Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A comparative transcriptomics and eQTL approach identifies SlWD40 as a tomato fruit ripening regulator

A comparative transcriptomics and eQTL approach identifies SlWD40 as a tomato fruit ripening... Although multiple vital genes with strong effects on the tomato (Solanum lycopersicum) ripening process have been identi- fied via the positional cloning of ripening mutants and cloning of ripening-related transcription factors (TFs), recent studies suggest that it is unlikely that we have fully characterized the gene regulatory networks underpinning this process. Here, combining comparative transcriptomics and expression QTLs, we identified 16 candidate genes involved in tomato fruit ripening and validated them through virus-induced gene silencing analysis. To further confirm the accuracy of the ap- proach, one potential ripening regulator, SlWD40 (WD-40 repeats), was chosen for in-depth analysis. Co-expression network analysis indicated that master regulators such as RIN (ripening inhibitor) and NOR (nonripening) as well as vital TFs includ- ing FUL1 (FRUITFUL1), SlNAC4 (NAM, ATAF1,2, and CUC2 4), and AP2a (Activating enhancer binding Protein 2 alpha) strongly co-expressed with SlWD40. Furthermore, SlWD40 overexpression and RNAi lines exhibited substantially accelerated and delayed ripening phenotypes compared with the wild type, respectively. Moreover, transcriptome analysis of these transgenics revealed that expression patterns of ethylene biosynthesis genes, phytoene synthase, pectate lyase, and branched chain amino transferase 2,in SlWD40-RNAi lines were similar to those of rin and nor fruits, which further demonstrated that SlWD40 may act as an important ripening regulator in conjunction with RIN and NOR. These results are discussed in the context of current models of ripening and in terms of the use of comparative genomics and transcriptomics as an effective route for isolating causal genes underlying differences in genotypes. Received November 19, 2021. Accepted March 28, 2022. Advance access publication May 4, 2022 V The Author(s) 2022. Published by Oxford University Press on behalf of American Society of Plant Biologists. Open Access This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Research Article Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 251 Introduction 2013; Irfan et al., 2016). Furthermore, the transcriptional be- havior of 1,000 TFs has been established during tomato rip- Given that seed dispersal is of major ecological and evolu- ening (Rohrmann et al., 2011) and gene regulator networks tionary importance for all plants and the fact that fleshy have been modeled on the basis of these data (Rohrmann fruit plays a vital role in this process, fruit ripening assumes et al., 2012). The stronger acting ripening genes mentioned a central importance in the plant life-cycle. It is well docu- above were, in contrast, identified in mutant screens intent mented that hundreds of genes display altered expression on isolating strong mutants in order to enhance tomato during this process (Karlova et al., 2011; Osorio et al., 2011), shelf-life (Tigchelaar, 1973; Lanahan et al., 1994; Vrebalov and that metabolism also undergoes concurrent dramatic et al., 2002; Manning et al., 2006). While recent years have shifts to form fruit quality (Carrari and Fernie, 2006). As one resulted in the identification of many additional genes with of the most important appearance qualities, the accumula- ripening consequences (Li et al., 2021; Shi et al., 2021), it is tion of carotenoids, when combined with naringenin chal- probable that additional contributors to this complex pro- cone tainted yellow peel, forms the reddish color of tomato cess remain to be found. As such, further genome-wide fruits (Solanum lycopersicum)(Ballester et al., 2010; analysis is required to mine the regulator affecting this pro- Fernandez-Moreno et al., 2016; Zhu et al., 2018). Among the cess and provide more comprehensive knowledge on this carotenoid biosynthesis pathway, phytoene synthase (PSY) is process. the key rate-limiting enzyme of the whole pathway; it cata- Time-series, species/accession specific or tissue specific lyzes two molecules of GGPP to form the colorless phy- and comparative transcriptomics studies have previously toene. Subsequently, under the catalysis of a series of deciphered gene regulatory networks underlying plant devel- enzymes, phytoene undergoes dehydrogenation and isomeri- opmental pathways allowing the identification of additional zation reactions to form lycopene, which is the dominant functional genes (Breschi et al., 2017; Chang et al., 2019; carotenoid of tomato fruit (Bird et al., 1991; Bartley and Batyrshina et al., 2020; Baranwal et al., 2021). For example, Scolnik, 1993; Cazzonelli and Pogson, 2010; Chen et al., Bolger et al. (2014) compared the transcriptomes of M82 2019). In addition, the texture of fruits is affected by the and Solanum pennellii, and identified 100 key candidate modification of cell walls and pectate lyase (PL), which genes related to salt and drought stress. Additionally, tran- hydrolyzes pectin and is the most substantially cell wall scriptomics studies on genetic mapping populations have gene contributor to this process identified to date (Yang defined expression QTLs (eQTLs) as genomic loci controlling et al., 2017). Besides appearance and textural qualities, an- variation in steady state levels of transcript between individ- other important quality aspect is taste, which has been at- uals (Sønderby et al., 2007) of what has subsequently be- tributed to the sugar/organic acid ratio, and volatile and come a well-characterized mapping population (Ofner et al., secondary metabolite accumulation. The key genes underly- 2016; Szymanski et al., 2020). During tomato domestication, ing the levels of these metabolites have been uncovered by many phenotypes (such as the leaf structure and ripening a range of quantitative trait loci (Fridman et al., 2004; process) of wild species, such as S. pennellii were under Tieman et al., 2006; Schauer et al., 2008; Centeno et al., strong selection and were substantially different to that of 2011) and genome-wide association studies (Sauvage et al., in the cultivar species S. lycopersicum. Based on the eQTL 2014; Tieman et al., 2017; Ye et al., 2017; Gao et al., 2019). analysis of the 76 introgression lines (ILs) from S. pennellii in Moreover, the considerable metabolic changes are coordi- the background of S. lycopersicum, Ranjan et al. (2016) iden- nated and mediated by transcription factors (TFs) and epi- tified important genetic regulators of leaf development on genome dynamics on the metabolic structural genes’ chromosomes 4 and 8. The above-mentioned studies dem- expression (Centeno et al., 2011; Rohrmann et al., 2011; onstrate the power of comparative transcriptomics in com- Giovannoni et al., 2017; Lu et al., 2018; Li et al., 2020). Over the last 50 years, several mutants, such as ripening-inhibitor bination with ILs; however, limited studies have been carried out using their integrated approach to mine the genes in- (rin), nonripening (nor), Never ripe (Nr), and Colorless nonrip- volved in tomato fruit ripening. ening (Cnr) mutations, have been identified as severely As a distant relative of the cultivated tomato S. lycopersi- impacting the tomato ripening process (Tigchelaar, 1973; cum, S. pennellii has many substantially different phenotypes Lanahan et al., 1994; Vrebalov et al., 2002; Manning et al., with the cultivated tomato and one of these is the mature 2006). Among these mutants, rin is the one of the most fa- mous ripening delaying mutants substantially lacking the fruit morphology. The mature fruit of S. lycopersicum is red ethylene burst and hindering the color change and softening and soft while the mature fruit of S. pennellii is green and hard, which renders this pair the ideal parents to cross and processes, which results from the repression of the ripening illustrate the genetic landscape of fruit ripening. The core inhibitor-macrocalyx (RIN-MC) chimeric protein (Robinson, set of 76 S. pennellii ILs, which represent whole-genome cov- 1968; Vrebalov et al., 2002; Ito et al., 2017). The integrated analysis of chromatin immunoprecipitation (ChIP)-chip and erage of S. pennellii in overlapping segments in the back- transcriptome indicated that RIN can directly induce the ex- ground of M82, have been widely used to identify the key pression of the key ripening-related structural and regulator genes of many traits such as yield and metabolic composi- genes, ACS2/4, SGR1, PSY, Cel2, EXP1, PAL1, C4H, LoxC, tion (Semel et al., 2006; Alseekh et al., 2013, 2015). In the AAT1, CNR, NOR, AP2a, and itself (Fujisawa et al., 2012, present study, to identify key candidates regulating tomato 252 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. fruit ripening, an integrated comparative transcriptomics not show such specificity and have expression in all ILs. and eQTL approach was taken utilizing S. pennellii ILs With the specific eQTL and nonspecific eQTL information, the candidates of Lyco data set were classified into 119 spe- (Eshed and Zamir, 1995). We isolated 16 candidates and provided primary validation of eight of them as being in- cific and 223 nonspecific eQTLs (Figure 1B). For the Penn volved in the ripening process via the virus-induced gene si- data set, around 105 specific and 202 nonspecific candidates were classified (Figure 1C). Here, the high number of non- lencing (VIGS) method. Following this screen, one candidate, SlWD40, was taken for further study. For this candidate sta- specific eQTL is attributed to the epistatic interactions be- ble RNAi and overexpression (OE) lines were generated and tween S. pennellii alleles and M82 alleles or, alternatively, the presence of a large number of trans-QTL as previously characterized. The OE of SlWD40 promoted ripening while reported for leaf expression analysis (Chitwood et al., 2013) its inhibition inhibited it. The co-expression networks, and fruit enzyme abundance analysis (Steinhauser et al., metabolome and transcriptome analysis indicated that 2010). Previous studies have indicated that the candidates SlWD40 acted as a positive regulator of tomato ripening whose functional categories belong to transcription regula- with the key ripening TFs such as RIN, NOR, AP2a,and tors, oxidase and cytochrome P450 may be involved in regu- SlWRKYs. These results are discussed within the context of lating tomato fruit ripening and secondary metabolism; their implications regarding fruit ripening as well as with re- therefore, we chose seven candidate genes that were of spect to the utility of genomic information in filling our these three functional categories among the specific eQTL knowledge gaps in important biological processes. candidates. Results In a parallel approach, given that TFs act as important regulators in fruit ripening, we also adopted a TF-centric ap- Integrating comparative transcriptomics and eQTL proach (Figure 1D). From the total of 34,727 genes, candi- mapping to mine for genes involved in tomato dates annotated as TFs and displaying more than five times ripening higher expression in S. lycopersicum were selected. Next, Our earlier work described a high-quality genome assembly eQTL mapping thinned the list to 127 candidates which of the parents of the Solanum pennellii IL population as well were then arranged with respect to the ratio of their expres- as identifying candidate genes involved in salt as well as sion in Breaker + 10 to that in Breaker stage. Based on puta- drought stress tolerance (Bolger et al., 2014). Surprisingly, tive ortholog information (Arabidopsis thaliana and S. the open-reading frame