Genome-Wide Identiﬁcation and Expression Analysis of Senescence-Associated Genes in Grapevine ( Vitis vinifera L.) Reveal Their Potential Functions in Leaf Senescence Order

: Natural leaf senescence is an acclimation strategy that enables plants to reallocate nutrients. In the present study, interestingly, we found that the basal mature leaves of grapevine primary shoots (P) exhibited the earliest senescence, followed by the apical young leaves of secondary shoots (ST), and then the basal mature leaves of secondary shoots (S). The Chl level decreased with the extent of leaf senescence. According to the genome-wide identiﬁcation and expression analysis, sixteen senescence-associated genes (SAGs) involved in Chl breakdown were identiﬁed in the grapevine genome. Their expression patterns showed that the transcript changes in VvSGR , VvPPH2 , and VvFtsH6-2 corresponded to the changes in Chl content among P, S, and ST. The changes in the transcription of VvNYC1 , VvSGR , VvPAO1 , VvPAO2 , VvPAO4 , VvPPH1 , VvPPH3, and VvFtsH6-1 only contributed to low Chl levels in P. The cis-element analysis indicated that these SAGs possessed several light- and hormone-responsive elements in their promoters. Among them, ABA-responsive elements were found in twelve of the sixteen promoters of SAGs. Correspondingly, ABA-signaling components presented various changes in transcription among P, S, and ST. The transcription changes in VvbZIP45 and VvSnRK2.1 were similar to those in VvSGR , VvPPH2 , and VvFtsH6-2 . The other nine ABA-signaling components, which included VvRCAR2 , VvRCAR4 , VvRCAR6 , VvRCAR7 , VvRCAR2 , VvPP2C4 , VvPP2C9 , VvbZIP25 , and VvSnRK2.3 , were highly expressed in P but there was no difference between S and ST, with similar expression patterns for VvNYC 1, VvSGR , VvPAO1 , VvPAO2 , VvPAO4 , VvPPH1 , VvPPH3 , and VvFtsH6-1 . These results suggested that the senescence of P and ST could be regulated by different members of Chl breakdown-related SAGs and ABA-signaling components. These ﬁndings provide us with important candidate genes to further study the regulation mechanism of leaf senescence order in grapevine.


