Identiﬁcation of a Novel Chloroplast Protein AtNYE1 Regulating Chlorophyll Degradation during Leaf Senescence in Arabidopsis 1[C][W][OA]

A dramatic increase of chlorophyll (Chl) degradation occurs during senescence of vegetative plant organs and fruit ripening. Although the biochemical pathway of Chl degradation has long been proposed, little is known about its regulatory mechanism. Identiﬁcation of Chl degradation-disturbed mutants and subsequently isolation of responsible genes would greatly facilitate the elucidation of the regulation of Chl degradation. Here, we describe a nonyellowing mutant of Arabidopsis ( Arabidopsis thaliana ), nye1 - 1 , in which 50% Chl was retained, compared to less than 10% in the wild type (Columbia-0), at the end of a 6-d dark incubation. Nevertheless, neither photosynthesis- nor senescence-associated process was signiﬁcantly affected in nye1 - 1 . Characteristically, a signiﬁcant reduction in pheophorbide a oxygenase activity was detected in nye1 - 1 . However, no detectable accumulation of either chlorophyllide a or pheophorbide a was observed. Reciprocal crossings revealed that the mutant phenotype was caused by a monogenic semidominant nuclear mutation. We have identiﬁed AtNYE1 by positional cloning. Dozens of its putative orthologs, predominantly appearing in higher plant species, were identiﬁed, some of which have been associated with Chl degradation in a few crop species. Quantitative polymerase chain reaction analysis showed that AtNYE1 was drastically induced bysenescence signals. Constitutive overexpression of AtNYE1 could result in either pale-yellow true leaves or even albino seedlings. These results collectively indicate that NYE1 plays an important regulatory role in Chl degradation during senescence by modulating pheophorbide a oxygenase activity.

Ever since the nonyellowing mutant was identified, chlorophyll (Chl) catabolism has been known as a separable event from other senescence processes in senescing vegetative plant organs and ripening fruits (Thomas and Stoddart, 1975;Thomas and Howarth, 2000). Extensive chemical and biochemical analyses have led to the construction of a major Chl catabolism pathway, i.e. pheophorbide a oxygenase (PaO) pathway (Hö rtensteiner, 2006;Tanaka and Tanaka, 2006). However, very little is known about its regulation. According to the proposed pathway, phytol and magnesium (Mg) are first removed from Chl a by chlorophyllase (Chlase) and a putative Mg-dechelatase or Mg-dechelating substance, respectively, and the resultant pheophorbide a (Pheide a) is then converted to the primary fluorescent Chl catabolites (pFCCs) through the sequential functioning of two enzymes: PaO and red Chl catabolite reductase (RCCR). The pFCC is subsequently transported out of the chloroplast and converted to the colorless nonfluorescent Chl catabolites in the vacuole. Chl b is proposed to be transformed to Chl a with a two-step reduction, catalyzed by Chl b and 7 1 -OH-Chl a reductases, respectively, before being further degraded (Tanaka and Tanaka, 2006). Over the past few years, the identification of genes encoding key enzymes, including CLHs (for chlorophyllchlorophyllido hydrolases), PaO, and RCCR, offered a possibility of exploring the regulation of Chl catabolism (Gray et al., 1997;Jacob-Wilk et al., 1999;Tsuchiya et al., 1999;Wü thrich et al., 2000;Pružinská et al., 2003). Of the three genes identified, PaO is thus far shown to have a regulatory effect on Chl degradation (Pružinská et al., 2003;Pružinská et al., 2005;Chung et al., 2006). Nevertheless, a mutation in PaO or even a down-regulation of PaO expression may cause a lesion mimic phenotype, a reaction resembling the hypersensitive response induced by pathogens.
An alternative way to explore the regulatory mechanism of Chl catabolism is to identify genes directly regulating Chl degradation during senescence through forward genetics. Mutants, with Chl degradation being disturbed but other senescence-associated processes not being significantly affected, have been reported in diversified species (Thomas and Stoddart, 1975;Guiamét et al., 1990;Akhtar et al., 1999;Cha et al., 2002;Guiamét et al., 2002;Luquez and Guiamét, 2002;Oh et al., 2003Oh et al., , 2004Efrati et al., 2005). But only quite recently have the identities of corresponding genes been indicated in two crop species, Festuca pratensis and rice (Oryza sativa; Armstead et al., 2006;Hö rtensteiner, 2006). However, there has been so far no genetic confirmation of cloning of the responsible genes.
Here we report the isolation of a nonyellowing mutant (nye1-1) from a fast-neutron mutagenized M2 population of Arabidopsis (Arabidopsis thaliana). It was shown that the mutation primarily affected Chl catabolism, mainly by reducing PaO activity, during senescence. However, no detectable accumulation of major Chl degradation intermediates was observed in nye1-1. We isolated the causal gene AtNYE1 by positional cloning and confirmed its identity by genomic complementation. Overexpression of AtNYE1 may cause either pale-yellow true leaves or even albino seedlings.