sequence of most well-characterized lycopersicum) and literature survey concerning their putative ripening-related genes is identical between S. pennellii and S. functions, a final set of 20 candidates was selected. Finally, lycopersicum. We therefore thought to try a comparative on the basis of tissue specific expression and the S. lycopersi- transcriptomics approach of the S. pennellii IL population cum to S. pennellii expression ratio, the 7 candidates from since studies on fruit gene expression of a subset of the ILs eQTL approach and 20 candidates from TF approach were has proven highly informative (Baxter et al., 2005; Alseekh narrowed down to the 16 potential candidates described in et al., 2015)as well as in leaves (Chitwood et al., 2013). For Supplemental Data Set S1. this purpose, as an initial approach, transcriptome data for M82 and S. pennellii mature fruits were sorted as follows: VIGS analysis of candidate genes the total 34,727 genes in transcriptome sorted into two dif- To provide preliminary analysis of the function of the candi- ferent data sets named as Lyco and Penn (Figure 1A)based date genes in tomato ripening, we carried out VIGS experi- on the ratio of their expression values. The Lyco data set ment using purple Microtom cv. tomato fruit which contained genes that are highly expressed in M82 (13,521), accumulate high amount of anthocyanin resulting from the while the Penn data set contained genes that are highly introduction of Del/Ros1 petunia (Petunia hybrida)TFs expressed in S. pennellii (11,781) (Figure 1A). For the Lyco (Orzaez et al., 2009). Partial fragments of the 16 candidate genes which we reasoned would be more likely to harbor genes were cloned into pTRV2-Ros/Del/GW vector. Around genes underlying the “red” ripe phenotype of cultivated to- 10–15 fruits per plant were infected with agrobacterium car- matoes, around 300 candidates could be narrowed down by rying the respective VIGS vector. After silencing Del/Ros1 using three independent filters. Firstly, we chose to focus on (empty vector) in Microtom Del/Ros1 fruits, there was de- genes for which expression was at least five times higher in pletion in purple pigments but not in lycopene content due S. lycopersicum with respect to S. pennellii. Secondly, based to the silenced part accumulating less purple anthocyanin on the transcriptome profiling of red ripe fruit from S. lyco- pigments and thereby being easy to discriminate from non- persicum (M82) parent and a set of lines with distinct silenced (purple pigment rich) tissues (Figure 2). Phenotypes introgressed S. pennellii segments (http://ted.bti.cornell.edu/ were scored visually after 15 d of infection for all the 16 vali- cgi-bin/TFGD/array_data/home.cgi), large numbers of spe- dated candidates (Supplemental Data Set S1). cific eQTL and nonspecific eQTL have been identified as the Red color of western tomato cultivars represents the accu- former definition that specific eQTL candidates are the can- mulation of lycopene, which is an important indicator of to- didates whose expression are sharply (exponentially) in- mato ripening. VIGS for the structural genes encoded by creased or decreased in its located IL compared with other Solyc01g094080 and Solyc03g095900 displayed a red pheno- 75 ILs while nonspecific eQTL candidates’ expression does type, indicating that these genes are not associated with Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 253 Figure 1 Candidate gene filtration by integrating comparative transcriptomics and eQTL mapping. A, Pipelines for candidate gene filtration of eQTL approach. Filters are shown in bullet points. B, Heat map of relative expression level of filtered candidates. Lyco/Penn, Genes were sorted by ratio of expression value for S. lycopersicum and S. pennellii. ILs are arranged as per the number of chromosome (X-axis). Genes are arranged according to their Gene IDs (Y-axis). Regions of red or blue indicate that the gene expression is increased or decreased, respectively, over that of M82. Chr, chromosome. C, Heat map of relative expression level of filtered candidates. Penn/Lyco, Genes were sorted by the ratio of expression value for S. pennellii and S. lycopersicum. ILs are arranged as per the number of chromosome (X-axis). Genes are arranged according to their Gene IDs (Y-axis). Regions of red or blue indicate that the gene expression is increased or decreased, respectively, over that of M82. D, Pipelines for can- didate gene filtration of TFs approach. Filters are shown in bullet points. Br, Break. Figure 2 VIGS of empty vector (silencing of Del/Ros1) and SlWD40 in Microtom Del/Ros1 fruit. 254 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. lycopene biosynthesis or the pathways that fuel it. However, the master ripening-related TFs, SlWD40 also highly co- as the TFs regulating ripening are generally reported to hin- expressed with key carotenoid-related genes (PSY1), as well as ethylene (ACS4) and abscisic acid (NCED3) biosynthesis der carotenoid biosynthesis, the yellowish phenotypes of Solyc11g010710 (ethylene response factors, ERF TF) and genes and cell wall modification genes (PL and PMEI) Solyc07g052700 (MADS-box TF, AGL66) VIGS fruits indicate (Supplemental Table S1). All of these results indicate that SlWD40 may act in concert with the better characterized that they may function as a ripening regulators in line with ripening TFs to regulate the ripening processes, including former studies that implicated SlERFs and MADS-box TFs in those dependent on changes in pigmentation, hormone lev- tomato fruit ripening (Wang et al., 2014; Liu et al., 2016; els, and signaling and cell wall modification. Supplemental Data Set S1 and Supplemental Figure S1). Interestingly, the VIGS fruits of a transcription regulator, SlWD40 affects the tomato fruit transcriptome SlWD40 (Solyc04g005020) also exhibited a yellowish pheno- To confirm the accuracy of our approach and to assess in type. Given that ERF and MADS box family TFs are already detail the function of SlWD40 in the tomato ripening pro- well-known to be involved in tomato fruit ripening and that cess, we chose the fruit specific patatin B33 promoter which SlWD40 was identified as a downstream target gene of RIN has been widely used for fruit specific expression to carry out (Fujisawa et al., 2013), we selected SlWD40 for in-depth the stable transformation (Rocha-Sosa et al., 1989; Vallarino analysis here (Figure 2). et al., 2020). T0 transformants of RNAi and OE lines were characterized by NPT-II-specific polymerase chain reaction Co-expression network and VIGS of SlWD40 (PCR). Real-time quantitative polymerase chain reaction confirmed its role in tomato fruit ripening (RT-qPCR) was also carried out using fruit samples from In order to analyze the function of SlWD40 on fruit ripening, promising T0 transgenics to select high OE and knockdown we initially identified its potential regulators following cis- lines to raise T1 generation (Supplemental Figure S3). Fruits regulatory element analysis of the promoter of SlWD40.This from all generations were analyzed and phenotype was stable analysis indicated that the promoter contained several ethyl- over T0 and T1 generations. Before the T1 generation plant ene (AP2, B3, EIN3, and EIL) and ripening-related elements transplant to soil, we also used the NPT-II-specific PCR to (C2H2, MADS, NF-YB, NF-YA, and NF-YC) in the 1-kb pro- confirm that the plants are transgenic. Based on the expres- moter region, which indicated that it may well be induced sion of SlWD40, two independent lines of RNAi (RNAi-1 and by the ripening and ethylene burst (Supplemental Figure -2 lines) and OE (OE-1 and -2 lines) were chosen for subse- S2). Moreover, the evaluation of publicly available expression quent experiments (Supplemental Figure S3). data with tissue-specific expression analysis of SlWD40 con- To analyze fruit phenotype at the identical stage, fruits of firmed the hypothesis that SlWD40 was only slightly each genotype were labeled upon anthesis and harvested for expressed in the leaf, bud, flower, root, and young fruit but phenotyping, transcriptome, and metabolite profiling at ma- that its expression increases exponentially following mature ture green (MG, 34 DPA), breaker (Br, 37DPA), and pink green stage (Tomato Genome Consortium, 2012). (Pink, 45 DPA) stages of the wild type (WT). As seen in Intriguingly, its expression in different cell types of the to- Figure 4, the development and ripening process were sub- mato fruit revealed that it is highly similar to that of the stantially hindered in the RNAi fruit while the ripening pro- other known ripening regulators, such as RIN and NOR cess was significantly accelerated in comparison to the OE (Shinozaki et al., 2018). lines. Moreover, especially at the Br stage of WT fruits, the Moreover, given that assembly of co-expression networks size of RNAi fruits was significantly smaller than that of the is an efficient method to identify the important interactions OE and WT lines and the RNAi fruits were still at the ma- and relationship among different genes (Mutwil et al., 2011), ture green stage while the OE fruit were almost at the pink available transcriptome data of different organ and fruit de- stage. The contents of chlorophylls and carotenoids, some velopment stages were used to construct tomato co- of the most important parameters of fruit ripening, also expression network (Figure 3 and Supplemental Table S1; confirmed the positive function of SlWD40 on tomato ripen- Tomato Genome Consortium, 2012). The co-expression sub- ing process: The degradation of chlorophylls and synthesis network containing SlWD40 included 171 structure genes/ of the predominant carotenoid, lycopene, were significantly regulators, which are involved in chlorophyll and carotenoid hindered in RNAi fruits but accelerated in the OE fruits metabolism as well as tomato fruit ripening and cell wall (Figure 4B). metabolic pathways. Among the 171 genes, a total of 62 In order to estimate the effect of SlWD40 on the global genes exhibited high co-expression phenotype (jCo-expres- difference of gene expression during the different fruit devel- sion Coefficientsj 40.6, P5 0.05) with SlWD40 opmental stages, we additionally analyzed the differentially (Supplemental Table S1). Consistent with the results of expressed genes (DEGs) among the RNAi, OE lines, and WT cis-regulatory element analysis, three MADS TFs, including fruit at MG, Br, and Pink stages. For this purpose, we used RIN, two AP2s TFs, and one ARF TF were significantly posi- FPKM (fragments per kilobase per million mapped frag- tively co-expressed with SlWD40. Moreover, another vital ments) and identified genes with jlog (fold change) j5 1 ripening-related TF, NOR, exhibited a co-expression coeffi- and false discovery rate (FDR) (corrected P value)5 0.05 cient of 0.86 with SlWD40 (Supplemental Table S1). Besides (Supplemental Table S2). Firstly, we checked the SlWD40 Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 255 Figure 3 Co-expression network of SlWD40 with tomato ripening pathway-specific genes. Well-characterized key regulators such as RIN (ripening inhibitor), NOR (nonripening), and FUL1 (FRUITFUL1) (labeled in blue) strongly co-expressed with SlWD40. ELIP, early light induced protein; NF-Y, nuclear factor Y; MADS, MADS domain protein. Figure 4 Photographs and pigments content of WT and T1 generation RNAi (lines 1 and 2) and OE (lines 1 and 2) lines at 28, 34 (MG), 37 (Br), and 45 (Pink) DPA fruits. A, Photographs of WT and transgenic SlWD40 fruits. Images were digitally extracted for comparison. B, Chlorophylls and carotenoids of WT and transgenic SlWD40 fruits. The values in each column are the mean of at least three biological replicates. Error bars