Introduction
Leaf senescence is a complex but orderly physiological process that enables plants to remobilize nutrients from old leaves to developing organs [1]. It is induced both by endogenous signals such as aging and by various exogenous factors such as stress [2]. However, a series of senescence syndromes occur irrespective of stimuli that cause leaf senescence, including programmed cell death and the ordered degradation of the macromolecules [3]. Among the other degradation processes, chlorophyll (Chl) catabolism into non-green derivatives leads to a visual manifestation of senescence syndromes, i.e., leaf yellowing via a multi-step catabolic pathway [4]. The Chl b-to-a conversion is the first step of Chl degradation; it is catalyzed sequentially by two Chl catabolic enzymes (CCEs), i.e., Chl b reductase (CBR) and droxymethyl Chl a reductase (HCAR) [5,6]. The former is encoded by two genes, NON-YELLOW COLORING1 (NYC1) and NYC1-LIKE (NOL), in rice (Oryza sativa) [7] and Arabidopsis [5]. The next step is the removal of the Mg atom from Int. J. Mol. Sci. 2022, 23, 12731 2 of 13 Chl a, followed by the hydrolysis reactions. The PPH gene has been identified to encode pheophytinase, which catalyzes hydrolysis reactions in Arabidopsis [8]. The resulting pheophorbide a is oxygenolytically opened and further reduced to a primary nonfluorescent Chl catabolite by pheophorbide a oxygenase (PAO) and red Chl catabolite reductase (RCCR) [9]. Specifically, mutants deficient in HCAR (hcar) maintained green leaves during dark-induced and/or natural senescence, while HCAR-overexpressing plants exhibited more tolerance to reactive oxygen species than the wild type in rice (Oryza sativa) [10] and accelerated leaf yellowing in Arabidopsis [11]. The pph-1 mutant showed a stay-green phenotype since it was unable to degrade chlorophyll during senescence in Arabidopsis [12]. LpNOL RNA interference (NOLi) in perennial ryegrass (Lolium perenne L.) has been shown to significantly block Chl degradation in senescent leaves and also delayed the initiation and progression of leaf senescence [13]. In addition to CCEs, chloroplast protein NYE1, also known as STAY-GREEN1 (SGR1) has been shown to interact with CCEs at light-harvesting complex II (LHC II) for chlorophyll detoxification during leaf senescence in Arabidopsis [14,15]. The Arabidopsis genome processes three SGR homologs, i.e., SGR1/NYE1, SGR2, and SGRL. SGR1/NYE1 has been characterized to promote Chl breakdown during leaf senescence in many plants [14], while SGRL contributes to Chl degradation in pre-senescent leaves [16]. By contrast, SGR2 negatively regulates the degradation of Chl under age-, dark-, and stress-induced senescence conditions [17]. FtsHs are a family of ATP-dependent metalloproteases [18]. The chloroplast-located protease FtsH6 has been identified as being responsible for the degradation of LHC II during high-light acclimation and dark-induced senescence by using a reversed genetic approach [19]. These genes, such as CCEs, chloroplast protein, and FtsHs protease, are thought to be a type of senescence-induced senescence executers (SIEs), whose high expression levels directly result in senescence syndromes [20]. Recently, a study confirmed three abscisic acid (ABA)responsive element (ABRE)-binding transcription factors (ABF2, ABF3, and ABF4) that functioned as the direct positive regulators of two CCEs (PAO and NYC1) to promote ABA-mediated Chl breakdown and leaf senescence [21]. Zhao et al. (2016) [22] found that PYL9 over-expressing plants exhibited dramatically increased drought-induced leaf senescence and ABA-promoted leaf senescence via activating core ABA-signaling pathways (PP2Cs-SnRKs-ABFs). These genes belong to the senescence-induced senescence regulators (SIRs) group, in the context that they are activated during leaf senescence and regulate leaf senescence [20].
Grapevine, a worldwide important crop with highly commercial products such as table berries, juice, and wines, presents a full seasonal senescent syndrome. Leaf senescence is tightly related to crop yield and biomass production and needs massive changes in the expression level of SAGs [23]. A number of SAGs, including SIEs and SIRs, have been characterized in model plants [11,17,21]. Most recently, the A NAC transcription factor (DRL1) has been reported to be involved in the leaf senescence of grapevine [24]. CitPPH was directly regulated by an ethylene response factor involved in chlorophyll degradation during citrus fruit ripening [25]. However, genome-wide studies on a number of SAGs have still not been conducted in fruit trees, such as grapevine. During our experiment, we found that there was a certain order of leaf senescence in grapevine during autumn. The order of leaf yellowing depended on the leaf position. The aim of this study was to further investigate leaf senescence order in grape shoots, focusing on the changes in Chl contents and the genomic identification and expression analysis of Chl-related SAGs. The expression profiles of these genes in different nodes of leaves were indicative of distinct sets of Chl-related SAGs controlling the order of leaf yellowing. In addition, evidence was provided that ABA-related SAGs were also involved in the regulation of the senescence order of grapevine leaves.

The Impact of Leaf Position on Senescence of Grapevine Leaves
Generally, leaf senescence in shoots occurs gradually from bottom to top. In the present study, the bottom leaves on primary shoots, indeed, turned yellow first. Interestingly, we observed senescence in the top leaves of secondary lateral shoots prior to the bottom leaves of secondary lateral shoots during natural autumn leaf senescence ( Figure 1A,B). The leaves were collected and divided into three groups: the leaves on the first to second nodes of primary shoots (P group), the leaves on the first to second nodes of secondary shoots (S group), and the leaves on the tip of secondary shoots (ST group) ( Figure 1C). Consistent with their senescence order, the chlorophyll a (Chl a) level in the S group was the highest, followed by the ST group, and the Chl a content in the P group was the lowest ( Figure 1D). The Chl b content of the S group was obviously higher than that of the P and ST groups, while no difference was observed between the P and ST groups ( Figure 1E).

