The NAC Transcription Factor Gene OsY37 (ONAC011) Promotes Leaf Senescence and Accelerates Heading Time in Rice

Abstract

Leaf senescence is an important physiological process involving the degradation of a number of metabolites and their remobilization to new reproductive and storage organs. NAC (NAM, ATAF, and CUC) transcription factors are reported as important regulators of the senescence process. Here, we describe the identification and functional characterization of the NAC transcription factor gene, OsY37 (Oryza sativa Yellow37, ONAC011) obtained from Oryza sativa cv. indica, and japonica. We created transgenic plants expressing the OsY37 gene under the control of a strong and constitutive CaMV35S promoter. The resulting transgenic plants overexpressing OsY37 gene showed early heading and precocious senescence phenotype of flag leaves compared with wild-type plants. By contrast, blocking the function of this gene via RNAi (RNA interference) and CRES-T (Chimeric Repressor Silencing Technology) technology, delayed both heading time and leaf senescence. Furthermore, knockdown of OsY37 expression caused dwarfism and high accumulation of chlorophyll during the vegetative phase. Irrespective of early or delayed senescence, transgenic plants showed reduced grain yields. Our results indicate that OsY37 acts as a positive regulator of heading and senescence during the reproductive phase in rice. In addition, OsY37 may be involved in plant development and grain yield.

1. Introduction

Plants undergo distinct developmental phases during their life cycle; mainly vegetative, reproductive, ripening, and the senescence phase. Senescence is the final stage of development in plants during which metabolic and morphological changes occur, leading to the death of the whole plant. During leaf senescence, chlorophyll degradation takes place in the earliest stages [1], and proteins and nucleic acids are degraded by newly synthesized proteases [2,3] and nucleases [4,5] respectively. The resulting catabolic products are mobilized from leaves for use in the growing parts of the plant or to be stored in developing seeds. Therefore, senescence is characterized by the transition from nutrient assimilation to nutrient remobilization [6,7,8,9]. In particular, flag leaves surviving close to the end of senescence, may provide carbon source to sink organs and play a major role in grain yield compared to the other leaves [10]. Therefore, natural leaf senescence is recognized not only as a negative, aging process but also as involving active biochemical and physiological events. All these events have strong effects on the relocation of nutrients and as a consequence, on grain quality in wheat [11] and crop yield in rice [12] as well as in other crop plants [13]. In rice plants, flowering immediately follows heading and these two processes precede leaf senescence progress. Flowering and senescence are strictly defined serial events and intimately associated. For example, Arabidopsis ELF3 (early-flowering 3) is a regulator of floral induction [14] and its homolog OsELF3.1 promotes leaf senescence in rice [15]. Furthermore, natural senescence is also closely associated to oxidative and abiotic stresses. Hydrogen peroxide triggers production of reactive oxygen species, which induce the expression of a number of senescence associated genes [16], and furthermore the induced senescence promotes accumulation of hydrogen peroxide as a positive feedback mechanism [17].

Comprehensive analysis for metabolic, physiological, and genetic alterations during leaf senescence and its close relationship to senescence-associated gene (SAG) expression have been extensively studied [16,18,19,20,21,22]. Alteration of gene expression profile is primarily controlled by transcription factors. NAC (NAM, ATAF, and CUC) transcription factor genes were originally discovered as morphological determinants in plants. For example, the nam (no apical meristem) mutant of petunia is unable to develop a shoot apical meristem [23], and double mutations in cuc1 and cuc2 (cup-shaped cotyledon) of Arabidopsis cause cup-shaped cotyledons [24]. NAC family genes encode the polypeptides consisting of a highly conserved N-terminal half segment (NAC domain) and a unique C-terminal half region. The NAC domain contains a nuclear localization signal and a characteristic DNA-binding domain, whereas the C-terminal motif is considered to determine the specificity of the target genes and hence to participate in diverse aspects of physiological functions. Based on the subdomain configurations, NAC family genes are classified into several classes [25,26,27]. It was later found that NAC genes regulate positively and negatively the expression of a number of SAGs. Additionally, previous studies demonstrated that overexpression of OsNAP, OsNAC2 from rice and NAC016 from Arabidopsis accelerates leaf senescence, whereas lowered expression of these genes in mutants or via RNAi (RNA interference) inhibition results in delayed senescence [12,28,29]. Similarly, Arabidopsis ANAC032 [17] and ORS1 (PRE1 SISTER1) a paralog of ORE1 (ORESARA1)/ANAC092 [30] promote leaf senescence in Arabidopsis. In contrast, similar experiments indicate that the NAC genes, JUB1 (JUNGBRUNNEN1) and VNI2 (VND-INTERACTING2) of Arabidopsis delay and hence suppress senescence [31,32].

Here we report that the expression of the novel NAC gene of rice (ONAC011, Os06g0675600), which we designated OsY37^I (Oryza sativa Yellow37) from indica rice and OsY37^N from japonica rice, is induced during natural leaf senescence. Furthermore, transgenic rice plants overexpressing the OsY37 gene showed promoted heading and natural senescence. In contrast, knockdown of the gene delayed heading and senescence and also affected normal vegetative growth and development.

2. Results

2.1. Isolation of the OsY37^I Gene

To identify suitable leaf materials for isolation of SAGs from rice, we first measured the chlorophyll content of the flag leaf, second and third leaves, from the top, after the heading date. These leaves were excised, extracted, and chlorophyll was measured using a spectrometer. Chlorophyll contents decreased gradually starting from the heading date to reach their lowest value 60 days later. At 34 days after heading, the chlorophyll content was reduced to 50% of the initial concentration in the flag leaf and second leaf. At this time point, equal amounts of these leaves were mixed and used for RNA isolation. The first and second green leaves from the top were harvested 10 days before heading and used as control. Total RNA was subjected to differential display using random forward 10-mer primers and an oligo-dT reverse primer. One forward primer (CAGGCCCTTC) preferentially amplified a fragment of 251 bp from senescent leaves compared with green leaves (Figures S1 and S2). Other amplified fragments were similarly amplified in both tissues, but the intensity of some fragments was larger in green leaves compared with senescent leaves. The full-length clone of the 251 bp-segment was obtained from the cDNA library from indica rice, and sequence analysis revealed that the gene corresponds to the NAC family gene ONAC011 (Os06g0675600) from japonica rice [25]. Thus, we designated the cloned genes as OsY37^I and OsY37^N from japonica respectively. Both coding regions showed identical nucleotide sequences except that a 9 bp stretch is absent in OsY37^I and three nucleotide positions are different between the two genes. OsY37^I and OsY37^N encode polypeptides comprising 301 and 304 amino acid (aa) residues, respectively, and they differ from each other in 2 aa residues as a result of base substitution (Figure S2). Based on the close resemblance in gene structure between OsY37^I and OsY37^N, functional characteristics and regulatory mechanisms of the two genes are believed to be almost identical to each other.

