figure 7d). SAG12 and SAG13 are leaf senescence-specific marker genes. These data suggest that AtNAP is sufficient to promote leaf senescence.

Figure 7.

 Inducible overexpression of AtNAP causes precocious senescence.
(a) The modified glucocorticoid-inducible gene expression system consisting of pTA7001 and pGL1167. pTA7001 provides the recombinant transcription factor GVG (GAL4 binding domain + VP16 activation domain + GR or glucocorticoid receptor), and pGL1167 contains the GAL4 cis elements and the AtNAP coding region.
(b) Phenotypes of wild-type (WT) and transgenic plants harboring different constructs. The picture was taken 4 days after treatment with 30 μm DEX inducer.
(c) Fv/Fm ratios of leaves from different plants that were treated with or without DEX.
(d) RNA gel blot analysis of the expression of AtNAP, SAG12, SAG13 and RBCS in leaves of plants that were treated with or without DEX. C, control, no treatment; D, DEX treatment.

AtNAP homologs in rice and kidney bean are specifically expressed in senescing leaves

The amino acid sequence of AtNAP was analyzed using various genomic databases such as the GenBank (, the TIGR plant genome databases (, and PopulusDB (, and genes with high sequence similarity were identified from many different plant species (Figure 8a). Among them are the NAC family transcription factor PvNAP (256 amino acids; AAK84884) from the dicot kidney bean (Phaseolus vulgaris) with 66% identity, and the NAC family transcription factor OsNAP (392 amino acids; NP_912423) from the monocot rice (Oryza sativa) with 70% identity (Figure 8b). We hypothesized that those homologs are functional orthologs of AtNAP. To test this hypothesis, first we examined whether PvNAP and OsNAP shared the same leaf senescence-specific expression pattern as AtNAP. Kidney bean leaves at five distinct developmental stages, ranging from young leaves to entirely yellow leaves (Figure 8c), were used for RNA gel blot analysis of PvNAP expression. As shown in Figure 8c, PvNAP transcript was detected in senescing leaves only. The expression of OsNAP was also shown to be senescence-specific in rice leaves (Figure 8d).

Figure 8.

 Homologs of AtNAP in kidney bean (Phaseolus vulgaris) and rice (Oryza sativa japonica cultivar group) and their senescence-specific expression patterns.
(a) A phylogenetic tree of NAP proteins from different plant species.
(b) Alignment of amino acid sequences of NAP proteins from Arabidopsis, kidney bean and rice.
(c) Expression of PvNAP in senescing leaves of kidney bean.
(d) Expression of OsNAP in senescing leaves of rice. Y, young leaf; S, senescing leaf.

AtNAP homologs in rice and kidney bean are able to restore the Arabidopsis atnap null mutant to wild-type

To further test the hypothesis that the homologs are functional orthologs of the Arabidopsis AtNAP, we performed heterogeneous complementation tests. The 2 kb AtNAP promoter was used to direct the expression of the coding region of OsNAP or PvNAP. These genes were expressed in senescent leaves of respective complementation lines (Figure 6a). Phenotypically, fully expanded non-senescing leaves detached from wild-type plants became senescent after being incubated in darkness for 4 days. In contrast, age-matched leaves from the atnap null mutant (line 1) remained green (Figure 6b) and photosynthetically active (Figure 6c). However, the leaves of the null plants complemented with OsNAP or PvNAP senesced like those of wild-type (Figure 6b). Similar observations were obtained when natural leaf senescence was examined in intact plants (Figure 6d). Similarly to AtNAP, OsNAP and PvNAP were able to restore the Arabidopsis null mutant to wild-type, which suggests that OsNAP and PvNAP are functional orthologs of AtNAP.