The Impact of Leaf Position on Senescence of Grapevine Leaves
Generally, leaf senescence in shoots occurs gradually from bottom to top. In the present study, the bottom leaves on primary shoots, indeed, turned yellow first. Interestingly, we observed senescence in the top leaves of secondary lateral shoots prior to the bottom leaves of secondary lateral shoots during natural autumn leaf senescence ( Figure 1A,B). The leaves were collected and divided into three groups: the leaves on the first to second nodes of primary shoots (P group), the leaves on the first to second nodes of secondary shoots (S group), and the leaves on the tip of secondary shoots (ST group) ( Figure 1C). Consistent with their senescence order, the chlorophyll a (Chl a) level in the S group was the highest, followed by the ST group, and the Chl a content in the P group was the lowest ( Figure 1D). The Chl b content of the S group was obviously higher than that of the P and ST groups, while no difference was observed between the P and ST groups ( Figure 1E).

Cis-Acting Regulatory Element Analysis of SAG Promoters
The cis-acting regulatory elements of the SAG promoters were analyzed ( Table 2). Except for a common cis-acting element in promoter and enhancer regions, different numbers and types of light-response elements were found in promoters of most SAG members, such as ACE, AE-box, G-box, TCT-element, Box4, GATA-motif, and MRE. The hormone-responsive elements were also identified in the SAG members. ABRE, the cis-acting element involved in abscisic acid responsiveness, was found in promoters of NYC1, NOL2, SGRL, SGR, RCCR, HCAR, PAO1, PAO2, PAO4, PPH2, FTsH6-1, and FTsH6-2. One auxinresponsive element (AuxRR-core) was observed in promoters of PAO3 and PAO4; two MeJA-responsive elements (CGTCA-motif and TGACG-motif) presented in promoters of NOL2, SGRL, PAO3, PAO4, and PPH2. Gibberellin-responsive elements, i.e., P-box and

Chl Degradation-Related SAGs Are Characterized by Diverse Expression Profiles in Different Leaf Positions of Grape Shoots
To identify the causal relationships between the order of grapevine leaf senescence and the transcript abundance of SAGs involved in Chl degradation, the expression level of 16 Chl degradation-related SAGs were quantified in different leaf nodes on grape shoots (Figure 3). In total, 15 SAGs were found to be differently expressed at one or three leaf positions: VvSGRL and VvRCCR were characterized by high expression levels in basal leaves on the secondary shoots (S) and low expression levels in basal leaves on primary shoots (P); VvSGR, VvPPH, VvFtsH6-1, and VvFtsH6-2 presented the highest expression level in P, but the lowest expression level in S; and VvHCAR showed increasing expression from P to ST. The expression level of the other six SAGs, including VvNYC1, VvPAO1, VvPAO2, VvPAO4, VvPPH1, and VvPPH1, were obviously higher in P as compared with S and ST, while VvNOL2 presented the opposite expression pattern. VvPAO3 exhibited a significantly higher expression level in ST but no differences between P and S. The diverse expression profiles of these genes in different leaf positions could suggest the specific function of each gene during leaf senescence.

Chl Degradation-Related SAGs Are Characterized by Diverse Expression Profiles in Different Leaf Positions of Grape Shoots
To identify the causal relationships between the order of grapevine leaf senescence and the transcript abundance of SAGs involved in Chl degradation, the expression level of 16 Chl degradation-related SAGs were quantified in different leaf nodes on grape shoots (Figure 3). In total, 15 SAGs were found to be differently expressed at one or three leaf positions: VvSGRL and VvRCCR were characterized by high expression levels in basal leaves on the secondary shoots (S) and low expression levels in basal leaves on primary shoots (P); VvSGR, VvPPH, VvFtsH6-1, and VvFtsH6-2 presented the highest expression level in P, but the lowest expression level in S; and VvHCAR showed increasing expression from P to ST. The expression level of the other six SAGs, including VvNYC1, VvPAO1, VvPAO2, VvPAO4, VvPPH1, and VvPPH1, were obviously higher in P as compared with S and ST, while VvNOL2 presented the opposite expression pattern. VvPAO3 exhibited a significantly higher expression level in ST but no differences between P and S. The diverse expression profiles of these genes in different leaf positions could suggest the specific function of each gene during leaf senescence.