2.2. Increased Accumulation of OsY37 Transcripts during Natural Senescence

Age-dependent increase of OsY37 transcript levels in the flag leaf was determined by Northern hybridization analysis. We used, as a probe, a 251 bp segment comprising the 3′-end of the gene with a unique sequence among the NAC family genes (Figure S2). Basal levels of transcripts of 1 kb and 3–4 kb were observed in young green top leaves at 47 days before heading (Figure 1a). Comparable transcript levels of both sizes accumulated as long as 23–30 days after heading (Figure 1b), indicating that the basal levels of OsY37 transcripts accumulated irrespective of developmental stage. Moreover, genomic information indicates that the total length of the genomic OsY37 comprises 4 kb, including three exons separated by two introns, whereas matured mRNA extended 906 bp in length (Figure S2).

Both of the 1 and 3–4 kb transcripts, started to accumulate to increased levels at 37 days after heading (Figure 1b). Accumulation of OsY37 transcripts increased during the following developmental stages to attain their maximum levels at 56 days corresponding to the mid-senescence phase (Figure 1b).

2.3. Induction of OsY37 Expression Under Nitrogen Starvation Conditions

Application of nitrogen fertilizer is crucial to obtain appropriate crop yields and hence limitation of the nitrogen source promotes leaf senescence [33]. Considering the possible involvement of the OsY37 gene in this process, we determined OsY37 transcript accumulation upon nitrogen starvation. Rice seedlings were grown in natural soil containing complete nutrients, and then transferred to hypotonic growth conditions with or without a nitrogen source. Total RNA was isolated and subjected to Northern hybridization analysis. OsY37 transcripts of both 1 and 3–4 kb sizes were highly accumulated at 12 and 19 days after nitrogen starvation, whereas the transcript levels remained low when a nitrogen source was fully supplied (Figure 2).

2.4. Nuclear Localization of GFP/OsY37 Fusion Polypeptide

To substantiate the relevance of OsY37 as a transcriptional regulator, we constructed the chimeric gene fusion GFP/OsY37^I under the control of the CaMV35S promoter, and then examined whether the fusion polypeptide is localized in the nucleus of plant cells. The chimeric construct, pGFOSY containing CaMV35S::GFP:OsY37^I fusion as well as psGFP(S65T) containing CaMV35S::GFP constructs were introduced into onion epidermal cells by particle bombardment. The fluorescence signal of GFP:OsY37^I polypeptide was localized exclusively in the nucleus (Figure 3a,b), whereas the signal of GFP alone was distributed throughout the cell including the nucleus (Figure 3c).

2.5. Effects of Overexpression and Knockdown of OsY37 on Heading Date and Leaf Senescence

To explore the physiological function of the OsY37 gene in regulating leaf senescence in rice, we created a chimeric construct in which the OsY37 gene expression is driven by a strong, constitutive CaMV35S promoter and transferred the construct to rice cv. Nipponbare and Kinmaze plants. The resulting transgenic plants were overexpressing (OE) the OsY37 gene. The number of days from sowing to heading of wild-type plants ranged from 95 to 97 days for Nipponbare while it ranged from 108 to 110 days for Kinmaze (Figure 4). We employed the two rice cultivars since they exhibit different heading dates, therefore effects of the overexpression could be better observed. Indeed, after 2–3 days from heading, rice plants set flowers in both cultivars and senescence symptoms started upon flowering. The OE transgenic plants grew normally without noticeable phenotypic alterations during the vegetative growth phase. The transgenic lines of both cultivars, however, showed a significant reduction in the number of days to heading compared with wild-type plants. Nipponbare OE lines started heading 5–12 days earlier, while in Kinmaze OE lines, heading occurred 9–13 days earlier than in the respective wild-type (Figure 4a,b).

Later, OE lines showed an early leaf senescence compared with wild-type plants. Visual inspection of the plants revealed that early senescence is most clearly recognized in the flag leaves (Figure 5a,b). The yellowing color of the flag leaves started from the distal end and progressively extended proximally with time.

To examine the progress of the senescence process in more detail, we determined the profile of the chlorophyll content in flag leaves for wild-type and transgenic rice plants upon heading.

Chlorophyll contents decreased in the wild-type and the pattern of declining chlorophyll shows a biphasic profile (Figure 6). At the first stage, chlorophyll levels started to decrease immediately in Nipponbare and 20 days after heading in Kinmaze respectively, but the rate of decline is gradual in both plants (Figure 6). In contrast, at the second stage (30–40 days after heading), chlorophyll contents started to decrease rapidly to reach their lowest levels at the end of this stage (Figure 6).

Our results also showed that chlorophyll levels were comparable among transgenic and wild-type plants at heading time. However, chlorophyll degradation of OE lines started at an early stage compared with wild-type (Figure 6a,b). Indeed, at 20 days after heading, chlorophyll content was reduced by 10% to 20% in OE plants of both Nipponbare and Kinmaze compared with wild-type. Consequently, at 50 days after heading, 50% of chlorophyll content was degraded in most of the OE plants, whereas it was maintained at 80% for wild-type. Induced senescence by the gene overexpression continued to proceed and was manifested by continuous chlorophyll degradation until the end of this process (Figure 6a,b). Our results further show that early senescence phenotypes exhibited by the OE plants are not merely the consequence of the early heading. Indeed, significant difference was observed between the rate of chlorophyll degradation and that predicted from the differential heading date between wild-type and OE plants.