indicate SD. The asterisks indicate statistically significant differences determined by the Student’s t test (two-tail): *P5 0.05; **P5 0.01. ND, not detected. expression among the different genotypes at the MG stage. the endogenous SlWD40 expression, the OE effect of B33 Given the low expression level of SlWD40 of WT fruit at promoter was concealed and SlWD40 expression was not MG stage, RNAi fruit did not exhibit significantly different significantly different between the OE and WT fruit at the expression from WT fruit. However, as the ripening process Br and Pink stages (Supplemental Table S2). These results was initiated, the expression of SlWD40 was significantly in- were further confirmed by the principal component analysis duced and its expression was remarkably lower in the RNAi (PCA) and cluster analysis based on the transcriptome data fruit compared with that of WT fruit at Br and Pink stages of different samples. RNAi samples were closely grouped without affecting the expression of other WD40 family genes with WT at MG stage but substantially separated samples at (Supplemental Table S2). In the OE fruits, SlWD40 expres- Br and Pink stages. Conversely, OE samples were clustered sion was 5.64- and 5.71-fold higher than that of WT fruit at with WT sample especially at the Pink stage and subse- the MG stage. That said owing to the massive induction of quently separated from the WT sample at the MG stage 256 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. (Figure 5, A and B). In order to further mine the important SlWRKY6 (Solyc02g080890.3.1), SlWRKY17 (Solyc07g051840. DEGs under the effect of SlWD40, we further analyzed the 4.1), SlWRKY31 (Solyc06g066370.4.1), and SlWRKY79 overlapping DEGs of OE-WT fruit at MG stage and RNAi- (Solyc02g072190.4.1) was increased by 5.7-, 3.4-, 2.6-, 2.6-, WT fruits at Br and Pink stage. The results indicate that 244 2.4-, 2.2-, 2.2-, and 1.8-fold, respectively. Moreover, since fruit genes were stably downregulated in the OE-MG and upregu- size of SlWD40-RNAi fruits was smaller and IAA content di- lated in the RNAi Br and Pink stages, while 60 genes were rectly affects organ size, we found that the expression of stably upregulated in the OE-MG and downregulated in the SlGH3-2 (gene regulating auxin homeostasis) was increased RNAi Br and Pink stages (Supplemental Table S3). To further by three-fold in OE lines while the level of the same gene mine the functional categorization of DEGs, AgriGO v2.0 was decreased in RNAi lines by three- to six-fold. analysis tools (http://bioinfo.cau.edu.cn/agriGO/) by singular Additionally, given that SlWD40 was strongly co-expressed enrichment analysis has been used based on the conserved with RIN and NOR, the conserved DEGs of rin, nor mutants, DEGs (Wang et al., 2017; Supplemental Table S4). Among and SlWD40 transgenic fruits were analyzed and 31 genes were found as conserved DEGs across the genotypes the GO terms included in the “Molecular Function” category of the DEGs upregulated in the OE-MG and downregulated (Figure 5C and Supplemental Table S5; Fujisawa et al., 2012; in the RNAi Br and Pink stages, the pathways that affected Gao et al., 2020). Among the 31 genes, several important the lysis and enzyme activity, such as lyase activity ripening-related genes, such as ACS4 for ethylene biosynthe- (FDR = 0.0014), oxidoreductase activity (FDR = 0.011), and sis, PSY and Z-ISO for carotenoid biosynthesis, PL and PMEI monooxygenase activity (FDR = 0.013), were enriched. for cell wall modification, INV for sugar metabolism, and branched chain amino transferase 2 (BCAT2)and THA1 for Moreover, in the “Biological Process” category of DEGs downregulated in the OE-MG and upregulated in the RNAi amino acid metabolism, were significantly downregulated in Br and Pink stages, several cell-wall-related pathways were the SlWD40-RNAi, rin,and nor fruits and upregulated in the significantly enriched (Supplemental Table S4). Interestingly, SlWD40-OE fruits, which indicates that SlWD40 may partici- in the SlWD40-OE lines, the expression of SlWRKY75 pate in some or all of the primary regulatory functions of (Solyc05g015850.4.1), SlWRKY37 (Solyc02g021680.3.1), SlWR these important regulators, such as RIN and NOR.As the transcriptome data indicated that SlWD40 exhibited KY23 (Solyc01g079260.4.1), SlWRKY30 (Solyc07g056280.3.1), Figure 5 The effect of SlWD40 on fruit transcriptome. A, PCA of transcriptome of different genotypes in the three fruit stages. PC, principal com- ponent. B, Heat map of transcriptome profiles of pericarp tissues of T1 SlWD40 RNAi, OE lines, and WT. X-axis indicates different transgenic lines along with different fruit ripening stages. The hierarchical clustering based on transcriptome indicated that the development of RNAi samples is delayed especially on Br stage which are cluster with MG stage samples. C, Venn diagram showing the overlap DEGs between SlWD40 transgenic fruits, rin and nor mutant fruits. Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 257 substantially changed expression in the MG stage of two OE lines and Br and Pink stage of two RNAi lines compared with that of in the WT, to further confirm the expression difference, we checked the expression of SlWD40 and other 17 ripening-related genes such as RIN, NOR, PSY,and PL in these samples. The results indicate that all of these genes were significantly upregulated in the two OE lines at the MG stage but significantly downregulated in the two RNAi lines at Br and Pink stages, which further confirm the posi- tive function of SlWD40 on tomato ripening with the strongly co-expressed TFs (Figure 6). Metabolome analysis of SlWD40 transgenic fruits Given that metabolites are important parameters for esti- mating progression of the fruit ripening process (Carrari et al., 2006), we analyzed their levels in the different geno- types at a time concurrent to the MG, Br, and Pink stages of WTs. To obtain the global metabolome variation of differ- ent samples, a PCA analysis was carried out based on the primary metabolites, lipids, and secondary metabolites. The results indicate substantial differences among RNAi, OE, and WT fruits, which were particularly prominent at the Pink stage. RNAi pink fruits were closest to Br fruits and sepa- rated clearly from OE and WT fruits (Figure 7, A and B). Given that SlWD40 was strongly co-expressed with RIN and NOR, we analyzed the overlapping differentially enriched pri- mary metabolites between SlWD40-RNAi lines, rin and nor mutants at the Pink stage (Osorio et al., 2011). The results indicated that as the principal free amino acid that most contributes to the “umami” flavor of ripe tomato fruits, glu- tamate exhibited a conserved lower accumulation in the SlWD40-RNAi lines, rin and nor pink fruits (Figure 7C). In detail, given that several amino acids are increased and or- ganic acid decreased during WT tomato ripening (Carrari Figure 6 The different expression of SlWD40 and other 17 ripening-re- et al., 2006), it is of interest that we observed that aspartic lated genes of T1 transformants of RNAi and OE lines compared with acid, glutamic acid, and tryptophan levels were significantly WT fruits. The values in each column are the mean of three biological replicates. Error bars indicate SD. lower in both RNAi lines compared with those exhibited by the OE and WT pink fruits (Figure 8 and Supplemental Table S6). Moreover, nicotinic acid and glyceric acid levels confirmed the positive influence of SlWD40 on the fruit rip- were significantly higher in at least one RNAi line compared ening process (Figure 9 and Supplemental Table S6). with those observed in OE and WT pink fruits (Figure 8 and Supplemental Table S6). Moreover, the former results indi- Discussion cated that the representative lipid components, triacylglycer- We adopted a computational approach to mine key candi- ols (TAGs) are significantly decreased in avocado fruit date genes involved in tomato fruit ripening by integrating ripening (Rodriguez-Lopez et al., 2017). The same trends also comparative transcriptomics and eQTL analysis. In doing so, have been found in the present study that the content of we identified 16 previously uncharacterized candidate genes TAG 48:0, TAG 52:5, TAG 52:6, TAG 54:6, TAG 54:7, and for involvement in ripening. Among them, we chose to fol- TAG 54:8 was remarkably higher in the RNAi fruits than low up on SlWD40 since it was identified in the ChIP seq that of OE and WT pink fruits (Figure 9 and Supplemental screen as a direct target gene of RIN and relatively little is Table S6). As the secondary metabolites (such as naringenin known concerning its regulatory role (Fujisawa et al., 2013). chalcone, naringenin hexose, and chlorogenic acid deriva- Intriguingly, cross-tissue co-expression networks for SlWD40 tives) dramatically accumulate during ripening and represent an important quality of fruit, the significantly higher and strongly suggest that it may act with the other important lower content of them in the OE MG fruits and RNAi pink ripening regulators, such as RIN and NOR in affecting to- fruits compared with that of WT fruits, respectively, further mato fruit ripening, and the detailed analysis of DEGs and 258 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. Figure 7 The effect of SlWD40 on fruit metabolism. A, PCA of metabolite levels of different genotypes in the three fruit stages. PC, principal com- ponent. B, Heat map of metabolite profiles of pericarp tissues of T1 SlWD40 RNAi, OE lines, and WT. X-axis indicates different transgenic lines along with different fruit ripening stages. Y-axis indicates metabolites. The hierarchical clustering based on all metabolites indicated that the devel- opment of RNAi samples is delayed especially in the Pink stage which are cluster with Br stage samples of OE and WT. C, Venn diagram showing the overlap different primary metabolite between SlWD40 RNAi fruits, rin and nor mutant fruits at the Pink stage. differentially abundant metabolites between SlWD40 trans- candidate gene filtration pipeline. Moreover, on the basis of genic fruits further illuminate the important function of the distance between QTL and the target transcript, eQTL SlWD40 on tomato ripening process. can either be classified as cis-eQTL (where the gene encod- ing the transcript resides within the QTL interval) or trans- Integrating comparative transcriptomics with the eQTL with both types being prominent in tomato eQTL approach to mine for ripening regulators (Rockman and Kruglyak, 2009; Zhu et al., 2018). Combining Several tomato genes with strong ripening phenotypes have the gene filtration pipeline and the comparative transcrip- been identified via mutagenesis-based breeding programs in- tomics, we filtered genes responsible for the red ripe pheno- cluding rin, nor, NR,and Cnr (Tigchelaar, 1973; Lanahan type of tomato fruits (i.e. genes highly expressing only in et al., 1994; Vrebalov et al., 2002; Manning et al., 2006). “red” ripe fruits). Since gene duplication and subsequent Moreover, the recombinant inbred line (RILs) population is functional diversification creates novel metabolic pathways also useful to identify key loci regulating quantitative traits and regulation, we focused on TF families that were evolved including fruit ripening as well as aroma, color, and disease by gene duplication and already reported to be regulating resistance (Kimbara et al., 2018). However, since RILs usually tomato fruit ripening (e.g. MADS box and basic helix loop harbor more than one introgression, the possibility of the helix (bHLH) TF families) (Hileman et al., 2006; Waseem introgressed loci having beneficial/inhibitory interactions et al., 2019). Out of the 16 candidate genes, 10 were found with the genes in the genetic background renders gene to be generated either through tandem or block duplication function elucidation more complex in RILs. In contrast, ILs (https://bioinformatics.psb.ugent.be/plaza). In the VIGS ex- are high resolution in that they normally carry only single periment, green and yellow phenotypes have been obtained introgressions and as such epistatic interactions masking sin- for the candidates Solyc11g010710 (AP2 like) and gle gene effects are largely minimized (Ofner et al., 2016). Solyc07g052700 (MADS TF, AGL66). Our previously pub- Here, we used available transcriptome data from red ripe lished work implicated AP2a in regulating tomato fruit rip- fruit of S. lycopersicum (M82) parent and a set of lines with ening via regulation of ethylene biosynthesis and signaling distinct introgressed S. pennellii segments (http://ted.bti.cor (Chung et al., 2010; Karlova et al., 2011). MADS box TFs nell.edu/cgi-bin/TFGD/array_data/home.cgi) and optimized a such as SlCMB1 (Solyc04g005320), TAGL1 (Solyc07g055920), Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 259 Figure 8 The scheme of major metabolic changes of the transgenic lines. The difference of sugar and amino acid-related metabolites between transgenic lines and WT fruit at MG (A), Br (B), and Pink (C) stage. Blue and red color depicts a decrease and increase in metabolic levels com- pared with the WT fruit samples, respectively. Figure 9 The difference of representative lipids and secondary metabolites of T1 transformants of RNAi and OE lines compared with WT fruits. The values in each column are the mean of at least three biological replicates. Error bars indicate SD. The asterisks indicate statistically significant differences determined by the Student’s t test (two-tail): *P5 0.05; **P5 0.01. 