ABA-Related SAGs Are Characterized by Consistent Expression Profiles with Chl Degradation-Related SAGs in Different Leaf Positions of Grape Shoots
ABA is a well-known regulator of plant senescence. Twelve out of sixteen Chl degradation-related SAGs were found to contain ABA-responsive elements in their promoters. Therefore, 14 orthologs of ABA-signaling components were focused on the transcription changes in the different leaf positions of shoots ( Figure 4). Five out of seven ABA receptors presented obviously high expression levels in P, but low levels in S and ST, including VvR-CAR2, VvRCAR4, VvRCAR5, VvRCAR6, and VvRCAR7. The ABA-signaling components, such as VvPP2C4, VvPP2C9, VvSnRK2.3, and VvZIP25, showed similar expression patterns. VvRCAR3 exhibited no significant differences among the three leaf positions, while VvRCAR1 was most abundant in S. VvSnRK2.3. VvZIP25 was significantly differentially expressed among the three leaf positions and presented the highest abundance in P but the lowest abundance in S. VvZIP08 showed a high expression level in ST but low expression levels in S and P. Therefore, 14 orthologs of ABA-signaling components were focused on the transcription changes in the different leaf positions of shoots (Figure 4). Five out of seven ABA receptors presented obviously high expression levels in P, but low levels in S and ST, including VvRCAR2, VvRCAR4, VvRCAR5, VvRCAR6, and VvRCAR7. The ABA-signaling components, such as VvPP2C4, VvPP2C9, VvSnRK2.3, and VvZIP25, showed similar expression patterns. VvRCAR3 exhibited no significant differences among the three leaf positions, while VvRCAR1 was most abundant in S. VvSnRK2.3. VvZIP25 was significantly differentially expressed among the three leaf positions and presented the highest abundance in P but the lowest abundance in S. VvZIP08 showed a high expression level in ST but low expression levels in S and P.

Discussion
Research on age-dependent leaf senescence has been mostly focused on the canopy or season level. However, there are few relevant studies on the senescence order of individual leaves. In this study, we found a type of branch group in which the leaves presented a certain senescence order. Specifically, leaf senescence was first exhibited on the mature basal leaves of primary shoots, followed by the top leaves of the secondary shoots, and then the basal leaves of the secondary shoots ( Figure 1A,B). Therefore, we used the leaves of these shoots as experimental materials to explore the potential genes involved in the leaf senescence order of grapevine to therefore provide a theoretical basis for further research on the regulation mechanism of leaf senescence order in grapevine. The loss of chlorophyll is a significant hallmark of leaf senescence. Several findings have suggested that the Chl cycle activities are maintained at a low level unless the Chl b-to-a conversion is maintained under specific conditions or at specific developmental stages [27]. Indeed, high levels of Chl a and Chl b occurred in mature leaves (S) but sharply decreased in