Gene expression of the OE lines was investigated by semi-quantitative reverse transcription polymerase chain reaction (Clontech) analysis. In the case of wild-type Kinmaze and Nipponbare, transcript accumulation was non-detectable at the heading date, but increased gradually with time to reach the maximum levels at 70 days when chlorophyll had declined to the lowest levels (Figure 7a,b). In contrast, transgenic OE plants showed high accumulation of transcripts even at 0 day and remained constant or slightly increased (the Kinmaze line, OEK8) throughout the senescence periods (Figure 7a,b). These results indicate that the promoted senescence as well as the heading time of the OE transgenic plants are the consequence of stably increased expression of OsY37.

We next investigated physiological alteration caused by repressing the internal OsY37 gene expression by introducing CaMV35S::OsY37^I/RNAi construct to rice. Small interfering RNAs generated from RNAi construct knockdown the endogenous target transcript levels. We additionally examined for the hygromycin-sensitive (Hgr⁻) segregates derived from the primary transformed plants of the T1 generation. The RNAi lines from both cultivars showed slightly reduced growth rate during the vegetative phase (Figure 6c,d). When they reached maturation, all the RNAi lines showed delayed heading by 12–22 days compared with respective wild-type plants (Figure 4c,d). Senescence of RNAi lines is delayed by approximately 30 days compared with wild-type. RNAi lines exhibited a dark green color phenotype (Figure 5c–f). By contrast, flag leaves of wild-type and Hgr⁻ segregates lines, N5 (Nipponbare) and K7 (Kinmaze), exhibited yellow color at an early stage after heading (Figure 5e,f). Correspondingly, for both rice cultivars, chlorophyll contents started to decline 50 days after heading in RNAi lines, whereas they started decline immediately for wild-type Nipponbare and 20 days after heading for wild-type Kinmaze (Figure 6c,d).

RT-PCR analysis revealed that accumulation of OsY37 transcripts varied among plants. At the interval of 50–70 days after heading, characterized by significant differential chlorophyll content between wild-type and transgenic rice plants, wild-type and Hgr⁻ segregates of Nipponbare showed almost equal levels of transcript accumulation. By contrast, the RNAi transgenic lines except for RNN12, showed no or substantially less transcript levels (Figure 7c). The RNN12 line showed only slightly less amounts of transcripts but comparable to wild-type levels at 60 days. At the end of the senescence process, 70 days after heading, the transcripts were uniformly accumulated among all plants (Figure 7b). This result can be explained by the reduction of the creation of small fragment derivatives from RNAi at the latest stages of senescence.

Similarly, for Kinmaze, transgenic lines displayed low accumulation level of OsY37 transcripts in line RNK18, and no accumulation in line RNK19. At 60 days, transcripts were reduced in both lines and did not resume even at 70 days. However, in wild-type and Hgr⁻ segregates (K7) transcripts were continuously present from 50 to 70 days (Figure 7d). These results indicate RNAi construct effectively suppressed intragenic OsY37 transcript accumulation and that interference of OsY37 gene expression is responsible for the delay of heading time and leaf senescence in rice.

To confirm the results of knockdown of the OsY37 gene expression using RNAi, we employed the CRES-T method. In this technology, the repression polypeptide domain termed SRDX(SUPRD)X (where X denotes any amino acid) is fused to the C-terminal end of the transcription factor. Upon the introduction of this fused construct to plants, the SRDX-tagged polypeptide suppresses dominantly and specifically the corresponding intragenic transcriptional activity [34].

We created the SRDX construct where OsY37^N-SRDX is regulated under the control of its own promoter and transferred it to Nipponbare, aiming at low expression levels of the chimeric gene during the vegetative phase and activation during senescence. Hence, endogenous OsY37 activity is specifically controlled irrespective of the developmental stages of the plants. The transgenic plants exhibited varying degrees of dwarf phenotype. In particular, SRN3 line showed extremely defective growth with only 40% height of wild-type plant at full maturity. In addition, this line developed a single tiller and a single panicle with a limited number of spikelets (Figure 5g). Heading date of the SRDX lines was delayed by 15–25 days compared with wild-type (Figure 4e and Figure 5g). At the heading date, transgenic plants maintained a level of chlorophyll 5–15% higher than wild-type, resulting in dark green leaf phenotype as described above. Subsequently, all the transgenic lines showed a reduced rate of chlorophyll degradation profile. Indeed, for wild-type, the chlorophyll content declined to the lowest level 70 days after heading, while transgenic plants exhibited stay-green phenotype and maintained chlorophyll at 55% and 75% in SRN2 and SRN3 at 70 days respectively (Figure 6e).

OsY37-SRDX transcripts were determined by RT-PCR using the OsY37-specific forward primer (OsyRri1) and SRDX-specific reverse primer (SRDX-R). Basal transcript levels were recognized at the onset of heading (0 day) in both lines SRN2 and SRN4 and slightly higher levels in SRN3 (Figure 7e). The transcript levels increased gradually over time in a senescence-dependent manner in the lines SRN2 and SRN4 (Figure 7e). This increment reflects that OsY37-SRDX is regulated under the control of OsY37 promoter which is activated during senescence. RT-PCR for the SRN3 line was performed at 0 day only since this dwarf plant developed a single small flag leaf used for RNA preparation at this time. Nonetheless, OsY37 expression levels after the heading time correlated well with the extent of delayed senescence.

2.6. Effects of Modulated OsY37 Versions on Vegetative Growth and Plant Development

To investigate the involvement of OsY37 in plant growth and development, we created SRDX-tagged OsY37^I under the control of strong and constitutive promoters, CaMV35S and Act1. Subsequently, we attempted to introduce the resulting chimeric gene to rice plants through Agrobacterium-mediated gene transfer. Agrobacterium harboring the chimeric gene was infected to embryonic unorganized rice callus, selected for Hgr^R. The resulting callus was transferred to regeneration medium. Although Hgr^R callus were successfully obtained, the regeneration of plants from the callus to shoots failed. In contrast, 10–20% regenerated transgenic plants were consistently obtained upon transformation by the vector alone. We performed the same transformation experiments three times; however no regeneration of shoot was obtained. These results indicate that the repressor version of the OsY37 protein, under the control of CaMV35S and Act1 in rice cells interferes with the normal regeneration process. In addition, we transferred these constructs to Arabidopsis thaliana to observe phenotypic alteration. All of the resulting transgenic seedlings showed dwarfing phenotype with less than 80% in height compared with wild-type plants (Figure S3). These results are a first indication that OsY37 might be involved in normal growth and development of rice plants in addition to controlling the senescence process.