260 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. and the canonical TF RIN (Solyc05g012020) have previously enhances the regulation of the complex during anthocyanin been reported as positive regulators of tomato fruit ripening biosynthesis (Ramsay and Glover, 2005; Zhang et al., 2014). (Zhang et al., 2018). Therefore, VIGS for AP2 like and MADS In the present study, the green phenotype of fruits of box candidates (Solyc11g010710 and Solyc07g052700) acted SlWD40-RNAi lines clearly demonstrated a positive role of as positive control supporting the efficacy of our approach. SlWD40 in tomato fruit ripening (Figure 4). Moreover, given Moreover, among the 16 genes, VIGS for two bHLH TFs that the global gene co-expression analysis is a powerful ap- (Solyc03g044460 and Solyc12g098620) showed green and proach to identify the important interactions among differ- light red phenotypes (Supplemental Figure S1). Here, light ent genes during the development of certain organs, it is red phenotype for Solyc12g098620 is in line with the re- important to note that the co-expression network of cently published work by D’Amelia et al. (2019) in which SlWD40 revealed links with RIN and NOR, while transcrip- the authors reported that this bHLH TF regulates carotenoid tome analysis indicated they shared conserved regulated biosynthesis. Additionally, one of the candidates, genes associated with ethylene (ACSs) and carotenoid (PSY Solyc12g010950 (alcohol dehydrogenase), obtained a whitish and Z-ISO) biosynthesis as well as cell wall degradation (PL green fruit phenotype, and another candidate, SlWD40, and PMEI) and sugar (INV) and amino acid (BCAT2 and THA1)metabolism (Supplemental Table S5). Interestingly, obtained a yellowish fruit phenotype (Figure 2 and Supplemental Figure S1). While the link between the alcohol based on available ChIP and transcriptome data, the RIN dehydrogenase and ripening is currently unclear, that for protein can directly bind to the promoter of SlWD40 and SlWD40 is not without precedence since it is a transcrip- thereby increase its expression (Martel et al., 2011; Fujisawa tional regulator which has been linked to rin and nor in et al., 2013). Moreover, as an important TF family member involved in ChIP experiments but not characterized in detail in its own ethylene signal, AP2a belonging to the AP2/ERF superfamily right. The eQTL for SlWD40 can bedefined as acis-eQTLsince has been shown to inhibit ethylene biosynthesis as well as SlWD40 is located in the region which has been introgressed positively regulate chlorophyll degradation and carotenoid by S. pennellii chromosomal architecture in the IL4-1 and biosynthesis in tomato (Wang et al., 2019). And in Setaria italica, SiAP2 can bind to the SiWD40 promoter to mediate IL4-1-1 and SlWD40’s expression is only dramatically lowered for IL4-1 and 4-1-1 (Supplemental Table S7). At the onset of abiotic stress responses (Mishra et al., 2012). In the present fruit ripening process (MG stage), SlWD40 is moderately study, the promoter region of SlWD40 was found to harbor expressed in S. lycopersicum (RPKM mean value 31) but neg- eight different AP2 TF binding sites and AP2a was signifi- ligibly expressed in S. pennellii (RPKM mean value 0.17) and cantly induced in SlWD40-OE-MG fruits while repressed in the WD40-RNAi-Breaker fruits (Supplemental Table S2). consistent with the ripening stage, the expression of SlWD40 was substantially induced in concert with fruit ripening. Additionally, it has been well-documented that several Considering the dramatically lowered SlWD40 expression in SlWRKYs regulate tomato fruit ripening and lycopene accu- IL4-1 and 4-1-1 and substantial difference of the red/green mulation (Cheng et al., 2016; Wang et al., 2017). In our colored fruits of S. lycopersicum and S. pennellii,respectively, analysis, SlWD40 OE resulted in upregulation (by two- to five-fold) of a broad number of SlWRKYs (SlWRKY75, we hypothesized that SlWD40 may participate in the S. lyco- persicum ripening process. Furthermore, VIGS for SlWD40 in SlWRKY37, SlWRKY23, SlWRKY30, SlWRKY6, SlWRKY17, MicroTom inhibited normal red coloration. These results SlWRKY31,and SlWRKY79) in mature green fruits of OE demonstrate the utility of our candidate gene filtration pipe- lines. Moreover, SlWRKY17, which was shown to interact line integrating comparative transcriptomics with eQTL in with RIN, SlERF2b, and SlERF7, is strongly co-expressed with ELIP2 and RIN in our analysis (Wang et al., 2017; identifying candidate ripening genes. Supplemental Table S1). All of these results indicate that SlWD40 is an important regulator of tomato fruit SlWD40 might act as a junction point and facilitate binding ripening of one or more above mentioned co-expressed TFs such as Among the eight genes which we validated by VIGS, the in- RIN, NOR, AP2a,and SlWRKYs to be involved in tomato rip- formation concerning the role of the WD40 family in regu- ening process. lating tomato ripening is the most limited. Ripening Besides the important ripening-related TFs, hormone sig- function of this gene was further confirmed by the stable naling also plays a vital role in the ripening process. As an OE and RNAi transformation (Figures 4 and 5). WD40 pro- important regulator of auxin-ethylene homoeostasis which teins contain a signature WD (Trp-Asp) dipeptide and 40 affects fruit ripening (Kumar et al., 2012; Sravankumar et al., amino acids in single repeats that then fold into four- 2018), the expression of SlGH3-2 (Solyc01g107390) increased stranded anti-parallel b-propeller sheets and are highly pro- in SlWD40-OE lines by three-fold while decreased in RNAi miscuous interactors, being both platforms for protein– lines by three- to six-fold (Figure 6 and Supplemental Table DNA and protein–protein interactions (Xu and Min, 2011; S2). Combined with the induction and repression effect of Mishra et al., 2012; Chen et al., 2022). In the canonical ACS4 in the SlWD40-OE and -RNAi fruit, respectively, these MYB–bHLH–WD40 protein complex, WD40 acts as a re- results collectively indicate that SlWD40 may also affect the cruiter and stabilizer of the MYB and bHLH protein which tomato ripening process through the regulation of the Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 261 ripening-related hormone homoeostasis. However, consider- (2) the presence of an eQTL for the gene, and (3) the pres- ably further experimentation will be required in order to ence of a functional annotation. Subsequently, these candi- test this hypothesis. dates were sorted into specific and nonspecific eQTL based on their expression in respective IL. Here, specific and non- Conclusion specific eQTL were defined based on the expression of a Our study aimed at better understanding the molecular particular S. pennellii candidate in a specific IL or several ILs, mechanisms underlying tomato fruit ripening. Comparative respectively. Specific eQTL candidates were then focused transcriptomics of small green fruited wild species with red and classified based on their function. For promoter analysis, fruited S. lycopersicum alongside eQTL mapping allowed the the 1-kb upstream sequence from the start codon of identification of key candidate genes involved in tomato SlWD40 was retrieved from the SGN tomato genome web fruit ripening. Utilizing co-expression networks alongside de- browser and the cis-regulatory elements were analyzed using tailed metabolome and transcriptome analysis indicated PlantCARE (http://bioinformatics.psb.ugentbe/webtools/ that SlWD40 has a positive impact on tomato ripening pro- plantcare/html/) and PlantPAN 2.0 (http://plantpan2.itps. cess and suggest that it may act in concert with strongly ncku.edu.tw/index.html) web tools. co-expressed TFs such as RIN, NOR, AP2a,and SlWRKYs (Figure 10). Beyond these insights into ripening, we believe SlWD40 co-expression network construction with our study also acts as a proof-of-concept study whereby the tomato ripening pathway genes transcriptome of phenotypically divergent wild relatives, We listed 171 target genes involved in carotenoid biosynthe- alongside eQTL mapping, can be used to identify causal sis, tomato fruit ripening, and cell wall metabolic pathways genes underlying trait variance. and their regulation. Some of these genes were well charac- terized. For all 171 genes, expression values were extracted Materials and methods from Tomato Genome Consortium (2012). The R script Narrowing down candidate genes involved in fruit written by Contreras-Lopez et al. (2018) was used to calcu- late correlation values and P values, both positive and nega- ripening tive correlation values were calculated and cytoscape was For M82 and S. pennellii (Penn) fruit, RNA-seq data are used to visualize network (Shannon et al., 2003). available (Bolger et al., 2014). Moreover, RNA-seq data for S. pennellii ILs fruit were also available (http://ted.bti.cornell. Virus-induced gene silencing edu/cgi-bin/TFGD/array_data/home.cgi). Starting with the Vector construction, infiltration, and fruit harvesting proce- RNA-seq data for M82 and S. pennellii fruit (Bolger et al., dures were performed as previously described (Orzaez et al., 2014) all 34,727 genes in the transcriptome were sorted in 2006, 2009). Briefly, an approximately 300-bp fragment of two ways. Firstly, the ratio of their expression value in M82 the candidate gene was amplified from tomato M82 fruit relative to that in S. pennellii and secondly, by ratio of their expression value of S. pennellii to that of M82 in order to cDNA using gateway compatible primers and recombined get genes that are highly expressed in M82 and S. pennellii, into the GATEWAY vector pDONR207 (Invitrogen, http:// www.invitrogen.com/) by the BP reaction following the respectively. Candidates from these two lists were further fil- manufacturer’s protocol to generate an entry clone. An er- tered by using three different criteria namely (1) that the relative fold change (FC) in the expression was at least 45, ror-free entry vector was confirmed by sequencing and then Figure 10 Proposed schematic overview of network of regulatory factors controlling tomato fruit ripening. 