Discussion
Research on age-dependent leaf senescence has been mostly focused on the canopy or season level. However, there are few relevant studies on the senescence order of individual leaves. In this study, we found a type of branch group in which the leaves presented a certain senescence order. Specifically, leaf senescence was first exhibited on the mature basal leaves of primary shoots, followed by the top leaves of the secondary shoots, and then the basal leaves of the secondary shoots ( Figure 1A,B). Therefore, we used the leaves of these shoots as experimental materials to explore the potential genes involved in the leaf senescence order of grapevine to therefore provide a theoretical basis for further research on the regulation mechanism of leaf senescence order in grapevine. The loss of chlorophyll is a significant hallmark of leaf senescence. Several findings have suggested that the Chl cycle activities are maintained at a low level unless the Chl b-to-a conversion is maintained under specific conditions or at specific developmental stages [27]. Indeed, high levels of Chl a and Chl b occurred in mature leaves (S) but sharply decreased in senescent leaves (P and ST) ( Figure 1D). Chl metabolism pathways have been well studied; a number of genes related to Chl degradation have been identified and characterized as senescence-associated genes in several plant species, such as Arabidopsis [27], rice [10], and perennial ryegrass [13], but not previously reported in perennial fruit species. The Arabidopsis genome has identified two NYC1 homologs (NYC1 and NOL) [27], one HCAR [11], one PPH [12], one PAO/ACD1 [28], one RCCR/ACD2 [29], three SGR homologs (SGR1/NYE1, SGR2, and SGRL) [14], and one ATP-dependent metalloprotease number (AtFtsH6) [19] encoding CBR, HCAR, PPH, PAO, RCCR, SGR, and AtFtsH6, respectively, and playing roles as senescence executers during leaf senescence. In the present study, twelve genes encoding CCEs (one NYC, two NOLs, one HCAR, three PPH, four PAO, and one RCCR), two encoding chloroplastlocated proteases (VvFtsH6-1 and VvFtsH6-2), and two encoding chloroplast proteins (SGR and SGRL) were identified in the grapevine genome by using the above Arabidopsis orthologs as queries (Table 1). Significantly, of three SGR1/SGRL homologs (SGR1/NYE1, SGR2, and SGRL) present in the Arabidopsis genome, SGR1/NYE1 [14] and SGRL [16] contribute to Chl breakdown in senescent leaves and pre-senescent leaves, respectively, while SGR2 [17] negatively regulates Chl degradation under different senescence conditions. In this study, three candidate SGR1/SGRL sequences were obtained from the BLAST process; however, one of them was deleted because of its short length. Therefore, only two potential SGR1/SGRL sequences were obtained ( Table 1). The phylogenetic tree analysis showed that the sequence named VvSGR clustered with AtSGR1, and VvSGRL clustered with AtSGRL ( Figure 2). In addition, the expression level of VvSGR increased with leaf senescence, while VvSGRL showed high abundance in pre-senescent leaves (S) but low abundance in senescent leaves (P). Therefore, VvSGR could be the ortholog of Arabidopsis SGR1, and VvSGRL could be the AtSGRL ortholog. Further tests of gene expressions and Chl changes at different nodes of leaves showed that Chl content decreased with the extent of leaf senescence (Figure 1), and sixteen SAGs were characterized by distinct expression profiles at different nodes of leaves (Figure 3). The expression levels of ten genes, i.e., VvNYC1, VvSGR, VvPAO1, VvPAO2, VvPAO4, VvPPH1, VvPPH2, VvPPH3, VvFtsH6-1, and VvFtsH6-2 were significantly high in P but low in S (Figure 3), which corresponded to low levels of Chl in P and high levels of Chl in S, respectively ( Figure 1). However, only four genes (VvSGR, VvPAO3, VvPPH2, and VvFtsH6-2) had significantly different expressions between S and ST; their expression patterns corresponded to a low level of Chl in ST and a high level of Chl in S. This could suggest a complex regulation of cellular Chl degradation during leaf senescence. Among the differently expressed CCEs, the PPH expression level has been found to positively correlate with the extent of leaf senescence during the natural leaf maturation process [30]. Two CBR isoforms, NYC1 and NOL, have been suggested to have a similar function in Chl metabolism in Arabidopsis but act at different developmental stages, resulting from their different mutant phenotypes and expression patterns [5,11]. Similarly, VvNYC expression patterns differed considerably from VvNOL1 and VvNOL2 in different nodes of grapevine leaves (Figure 3). A significant difference in VvNYC1 expression was observed with the extent of leaf senescence, corresponding to the