2.7. Effects of Modified OsY37 Genes on Grain Yield in Rice

Since overexpressing and repressing versions of OsY37 affected heading time and senescence progression, grain yields are susceptible to be affected as a result. Mature rice grains were harvested and counted. Compared with wild-type, OE lines showed reduced yield by 85–93% and RNAi lines showed 35–75% reduction. Thus, we found that both OE and RNAi lines showed reduced grain yields of different degrees (Figure 8). For SRDX lines, grain yield was severely affected, so that only a few seeds were harvested. These results show that the precocious senescence of flag leaves caused an incomplete grain-filling stage for OE lines. In contrast, for RNAi plants the suppression of OsY37 generated a reduced number of tiller compared with wild type resulting in less grain yield per plant (Figure 5c,d).

3. Discussion

3.1. OsY37 Gene Expression Profile

Northern hybridization analysis using 3′ end side of the OsY37 gene segment as a probe showed two distinct transcripts; 1 kb mRNA and an additional large 3–4 kb transcript. The 3–4 kb transcript most likely represents the precursor molecules of matured OsY37 transcript, since the used probe is highly specific to OsY37 transcript. Additionally, the size of the recognized transcript corresponds to the predicted size based on genome organization of the OsY37 gene. Both transcripts accumulated at basal levels in vegetative, healthy green leaves as well as in leaves undergoing early senescence. Accumulation of transcripts was highly induced during subsequent progression of senescence. This OsY37 gene expression profile is consistent with the existing database RiceXpro (http://ricexpro.dna.affrc.go.jp/), in which the OsNAC011 is preferentially expressed in the leaf blade during the grain ripening period, vegetative and reproductive phase of roots, the anthers, and ovary, whereas this expression is limited in vegetative leaves.

As we discuss below, OsY37 expression promotes heading and hence flowering. Since these events occur before the senescence phenotype is observed, basal levels of the transcripts appear to be required to mediate signals leading to heading. By contrast, high transcript levels are associated with chlorophyll degradation and therefore, responsible for the induction of the irreversible senescence phase [33].

Expression of OsY37 gene was also promoted by nitrogen starvation. This transcriptional response is consistent with a previous finding that nitrogen limitation induces senescence in Arabidopsis [35] and sunflower [33]. Based on the observation of expression profile of the OsY37, we conclude that this gene possesses dual functions; to incorporate or mediate internal triggers towards heading/flowering and to promote senescence.

3.2. Dual Functions of the OsY37 Gene in the Reproductive Phase

In the OE lines, the vegetative phase was restricted and transition to the reproductive phase occurred utmost 2 weeks earlier than for wild-type. Hence OE lines showed early heading/flowering. Precocious leaf senescence followed heading/flowering as evidenced by the immediate chlorophyll degradation upon heading. Conversely, heading/flowering time and senescence were significantly delayed in RNAi and SRDX lines. Consistent with these phenotypic alterations, OsY37 gene expression levels are constantly high in OE lines. In contrast, intragenic OsY37 transcripts were maintained at basal levels by RNAi construct. However, comparable levels of transcripts were observed at 70 days after heading in wild-type and RNAi lines of Nipponbare, while in Kinmaze RNAi lines, no transcripts were observed. Nonetheless, the resumption of OsY37 transcript accumulation in Nipponbare RNAi lines did not affect the results of senescence since this resumption occurred after establishment of the senescence process.

Taken together, our results suggest that OsY37 positively regulates both heading/flowering and senescence in rice. It has been reported that the other NAC genes, rice OsNAC2 [12] and Arabodpsis ANAC046 [36] are positive regulators for chlorophyll catabolic genes and hence promote leaf senescence. Although whether the OsY37 gene is involved in direct catabolism of chlorophyll remains to be determined, the OsY37 gene expression levels are associated remarkably with chlorophyll degradation. Overall, our study shows that the OsY37 gene product can be involved in the modulation of common or mutually independent signaling pathways leading to heading, chlorophyll degradation, and senescence in rice.

Little literature describes molecular mechanisms underlying the linkage between flowering and leaf senescence. However, it is known that stress-induced senescence is closely linked to reactive oxygen species (ROS) [37] and that the NAC gene ANAC032 mediates ROS-induced senescence [17]. Further, circadian clock regulates ROS homeostasis in Arabidopsis [38]. These results suggest that physiological transition is related to ROS status. Furthermore, ascorbic acid, which is a scavenger of ROS, promotes both flowering and senescence [39]. Further study is necessary to determine whether ROS involved in both flowering time and senescence are regulated by the OsY37 gene in rice.

Similar to wild-type, heading/flowering precedes leaf senescence in transgenic plants. However, modulated OsY37 gene expression affected both processes in transgenic plants. This result suggests that additional factors are required to primarily determine these two events and that the OsY37 gene can promote the activities of these putative factors.

Finally, both overexpressing and repressing OsY37 versions showed reduced grain yield compared with wild-type. In OE lines, the seed sink-strength may have not yet been established due to restricted vegetative phase. Furthermore, OE lines showed severely damaged flag leaves resulting from an early senescence observed immediately after heading. Hence the source strength of flag leaves to seeds may be weakened, due to deprivation of nutrient remobilization, the key step for grain filling. In agreement with these findings, it was reported that flag leaf contributed to 45% of rice grain yield and, when removed, was the major component for yield loss [10]. Contrary to OE versions, RNAi, and SRDX repressing versions prolonged the vegetative phase compared with the wild type plants. Although RNAi lines showed reduced yield, the grain yield of OE lines was more severely affected than RNAi lines. Reduced yield is most likely due to limited tiller number compared with wild-type. Furthermore, few grains were obtained from SRDX lines, where the number of tillers was severely affected, resulting in reduced number of panicles as exemplified by the fact that only a single panicle emerged in the SRN3 line. In addition, Vergara et al. showed that rice plants with a duration time of the vegetative phase longer than the critical duration time produced the same number of panicles, but formed fewer spikelets per panicle compared with wild-type plants [40].