262 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. recombined with the pTRV2-Ros/Del/GW destination vector available in Zenodo (https://zenodo.org/) (doi:10.5281/zen- using an LR reaction to produce the expression clones odo.5525948 and 10.5281/zenodo.5525946). The RNA seq pTRV2-Ros/Del/GW-Respective Gene ID. Agrobacterium data were analyzed using LSTrAP (Proost et al., 2017). The tumefaciens strain GV3101:pMP90 was then transformed clean reads of each sample were aligned to the Tomato with sequenced expression vectors by electroporation. In or- Genome version SL4.0 and Annotation ITAG4.0 (ftp://ftp.sol der to infiltrate fruit for VIGS, purple MicroTom tomato genomics.net/tomato_genome/annotation/ITAG4.0_release/). was used and agroinfiltration was performed as previously The DEGs between transgenic fruit and WT fruit were identi- described (Alseekh et al., 2015). fied under the parameter of FC 5 2and FDR 5 0.05. Development of OE and RNAi lines Metabolic profiling The sequence encoding Solyc04g005020 was amplified from Fruit pericarp samples were harvested, immediately frozen in S. lycopersicum cv. Moneymaker (MM) cDNA by using gene- liquid nitrogen, and stored at –80 C until further analysis. -1 specific primers and inserted into the pDONR207 by attB re- Samples were then powdered by using retsch mill at 30 Ls , combination to generate entry clone. Primer sequences are for 30 s. Extraction of pigments, primary metabolites, lipid, provided in Supplemental Table S8.An error-free entry and secondary metabolites was performed as described previ- clone was confirmed by sequence analysis before recombina- ously (Salem et al., 2016). In brief, 500mL of the upper lipid tion into destination vector B33BinAR for fruit-specific OE and pigments containing phase was dried in a SpeedVac and named as B33BinAR_SlWD40. Additionally, artificial concentrator and resuspended in 250mL acetonitrile: 2-prop- miRNA (amiRNA) cassette was designed for Solyc04g005020. anol (7:3, v/v) solution. Two microliters of the solution were For this, Solyc04g005020 cDNA sequence was used as target analyzed by the Waters Acquity ultra-performance LC system sequence, employing the WMD3 program (http://wmd3.wei coupled with Fourier transform MS in positive ionization gelworld.org/cgi-bin/webapp.cgi) to design corresponding mode. Moreover, 150 and 300 lL of the polar phase were amiRs. An overlapping PCR (polymerase chain reaction) dried in a centrifugal vacuum concentrator for primary and strategy was employed with in-hand precursor DNA, follow- secondary metabolite profiling. The primary metabolite ing the WMD3 protocol (http://wmd3.weigelworld.or/down pellet was resuspended in 40 lL of methoxyaminhydrochlor- loads/CloningofartificialmicroRNAs.pdf). The pre-amiRs -1 ide (20 mgmL in pyridine) and derivatized for 2 h at obtained from overlapping PCR (using the athmir-319a 37 C. Afterward, 70 lLof N-methyl-N-[trimethylsilyl] trifluor- backbone) were cloned into the pENTR/D-TOPO vector and -1 oacetamide was added containing 20 lLmL fatty acid the clones were confirmed by DNA sequencing. methyl esters mixture as retention time standards. The mix- Subsequently, these sequences were cloned into B33BinAR ture was incubated for 30 min at 37 C at 400 rpm. A vol- via Asp718 and BamHI digestion and cohesive end ligation. ume of 1 lL of this solution was used for injection. The gas Primer sequences are provided in Supplemental Table S8. chromatography–mass spectroscopy system comprised a This and other final LR plasmids were then introduced into CTC CombiPAL autosampler, an Agilent 6890N gas chro- A. tumefaciens strain GV2260 by electroporation and subse- matograph, and a LECO Pegasus III time of flight mass spec- quently submitted for transformation into MM plants using trometry (TOF-MS) running in EI + mode. The secondary the leaf disc transformation method (McCormick et al., metabolite pellet was resuspended in 200-mL50% (v/v) 1986). methanol in water and 2 mL was injected on RP high strength silica T3 C column using a Waters Acquity UPLC Plant material and growth conditions system. The analysis workflow included peak detection, re- Transgenic plants for each genotype were selected on kana- -1 tention time alignment, and removal of chemical noise fol- mycin containing MS medium (50 mgL ). SNN and MM lowing the method of Salem et al. (2016). For metabolites (WT) were germinated on MS medium without kanamycin. and transcriptome data processing, the PCA and heat map Both transgenic lines and WT were selected and transferred analysis were performed by MetaboAnalyst 5.0 (https://www. to soil pot for cultivation under long-day conditions (16-h/ metaboanalyst.ca/). 8-h day/night cycle) at 22 C and 50% humidity, as described previously in the literature (Carrari et al., 2003). Upon anthe- RT-qPCR analysis sis, flowers were labeled with that particular date. Total RNA was extracted from fruit using TRIzol reagent (Invitrogen, Waltham, MA, USA). And the first-strand cDNA Transcriptome analysis synthesis was carried out as the manufacturer’s instructions Two biological replicate samples from two independent of PrimeScript RT Reagent Kit with gDNA Eraser (Takara, plants of each genotype of MG, Br, and Pink stages have Shiga, Japan). RT-qPCR was analyzed on an ABI Prism 7900 been harvested. Total RNA was extracted using the HT real-time PCR system (Applied Biosystems/Life NucleoSpin RNA Plant kit (Macherey-Nagel) and sent to the Novogene Company (Beijing, China) for Illumina HiSeq Technologies, Darmstadt, Germany) in 384-well PCR plates. –DDCt PE150 sequencing. The cDNA library was constructed follow- The RT-qPCR data were analyzed using the 2 analysis ing the manufacturer’s recommendations and then purified method according to Bustin et al. (2009) and all primers are to remove the low-quality sequences. The clean data are listed in Supplemental Table S8. Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 263 Statistical analysis FPA No. 664620). S.S.J. acknowledges funding by ICAR, India, Student’s paired t test was performed to assess whether the in the form of ICAR-International Fellowship. differences between different genotypes were statistically sig- Conflict of interest statement. None declared. nificant. The asterisks indicate statistically significant differ- ences determined by the Student’s t test (two-tail): *P5 0.05; **P5 0.01. References Accession numbers Alseekh S, Ofner I, Pleban T, Tripodi P, Di Dato F, Cammareri M, Mohammad A, Grandillo S, Fernie AR, Zamir D (2013) Sequence data from this article can be found in the Resolution by recombination: breaking up Solanum pennellii intro- GenBank/EMBL data libraries under accession numbers gressions. Trends Plant Sci 18: 536–538 SlWD40, Solyc04g005020. Alseekh S, Tohge T, Wendenberg R, Scossa F, Omranian N, Li J, Kleessen S, Giavalisco P, Pleban T, Mueller-Roeber B, et al. Supplemental data (2015) Identification and mode of inheritance of quantitative trait loci for secondary metabolite abundance in tomato. Plant Cell 27: The following materials are available in the online version of 485–512 this article. Ballester AR, Molthoff J, de Vos R, Hekkert B, Orzaez D, Supplemental Figure S1. VIGS phenotype of candidate Fernandez-Moreno JP, Tripodi P, Grandillo S, Martin C, Heldens J, et al. (2010) Biochemical and molecular analysis of pink genes. tomatoes: deregulated expression of the gene encoding transcrip- Supplemental Figure S2. Promoter analysis of SlWD40 for tion factor SlMYB12 leads to pink tomato fruit color. Plant Physiol thepresenceof ethylene (C2H2, AP2, EIN), auxin, and 152: 71–84 MADS-box binding-related cis-regulatory elements. Baranwal VK, Negi N, Khurana P (2021) Comparative transcriptom- Supplemental Figure S3. Genotyping of the SlWD40 of ics of leaves of five mulberry accessions and cataloguing structural and expression variants for future prospects. PLoS ONE 16: T0 transformants. e0252246 Supplemental Table S1. Co-expression network of Bartley GE, Scolnik PA (1993) cDNA cloning, expression during de- SlWD40 with tomato ripening pathway specific genes. velopment, and genome mapping of PSY2, a second tomato gene Supplemental Table S2. Transcriptome profiling of encoding phytoene synthase. J Biol Chem 268: 25718–25721 SlWD40 transgenic fruits. Batyrshina ZS, Yaakov B, Shavit R, Singh A, Tzin V (2020) Comparative transcriptomic and metabolic analysis of wild and Supplemental Table S3. The overlapped DEGs of SlWD40 domesticated wheat genotypes reveals differences in chemical and OE and RNAi fruit. physical defense responses against aphids. BMC Plant Biol 20:19 Supplemental Table S4. Functional categorization of Baxter CJ, Carrari F, Bauke A, Overy S, Hill SA, Quick PW, Fernie DEGs of SlWD40. AR, Sweetlove LJ (2005) Fruit carbohydrate metabolism in an in- Supplemental Table S5. The overlap DEGs of SlWD40, trogression line of tomato with increased fruit soluble solids. Plant Cell Physiol 46: 425–437 rin,and nor mutants. Bird CR, Ray JA, Fletcher JD, Boniwell JM, Bird AS, Teulieres C, Supplemental Table S6. Metabolite profiling of SlWD40 Blain I, Bramley PM, Schuch W (1991) Using antisense RNA to transgenic fruits. study gene-function—inhibition of carotenoid biosynthesis in Supplemental Table S7. Expression of SlWD40 in ILs. transgenic tomatoes. Bio-Technology 9: 635–639 Supplemental Table S8. Primer sequences used in this Bolger A, Scossa F, Bolger ME, Lanz C, Maumus F, Tohge T, Quesneville H, Alseekh S, Sorensen I, Lichtenstein G, et al. study. (2014) The genome of the stress-tolerant wild tomato species Supplemental Data Set S1. Finalized potential candidates Solanum pennellii. Nat Genet 46: 1034–1038 from both eQTL and TF approaches and their VIGS Breschi A, Gingeras TR, Guigo R (2017) Comparative transcriptom- phenotypes. ics in human and mouse. Nat Rev Genet 18: 425–440 Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista Acknowledgments M, Mueller R, Nolan T, Pfaffl MW, Shipley GL, et al. (2009) The MIQE guidelines: minimum information for publication of quanti- We thank Dr Youjun Zhang and Regina Wendenburg from tative real-time PCR experiments. Clin Chem 55: 611–622 Max-Planck-Institut fu ¨r Molekulare Pflanzenphysiologie for Carrari F, Baxter C, Usadel B, Urbanczyk-Wochniak E, Zanor MI, useful discussion and experiment assistance. Nunes-Nesi A, Nikiforova V, Centero D, Ratzka A, Pauly M, et al (2006) Integrated analysis of metabolite and transcript levels Funding reveals the metabolic shifts that underlie tomato fruit develop- ment and highlight regulatory aspects of metabolic network be- F.Z. acknowledges funding of The Key R&D Program havior. Plant Physiol 142: 1380–1396 of Hubei Province (2021BBA095) and the National Carrari F, Fernie AR (2006) Metabolic regulation underlying tomato Natural Science Foundation of China (32002102) and fruit development. J Exp Bot 57: 1883–1897 the work in Fernie Lab was supported by the Deutsche Carrari F, Nunes-Nesi A, Gibon Y, Lytovchenko A, Loureiro ME, Fernie AR (2003) Reduced expression of aconitase results in an en- Forschungsgemeinschaft in the framework of Deutsche hanced rate of photosynthesis and marked shifts in carbon parti- Israeli Project FE 552/12-1. In addition, S.A. and A.R.F. ac- tioning in illuminated leaves of wild species tomato. Plant Physiol knowledge funding of the PlantaSYST project by the 133: 1322–1335 European Union’s Horizon 2020 Research and Innovation Cazzonelli CI, Pogson BJ (2010) Source to sink: regulation of carot- Programme (SGA-CSA No. 664621 and No. 739582 under enoid biosynthesis in plants. Trends Plant Sci 15: 266–274 264 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. Centeno