Various molecular, genetic and genomic strategies have been used to isolate genes that are differentially expressed during senescence, and, as a result, thousands of SAGs have been identified. The structure and function of most SAGs have been predicted bioinformatically. There are only a few of those genes whose enzymatic activities have been shown biochemically, including several RNases (Lers et al., 1998), a phospholipase D (Fan et al., 1997) and an acyl hydrolase (He and Gan, 2002). Similarly, there are only a few genes whose role in leaf senescence has been investigated genetically. For example, the ABA-promoted senescence in detached leaves of PLDα-antisense Arabidopsis plants was delayed (Fan et al., 1997). The Arabidopsis F-box gene ORE9 has also been shown to play a role in leaf senescence because the ore9 mutant plants displayed increased leaf longevity (Woo et al., 2001). It has previously been shown that SAG101, a gene encoding an acyl hydrolase, plays a significant role in leaf senescence in Arabidopsis; leaf senescence is delayed for 4–5 days in the SAG101 antisense plants (He and Gan, 2002). In this report, we show that leaf senescence in the atnap null mutant line is delayed for up to 10 days (Figures 4 and 5). The null phenotype can be rescued by the wild-type AtNAP, confirming that the lack of AtNAP is responsible for the retardation in leaf senescence in the null mutant plants. The role of AtNAP in controlling leaf senescence is further confirmed by our gain-of-function analysis. Young leaves begin yellowing as early as 2 days after the initial induction of the AtNAP overexpression, and become completely senescent 4 days after the initial induction (Figure 7), suggesting that AtNAP is sufficient to cause senescence.

Our RNA gel blot analysis revealed that AtNAP is expressed in senescing leaf cells (Figure 1). Microarray data available at Genevestigator (Zimmermann et al., 2004) also show that AtNAP is mainly expressed in senescent rosette leaves, cauline leaves, sepals and petals (Figure 9a), although very low levels of AtNAP expression in young seedlings and other parts of adult plants have been detected (Figure 9b). The AtNAP expression can be strongly induced by inducers of programmed cell death (PCD) in cell suspension, a process similar to leaf senescence. In contrast, other senescence-promoting factors such as ethylene and ABA, osmotic and salt stress only moderately induced the expression of AtNAP, while darkness, drought, oxidative stress, jasmonic acid and salicylic acid did not have significant effects on AtNAP expression levels (Figure 9b). In the absence of external stressors, initiation of leaf senescence is dependent on age and developmental stage (Hensel et al., 1993; Nooden and Penney, 2001). Our study and the microarray data suggest that AtNAP may be primarily upregulated by age.

Figure 9.

 Microarray analysis of the expression profile of AtNAP.
(a) Expression levels of AtNAP in different plant tissues. The highest value shown in ‘44 senescent leaf’ is 21 790 ± 391.
(b) Effect of various treatments on the AtNAP expression. The ratios of expression change (numbers in colored squares) and expression levels after different treatments are presented. The highest value shown in ‘PCD: senescence’ is 26 597 ± 1957.
The data were extracted from the Genevestigator microarray database (Zimmermann et al., 2004).

Although AtNAP is not readily induced by darkness (Buchanan-Wollaston et al., 2005; Lin and Wu, 2004), dark-induced senescence of detached leaves was delayed in the atnap null mutant (Figure 6b,c), suggesting that AtNAP may function in dark-induced senescence downstream of the dark-responsive signaling pathway. During natural senescence of leaves on intact plants, AtNAP is only expressed in leaf tissues that are already senescent (Figure 1). These data suggested that AtNAP is likely to play a key role in regulating the common execution process of leaf senescence downstream of various senescence-inducing pathways. Even though expression of AtNAP may not necessarily be responsive to a particular senescence-inducing factor, some of these factors may need AtNAP to trigger the senescence syndrome. As a transcription factor, AtNAP might control the leaf senescence process by transcriptionally activating/repressing genes involved in the execution of senescence.

Plant transcription factors of the same family often have similar functions. At some developmental stages or cellular processes, certain families of transcription factors may play predominant roles (Liu et al., 1999; Riechmann and Ratcliffe, 2000), such as the MADS box genes in flowering development (Saedler et al., 2001) and the WRKY genes in defense response (Ulker and Somssich, 2004). The senescence-associated expression pattern of >20 other NAC family members (Buchanan-Wollaston et al., 2005; Guo et al., 2004) suggests a general role of the NAC family genes in leaf senescence. The evidence of transcriptional self-regulation (Xie et al., 2000) and inter-regulation (Vroemen et al., 2003) between NAC members, as well as homodimerization (Ernst et al., 2004; Xie et al., 2000) and heterodimerization (Hegedus et al., 2003) among NAC proteins, suggest possible regulatory networks of leaf senescence involving many NAC genes.