In contrast to our findings, RNAi lines of OsNAC2 and OsNAP delayed senescence in rice and produced high yield via modulation of abscisic acid metabolism [12,41]. It is not known, however, if these genes also affect vegetative growth and heading time. Appropriate physiological states such as duration time of the vegetative phase, can be differentially regulated by the OsNAC2, OsNAP, and OsY37 genes, which are influential in determining the final grain yield. Since the time required for seed maturation remained unchanged irrespective of OE or RE OsY37 versions, coordination of grain filling and senescence seems to be an important factor to obtain full grain yield.

3.3. Involvement of OsY37 in Vegetative Growth and Development

Our results showed that the OsY37 gene is possibly involved in vegetative growth. Many arguments support our hypothesis. Initially, the OsY37-SRDX construct caused dwarfism when it was transferred to Arabidopsis. Although all transformations were successful, regeneration to shoots in OsY37-SRDX with a strong promoter failed in rice. We cannot strictly exclude that transformation of the OsY37-SRDX construct to rice requires different growth conditions to be accomplished successfully. Second, the OsY37 promoter-driven OsY37^N-SRDX generated dwarfing plants with a limited number of tillers. Therefore, repression of the OsY37 gene activities via CRES-T prevents normal growth and development during the vegetative phase. Furthermore, leaves of the SRDX lines showed a dark green phenotype even at heading time, suggesting that knockdown of the gene expression prevented chlorophyll degradation during the vegetative phase. Therefore, it appears that the OsY37 gene is involved in chlorophyll degradation not only during senescence but also during the vegetative phase.

3.4. Functional Comparison with the Other NAC Genes

To date, senescence associated NAC genes comprise both types; promoting senescence (ANAC032 [17], ANAC046 [37], AtNAP/ANAC029 [42], ORE1/ANAC092 [43], ORS1/ANAC059 [30], OsNAP [41], OsNAC2 [12], ONAC106 [44]) and delaying senescence (JUB1/ANAC042 [31], VNI2/ANAC083 [32]). Among these genes, ANAC032, ANAC046 from Arabidopsis and OsNAP and OsNAC2 from rice positively regulate chlorophyll catabolism genes. Since OsY37 promotes chlorophyll degradation, the gene can be classified in this group of NAC genes. Among these genes, ANAC032 as well as OsNAP expression is inducible by the phytohormone abscisic acid (ABA), suggesting that these genes promote senescence via ABA signaling pathway. In contrast, ANAC046 is not inducible by ABA despite the senescence promoting characteristics of the gene. Therefore, it is suggested that triggering senescence is not limited to a single pathway. Further, VNI2/ANAC083 shows ABA responsive expression but delays senescence. Moreover, oxidative stress is shown to promote senescence [45,46]. It was reported that nitrogen starvation in Arabidopsis roots induces accumulation of H₂O₂, a reactive oxygen species [47]. Our results show that OsY37expression was induced upon nitrogen starvation. On one hand, both ORS1/ANAC059 and JUB1 = ANAC042 are responsive to H₂O₂, a reactive oxygen species, but on the other hand, ORS1 gene promotes senescence, whereas JUB1 delays it. Therefore, some genes are apparently involved in common signaling pathways, but phenotypic consequence is not limited to a single type. Considering that OsY37 expression was induced upon nitrogen starvation and that some NAC genes regulate senescence via ABA and ROS pathway, we conclude that OsY37 can be involved in a similar senescence regulation pathway. Nevertheless, further analysis is required to identify the OsY37 specific pathway.

4. Materials and Methods

4.1. Plant Materials and Growth Conditions

Wild-type rice (Oryza sativa) subsp. japonica cv. “Akitakomachi”, “Nipponbare” and “Kinmaze” were grown in the paddy field of Akita Prefectural University from April to October. For the laboratory work, wild-type and transgenic rice plants were grown in a greenhouse under 16 h light/8 h dark cycle at 28 and 25 °C respectively.

4.2. Plasmid Vectors and Primers

The pGEM-T (Promega, Madison, WI, USA) and pGEM-T Easy vectors were used to clone PCR-amplified DNA segments. We used the psGFP (S65T) expression vector which carries a modified green fluorescent protein (GFP) gene sGFP (S65T) under the control of the cauliflower mosaic virus 35S promoter; CaMV35S [48]. To clone the suppressing version of OsY37, we used the plasmids pActSRDXG and p35SSRDXG carrying rice actin, Act1 promoter, and CaMV35S promoter respectively together with the transcription factor repression domain; SRDX comprising amino acid residues, GLDLDLELRLGFA. The japonica OsY37^N full-length clone in the vector pME18SFL3, was obtained from the National Institute of Agrobiological Sciences (http://www.dna.affrc.go.jp/jp/). Versatile plant binary vector, pIG121Hm, confers hygromycin resistance (Hgr^R) to rice plants and Agrobacterium tumefaciens [49] and kanamycin resistance to A. tumefaciens and Escherichia coli. Primers used in the present report are listed in

Table 1. List and sequence of primers.