Isolation and Phenotypic Characterization of nye1-1
To explore the regulatory mechanism of Chl degradation during leaf senescence, a nonyellowing mutant was identified from the M2 population of fast-neutron mutagenized Arabidopsis ). It was subsequently backcrossed with Col-0 four times and designated as nye1-1 (nonyellowing). The nye1-1 mutant exhibited a stable nonyellowing phenotype during both natural and dark-induced senescence of leaves either attached, detached, or in planta (Figs. 1,and 2A). Its siliques also stayed lightly green at harvesting time (Supplemental Fig. S1). In all the follow-up experiments, incubating detached leaves in darkness was therefore adopted constantly as the simplest way to identify the mutation phenotype. Under long-day (16-h light/8-h dark) growth condition, an approximately 9-d delayed bolting and a slightly prolonged flowering period were observed in nye1-1, as compared with Col-0 ( Fig. 1F). No other significant morphological differences were observed between nye1-1 and Col-0, and the etiolation/deetiolation behavior was not altered either (data not shown).
In reciprocal crosses with the wild types of both the Col-0 and Landsberg erecta (Ler), all the leaves of F1 plants showed a partially stay-green phenotype after 4-d dark treatment ( Fig. 2A). Total Chl content of F1 plants was significantly higher (P , 0.01) than that of Figure 1. Phenotypic characterization of nye1-1. A, Three-week-old plants were incubated in permanent darkness for 4 d. B, Differential changing patterns in yellowing phenotypes between nye1-1 and Col-0 plants, as well as between young and old leaves, with roots removed, during a 7-d incubation in darkness, as described in ''Materials and Methods.'' C and D, Nonyellowing senescent leaves (red arrows) of nye1-1, compared to the yellowing ones of Col-0, 30 (C) and 45 (D) d after germination, respectively. Bar 5 1 cm. E, Changes in the levels of total Chl in nye1-1 and Col-0 leaves during a 6-d dark incubation. Data are means of three replicates. Error bars indicate SD. The experiment was repeated twice with almost the same results. DAD represents day(s) after darkness. Statistical analysis of differences in nye1-1 with respect to Col-0 was performed using two-tailed t test. The significant difference is denoted with two asterisks (P , 0.01). F, nye1-1 bolts some a week later than Col-0 under long-day conditions. wild-type plants after 4-d dark treatment and lower (P , 0.01) than that of mutants (Fig. 2B). A statistical 1:2:1 segregation ratio was detected in the F2 population (backcrossed to Col-0; 44 wild-type phenotype:64 mutant phenotype:131 intermediate phenotype; x 2 5 5.56, P 5 0.062 . 0.05), and a consistent result was also observed in the F2 population of crosses to Ler (data not shown). Thus, we concluded that the mutant phenotype was caused by a monogenic semidominant nuclear mutation. To detect the lesion step of Chl degradation, total pigments were extracted from the leaves of nye1-1 as well as Col-0, and initially quantified by spectrophotometry. Over the course of a 6-d dark treatment, the Chl content declined dramatically in the leaves of Col-0, down to approximately 10% of that in the untreated leaves by the end of the treatment. By contrast, the Chl content in nye1-1 dropped much more slowly, 50% of that in the untreated leaves being retained at day 6 ( Fig. 1E). The result was further validated by HPLC analysis (data not shown). However, no significant accumulation of either chlorophyllide a (Chlide a) or Pheide a was observed either in nye1-1 or in Col-0 in the chromatogram assay, although trace levels of the above catabolites were detected in both of the genotypes after 0-, 2-, and 4-d dark treatment (Fig. 3 indicates representative data of 4-d samples). To gain insight into the mechanism of the delayed Chl degradation, time-course changes in the transcript levels of AtCLH1, PaO, and RCCR were examined during dark treatment and no differences were detected either between the two genotypes ( Fig. 4). Nevertheless, subsequent analyses of in vivo key enzyme activities known for green pigment degradation revealed that the function of PaO in nye1-1 was somehow reduced, while Chlase remained fully competent (Fig. 5,A and B). pFCC-1 was hardly detected at 0 d after darkness in both nye1-1 and Col-0 in a coupled PaO-RCCR enzymatic assay. The PaO activity in nye1-1, although also dramatically induced, was significantly lower (20% to approximately 50% residual activity relative to that in Col-0 in repeated experiments) than that in Col-0 ( Fig.  5B; Table I). A crossover experiment (Roca et al., 2004) was also employed to exclude the possibility that the nonyellowing phenotype in nye1-1 was due to a lesion in the RCCR activity. Our results clearly demonstrated that the RCCR in nye1-1 was fully competent (Table I). These data suggest that NYE1 may have a function in modulating the PaO activity, either directly or indirectly, on its cleavage of the porphyrin ring.