Sequence homologs of AtNAP in kidney bean (a dicot) and rice (a monocot) also displayed a leaf senescence-specific expression pattern (Figure 8c,d). PvNAP and OsNAP were able to restore the Arabidopsis atnap null mutant to wild-type (Figure 6). In addition to rice and kidney bean, sequence homologs exist in a variety of other plant species including soybean (Glycine max), nightshade (Solanum demissum), Medicago truncatula, Populus trichocarpa, wheat (Triticum aestivum), maize (Zea mays), peach (Prunus persica), tomato (Lycopersicon esculentum), potato (Solanum tuberosum) and petunia (Petunia × hybrida; Figure 8a). These suggest that NAP may be a universal regulator in plant leaf senescence. It is likely that knocking NAP out in other plant species will cause a significant delay of leaf senescence, which may be a new strategy for manipulating leaf senescence in agriculturally important crops.

It should be noted that AtNAP was previously identified as an immediate target of the floral homeotic genes APETALA3/PISTILLATAL that are essential for petal and stamen formation (Sablowski and Meyerowitz, 1998). In AtNAP antisense lines, the first 2–4 flowers of the main and lateral inflorescences had short stamens and their anthers often did not dehisce. No leaf senescence phenotype was described (Sablowski and Meyerowitz, 1998). In this study we did not observe any developmental abnormities other than the delayed leaf senescence in our two T-DNA mutant lines. This discrepancy may be due to the different Arabidopsis ecotypes that were used: Landsberg erecta (Sablowski and Meyerowitz, 1998) versus Columbia in this study. Leaf senescence is a trait with great variations within Arabidopsis ecotypes, and the molecular regulation of leaf senescence may differ in different genetic backgrounds (Levey and Wingler, 2005). This discrepancy may also result from different research approaches employed; T-DNA insertion mutation and the antisense approach have shown completely different roles for the phytochrome interacting factor 3 (Kim et al., 2003; Ni et al., 1998).

Experimental procedures

Plant materials and growth conditions

Arabidopsis plants were grown at 23°C with 60% relative humidity under constant light (150 μmol m−2 sec−1 light from a mixture of fluorescent and incandescent bulbs). Seeds were sown on Petri dishes containing one-half strength of Murashige and Skoog salts, 0.8% w/v phytoagar (Sigma, St. Louis, MO, USA), and appropriate antibiotics. After imbibition, the seeds were kept at 4°C overnight. Two-week-old seedlings were transplanted to Cornell mix soils (3 parts peat moss: 2 parts vermiculite: 1 part perlite; Tower Road Green house, Cornell University, Ithaca, NY, USA).

Arabidopsis thaliana ecotype Columbia-0 was used. The T-DNA insertion lines, the inducible overexpression lines and the complementation lines were grown side by side with wild-type and other control lines unless indicated otherwise.

Isolation of T-DNA insertions within AtNAP

Two Arabidopsis lines for T-DNA insertions in AtNAP, SALK_005010 (line 1) and SALK_004077 (line 2) were obtained from the Salk T-DNA collection (Alonso et al., 2003). A PCR-based method was used to identify homozygous mutant plants. Genomic DNA was prepared from a small piece of leaf using a modified CTAB method (Murray and Thompson, 1980). Briefly, 50–100 mg fresh leaf tissue was ground in a 1.5 ml microcentrifuge tube with a Craftsman 9-inch drill press (Sears, Roebuck and Co., Hoffman Estates, IL, USA). The powdered samples were incubated at 55°C for 30 min after 500 μl 2X extraction buffer (0.7 m NaCl, 1% w/v CTAB, 50 mm Tris (pH 8.0), 10 mm EDTA, 1% beta-ME added fresh) had been added. After incubation, 500 μl chloroform:isoamyl alcohol 24:1 was added and mixed, and the samples were centrifuged for 10 min at 13 000 g. The aqueous phase (approximately 500 μl) was transferred to a new microfuge tube and 500 μl isopropanol was added to precipitate genomic DNA. PCR was used to amplify the genomic DNAs. The PCR conditions were as follows: 35 cycles with each cycle consisting of 94°C for 30 sec, 55°C for 30 sec and 72°C for 1 min. The T-DNA left border primer G1099 (5′-GCGTGGACCGCTTGCTGCAACT-3′) and gene-specific primers G1027 (5′-ATCATGGAAGTAACTTCCCAATC-3′) and G1028 (5′-TTCAGTTCTTCTCTCTGCTTC-3′) for line 1, G1273 (5′-GGCCATTTTCTACGCTACCT-3′) and G1123 (5′-CTTCCATGGTTTTCAGACAATTTAG -3′) for line 2 were used in the PCR reactions.