Primer	Sequce
FwOsGFP	AGGAGGTGTACAAGATGAGCGGGATGAATTCGCTT
M13Rv	CAGGAAACAGCTATGAC
IndFw	TCGAGGAGGAGGGGGTGGGC
IndRv	GCAAATAATGTGAACTAGAATGGTAC
OsYIFw	ATGAGCGGGATGAATTCGCT
OsYIRv	ACTGAGTGAGTTCCACATTT
Fw24	TCGAGGAGGAGGGGGTGGGCATGA
Rv869	AGTTCCACATTTGTGTGAAGGCATGCTCAGCCAAGTAATCAAATCCATC
FwOsSR	ATGAGCGGGATGAATTCGCT
RvSRDX	ACTGAGTGAGTTCCACATTT
osy-2	GCATGAGCGGGATGAATTCG
osy752	GATCGCCCTACCCATATTGT
PromRv	GAATTCATCCGCTCATCCCGGGCCCCTCCTCCTCGAATTCCT
UPSPE	GGATTGTAAGATAAGTCCTC
OsyCRESTFw	GAATTCGAGGAGGAGGGGCCCGGGATGAGCGGGATGAATTCGCT
OsyCRESTRv	TATATATCTCCTTGGAAATTCCCGGGACTGAGTGAGTTCCACAT
hph1	ATGTCCTGCGGGTAAATAGC
hph2	CGTCTGCTGCTCCATACAAG
OsActinF318	GGTATCGTCAGCAACTGGGATGATATGG
OsActinR675	GCTCCGTCAGGATCTTCATGAGGTAATC
OsyPri1	TGCCCGGTGATGGTCGACGT
SRDX-R	TTAAGCGAAACCCAAACGGAGTTCTAG

.

4.3. Measurement of Chlorophyll Content

Leaves were excised from rice plants at different time interval after heading. Leaves were cut into pieces and were weighed immediately. The middle pieces weighing 100 mg were ground to a fine powder with liquid nitrogen and then used for extraction and measurement of chlorophyll content. Extracts were made from the leaf tissue by adding 1 mL extraction buffer composed of 0.4 M sucrose, 50 mM Tricine-KOH, 10 mM NaCl, 2 mM MgCl₂, 0.1% (w/v) bovine serum albumin, and 0.1% (w/v) polyvinylpyrrolidone at pH 7.8, and placed on ice. Acetone was added to the samples to give a concentration of 80% (v/v) and the solution was centrifuged at 800× g for 10 min at 4 °C. The supernatant was saved and absorbance at 663 nm (A₆₆₃) and 645 nm (A₆₄₅) was measured with a spectrophotometer (Gene Spec III, Hitachi Genetic Systems, Alameda, CA, USA). Chlorophyll content (µg mL⁻¹) was calculated using the formula 8.05 × A₆₆₃ + 20.29 × A₆₄₅ [50].

4.4. RNA Extraction and Northern Hybridization Analysis

Rice leaves (1 g) were ground to a fine powder in liquid nitrogen and immersed in 5 mL extraction buffer (50 mM Tris-HCl pH 8, 300 mM NaCl, 5 mM EDTA (ethylenediaminetetraacetic acid), 2% (w/v) SDS (sodium dodecyl sulfate), 10 mM β-mercaptoethanol and 1 mM aurin tricarboxylic acid). Total RNA was precipitated by LiCl [51]. Afterward, total RNA (40 µg) samples were separated by 1% agarose/formaldehyde gel electrophoresis and then blotted onto Hybond-N⁺ nylon membranes (Amersham/GE Healthcare, Chicago, IL, USA). Northern hybridization was carried out in a rapid hybridization buffer (Amersham). Membranes were washed under high stringency washing conditions in 0.1× SSC (standard saline citrate) and 0.1% SDS at 65 °C. A 251 bp OsY37 fragment from the pGEM200 vector was radiolabeled with [α-³²P]-dCTP (50 µCi) and used as a probe. Alternatively, for RT-PCR analysis, RNA was extracted from 100mg of green or senescing leaves using RNeasy Plant Mini Kit (QIAGEN, Hilden, Germany), then quantified fluorometrically using Qubit RNA BR Assay Kit (ThermoFisher Scientific, Waltham, MA, USA).

4.5. Differential Display

To identify genes differentially expressed in green leaves and naturally senescent leaves, we used differential-display reverse transcription kit (Sawady Technology, Tokyo, Japan). The upstream primers used were 10-mer oligonucleotides with an arbitrary sequence (Operon Technologies Inc., Alameda, CA, USA) and the downstream primers were 11-oligo-dT nucleotides with an additional two degenerate nucleotides at the 3′ end. Total RNA (1 µg) isolated from young, green and yellow leaves were subjected to reverse transcription to generate cDNA using the downstream primers. Subsequent PCR amplification of the cDNA using the upstream and downstream primers was performed in the presence of [α-³⁵S]-dATP (3.7 × 10⁴ Bq µL⁻¹) under the condition of 40 cycles of 30 s at 94 °C, 90 s at 40 °C, and 60 s at 72 °C, and one cycle of 1 min at 72 °C. The amplified radiolabeled DNA was separated by 5% polyacrylamide/8 M urea sequencing gel electrophoresis, followed by autoradiography. The portion of the gel containing the fragments showing higher intensity in senescent leaves than in green leaves was excised, immersed in 400 µL of 10 mM Tricine (pH 9.5), 0.2 mM EDTA and incubated at 94 °C for 5 min to dissolve the corresponding fragments. Aliquots were used for re-amplification of the fragment using the same primers as in the first PCR. The final product (251 bp length) was cloned into the pGEM-T vector to generate pGEM200.

4.6. Full-Length cDNA Library Screening and Sub-Cloning

A Rice 5′-STRETCH cDNA library was purchased from Clontech Laboratories (Palo Alto, CA, USA). The library carried total full-length cDNA and was constructed from etiolated seedlings of rice (O. sativa subsp. indica “IR36”) and cloned at the EcoRI restriction site of the λgt11 vector. Phages were propagated in Escherichia coli Y1090r⁻ cells. Approximately 10⁶ phage plaques were screened for the presence of the appropriate cDNA using, as a probe, the 251 bp senescence-specific fragment obtained from pGEM200 [52]. To clone the full-length cDNA into a plasmid vector, DNA was prepared from the λgt11 recombinant phage carrying the full-length segment, and was further amplified by PCR using the forward primer IndFw situated 20 bp upstream of the initiation codon, and the reverse primer IndRv situated 102 bp downstream of the termination codon and extended immediately upstream of a polyA sequence. The amplified segment was cloned into the pGEM-T vector to generate pGE37 (Figure S4).