Photosynthesis-and Senescence-Associated Processes
Largely Unaffected in the Leaves of nye1-1 To elucidate the possible relationship of the nonyellowing mutation with photosynthesis, the photosynthetic capacity was initially examined in both genotypes. Compared with that in Col-0, a substantially identical changing pattern of the net photosynthetic rate in the fourth to sixth leaves was observed in nye1-1 in a timecourse monitoring experiment. The photosynthetic rates started to decline slightly 15 d after emergence (DAE), dramatically 21 DAE and sharply 27 DAE in the leaves of both genotypes (Fig. 6A). To explore the relationship further, the transcriptions of two photosynthesisassociated genes, Chl a/b-binding (CAB) protein gene and ribulose bisphosphate carboxylase/oxygenase small subunit (RBCS) gene, were examined by quantitative PCR (qPCR). The transcript levels of CAB declined sharply during dark-induced senescence, but indistinguishably between nye1-1 and Col-0 (Fig. 6B). The transcript levels of RBCS also declined rapidly in Col-0, but less dramatically in nye1-1 (Fig. 6C). To determine whether senescence processes are affected by the mutation, the transcript levels of two well-characterized senescence-associated genes, SEN1 and SAG12, were examined in both nye1-1 and Col-0. Increased abundances of both SEN1 and SAG12 transcripts were detected in both the genotypes after a 2-d dark treatment in a similar way (Fig. 6, D and E). These results indicated that neither photosynthesis-nor senescenceassociated processes were significantly affected by the mutation. The nye1-1 could therefore be classified as a nonfunctional (type C) stay-green mutant (Thomas and Howarth, 2000).

Mapping and Isolation of AtNYE1
Twenty-two simple sequence length polymorphic (SSLP) markers (http://www.arabidopsis.org), distributed evenly on five chromosomes, were employed  Chlase activity in crude enzyme extracts from both nye1-1 and Col-0 was measured on day 0, 3 and 6 d after incubation in darkness and normalized to their protein contents. Values are means of three replicates. Error bars indicate SD, and DAD represents day(s) after darkness. B, HPLC analysis of PaO-RCCR coupled assay with crude extraction of PaO and RCCR from nye1-1 and Col-0 leaves during a 4-d dark treatment; pFCC-1 was eluted after 5.8 min, as indicated. U, Unknown peak. The large early peak was due to NADPH present in the reaction mixture. The experiment was repeated once with qualitatively identical results.
to map NYE1 locus, with a mixed DNA sample extracted separately from 100 homozygous individuals identified from an F2 population derived from a cross with wild-type Ler. The locus was roughly located between F7J7 and F10N7, with an interval of approximately 8.0 cM on the long arm of chromosome IV. New SSLP markers within this region were generated to narrow down the interval. With a mapping population of 828 plants, the locus was further located between T12H17 and F7H19, with an interval of 0.5 cM (Fig. 7).
Sequencing the middle region of approximately 40 kb within this interval led to the identification of a single base pair change, an A to T substitution, at position 29 of the first exon in the open reading frame (ORF) of at4g22920. This mutation converts the 10th Leu to a stop codon, which leads to an early termination of the translation. The at4g22920 was therefore tentatively designated as AtNYE1.
Complementation of nye1-1 Phenotype by the Genomic Copy of AtNYE1 A 4.1-kb PstI-BamHI genomic fragment, containing AtNYE1 as well as its 1.9 kb upstream and 1.0 kb downstream sequences, was amplified and constructed into pPZP221 (Hajdukiewicz et al., 1994), which was subsequently introduced into Agrobacterium tumefaciens strain LBA4404. Plants of nye1-1 were transformed using standard dipping method (Clough and Bent, 1998). In the first batch of selection, 11 gentamycinresistant plants were selected and their transgenic status was confirmed by PCR (data not shown). The leaves of these transgenic plants showed a rapid yellowing phenotype during natural aging and darkinduced senescence, just as those of Col-0 did (Fig. 8, A and B). The complementation results not only confirmed the genetic function of AtNYE1, but also indicated that the 4.1-kb genomic fragment was sufficient to regulate Chl degradation.