Plasmid construction

The GFP–AtNAP expression plasmid pGL1185 was generated by cloning AtNAP coding region into pRTL2-S65TGFP (Lin and Wang, 2004). The coding region without the stop codon was amplified via PCR reaction using primers G1526 (5′-TAGTCGACAGTTCCTGTTCTATTAGATTG-3′; the underlined section is an engineered SalI site) and G1527 (5′-TATCATGAACTTAAACATCGCTTGACG-3′; the underlined section is an engineered BspHI site). Pfu polymerase (Stratagene, La Jolla, CA, USA) was used and the PCR product was sequenced. The PCR product cut with SalI and BspHI was cloned into pRTL2-S65TGFP at the XhoI and NcoI sites.

For inducible overexpression of AtNAP, the 320 bp fragment of 6x GAL4 UAS and the 35S TATA region from pTA7001 (Aoyama and Chua, 1997) was cloned into a binary vector called pPZP211 (Hajdukiewicz et al., 1994) to form pGL1152. The full-length cDNA of AtNAP (907 bp, including 43 bp 5′-UTR region and 57 bp 3′-UTR region) was amplified with primers G1100 (5′-CACTAGTTCCTGTTCTATTAGATTG-3′; the underlined section is an engineered SpeI site) and G1101 (5′-GCTGCAGTAACTTTTCAAGCACATC-3′; the underlined section is an engineered PstI site) using Pfu polymerase. The PCR product, after an A-tailing procedure described by the manufacturer (Promega, Madison, WI, USA) was cloned into the pGEM-T vector (Promega) to form pGL1165. The plasmid was then sequenced. pGL1165 was digested with SpeI and PstI, and the released AtNAP coding region was subcloned into pGL1152, resulting in pGL1167.

For the complementation test involving the Arabidopsis wild-type AtNAP, a 3166 bp genomic DNA containing the promoter (1961 bp) and coding region (1205 bp) of AtNAP was PCR-amplified with primers G1628 (5′-GCGTCATCTCATCCTAATCCTCAT-3′) and G1629 (5′-CGTGACTTCGTCTTATCATGCTG-3′) using Pfu polymerase, and cloned into pGEM-T after A-tailing, to form pGL1186, which was subsequently sequenced. pGL1186 was digested with SacII, followed by treatment with T4 DNA polymerase (NEB, Beverly, MA, USA) to remove the 3′ overhangs to form blunt ends. The plasmid was further digested with SacI, and the released AtNAP was cloned into the binary vector pPZP221 (Hajdukiewicz et al., 1994) at the SacI and SmaI sites. The construct is named pGL1199.

When using the rice (Oryza sativa, japonica cultivar group) homolog of AtNAP (OsNAP) for complementation, primers G1807 (5′-TTCTGCAGCGTCATCTCATCCTAATCCTCAT -3′; the underlined section is an engineered PstI site) and G1808 (5′-GTTACTTCCATGGTTTTCAGACAATTTAG-3′; the underlined section is an engineered NcoI site) were used to PCR amplify the AtNAP promoter region. After an A-tailing procedure, the 2 kb PCR product was cloned into pGEM-T to form pGL1193. The genomic fragment containing the coding region of OsNAP (NP_912423) was PCR-amplified using primers G1805 (5′-TTCCATGGTTCTGTCGAACCCG-3′; the underlined section is an engineered NcoI site) and G1666 (5′-GATCTAGACGAAGAACGAGCTATCAG-3′). The 1.8 kb PCR product was cloned into pGEM-T to form pGL1191. The plasmids were sequenced. OsNAP released from pGL1191 upon NcoI digestion was then cloned into pGL1193 to form pGL1195. The 3.8 kb chimeric gene was then released from pGL1195 after digestion with SacI and ApaI (3′ overhangs removed by T4 DNA polymerase treatment) and cloned into pZP221 at the SacI and SmaI sites to form pGL1197. A nos terminator was added to the end of the chimeric gene in pGL1197 at the XbaI site to form pGL1800.