4.7. Generation of Recombinant Clones

To generate a chimeric construct encoding the GFP/OsY37^I fusion polypeptide, the full-length OsY37^I cDNA was cloned into the vector plasmid sGFP (S65T). The vector contains the unique BsrG and NotI restriction sites situated at the 3′-end of the GFP coding region and 7 bp downstream of the BsrG site, respectively. OsY37^I in pGE37 was amplified using the forward primer FwOsGFP and the M13 universal RV primer. The FwOsGFP primer contained the BsrGI (TGTACA) restriction site and the ATG initiation codon located 3 bp downstream of the BsrG site. The amplified fragment was subjected to double digestion with BsrG and NotI (derived from the vector), and ligated to the expression vector that had been digested with BsrG and NotI, thus generating the clone pGFOsY. This clone harbored the OsY37^I polypeptide, which was immediately fused to the C-terminal end of the intact GFP polypeptide.

To obtain overexpression construct for OsY37^I, the full-length was cloned in the plant vector, pIG121Hm. In pGE37 plasmid, the restriction sites; SphI and SacI are closely located upstream and downstream of the full-length coding region respectively (Figure S4). The pGE37^I was digested with SphI, blunt-ended by the T4 DNA polymerase and partially digested with SacI to avoid additional cut at the SacI site, occurring in the intragenic region of the OsY37^I. The digests were separated by agarose gel electrophoresis and the appropriate segment was recovered from the gel. The plant vector pIG121Hm was digested with XbaI, filled-in by the Klenow fragment, and finally digested with SacI. The resulting vector was ligated to the above-mentioned OsY37^I segment generating the clone, pIG20 where the gene is regulated under the control of CaMV35S promoter and NOS (nopaline synthase) terminator.

The construction process for repressing version against intragenic OsY37^I gene activity, RNAi and CRES-T clones is diagramed in the supplemental figures. To construct RNAi-OsY37^I, the full-length clone pGE37 was cut with a unique site, Tth111I in the coding region (position 487 from the start codon, Figure S4), filled-in by the Klenow fragment and further digested with SphI. The resulting fragment comprises 3′-end side of pGE37 together with the vector (Figure S4, segment #1). Subsequently, the pGE37 was amplified by PCR with primers, forward Fw24 and reverse Rv847 to generate the segment #2 (Figure S4). The Fw24 primer is situated at 5′ flanking sequence of the OsY37^I and the Rv869 primer is located at position 847 and possesses additionally introduced SphI recognition sequence. The resulting fragment was digested with SacI, blunt-ended by T4 DNA polymerase, and then digested with SphI (Figure S4) segment #3 and ligated to the above-mentioned segment #1 to generate pRNAiOsY37. In this clone, two regions corresponding to C-terminal domain are situated in mutually reversed orientation and hence give rise to functional sequence for RNAi inhibition for intragenic OsY37. To clone into a plant vector, the pRNAiOsY37DNA was digested with SphI, blunt-ended by T4 DNA polymerase, then digested with SacI and ligated to pIG121Hm, previously digested with XbaI, filled-in by Klenow, and lastly digested with SacI. In this final clone, pRNAiOsY37, RNAi version of OsY37^I was regulated under the control of CaMV35S promoter (Figure S4).

The SRDX-tagged OsY3^I gene was driven either under the control of CaMV35S promoter or its own promoter. To this end, the full-length OsY37^I gene segment was amplified with the forward primer OsYFw and the reverse primer OsYRv, filled-in with T4 DNA polymerase and ligated to the SmaI restriction site of the pActSRDXG and p35SSRDXG vectors to generate pPAct37SRDXG and pP35S37SRDXG, respectively (Figure S5). These constructs were transferred to the plant vector pBIG-Hgr, conferring Hgr^R in plants, using the Gateway vector conversion system [53] (Invitrogen, Carlsbad, CA, USA) to produce pBIGAct37SRDXG and pBIG35S37SRDXG (Figure S4).

To obtain the SRDX version under the control of the OsY37 gene, promoter region of OsY37^I was isolated from the genomic library constructed in λ phage vector (Clontech) using the coding region as a probe. Insert DNA from the identified phage was digested with SalI and further subcloned in the SalI site of pUC19, generating the clone, pProm17 (Figure S6). Hence we obtained 3.8 kp fragment including 169 bp coding region with the initiation codon of OsY37^I. Upstream of the initiation codon, only small open reading frames with no more than 103 amino acid residues aside from OsY37^I were found in either forward or reverse orientation. To introduce SmaI site next to the initiation codon of the OsY37^I, pProm17 was amplified by PCR with UPSPE (located −884 bp upstream of the initiation codon) and PromRv (possessing SmaI site together with ATG initiation codon), then cloned in pGEM-T Easy vector to generate pProm17seg1 (Figure S6). Faithful amplification was verified by DNA sequencing. The fragment was excised from the clone pProm17seg1 by SpeI (in the promoter region) and SmaI, then ligated to the pProm17 formerly digested with SpeI (in the promoter region, 729 bp upstream of the initiation codon) and SmaI (in the vector region), generating the clone pProm17-10 (Figure S6). In this clone, SpeI/SalI segment of Prom17 was replaced by SpeI/SmaI segment. pProm17-10 was digested with HindIII (in pUC19) and SmaI, and then ligated to pActSRDXG which had been digested with HindIII and SmaI, generating pProm17-10-1. In this clone, OsY37^I promoter was immediately followed by SmaI site and SRDX domain region. Afterward, full-length OsY37^N was amplified by PCR with primers OsyCRESTFw and OsyCRESTRv, both containing SmaI sites, and cloned in pGEM-T Easy vector, generating CR3. Accurate sequence of CR3 was determined. Further, coding region was excised from the CR3 by SmaI digestion, and then ligated to SmaI-digested pProm17-10-1. Right orientation of the coding region relative to the promoter was isolated, and termed pPromOsY37SRDX7 (Figure S6). In this clone, promoter is followed by coding sequence of OsY37^N fused with the SRDX domain. Both the plant vectors, pIG121Hm, and pPromOs37^NSRDX7 were digested with HindIII and ligated, to obtain co-integrate clone pIG24, which was used to transform rice plants.