Overexpression of AtNYE1 Caused Precocious Chl Degradation
To examine its capacity in regulating Chl degradation, we expressed AtNYE1, driven by cauliflower mosaic virus 35S promoter, in nye1-1. Fifty transgenic seedlings were obtained in the first batch. Of them, 21 developed normally, 11 showed pale-yellow true leaves displaying varied degrees of degreening, and the rest (18) exhibited albino phenotype (Fig. 9, A and B). The severity of the degreening in the AtNYE1 overexpressors was not only reflected on retardation to their growth and development but also well correlated with the expression level of AtNYE1 (Fig. 9C). In addition, we also detected that PaO expression was not induced in these overexpression lines (Fig. 9C), indicating that a higher expression of AtNYE1 is sufficient to induce Chl degradation, provided that there is a basal level of PaO expression.
Searching of the GenBank database (http://www. ncbi.nlm.nih.gov/BLAST) and The Institute of Genomic Research plant genome database (http://www.tigr. org/plantProjects.shtml) revealed dozens of ESTs that are highly homologous to AtNYE1 in diversified species of angiosperms and gymnosperms, but not in yeast (Saccharomyces cerevisiae) and animals (Fig. 10A). Putative genes that encode proteins showing weak similarities to AtNYE1 are also present in a green algae, Ostreococcus tauri (CAL56489), and some bacterial species (ZP_01170511.1 of Bacillus sp., NP_563502 of Clostridium perfringens, and NP_350006 of Clostridium acetobutylicum; data not shown). An EST clone (AV618908) in Chlamydomonas reinhardtii and several ribonuclease II family proteins in some Cyanobacteria species, showing very low similarities to AtNYE1 though, were also revealed in the C. reinhardtii ESTs and Cyanobacteria databases (http://www.kazusa.or. jp/en/plant/database.html). Even though highly conserved among higher plants, domain sequences hardly provide any cues about its possible biochemical roles. However, a variety of potential modifications are suggested by the presence of three N-glycosylation sites, four protein kinase C phosphorylation sites, three casein kinase II phosphorylation sites, and an N-myristoylation site (Bairoch et al., 1997).
Twelve sequences from eight representative plant species, corresponding to the amino acids between 78 and 211 in AtNYE1, were adopted to construct a phylogenetic tree (Fig. 10B). The angiosperms-gymnosperms divergence and monocot-dicot divergence were well reflected on the tree. Furthermore, our data also suggest that the divergence of family members within a species is posterior to the divergence of species.

AtNYE1 Localized in Chloroplast
To verify the prediction that AtNYE1 was likely targeted to the chloroplast, its C terminus was fused to enhanced yellow fluorescent protein (eYFP) and the resultant fusion protein was introduced into nye1-1. Transgenic plants expressing the AtNYE1-eYFP fusion protein driven by the 35S promoter also could result in albino seedlings or pale-yellow true leaves (data not shown), suggesting that the AtNYE1-eYFP fusion protein was functioning properly. Expectedly, eYFP fluorescence in albino seedlings was predominantly observed in chloroplast (Fig. 11), while it was hardly detected in those phenotypically normal transgenic plants (data not shown). By comparison, eYFP in the 35S-eYFP control plants was present in both cytoplasm and nucleus (Fig. 11). These data confirm that AtNYE1 is localized in chloroplast.

Spatial and Temporal Transcriptions of AtNYE1
In attempt to elucidate its physiological function in Arabidopsis, relative transcriptions of AtNYE1 in different organs and in leaves at various developmental stages were determined by real-time PCR. The transcript of AtNYE1 was detected in all the organs examined, with relatively higher levels being detected in flowers and buds, medium to low levels in siliques, and the lowest in roots, seedlings, and young rosette leaves (Fig. 12A). As expected, AtNYE1 was strongly induced by dark treatment or natural senescence (Fig.  12, B and C; Guo et al., 2004;Armstead et al., 2006). Interestingly, in nye1-1, its expression was also sharply induced, as it was in Col-0, by dark treatment (Fig.  12C).