For complementation tests involving the kidney bean (Phaseolus vulgaris) NAP homolog (PvNAP), primers G1807 (see above) and G1809 (5′-AAGTCGACGATTTTCAGACAATTTAGAAAACAATC-3′; the underlined section is an engineered SalI site) were used to PCR amplify the AtNAP promoter region. The 2 kb PCR product was cloned into pGEM-T to form pGL1194. The genomic fragment containing the coding region of PvNAP (AAK84884) was PCR-amplified using primers G1806 (5′-AAGTCGACATGGACTACCACACCCTC-3′; the underlined section is an engineered SalI site) and G1668 (5′-GATCTAGATGGACGAAGCTTATCGTC-3′). The 1.3 kb PCR product was cloned into pGEM-T to form pGL1190. The plasmids were sequenced for sequence confirmation. The PvNAP coding region released from pGL1190 by SalI was then cloned into pGL1194 to form pGL1196. The 3.1 kb chimeric gene was released from pGL1196 by PstI and cloned into pPZP221, forming pGL1198. A nos terminator was added to the end of the chimeric gene in pGL1198 at the XbaI site to form pGL1801.

Agrobacterium and plant transformation

The above constructs in binary vectors (pGL1167, pGL1199, pGL1800 and pGL1801) were transferred into Agrobacterium tumefaciens strain ABI as previously described (He and Gan, 2002). Similarly, pTA7001 was transferred into A. tumefaciens strain LBA4404. The Agrobacterium cells containing the respective constructs were then used to transform Arabidopsis ecotype Columbia-0 or the atnap null mutant plants via vacuum infiltration (Bechtold et al., 1993). Transgenic plants were selected on plates containing 50 mg l−1 kanamycin (for pGL1167 transformants), 80 mg l−1 gentamycin (pGL1199, pGL1800 and pGL1801 transformants) or 25 mg l−1 hygromycin (pTA7001 transformants). Arabidopsis plants harboring pGL1167 were crossed with plants harboring pTA7001 and the hybrids were selected on plates containing both kanamycin (50 mg l−1) and hygromycin (25 mg l−1).

RNA gel blot and RT-PCR analyses

Total RNA extraction from Arabidopsis leaves and RNA gel blot analysis were performed as described previously (He and Gan, 2002). The hybridization was performed at 65°C. The Ambion RetroScript Kit (Ambion, Austin, TX, USA) was used to perform RT-PCR analysis according the manufacturer's instructions. The QuantumRNATM Universal 18S Internal Standard Kit (Ambion) was used for equal loading control. DNA fragments for creating related hybridization probes were PCR-amplified using the following primers: G1027 and G1028 (see above) for AtNAP, G10 (5′-CAGCTGCGGATGTTGTTG-3′) and G246 (5′-CCACTTTCTCCCCATTTTG-3′) for SAG12, G9 (5′-GCAACCAAAGGAGCCATG-3′) and G16 (5′-GTTTGGCCAACTAGTCTGC-3′) for SAG13, G1148 (5′-AGTAATGGCTTCCTCTATGC-3′) and G1149 (5′-GGCTTGTAGGCAATGAAACT-3′) for RBCS, G1665 (5′-ATCCCTTCCATT TCCGAC-3′) and G1666 (see above) for OsNAP, and G1667 (5′- CTGGGTCTTGTGCAGAAT-3′) and G1668 (see above) for PvNAP. Some of the primers were also used for related RT-PCR analysis.

Transient gene expression in onion epidermal cells

Onion (Allium cepa) epidermal cells were transfected with pGL1185 using a helium biolistic gun transformation system (Bio-Rad, Hercules, CA, USA) as described previously (Lin and Wang, 2004), and incubated in light or darkness for 24–48 h at 22°C. The subcellular localization of GFP fusion proteins was visualized with a fluorescence microscope.

Glucocorticoid treatments

The glucocorticoid treatments were performed as described by Aoyama and Chua (1997). Two-week-old plants grown in pots were sprayed with 30 μm dexamethasone (DEX). The plants were sprayed once a day for 2 days and incubated for two additional days. The Fv/Fm ratios for leaves of these plants were measured, and leaves were harvested for molecular analysis.