4.8. Nitrogen Starvation

Rice seeds with intact coats were surface-sterilized with 1.5% pefurazoate (Hokko Chemical Industry Co., Tokyo, Japan) for 10 min, thoroughly washed with water and imbibed for 48 h. The seeds were sown on soil containing 0.6 g nitrogen, 0.8 g phosphorous, and 0.5 g potassium sources, then were allowed to grow in a greenhouse until the seedlings attained about 5 cm in height. Single seedlings were transferred to 250 mL vermiculite in 300 mL pots containing 100 mL tap. Two hypotonic solutions (20 mL) were used; one contained a nitrogen source (N-medium) while the other was devoid of a nitrogen source (N-free medium) for nitrogen starvation. The N-medium was composed of 1 mM (NH₄)₂SO₄, 15 mM KH₂PO4, 20 mM KNO₃, and 39 mM NH₄NO₃, as major nutrient sources; 0.5 mM MgCl₂, 0.5 mM CaCl₂, 2mM FeNa (EDTA), and 5 mM citric acid as minor sources; and 16 µM H₃BO₄, 22 µM MnSO₄, 37 µM ZnCl₂, 5.9 µM CuCl₂, 0.8 µM (NH₄)₆Mo₇O₂₄, and 4.1 µM CoCl₂ as sources of trace elements. For the N-free medium, N sources, (NH₄)₂SO₄, KNO₃, and NH₄NO₃ were replaced by 20 mM K₂SO₄. Hence seedlings were incubated in a growth chamber under constant light (300 µmol m⁻² s⁻¹) at 28 °C. The hypotonic nutrient solution was refreshed every 2 weeks.

4.9. RT-PCR

Total RNA (6 µg) was extracted using RNeasy Plant Mini Kit (QIAGEN), then treated with RNase-free DNase I (RQ1, 5 U; Promega, Madison, WI, USA) in 40 mM Tris-HCl (pH 8.0), 10 mM MgSO₄, and 1 mM CaCl₂ for 30 min at 37 °C. RQ1 was inactivated by phenol treatment, and the RNA was precipitated with ethanol. cDNA was synthesized from the RNA (1 µg) by reverse transcriptase using the Transcriptor First Strand cDNA Synthesis Kit (Roche, Indianapolis, IN, USA) in the presence of an RNase inhibitor (Rnasin Plus RNase Inhibitor; Promega). The enzymes were heat inactivated at 85 °C for 5 min. Aliquots were withdrawn and amplified by PCR using Taq DNA polymerase (Titanium DNA Polymerase; Clontech) with the following protocol: 94 °C for 30 s (1 cycle), then 94 °C for 30 s, 60 °C for 30 s, and 72 °C for 1 min (30 cycles), and finally 72 °C for 7 min (one cycle). The primers used to amplify OsY37 were the forward primer osy-2, which extends from 2 bp upstream of the initiation codon, and the reverse primer osy752 starting from 780 bp downstream of the initiation codon.

4.10. Transformation of Rice and Arabidopsis

Japonica rice Oryza sativa cv. “Nipponbare” and “Kinmaze” were transformed using the callus method [49]. Briefly, seeds were husked, surface-sterilized in 2% sodium hypochlorite for 25 min, then cultured on medium supplemented with 2,4-D (2,4-Dichlorophenoxyacetic acid, 1 µg mL⁻¹) and allowed to form callus. The callus was infected with A. tumefaciens strain EHA101 carrying a chimeric plasmid clone in the presence of acetosyringone (10 µg mL⁻¹) and selected for Hgr^R (50 µg mL⁻¹) resistance. Subsequently, callus was transferred to regenerated to shoots on medium containing NAA (naphthalene acetic acid, 2 ng mL⁻¹), kinetin (2 µg mL⁻¹), and Hgr (50 µg mL⁻¹). T1 generation of OsY37-overexpressing (OE) and OsY37-RNAi lines and primary transformants of OsY37-SRDX lines were subjected to further study.

Transformation of Arabidopsis thaliana ecotype Columbia was performed according to the floral dip method [54] by infection with Agrobacterium tumefaciens strain GV3101 carrying the plasmid pBC35SOsY37SRDXG or pBCActOsY37SRDXG. A total of 4000 T₁ seeds were surface-sterilized in 1.0% (w/v; chlorine) sodium hypochlorite together with 0.02% (w/v) Triton X-100 for 7 min. The sterilized seeds were sown on MS medium [55] containing 30 µg mL⁻¹ Hgr to obtain 40–100 transgenic seedlings.

4.11. Particle Bombardment

DNA was adsorbed to gold particles (1.0 µm diameter) in 125 µL solution containing 5 µg DNA, 3 mg gold particles, 1 M CaCl₂, and 0.016 M spermidine for 3 min at room temperature. The DNA/gold particle complexes were recovered by centrifugation (7000× g, 10 s) and dissolved in ethanol. The entire mixture was placed onto a micro-carrier and air-dried. Onion bulb epidermal peels were placed on basal MSB5 agar plates, composed of MS salts [55] supplemented with Gamborg B5 vitamins [56], 3% (w/v) sucrose and 0.8% (w/v) BA10 agar (Ina Food, Nagano, Japan). DNA/gold particles were bombarded twice onto the surface of the onion epidermal cells at 1100 psi working with a 900 psi rupture disk under vacuum (28 in. Hg) using a Biolistic PDS-1000/He Particle Delivery System (Bio-RAD, Hercules, CA, USA). The plates were incubated for 16 h of dark condition at 30 °C. The presence of GFP protein was examined by fluorescence microscopy (BX41, OLYMPUS, Nishi-Shinjuku, Tokyo, Japan) under ultraviolet light irradiation with appropriate filters. To visualize nuclei, samples were stained with 10 mL of 1 µg mL⁻¹ 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) dissolved in water for 10 min then viewed using a fluorescence microscope under ultraviolet light irradiation with appropriate filters.

5. Conclusions

We isolated a member of the NAC transcription factor family genes OsY37^I from indica and identified the gene to be an orthologue of OsNAC011 from japonica rice. OsNAC011 was renamed OsY37^N. Our results show that OsY37 is a senescence associated gene encoding a novel protein involved in plant growth and development; implicated in heading and flowering times, and presumably regulating the senescence process by promoting chlorophyll degradation in rice. We conclude that OsY37 acts as a positive regulator of senescence, however further analysis is required to determine the specific pathway in which OsY37 is involved.