NYE1 Is a Key Regulator Directly Associated with Chl Degradation during Organ Senescence or Maturation
Recent studies showed that alterations in the sequences of AtNYE1 orthologs were associated with stay-green phenotypes in various crop species. In rice, a missense mutation was identified in its putative ortholog (SGR) through mapping a stay-green locus (Cha et al., 2002;Hö rtensteiner, 2006). In F. pratensis, a  four-base (ATAT) insertion in its putative ortholog (Y) was also reported to be genetically associated with the stay-green phenotype of the most extensively analyzed nonyellowing mutant, Bf993 (Thomas and Stoddart, 1975;Thomas et al., 1999;Roca et al., 2004;Armstead et al., 2006). Even more amazingly, the color (yellow versus green) trait of cotyledons in pea (Pisum sativum), first described by Mendel in 1866, was recently considered to be due to a variation in intron size of its putative ortholog (I; Armstead et al., 2007). However, the function of these orthologs has not been confirmed by genetic complementation in any of these crops, probably due to their recalcitrance to genetic transformation. In this study, AtNYE1 genomic copy can rescue the nye1-1 phenotype in more than 10 independent transformants, confirming that the NYE1 encodes the gene responsible for the nonyellowing phenotype. These data collectively indicate that NYE1 may play a key role in regulating Chl degradation during senescence in plants.
RNA interference and overexpression have been adopted to analyze the regulatory role of AtNYE1 on Chl degradation in Arabidopsis. In RNA interference plants, Chl degradation was delayed significantly during dark-induced senescence (L. Yang and B. Kuai, unpublished data; Armstead et al., 2007). Strong upregulation of AtNYE1 expression can result in either pale-yellow true leaves or albino seedlings, while slight up-regulation of its expression causes no significant changes in the phenotype (Fig. 9). These results, along with the semidominant nature, or haploinsufficiency, of the mutation, clearly indicate that the NYE1 transcript abundance is closely correlated with an effect on Chl degradation. Practically, this correlation implies that a regulatory effect on Chl catabolism could be imposed by modulating NYE1 transcript level.
In a series of related studies, Chl degradation is also found to be affected by plant hormones, such as cytokinin and ethylene. However, this kind of effect is mainly considered an indirect consequence of altered senescence (Smart, 1994;Gan and Amasino, 1995). In nye1-1, both photosynthetic capacity/process and senescence process are not significantly affected, indicating that Chl degradation is directly slowed down by the null mutation of AtNYE1. It is therefore fairly convincing to conclude that NYE1 directly regulates Chl degradation during organ senescence or maturation, and the identification of nye1-1 mutant in   Arabidopsis opens wide possibilities for exploring the genetic and molecular regulatory mechanisms of Chl degradation.

NYE1 Accelerates Chl Degradation by Maintaining Sufficient PaO Activity
Although highly conserved domains are present in its putative orthologs of various plant species, no informative implications about the molecular and biochemical roles of NYE1 have been revealed by any of the available bioinformatical analyses. Nevertheless, experimentally examining (chemical, biochemical, and molecular) parameters associated to green pigment (Chl, Chlide, and Pheide) breakdown in nye1-1 demonstrated that PaO activity was greatly reduced, with an estimated reduction of about 50% to approximately 80%. This detection is in accordance with that observed in Bf993 of F. pratensis , which provides further evidence to the postulation that the four-base insertion in the NYE1 ortholog might be responsible for the stay-green phenotype (Armstead et al., 2006). The reduction of PaO activity in these NYE1 mutants indicates that NYE1 normally accelerates Chl degradation during senescence by maintaining sufficient PaO activity to ensure no accumulation of the photodynamic Pheide a. One of the immediate speculations is that NYE1 might directly or indirectly interact with PaO or a complex of enzymes involving PaO (for a recent review, see Hö rtensteiner, 2006). This hypothesis is supported by the fact that the expression of PaO was not affected in nye1-1 (Fig. 4B), implying that the modulation likely occurs at the protein level. This implication is consistent with recent findings that the PaO activity might be posttranslationally regulated, mainly through PaO dephosphorylation (Chung et al., 2006). Interestingly, the expression of PaO was not found to be induced in albino seedlings expressing 35S-AtNYE1, which might imply that NYE1 plays a dual role in both initiating Chl degradation and modulating the residual PaO activity that appears to be sufficient for the degradation of Pheide a.
On the other hand, Chl in NYE1 mutants was still being degraded upon dark incubation, even in the null Figure 11. Localization of AtNYE1 in chloroplast. Micrographs of leaf tissue from nye1-1 (top), nye1-1 expressing the 35S-eYFP reporter (middle), and nye1-1 expressing the 35S-AtNYE1-eYFP reporter (bottom). First column: confocal micrographs of fluorescence from eYFP (green channel). Second column: Chl autofluorescence (red channel). Third column: merged micrographs of red and green images. Fourth column: transmitted light image. Of each construct, three or more independent lines were taken for observation. Bar 5 20 mm. mutant nye1-1, indicating that there might exist a paralog, less efficient though, of NYE1, or even an alternative Chl degradation pathway. Our HPLC analysis data, along with Roca's (Roca et al., 2004), seemingly supports the assumed existence of a less efficient paralog, which is presumably also less competent in occupying a putative acting site. In Bf993, Chlide a and Pheide a were found to accumulate sequentially, with the Chlide a accumulation being much more significant; in nye1-1, however, no detectable levels of such catabolites were accumulated, although a similar changing pattern in the dynamics of Chl degradation was observed (Thomas, 1987;Thomas et al., 1999). These observations indicated that the Chl degradation pathway in Bf993 suffered a more dramatic interference than that in nye1-1. However, according to Armstead's report (Armstead et al., 2006), a four-base insertion in the NYE1 ortholog in Bf993 resulted in a final protein consisting of 232 residues, with its amino acid sequence from position 100 being radically changed; whereas the nonsense mutation in AtNYE1 in nye1-1 caused an early termination of translation, which results nye1-1 as a null mutant. Only when a paralog, not only less efficient but also less competitive in occupying a putative acting site is assumed to exist, could these data be plausibly explained, i.e. the truncated protein in Bf993 was even less efficient in functioning but still more competitive in occupying a putative acting site than its wild-type paralog, while, in nye1-1, the wildtype paralog was functioning properly, presumably without being subject to such a competition. It has been indicated that the formation of pFCC-1 from Pheide a requires an unknown factor, RFF, which is present in Escherichia coli (Pružinská et al., 2005). Since no NYE1 homologs are found in E. coli, it can be excluded that NYE1 is identical with RFF. In Arabidopsis, at4g11910 could be a candidate paralog. However, both the genetic role of at4g11910 and the molecular nature of RFF have not been revealed.