DC, Osorio S, Nunes-Nesi A, Bertolo ALF, Carneiro RT, Giovannoni J, Nguyen C, Ampofo B, Zhong S, Fei Z (2017) The epi- Arau´jo WL, Steinhauser M-C, Michalska J, Rohrmann J, genome and transcriptional dynamics of fruit ripening. Annu Rev Geigenberger P, et al. (2011) Malate plays a crucial role in starch Plant Biol 68: 61–84 metabolism, ripening, and soluble solid content of tomato fruit Hileman LC, Sundstrom JF, Litt A, Chen M, Shumba T, Irish VF (2006) Molecular and phylogenetic analyses of the MADS-box and affects postharvest softening. Plant Cell 23: 162–184 gene family in tomato. Mol Biol Evol 23: 2245–2258 Chang YM, Lin HH, Liu WY, Yu CP, Chen HJ, Wartini PP, Kao YY, Irfan M, Ghosh S, Meli VS, Kumar A, Kumar V, Chakraborty N, Wu YH, Lin JJ, Lu MJ, et al. (2019) Comparative transcriptomics Chakraborty S, Datta A (2016) Fruit ripening regulation of method to infer gene coexpression networks and its applications alpha-mannosidase expression by the MADS box transcription fac- to maize and rice leaf transcriptomes. Proc Natl Acad Sci USA tor RIPENING INHIBITOR and ethylene. Front Plant Sci 7:10 116: 3091–3099 Ito Y, Nishizawa-Yokoi A, Endo M, Mikami M, Shima Y, Chen L, Li W, Li Y, Feng X, Du K, Wang G, Zhao L (2019) Nakamura N, Kotake-Nara E, Kawasaki S, Toki S (2017) Identified trans-splicing of YELLOW-FRUITED TOMATO 2 encoding Re-evaluation of the rin mutation and the role of RIN in the in- the PHYTOENE SYNTHASE 1 protein alters fruit color by duction of tomato ripening. Nat Plants 3: 866–874 map-based cloning, functional complementation and RACE. Plant Karlova R, Rosin FM, Busscher-Lange J, Parapunova V, Do PT, Mol Biol 100: 647–658 Fernie AR, Fraser PD, Baxter C, Angenent GC, de Maagd RA Chen W, Chen L, Zhang X, Yang N, Guo J, Wang M, Ji S, Zhao X, (2011) Transcriptome and metabolite profiling show that Yin P, Cai L, et al. (2022) Convergent selection of a WD40 protein APETALA2a is a major regulator of tomato fruit ripening. Plant that enhances grain yield in maize and rice. Science 375: eabg7985 Cell 23: 923–941 Cheng Y, Ahammed GJ, Yu J, Yao Z, Ruan M, Ye Q, Li Z, Wang R, Kimbara J, Ohyama A, Chikano H, Ito H, Hosoi K, Negoro S, Feng K, Zhou G, et al. (2016) Putative WRKYs associated with Miyatake K, Yamaguchi H, Nunome T, Fukuoka H, et al. (2018) regulation of fruit ripening revealed by detailed expression analysis QTL mapping of fruit nutritional and flavor components in to- of the WRKY gene family in pepper. Sci Rep 6: 39000 mato (Solanum lycopersicum) using genome-wide SSR markers and Chitwood DH, Kumar R, Headland LR, Ranjan A, Covington MF, recombinant inbred lines (RILs) from an intra-specific cross. Ichihashi Y, Fulop D, Jimenez-Go´mez JM, Peng J, Maloof JN, Euphytica 214: 210 et al (2013) A quantitative genetic basis for leaf morphology in a Kumar R, Agarwal P, Tyagi AK, Sharma AK (2012) Genome-wide set of precisely defined tomato introgression lines. Plant Cell 25: investigation and expression analysis suggest diverse roles of 2465–2481 auxin-responsive GH3 genes during development and response to Chung MY, Vrebalov J, Alba R, Lee J, McQuinn R, Chung JD, Klein different stimuli in tomato (Solanum lycopersicum). Mol Genet P, Giovannoni J (2010) A tomato (Solanum lycopersicum) Genomics 287: 221–235 APETALA2/ERF gene, SlAP2a, is a negative regulator of fruit ripen- Lanahan MB, Yen HC, Giovannoni JJ, Klee HJ (1994) The never ing. Plant J 64: 936–947 ripe mutation blocks ethylene perception in tomato. Plant Cell 6: Contreras-Lopez O, Moyano TC, Soto DC, Gutierrez RA (2018) 521–530 Step-by-step construction of gene co-expression networks from Li S, Chen K, Grierson D (2021) Molecular and hormonal mecha- high-throughput Arabidopsis RNA sequencing data. Methods Mol nisms regulating fleshy fruit ripening. Cells 10: 1136 Biol 1761: 275–301 Li Y, Chen Y, Zhou L, You S, Deng H, Chen Y, Alseekh S, Yuan Y, D’Amelia V, Raiola A, Carputo D, Filippone E, Barone A, Rigano Fu R, Zhang Z, et al. (2020) MicroTom metabolic network: rewir- MM (2019) A basic helix-loop-helix (SlARANCIO), identified from ing tomato metabolic regulatory network throughout the growth a Solanum pennellii introgression line, affects carotenoid accumula- cycle. Mol Plant 13: 1203–1218 tion in tomato fruits. Sci Rep 9: 3699 Liu M, Gomes BL, Mila I, Purgatto E, Peres LE, Frasse P, Maza E, Eshed Y, Zamir D (1995) An introgression line population of Zouine M, Roustan JP, Bouzayen M, et al. (2016) Comprehensive Lycopersicon pennellii in the cultivated tomato enables the identifi- profiling of ethylene response factor expression identifies cation and fine mapping of yield-associated QTL. Genetics 141: ripening-associated ERF genes and their link to key regulators of 1147–1162 fruit ripening in tomato. Plant Physiol 170: 1732–1744 Fernandez-Moreno JP, Tzfadia O, Forment J, Presa S, Rogachev I, Lu P, Yu S, Zhu N, Chen YR, Zhou B, Pan Y, Tzeng D, Fabi JP, Meir S, Orzaez D, Aharoni A, Granell A (2016) Characterization Argyris J, Garcia-Mas J, et al. (2018) Genome encode analyses re- of a new pink-fruited tomato mutant results in the identification veal the basis of convergent evolution of fleshy fruit ripening. Nat of a null allele of the SlMYB12 transcription factor. Plant Physiol Plants 4: 784–791 171: 1821–1836 Manning K, Tor M, Poole M, Hong Y, Thompson AJ, King GJ, Fridman E, Carrari F, Liu YS, Fernie AR, Zamir D (2004) Zooming Giovannoni JJ, Seymour GB (2006) A naturally occurring epige- in on a quantitative trait for tomato yield using interspecific intro- netic mutation in a gene encoding an SBP-box transcription factor gressions. Science 305: 1786–1789 inhibits tomato fruit ripening. Nat Genet 38: 948–952 Fujisawa M, Nakano T, Shima Y, Ito Y (2013) A large-scale identifi- Martel C, Vrebalov J, Tafelmeyer P, Giovannoni JJ (2011) The to- cation of direct targets of the tomato MADS box transcription fac- mato MADS-box transcription factor RIPENING INHIBITOR inter- tor RIPENING INHIBITOR reveals the regulation of fruit ripening. acts with promoters involved in numerous ripening processes in a Plant Cell 25: 371–386 COLORLESS NONRIPENING-dependent manner. Plant Physiol 157: Fujisawa M, Shima Y, Higuchi N, Nakano T, Koyama Y, Kasumi T, 1568–1579 Ito Y (2012) Direct targets of the tomato-ripening regulator RIN McCormick S, Niedermeyer J, Fry J, Barnason A, Horsch R, Fraley identified by transcriptome and chromatin immunoprecipitation R (1986) Leaf disc transformation of cultivated tomato (L. esculen- analyses. Planta 235: 1107–1122 tum) using Agrobacterium tumefaciens. Plant Cell Rep 5: 81–84 Gao L, Gonda I, Sun H, Ma Q, Bao K, Tieman DM, Burzynski- Mishra AK, Puranik S, Bahadur RP, Prasad M (2012) The Chang EA, Fish TL, Stromberg KA, Sacks GL, et al. (2019) The DNA-binding activity of an AP2 protein is involved in transcrip- tomato pan-genome uncovers new genes and a rare allele regulat- tional regulation of a stress-responsive gene, SiWD40, in foxtail mil- ing fruit flavor. Nat Genet 51: 1044–1051 let. Genomics 100: 252–263 Gao Y, Wei W, Fan Z, Zhao X, Zhang Y, Jing Y, Zhu B, Zhu H, Mutwil M, Klie S, Tohge T, Giorgi FM, Wilkins O, Campbell MM, Shan W, Chen JJ (2020) Re-evaluation of the nor mutation and Fernie AR, Usadel B, Nikoloski Z, Persson S (2011) PlaNet: com- the role of the NAC-NOR transcription factor in tomato fruit rip- bined sequence and expression comparisons across plant networks ening. J Exp Bot 71: 3560–3574 derived from seven species. Plant Cell 23: 895–910 Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 265 Ofner I, Lashbrooke J, Pleban T, Aharoni A, Zamir D (2016) Shi Y, Vrebalov J, Zheng H, Xu Y, Yin X, Liu W, Liu Z, Sorensen I, Solanum pennellii backcross inbred lines (BILs) link small genomic Su G, Ma Q, et al. (2021) A tomato LATERAL ORGAN bins with tomato traits. Plant J 87: 151–160 BOUNDARIES transcription factor, SlLOB1, predominantly regu- Orzaez D, Medina A, Torre S, Fernandez-Moreno JP, Rambla JL, lates cell wall and softening components of ripening. Proc Natl Fernandez-Del-Carmen A, Butelli E, Martin C, Granell A (2009) Acad Sci USA 118: e2102486118 A visual reporter system for virus-induced gene silencing in tomato Shinozaki Y, Nicolas P, Fernandez-Pozo N, Ma Q, Evanich DJ, Shi fruit based on anthocyanin accumulation. Plant Physiol 150: Y, Xu Y, Zheng Y, Snyder SI, Martin LBB, et al. (2018) 1122–1134 High-resolution spatiotemporal transcriptome mapping of tomato Orzaez D, Mirabel S, Wieland WH, Granell A (2006) Agroinjection fruit development and ripening. Nat Commun 9: 364 of tomato fruits. A tool for rapid functional analysis of transgenes Sønderby IE, Hansen BG, Bjarnholt N, Ticconi C, Halkier BA, directly in fruit. Plant Physiol 140: 3–11 Kliebenstein DJ (2007) A systems biology approach identifies a Osorio S, Alba R, Damasceno CMB, Lopez-Casado G, Lohse M, R2R3 MYB gene subfamily with distinct and overlapping functions Zanor MI, Tohge T, Usadel B, Rose JKC, Fei Z, et al. (2011) in regulation of aliphatic glucosinolates. PLoS ONE 2: e1322 Systems biology of tomato fruit development: combined transcript, Sravankumar T, Patel A, Naik N, Kumar R (2018) A protein, and metabolite analysis of tomato transcription factor ripening-induced SlGH3-2 gene regulates fruit ripening via adjust- (nor, rin) and ethylene receptor (Nr) mutants reveals novel regula- ing auxin-ethylene levels in tomato (Solanum lycopersicum L.). tory interactions. Plant Physiol 157: 405–425 Plant Mol Biol 98: 455–469 Proost S, Krawczyk A, Mutwil M (2017) LSTrAP: efficiently combin- Steinhauser M-C, Steinhauser D, Koehl K, Carrari F, Gibon Y, ing RNA sequencing data into co-expression networks. BMC Fernie AR, Stitt M (2010) Enzyme activity profiles during fruit de- Bioinformatics 18: 444 velopment in tomato cultivars and Solanum pennellii. Plant Physiol Ramsay NA, Glover BJ (2005) MYB–bHLH–WD40 protein complex 153: 80–98 and the evolution of cellular diversity. Trends Plant Sci 10: 63–70 Szymanski J, Bocobza S, Panda S, Sonawane P, Cardenas PD, Ranjan A, Budke JM, Rowland SD, Chitwood DH, Kumar R, Lashbrooke J, Kamble A, Shahaf N, Meir S, Bovy A, et al. (2020) Carriedo L, Ichihashi Y, Zumstein K, Maloof JN, Sinha NR Analysis of wild tomato introgression lines elucidates the genetic (2016) eQTL regulating transcript levels associated with diverse bi- basis of transcriptome and metabolome variation underlying fruit ological processes in tomato. Plant Physiol 172: 328–340 traits and pathogen response. Nat Genet 52: 1111–1121 Robinson R (1968) Ripening inhibitor: a gene with multiple effects Tieman D, Zhu G, Resende MF Jr, Lin T, Nguyen C, Bies D, Rambla on ripening. Rep Tomato Genet Coop 18: 36–37 JL, Beltran KS, Taylor M, Zhang B, et al. (2017) A chemical genetic Rocha-Sosa M, Sonnewald U, Frommer W, Stratmann M, Schell J, roadmap to improved tomato flavor. Science 355: 391–394 Willmitzer L (1989) Both developmental and metabolic signals ac- Tieman DM, Zeigler M, Schmelz EA, Taylor MG, Bliss P, Kirst M, tivate the promoter of a class I patatin gene. EMBO J 8: 23–29 Klee HJ (2006) Identification of loci affecting flavour volatile emis- Rockman MV, Kruglyak L (2009) Recombinational landscape and sions in tomato fruits. J Exp Bot 57: 887–896 population genomics of Caenorhabditis elegans. PLoS Genet 5: Tigchelaar E (1973) A new ripening mutant, non-ripening (nor). Rep e1000419 Tomato Genet Coop 35:20 Rodriguez-Lopez CE, Hernandez-Brenes C, Trevino V, Diaz de la Tomato Genome Consortium (2012) The tomato genome sequence Garza RI (2017) Avocado fruit maturation and ripening: dynamics provides insights into fleshy fruit evolution. Nature 485: 635–641 of aliphatic acetogenins and lipidomic profiles from mesocarp, idi- Vallarino JG, Kubiszewski-Jakubiak S, Ruf S, Rossner M, Timm S, oblasts and seed. BMC Plant Biol 17: 159 Bauwe H, Carrari