function that NYC1 was involved with in naturally induced senescence in Arabidopsis [5,11]. Similar to rice and Arabidopsis [6,10], a single HCAR homolog was identified in the grapevine genome database (Table 1). Its expression pattern was similar to VvNOL2 (Figure 3), which was consistent with the results reported for Arabidopsis [11]. Tanaka et al. (2003) [28] found three possible PAO-encoding genes in the Arabidopsis genome (Tic55, At2g24820; ACD1, At3g44880; and ACD1-like, At4g25650) and confirmed that only the ACD1 gene encodes PAO. Using the ACD1 protein as the query sequence to blast the grapevine genome database, four VvPAO homologs were searched in the grapevine genome database. Their expression patterns indicated that all of them were involved in Chl breakdown at different nodes of leaves. Red Chl catabolite reductase (RCCR) catalyzes the conversion of red Chl catabolite (RCC) to red Chl catabolite (RCC), which is essential in Chl breakdown [27]. However, the expression changes of VvRCCR in different-aged leaves seem not to be related to Chl breakdown ( Figure 3). Its function needs to be further characterized.
To explore potential SIRs regulating these SIEs, an analysis of the cis-element components of sixteen SIEs was performed. The results showed that some SIE members shared conserved core nucleotide sequences potentially recognized by abscisic acid (ABRE), auxin (AuxRR-core), methyl jasmonate (CGTCA-motif/TGACG-motif), gibberellin (TATCbox/GARE-motif/P-box), and salicylic acid (TCA-element) signaling (Table 2). Among them, ABRE is a downstream core nucleotide sequence recognized by ABRE-binding proteins (ABFs/AREBs) [31,32] and was identified in most (twelve out of sixteen) of the SIE promoters in the present study. A recent study reported that ABF4 activated the expression of AtNYC1 and AtPAO by directly binding their promoters in mediating ABA-triggered Chl degradation and leaf senescence [21]. A number of cis-elements involved in light responsiveness were also found in the SIE promoters in addition to hormone-responsive elements, including ACE, AE-box, G-box, Box4, GATA-motif, MYE, and TCCC-motif. Many genes associated with Chl breakdown have been reported to respond to dark-induced senescence [5,11]. However, there are still limited reports on the binding specification for light-responsive transcription factors that may activate or deactivate SIE expression. The presence of these conserved cis-elements, including hormones and light, provides us with a strong clue for identifying potential transcription factors that may directly regulate Chl breakdown during the natural senescence of grapevine leaves, such as ABFs. In an attempt to uncover potential ABA-related SIRs involved in leaf senescence order, the expression patterns of seven ABA receptors identified in the grapevine genome (VvRCAR1-7), two of the major binding partners of the ABA receptor (VvPP2C4 and VvPP2C9) [33], three ABF3/4 orthologs (VvbZIP08, VvbZIP25, and VvbZIP45) [34], and two VvSnRKs expressed highly in leaves (VvSnRK2.1 and VvSnRK2.3) [35] were investigated in leaves with different extents of leaf senescence. Similar to previous findings, ABA-signaling components are required for ABA-triggered Chl breakdown and leaf senescence [21]. In this study, ABA-signaling components showed differential expression patterns with the extent of Chl degradation and leaf senescence (Figure 4). It was observed that the transcription levels of VvRCAR2, 4, 6, 7, VvPP2C4, 9, VvbZIP25, and VvSnRK2.3 were obviously high in P but showed no difference between S and ST, which presented similar expression patterns with six Chl degradationrelated SAGs. The ZIP25 ortholog in Arabidopsis (ABF4) has been confirmed to mediate ABA-triggered Chl degradation and leaf senescence by directly activating the expression of AtNYC1 and AtPAO [21]. The expression patterns of VvbZIP45 and VvSnRK2.1 were similar to those of VvSGR, VvPPH2, and VvFtsH6-2, and their expression patterns corresponded to changes in Chl content in leaves with different extents of leaf senescence. The PPH orthologs in perennial ryegrass (LpPPH) have been shown to be positively related to the extent of leaf senescence and could be a direct downstream target gene of transcription factors in the ABA-signaling pathways [30]. It seems that ABA-signaling components are related to the leaf senescence order of grapevine, but the various changes in the transcription of Chl degradation-related SAGs and ABA-signaling components suggest that different regulatory networks could occur in leaves with different extents of leaf senescence. The integral regulatory network of Chl breakdown in leaves with different extents of leaf senescence remains to be discovered.