NYE1 Likely Promotes Higher Plant Growth and Development by Timely Remobilizing Nitrogen Nutrition from Senescing Organs to Nascent Ones
It has been reported that the growth and productivity of stay-green plants of Lolium prenne, with the putative NYE1 mutant ortholog introgressed from Bf993, are significantly reduced compared to wild-type plants under conditions of extreme nitrogen (N) limitation, and the extra retention of leaf N in its senescing leaves contributes a functionally significant proportion of the total N content of the plant with respect to growth potential (Macduff et al., 2002). In this study, although no experiments were deliberately designed to precisely examine the N nutrition remobilization, a significant 9-d delay in bolting and consequently flowering was observed in nye1-1 relative to that in the wild type, both of which were grown under standard Arabidopsis growth conditions, without applying nutrition after vegetative growth, a condition similar to N limitation at the later stages. An immediate significance of this coincidence in inhibition to plant growth and development would not only reinforce the speculation that Sid, or recently redesignated as Y in F. pratensis, was the ortholog of AtNYE1, but more importantly, also imply that NYE1 has been likely evolved as a promoter of higher plant growth and development by timely remobilizing N nutrition from senescing organs to nascent ones. This implication is evidenced by the fact that the putative orthologs of NYE1 predominantly appear in higher plant species, multigreen organs of which develop and senesce sequentially.
In conclusion, NYE1 has been convincingly identified as a key and direct regulator of Chl degradation during senescence, probably by modulating PaO activity. The successful identification of NYE1 mutation in Arabidopsis would not only provide an efficient system to analyze its molecular and biochemical roles, but also greatly facilitate further exploration for the regulatory mechanism of Chl degradation in general. Figure 12. Temporal and spatial expression patterns of AtNYE1. A, Spatial expression pattern of AtNYE1 in the wild type. B, AtNYE1 transcript levels in the leaves of different developmental stages in the wild type. Y, Young leaves; ES, early senescence; S2, progressive senescence stage 2 (Gepstein et al., 2003). C, Changes in AtNYE1 transcript levels during a 48-h incubation in darkness in Col-0 and nye1-1. cDNA abundances quantified by real-time RT-PCR, ACTIN2 as the reference. Values are means of three replicates. Error bars indicate SD. The experiment was repeated once.
All the wild-type and mutant Arabidopsis (Arabidopsis thaliana) lines used in this study are in the Col-0 background unless indicated otherwise. Seeds were sown in square pots (10 cm in length) with soil (v [peat soil]:v [vermiculite]:v [pearlite] 5 3:9:0.5; Shanghai Institute of Landscape Science) presoaked with plant nutrient medium including 0.5% Suc (Estelle and Somerville, 1987). Plants were grown at 24°C, with an approximately 100 mmol m 22 s 21 light intensity under a 16-h light /8-h dark photoperiod.
For dark treatment, leaves 4 to 6, or whole plants without roots, from 3-to 4-week-old Arabidopsis plants were excised and incubated on wet filter paper in total darkness at 25°C for various periods, as indicated. For the dark treatment of whole plants in soil, trays were covered with black bags and incubated in growth room for a period of time as indicated.