Dark-induced leaf senescence

Leaf number 6 from a 3-week-old Arabidopsis plant was excised and placed on moisturized filter papers in Petri dishes with adaxial side facing up. The plates were kept in darkness at 23°C for 4 days.

Measurements of chlorophyll content, fluorescence and ion leakage

Chlorophyll was extracted and quantified as described previously (He and Gan, 2002). Fluorescence in leaves was measured using a portable modulated chlorophyll fluorometer (model: OS1-FL) according to the manufacturer's instructions (Opti-Sciences, Tyngsboro, MA, USA). The ratio of variable fluorescence to maximal fluorescence (Fv/Fm) for each leaf was quantified directly using the fluorometer's test mode 1. For ion leakage, leaves were immersed into deionized distilled water, shaken in a 25°C water bath for 30 min, and the conductivity was measured using a digital conductivity meter (Fisher Scientific Traceable, Hampton, NH, USA). Samples were boiled for 10 min and then monitored for conductivity. The percentage of the first measurement over the second measurement was used as the membrane leakage indicator.

Data mining from the Genevestigator microarray database

The ‘Gene Atlas’ program of the microarray database Genevestigator ( was used to search expression levels of AtNAP (At1g69490) in different plant tissues. The program ‘Response Viewer’ was used to search expression change and expression levels of AtNAP under different treatments. To run both programs, chip type ‘ATH1:22k array’ for ‘wild-type only’ was used. When running ‘Response Viewer’, chips from all sources were selected in ‘ATH1:22k array’ for ‘wild-type only’ chip type. The database search was performed on 10 September 2005.

Molecular phylogenetic analyses

The amino acid sequence of AtNAP was used to search different genomic databases including GenBank (; for soybean, kidney bean, rice, nightshade, wheat, peach, tomato, petunia and potato), the TIGR plant genome databases (; for maize and Medicago), and PopulusDB (; for Populus). The NAC family genes with highest sequence similarity to AtNAP from different plant species including kidney bean (Phaseolus vulgaris; AAK84884), rice (Oryza sativa; NP_912423), soybean (Glycine max; AAY46121), nightshade (Solanum demissum; AAU90314), Medicago truncatula (AC140030_19.1), Populus trichocarpa (gene model gw1.X.1066.1), wheat (Triticum aestivum; AAU08785), maize (Zea mays; AZM5_18141), peach (Prunus persica; CAG28971), tomato (Lycopersicon esculentum; AAU43923), potato (Solanum tuberosum; AAU12055) and petunia (Petunia × hybrida; AAM34773), were used for molecular phylogenetic analyses. The kidney bean and rice NAP homologs, which were further studied for their expression patterns and used to transform atnap mutant in heterogeneous complementation tests, are referred to, in the text, as PvNAP and OsNAP, respectively. Predicted amino acid sequences of AtNAP homologs from different plant species were first aligned using the alignment program CLUSTALW (Chenna et al., 2003) with the default parameter values (alignment algorithm: full; CPU mode: single; Kimura correction: off; output: aln1; output order: aligned; score type: percent; ignore gaps in alignment: off; number of sequences: 13; Figure S1). The alignments were then used to produce the phylogeny using the phylogenetic analysis program MEGA3.1 (Kumar et al., 2004). Parameters used in this analysis were: data type: amino acid; analysis: phylogeny reconstruction; method: neighbor-jointing method; gaps/missing data: complete deletion; model: amino:Poisson correction; substitutions to include: all; pattern among lineages: same (homogeneous). The bootstrap values for nodes in the phylogenetic tree are from 1000 replications.


We thank Drs Rongcheng Lin and Haiyang Wang of Boyce Thompson Institute for help in onion transient expression and visualization of the AtNAP–GFP fusion proteins, Jessica Westbrook and Robert Bode for careful reading of the manuscript, and Drs Ray Wu and Donald Halseth of Cornell University, Ithaca, NY, USA for providing rice and kidney bean materials. We are also grateful to Drs William Millar, Jocelyn Rose and Steven Tanksley of Cornell University, for useful discussions. This work was supported by grants to S.G. from the US Department of Energy Basic Energy Sciences (grant number DE-FG02-02ER15341) and from the US-Israel BARD (grant number IS-3645-04). Arabidopsis seed stocks used were SALK_005010 and SALK_004077, from the Arabidopsis Biological Resource Center (ABRC) at Ohio State University.