F, Rentsch D, Bock R, Sweetlove LJ, et al. Rohrmann J, McQuinn R, Giovannoni JJ, Fernie AR, Tohge T (2020) Multi-gene metabolic engineering of tomato plants results (2012) Tissue specificity and differential expression of transcription in increased fruit yield up to 23%. Sci Rep 10: 17219 factors in tomato provide hints of unique regulatory networks dur- Vrebalov J, Ruezinsky D, Padmanabhan V, White R, Medrano D, ing fruit ripening. Plant Signal Behav 7: 1639–1647 Drake R, Schuch W, Giovannoni J (2002) A MADS-box gene nec- Rohrmann J, Tohge T, Alba R, Osorio S, Caldana C, McQuinn R, essary for fruit ripening at the tomato ripening-inhibitor (rin) locus. Arvidsson S, van der Merwe MJ, Riano-Pachon DM, Mueller- Science 296: 343–346 Roeber B, et al. (2011) Combined transcription factor profiling, Wang L, Zhang XL, Wang L, Tian Y, Jia N, Chen S, Shi NB, Huang microarray analysis and metabolite profiling reveals the transcrip- X, Zhou C, Yu Y, et al. (2017) Regulation of ethylene-responsive tional control of metabolic shifts occurring during tomato fruit de- SlWRKYs involved in color change during tomato fruit ripening. Sci velopment. Plant J 68: 999–1013 Rep 7: 16674 Salem MA, Juppner J, Bajdzienko K, Giavalisco P (2016) Protocol: Wang R, Tavano E, Lammers M, Martinelli AP, Angenent GC, de a fast, comprehensive and reproducible one-step extraction Maagd RA (2019) Re-evaluation of transcription factor function in to- method for the rapid preparation of polar and semi-polar metabo- mato fruit development and ripening with CRISPR/Cas9-mutagenesis. lites, lipids, proteins, starch and cell wall polymers from a single Sci Rep 9: 1696 sample. Plant Methods 12:45 Wang S, Lu G, Hou Z, Luo Z, Wang T, Li H, Zhang J, Ye Z (2014) Sauvage C, Segura V, Bauchet G, Stevens R, Do PT, Nikoloski Z, Members of the tomato FRUITFULL MADS-box family regulate Fernie AR, Causse M (2014) Genome-wide association in tomato style abscission and fruit ripening. J Exp Bot 65: 3005–3014 reveals 44 candidate loci for fruit metabolic traits. Plant Physiol Waseem M, Li N, Su D, Chen J, Li Z (2019) Overexpression of a ba- 165: 1120–1132 sic helix–loop–helix transcription factor gene, SlbHLH22, promotes Schauer N, Semel Y, Balbo I, Steinfath M, Repsilber D, Selbig J, early flowering and accelerates fruit ripening in tomato (Solanum Pleban T, Zamir D, Fernie AR (2008) Mode of inheritance of pri- lycopersicum L.). Planta 250: 173–185 mary metabolic traits in tomato. Plant Cell 20: 509–523 Xu C, Min J (2011) Structure and function of WD40 domain pro- Semel Y, Nissenbaum J, Menda N, Zinder M, Krieger U, Issman N, teins. Protein Cell 2: 202–214 Pleban T, Lippman Z, Gur A, Zamir D (2006) Overdominant Yang L, Huang W, Xiong F, Xian Z, Su D, Ren M, Li Z (2017) quantitative trait loci for yield and fitness in tomato. Proc Natl Silencing of SlPL, which encodes a pectate lyase in tomato, confers Acad Sci USA 103: 12981–12986 enhanced fruit firmness, prolonged shelf-life and reduced suscepti- Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, bility to grey mould. Plant Biotechnol J 15: 1544–1555 Amin N, Schwikowski B, Ideker T (2003) Cytoscape: a software Ye J, Wang X, Hu T, Zhang F, Wang B, Li C, Yang T, Li H, Lu Y, environment for integrated models of biomolecular interaction Giovannoni JJ, et al. (2017) An InDel in the promoter of networks. Genome Res 13: 2498–2504 Al-ACTIVATED MALATE TRANSPORTER9 selected during tomato 266 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. domestication determines fruit malate contents and aluminum Zhang Y, Butelli E, Martin C (2014) Engineering anthocyanin bio- tolerance. Plant Cell 29: 2249–2268 synthesis in plants. Curr Opin Plant Biol 19: 81–90 Zhang J, Hu Z, Yao Q, Guo X, Nguyen V, Li F, Chen G (2018) A to- Zhu G, Wang S, Huang Z, Zhang S, Liao Q, Zhang C, Lin T, Qin mato MADS-box protein, SlCMB1, regulates ethylene biosynthesis M, Peng M, Yang C, et al. (2018) Rewiring of the fruit metabo- and carotenoid accumulation during fruit ripening. Sci Rep 8: 3413 lome in tomato breeding. Cell 172: 249–261.e12 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png PLANT PHYSIOLOGY Oxford University Press

A comparative transcriptomics and eQTL approach identifies SlWD40 as a tomato fruit ripening regulator

Loading next page...
 
/lp/oxford-university-press/a-comparative-transcriptomics-and-eqtl-approach-identifies-slwd40-as-a-VSPMWh7M4i

References (99)

Publisher
Oxford University Press
Copyright
© The Author(s) 2022. Published by Oxford University Press on behalf of American Society of Plant Biologists.
ISSN
0032-0889
eISSN
1532-2548
DOI
10.1093/plphys/kiac200
Publisher site
See Article on Publisher Site

Abstract

Although multiple vital genes with strong effects on the tomato (Solanum lycopersicum) ripening process have been identi- fied via the positional cloning of ripening mutants and cloning of ripening-related transcription factors (TFs), recent studies suggest that it is unlikely that we have fully characterized the gene regulatory networks underpinning this process. Here, combining comparative transcriptomics and expression QTLs, we identified 16 candidate genes involved in tomato fruit ripening and validated them through virus-induced gene silencing analysis. To further confirm the accuracy of the ap- proach, one potential ripening regulator, SlWD40 (WD-40 repeats), was chosen for in-depth analysis. Co-expression network analysis indicated that master regulators such as RIN (ripening inhibitor) and NOR (nonripening) as well as vital TFs includ- ing FUL1 (FRUITFUL1), SlNAC4 (NAM, ATAF1,2, and CUC2 4), and AP2a (Activating enhancer binding Protein 2 alpha) strongly co-expressed with SlWD40. Furthermore, SlWD40 overexpression and RNAi lines exhibited substantially accelerated and delayed ripening phenotypes compared with the wild type, respectively. Moreover, transcriptome analysis of these transgenics revealed that expression patterns of ethylene biosynthesis genes, phytoene synthase, pectate lyase, and branched chain amino transferase 2,in SlWD40-RNAi lines were similar to those of rin and nor fruits, which further demonstrated that SlWD40 may act as an important ripening regulator in conjunction with RIN and NOR. These results are discussed in the context of current models of ripening and in terms of the use of comparative genomics and transcriptomics as an effective route for isolating causal genes underlying differences in genotypes. Received November 19, 2021. Accepted March 28, 2022. Advance access publication May 4, 2022 V The Author(s) 2022. Published by Oxford University Press on behalf of American Society of Plant Biologists. Open Access This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Research Article Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 251 Introduction 2013; Irfan et al., 2016). Furthermore, the transcriptional be- havior of 1,000 TFs has been established during tomato rip- Given that seed dispersal is of major ecological and evolu- ening (Rohrmann et al., 2011) and gene regulator networks tionary importance for all plants and the fact that fleshy have been modeled on the basis of these data (Rohrmann fruit plays a vital role in this process, fruit ripening assumes et al., 2012). The stronger acting ripening genes mentioned a central importance in the plant life-cycle. It is well docu- above were, in contrast, identified in mutant screens intent mented that hundreds of genes display altered expression on isolating strong mutants in order to enhance tomato during this process (Karlova et al., 2011; Osorio et al., 2011), shelf-life (Tigchelaar, 1973; Lanahan et al., 1994; Vrebalov and that metabolism also undergoes concurrent dramatic et al., 2002; Manning et al., 2006). While recent years have shifts to form fruit quality (Carrari and Fernie, 2006). As one resulted in the identification of many additional genes with of the most important appearance qualities, the accumula- ripening consequences (Li et al., 2021; Shi et al., 2021), it is tion of carotenoids, when combined with naringenin chal- probable that additional contributors to this complex pro- cone tainted yellow peel, forms the reddish color of tomato cess remain to be found. As such, further genome-wide fruits (Solanum lycopersicum)(Ballester et al., 2010; analysis is required to mine the regulator affecting this pro- Fernandez-Moreno et al., 2016; Zhu et al., 2018). Among the cess and provide more comprehensive knowledge on this carotenoid biosynthesis pathway, phytoene synthase (PSY) is process. the key rate-limiting enzyme of the whole pathway; it cata- Time-series, species/accession specific or tissue specific lyzes two molecules of GGPP to form the colorless phy- and comparative transcriptomics studies have previously toene. Subsequently, under the catalysis of a series of deciphered gene regulatory networks underlying plant devel- enzymes, phytoene undergoes dehydrogenation and isomeri- opmental pathways allowing the identification of additional zation reactions to form lycopene, which is the dominant functional genes (Breschi et al., 2017; Chang et al., 2019; carotenoid of tomato fruit (Bird et al., 1991; Bartley and Batyrshina et al., 2020; Baranwal et al., 2021). For example, Scolnik, 1993; Cazzonelli and Pogson, 2010; Chen et al., Bolger et al. (2014) compared the transcriptomes of M82 2019). In addition, the texture of fruits is affected by the and Solanum pennellii, and identified 100 key candidate modification of cell walls and pectate lyase (PL), which genes related to salt and drought stress. Additionally, tran- hydrolyzes pectin and is the most substantially cell wall scriptomics studies on genetic mapping populations have gene contributor to this process identified to date (Yang defined expression QTLs (eQTLs) as genomic loci controlling et al., 2017). Besides appearance and textural qualities, an- variation in steady state levels of transcript between individ- other important quality aspect is taste, which has been at- uals (Sønderby et al., 2007) of what has subsequently be- tributed to the sugar/organic acid ratio, and volatile and come a well-characterized mapping population (Ofner et al., secondary metabolite accumulation. The key genes underly- 2016; Szymanski et al., 2020). During tomato domestication, ing the levels of these metabolites have been uncovered by many phenotypes (such as the leaf structure and ripening a range of quantitative trait loci (Fridman et al., 2004; process) of wild species, such as S. pennellii were under Tieman et al., 2006; Schauer et al., 2008; Centeno et al., strong selection and were substantially different to that of 2011) and genome-wide association studies (Sauvage et al., in the cultivar species S. lycopersicum. Based on the eQTL 2014; Tieman et al., 2017; Ye et al., 2017; Gao et al., 2019). analysis of the 76 introgression lines (ILs) from S. pennellii in Moreover, the considerable metabolic changes are coordi- the background of S. lycopersicum, Ranjan et al. (2016) iden- nated and mediated by transcription factors (TFs) and epi- tified important genetic regulators of leaf development on genome dynamics on the metabolic structural genes’ chromosomes 4 and 8. The above-mentioned studies dem- expression (Centeno et al., 2011; Rohrmann et al., 2011; onstrate the power of comparative transcriptomics in com- Giovannoni et al., 2017; Lu et al., 2018; Li et al., 2020). Over the last 50 years, several mutants, such as ripening-inhibitor bination with ILs; however, limited studies have been carried out using their integrated approach to mine the genes in- (rin), nonripening (nor), Never ripe (Nr), and Colorless nonrip- volved in tomato fruit ripening. ening (Cnr) mutations, have been identified as severely As a distant relative of the cultivated tomato S. lycopersi- impacting