Mutant Screening
The fast-neutron mutagenized M2 seeds in Col-0 genetic background were purchased from Lehle Seeds (Round Rock). About 50,000 3-week-old M2 plants were incubated in darkness for 6 d to screen mutants with delayed Chl loss. The putative screened mutants were subjected to further confirmation for a couple of times during the remaining growth period by visually checking the phenotypes of their detached leaves incubated in permanent darkness for 4 to 6 d. A stable nonyellowing mutant, without obvious pleiotropic effects, was finally identified and designated as nye1-1. It was then backcrossed with the wild-type Col-0 four times, and the resulting homozygous nye1-1 plants were used for further analysis.

Map-Based Cloning
For genetic analysis of the mutation, nye1-1 plants were crossed with the wild-type Col-0 as well as Ler plants reciprocally, and the resulting F1 seedlings were allowed to self-pollinate to produce F2 populations. Phenotypes of F1 and F2 were visually classified and scored as mutant, intermediate, or wild type.
Mapping procedure was performed as described by Lukowitz and Jander (Lukowitz et al., 2000;Jander et al., 2002). From the F2 population of a cross between nye1-1 and wild-type Ler plants, 828 plants showing mutant phenotype were selected to construct a mapping population. The homozygosity of the selected plants was checked in F 3 generations. NYE1 locus was initially mapped using 22 SSLP markers (selected from the Arabidopsis Information Resource, http://www.arabidopsis.org/) with a mixed sample of DNAs extracted from 100 selected plants individually. The interval was narrowed down with newly designed markers (Table II). The candidate gene was identified by sequencing ORFs of putative genes within the expected region. PCR amplifications were performed using Primerstar DNA polymerase (TaKaRa Biotechnology).

Plasmid Construction
For the genomic complementation test, a 4.1-kb DNA fragment containing the predicted ORF as well as 1.9 kb promoter region and 1.0 kb downstream sequence was amplified from the wild type by PCR using two oligonucleotides, 5# CACGGATCCAACAAGATCTAACCCTTTTG 3# and 5# CAGCTGCAGTG-GAGGTGACAGGAGG 3# (the lowercase c is an introduced mismatch to create restriction site, underlined). The fragment was subcloned into PstI and BamHI sites of pPZP221 (Hajdukiewicz et al., 1994) and sequenced with self-designed primers (data not shown).
For overexpression of AtNYE1, the full-length coding region (807 bp) was amplified with a pair of primers 5# CATGGTACCATGTGTAGTTTGTCGGC-GATTATG 3# (the underlined sequence is a KpnI site) and 5# CATGGATCCCT-AGAGTTTCTCCGGATTTGGAG 3# (the underlined sequence is a BamHI site). The PCR products were cloned into pMD 19-T vector and sequenced. After digesting with KpnI and BamHI, the releasing fragment was subcloned into pCHF3 binary vector (Ge et al., 2005).

Plant Transformation
The above constructs were introduced into the Agrobacterium tumefaciens strain LBA4404. Plants were transformed via the floral-dipping method (Clough and Bent, 1998). Putative transgenic plants were initially selected with 90 mg L 21 Gentamycin (for pPZP221 derived constructs and pPZPY122 derived constructs) or 50 mg L 21 Kanamycin (for pCHF3 derived contructs), and their transgenic status was determined by PCR analysis.

Fluorescence Microscopy
Excised leaves were mounted on glass with a drop of water and then covered with coverslips. They were examined by confocal microscopy using a LSM 510 laser scanning confocal microscope (ZEISS) with a 403 oil immersion objective. Chl autofluorescence and YFP signal were sequentially excited by switching between the 488-and 514-nm laser lines, respectively. The fluorescence emission was selected by a band pass filter (BP530-600) for eYFP and a long pass filter (LP560) for Chl fluorescence. Recorded images from the microscopes were processed with Adobe Photoshop (Adobe Systems).