Plant Material and Sampling
Two-year-old vines of V. vinifera cv. 'Huangjinmi' were grown in potting soil conditions in a greenhouse at Yangzhou University, Yangzhou, Jiangsu Province of China (119 • 26 E, 32 • 24 N). The vines were grafted on rootstock varieties 'Beta' (Vitis riparia × V. labrusca). Seven uniform trees were selected and 18 branches arising from the lateral buds on the primary shoot and developing into primary and secondary lateral shoots were marked and used for sampling. Six branches served as one biological replicate. Leaves at the same positions were pooled and frozen in liquid nitrogen and then stored at −80 • C.

Measurement of Chl Content
For each replicate, 0.1 g of frozen leaves was extracted, and Chl a/b content was determined using the method reported by Burnison et al. (2011) [36].

Identification of Chl Degradation-Related SAGs in Grapevine Leaves
Arabidopsis thaliana Chl degradation-related SAGs were downloaded from the leaf senescence database LSD3.0 [26] and used as queries to identify genes encoding homologous proteins in the grape genome [37] using the BLAST program in the Phytozome comparative platform [38] with a 10 −5 E-value as the threshold. The obtained grape SAG proteins were then aligned using a conserved domain search tool (http://www.ncbi.nlm. nih.gov/Structure/cdd/wrpsb.cgi (accessed on 19 April 2020)). Genes lacking conserved domains or that were less than 200 bp were discarded.

Expression Analysis by Quantitative Real-Time Polymerase Chain Reaction (RT-qPCR)
Total RNA was extracted from frozen leaves using a CTAB-based method [42]. Firststrand cDNA was synthesized with a PrimeScriptTM RT reagent Kit with gDNA Eraser (Perfect Real Time) (TaKaRa, Dalian, China). The primer pairs of SAGs for RT-qPCR were designed using Primer 6.0 and are listed in Supplementary Table S1. The RT-qPCR protocol was performed based on the manufacturer's instructions with the SYBR Premix ExTaqTM II kit (Takara, Beijing, China) using the CFX Connect Real-Time PCR Detection System (Bio-Rad). ACTIN from grapes was used as a reference gene. The relative expression was calculated using the 2 −∆∆Ct method [43].

Statistical Analysis
The Chl content and RT-qPCR results were subjected to a one-way analysis of variance and the means were compared using the Tukey test with the SPSS software 22.0. Figures were constructed using Excel 2010.

Conclusions
In this work, we found that the grapevine leaves on primary and secondary shoots aged in a certain order in autumn. The Chl level decreased with the extent of leaf senescence and sixteen Chl breakdown-related SAGs were identified in grapevine leaves. The greater degree of Chl degradation in the basal mature leaves of grapevine primary shoots than that in the basal mature leaves of secondary shoots was associated with higher transcript levels of VvNYC1, VvSGR, VvPAO1, VvPAO2, VvPAO4, VvPPH1, VvPPH2, VvPPH3, and VvFtsH6-1 in the former leaves. The greater degree of Chl degradation in the apical young leaves of secondary shoots than that in the basal mature leaves of secondary shoots was related to the higher expression levels of VvSGR, VvPAO3, VvPPH2, and VvFtsH6-2 in the former leaves. ABA-responsive elements were mostly identified in the promoters of these SAGs. Further analysis of the expression pattern of ABA-signaling components suggested that ABA signaling was involved in regulating the order of leaf senescence, but the regulation mechanism of ABA-mediated Chl degradation could be different in different nodes of leaves. Further studies are required to fully understand the mechanism of the different extents of leaf senescence in different leaf orders of grapevine leaves.