Real-Time Reverse Transcription-PCR
Total RNA was extracted from control and dark-treated leaves of Col-0 using Trizol Reagent (Invitrogen), followed by chloroform extraction, isopropanol precipitation, and spectrophotometric quantification. After digestion with Rnase-free DNase (Promega), 2 mg total RNA was reverse transcribed with the Superscript reverse transcriptase (Shenenergy Biocolor). The products were subsequently used as templates for real-time PCR analysis. The realtime PCR was performed using SYBR Green I PCR kit (Toyobo) on an iCycler according to the manufacturer's suggestions, with ACT2 as a reference (Bio-Rad).

Photosynthetic Rate Measurement
Net photosynthetic rate was measured using LI-6400 (LI-COR) under a fixed LED light source (500 mmol m 22 s 21 ) at 25°C, as described in the manufacturer's instruction.

Measurement of Chl Contents and Detection of Its Catabolites
For pigment quantification, leaves 4 to 6 (about 0.1 g fresh weight) from 3-to 4-week-old plants were taken. After dark treatment, excised leaves were immediately frozen in liquid N and stored at 280°C. Chl contents were quantified by spectrometer according to Benedetti and Arruda (Benedetti and Arruda, 2002), and HPLC analysis was performed on Aglient 1100 HPLC by using Almela's method at 440 nm (Almela et al., 2000). Identification of pigments was based on their respective retention time or coinjection with standard pigment.
Standard Chl a was purchased from Sigma and Pheide a from Wako Chemicals. Chlide a was prepared by digesting Chl a with Chlase, as described  (Fang et al., 1998), and verified by its spectral pattern. To further verify the preparation, the prepared Chlide a was acidified with HCl to remove Mg, and the resultant Pheide a was confirmed by referring to the retention time of standard Pheide a. Chl b was prepared by the following procedure under darkness at 4°C. Chls were first extracted and partially purified from spinach (Spinacia oleracea) leaves according to a reported protocol (Iriyama et al., 1974). Suc was ground into fine powder and roasted at 70°C for 2 h. It was then loaded into a column with petroleum ether after being cooled down to room temperature. The purified Chls were then dissolved in petroleum ether and transferred into the column. After xanthin was eluted by petroleum ether, petroleum ether/acetone/benzene (20/2/1, v/v/v) was used to separate the remaining compounds. Chl b, showing a yellow-green color, was then collected, dried, and redissolved in acetone.

Chlase Assay
Chlase activity was measured according to a reported protocol (Todorov et al., 2003), with an exception that acetone power was dried and weighed. Chlase activity data were normalized to total protein content, determined by bicinchoninic acid method (Smith et al., 1985), and the total protein content was normalized to dry acetone powder weight.

PaO Assays
PaO was measured using a coupled assay with RCCR according to previous reports, with a slight modification (Hörtensteiner et al., 1995;Vicentini et al., 1995). Briefly, leaf samples were homogenized in 5 mL/g fresh weight of a medium containing 400 mM sorbitol, 25 mM Tricine-KOH, pH 8.0, 2 mM EDTA, 1 mM MgCl 2 , 0.1% bovine serum albumin (w/v), 5 mM polyethylene glycol 4,000, and 0.5 mM dithiothreitol, with chilled mortars and pestles. After filtration through two layers of cloth, the homogenate was centrifuged at 7,000g for 4 min. The green sediment was washed once with buffered medium (the above medium with EDTA, MgCl 2 , and polyethylene glycol 4,000 omitted). After centrifugation at 8,000g for 5 min, pellets were resuspended in the same buffered medium (2 mL/g fresh weight leaf tissue) and 1 mL aliquots were pipetted into 1.5 mL tubes. After centrifugation at 14,000g for 5 min, sediment in the tube was frozen in liquid N and stored at 280°C.
The PaO and RCCR were prepared strictly according to the previous protocols. Soluble stroma protein (S1 fraction) was used as the source of RCCR and solubilized membrane protein in osmotic buffer (Tris-MES, pH 8.0, 0.1% bovine serum albumin) with 1% Triton X-100 used as the source of PaO.
HPLC analyses were performed with a Waters 2690 HPLC coupled to a Waters 2475 fluorescence detector equipped with a ZorbaxSB C-18 column (4.6 mm 3 25 cm, 5-mm particle diameter). Twenty microliters were injected at a flow rate of 0.8 mL/min. Under these conditions, pFCC-1 was eluted after 5.8 min. Fluorescence was recorded at 320 nm (excitation)/450 nm (emission). The amount of pFCC-1 was given as integrated peak areas.

Accession Numbers
The GenBank accession numbers for AtNYE1 and At4g11910 are DQ437531 and DQ437532, respectively.

Supplemental Data
The following materials are available in the online version of this article.