the tomato ripening process (Tigchelaar, 1973; cum, S. pennellii has many substantially different phenotypes Lanahan et al., 1994; Vrebalov et al., 2002; Manning et al., with the cultivated tomato and one of these is the mature 2006). Among these mutants, rin is the one of the most fa- mous ripening delaying mutants substantially lacking the fruit morphology. The mature fruit of S. lycopersicum is red ethylene burst and hindering the color change and softening and soft while the mature fruit of S. pennellii is green and hard, which renders this pair the ideal parents to cross and processes, which results from the repression of the ripening illustrate the genetic landscape of fruit ripening. The core inhibitor-macrocalyx (RIN-MC) chimeric protein (Robinson, set of 76 S. pennellii ILs, which represent whole-genome cov- 1968; Vrebalov et al., 2002; Ito et al., 2017). The integrated analysis of chromatin immunoprecipitation (ChIP)-chip and erage of S. pennellii in overlapping segments in the back- transcriptome indicated that RIN can directly induce the ex- ground of M82, have been widely used to identify the key pression of the key ripening-related structural and regulator genes of many traits such as yield and metabolic composi- genes, ACS2/4, SGR1, PSY, Cel2, EXP1, PAL1, C4H, LoxC, tion (Semel et al., 2006; Alseekh et al., 2013, 2015). In the AAT1, CNR, NOR, AP2a, and itself (Fujisawa et al., 2012, present study, to identify key candidates regulating tomato 252 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. fruit ripening, an integrated comparative transcriptomics not show such specificity and have expression in all ILs. and eQTL approach was taken utilizing S. pennellii ILs With the specific eQTL and nonspecific eQTL information, the candidates of Lyco data set were classified into 119 spe- (Eshed and Zamir, 1995). We isolated 16 candidates and provided primary validation of eight of them as being in- cific and 223 nonspecific eQTLs (Figure 1B). For the Penn volved in the ripening process via the virus-induced gene si- data set, around 105 specific and 202 nonspecific candidates were classified (Figure 1C). Here, the high number of non- lencing (VIGS) method. Following this screen, one candidate, SlWD40, was taken for further study. For this candidate sta- specific eQTL is attributed to the epistatic interactions be- ble RNAi and overexpression (OE) lines were generated and tween S. pennellii alleles and M82 alleles or, alternatively, the presence of a large number of trans-QTL as previously characterized. The OE of SlWD40 promoted ripening while reported for leaf expression analysis (Chitwood et al., 2013) its inhibition inhibited it. The co-expression networks, and fruit enzyme abundance analysis (Steinhauser et al., metabolome and transcriptome analysis indicated that 2010). Previous studies have indicated that the candidates SlWD40 acted as a positive regulator of tomato ripening whose functional categories belong to transcription regula- with the key ripening TFs such as RIN, NOR, AP2a,and tors, oxidase and cytochrome P450 may be involved in regu- SlWRKYs. These results are discussed within the context of lating tomato fruit ripening and secondary metabolism; their implications regarding fruit ripening as well as with re- therefore, we chose seven candidate genes that were of spect to the utility of genomic information in filling our these three functional categories among the specific eQTL knowledge gaps in important biological processes. candidates. Results In a parallel approach, given that TFs act as important regulators in fruit ripening, we also adopted a TF-centric ap- Integrating comparative transcriptomics and eQTL proach (Figure 1D). From the total of 34,727 genes, candi- mapping to mine for genes involved in tomato dates annotated as TFs and displaying more than five times ripening higher expression in S. lycopersicum were selected. Next, Our earlier work described a high-quality genome assembly eQTL mapping thinned the list to 127 candidates which of the parents of the Solanum pennellii IL population as well were then arranged with respect to the ratio of their expres- as identifying candidate genes involved in salt as well as sion in Breaker + 10 to that in Breaker stage. Based on puta- drought stress tolerance (Bolger et al., 2014). Surprisingly, tive ortholog information (Arabidopsis thaliana and S. the open-reading frame sequence of most well-characterized lycopersicum) and literature survey concerning their putative ripening-related genes is identical between S. pennellii and S. functions, a final set of 20 candidates was selected. Finally, lycopersicum. We therefore thought to try a comparative on the basis of tissue specific expression and the S. lycopersi- transcriptomics approach of the S. pennellii IL population cum to S. pennellii expression ratio, the 7 candidates from since studies on fruit gene expression of a subset of the ILs eQTL approach and 20 candidates from TF approach were has proven highly informative (Baxter et al., 2005; Alseekh narrowed down to the 16 potential candidates described in et al., 2015)as well as in leaves (Chitwood et al., 2013). For Supplemental Data Set S1. this purpose, as an initial approach, transcriptome data for M82 and S. pennellii mature fruits were sorted as follows: VIGS analysis of candidate genes the total 34,727 genes in transcriptome sorted into two dif- To provide preliminary analysis of the function of the candi- ferent data sets named as Lyco and Penn (Figure 1A)based date genes in tomato ripening, we carried out VIGS experi- on the ratio of their expression values. The Lyco data set ment using purple Microtom cv. tomato fruit which contained genes that are highly expressed in M82 (13,521), accumulate high amount of anthocyanin resulting from the while the Penn data set contained genes that are highly introduction of Del/Ros1 petunia (Petunia hybrida)TFs expressed in S. pennellii (11,781) (Figure 1A). For the Lyco (Orzaez et al., 2009). Partial fragments of the 16 candidate genes which we reasoned would be more likely to harbor genes were cloned into pTRV2-Ros/Del/GW vector. Around genes underlying the “red” ripe phenotype of cultivated to- 10–15 fruits per plant were infected with agrobacterium car- matoes, around 300 candidates could be narrowed down by rying the respective VIGS vector. After silencing Del/Ros1 using three independent filters. Firstly, we chose to focus on (empty vector) in Microtom Del/Ros1 fruits, there was de- genes for which expression was at least five times higher in pletion in purple pigments but not in lycopene content due S. lycopersicum with respect to S. pennellii. Secondly, based to the silenced part accumulating less purple anthocyanin on the transcriptome profiling of red ripe fruit from S. lyco- pigments and thereby being easy to discriminate from non- persicum (M82) parent and a set of lines with distinct silenced (purple pigment rich) tissues (Figure 2). Phenotypes introgressed S. pennellii segments (http://ted.bti.cornell.edu/ were scored visually after 15 d of infection for all the 16 vali- cgi-bin/TFGD/array_data/home.cgi), large numbers of spe- dated candidates (Supplemental Data Set S1). cific eQTL and nonspecific eQTL have been identified as the Red color of western tomato cultivars represents the accu- former definition that specific eQTL candidates are the can- mulation of lycopene, which is an important indicator of to- didates whose expression are sharply (exponentially) in- mato ripening. VIGS for the structural genes encoded by creased or decreased in its located IL compared with other Solyc01g094080 and Solyc03g095900 displayed a red pheno- 75 ILs while nonspecific eQTL candidates’ expression does type, indicating that these genes are not associated with Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 253 Figure 1 Candidate gene filtration by integrating comparative transcriptomics and eQTL mapping. A, Pipelines for candidate gene filtration of eQTL approach. Filters are shown in bullet points. B, Heat map of relative expression level of filtered candidates. Lyco/Penn, Genes were sorted by ratio of expression value for S. lycopersicum and S. pennellii. ILs are arranged as per the number of chromosome (X-axis). Genes are arranged according to their Gene IDs (Y-axis). Regions of red or blue indicate that the gene expression is increased or decreased, respectively, over that of M82. Chr, chromosome. C, Heat map of relative expression level of filtered candidates. Penn/Lyco, Genes were sorted by the ratio of expression value for S. pennellii and S. lycopersicum. ILs are arranged as per the number of chromosome (X-axis). Genes are arranged according to their Gene IDs (Y-axis). Regions of red or blue indicate that the gene expression is increased or decreased, respectively, over that of M82. D, Pipelines for can- didate gene filtration of TFs approach. Filters are shown in bullet points. Br, Break. Figure 2 VIGS of empty vector (silencing of Del/Ros1) and SlWD40 in Microtom Del/Ros1 fruit. 254 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. lycopene biosynthesis or the pathways that fuel it. However, the master ripening-related TFs, SlWD40 also highly co- as the TFs regulating ripening are generally reported to hin- expressed with key carotenoid-related genes (PSY1), as well as ethylene (ACS4) and abscisic acid (NCED3) biosynthesis der carotenoid biosynthesis, the yellowish phenotypes of Solyc11g010710 (ethylene response factors, ERF TF) and genes and cell wall modification genes (PL and PMEI) Solyc07g052700 (MADS-box TF, AGL66) VIGS fruits indicate (Supplemental Table S1). All of these results indicate that SlWD40 may act in concert with the better characterized that they may function as a ripening regulators in line with ripening TFs to regulate the ripening processes, including former studies that implicated SlERFs and MADS-box TFs in those dependent on changes in pigmentation, hormone lev- tomato fruit ripening (Wang et al., 2014; Liu et al., 2016; els, and signaling and cell wall modification. Supplemental Data Set S1 and Supplemental Figure S1). Interestingly, the VIGS fruits of a transcription regulator, SlWD40 affects the tomato fruit transcriptome SlWD40 (Solyc04g005020) also exhibited a yellowish pheno- To confirm the accuracy of our approach and to assess in type. Given that ERF and MADS box family TFs are already detail the function of SlWD40 in the tomato ripening pro- well-known to be involved in tomato fruit ripening and that cess, we chose the fruit specific patatin B33 promoter which SlWD40 was identified as a downstream target gene of RIN has been widely used for fruit specific expression to carry out (Fujisawa et al., 2013), we selected SlWD40 for in-depth the stable transformation (Rocha-Sosa et al., 1989; Vallarino analysis here (Figure 2). et al., 2020). T0 transformants of RNAi and OE lines were characterized by NPT-II-specific polymerase chain reaction Co-expression network and VIGS of SlWD40 (PCR). Real-time quantitative polymerase chain reaction confirmed its role in tomato fruit ripening (RT-qPCR) was also carried out using fruit samples from In order to analyze the function of SlWD40 on fruit ripening, promising T0 transgenics to select high OE and knockdown we initially identified its potential regulators following cis- lines to raise T1 generation (Supplemental Figure S3). Fruits regulatory element analysis of the promoter of SlWD40.This from all generations were analyzed and phenotype was stable analysis indicated that the promoter contained several ethyl- over T0 and T1 generations. Before the T1 generation plant ene (AP2, B3, EIN3, and EIL) and ripening-related elements transplant to soil, we also used the NPT-II-specific PCR to (C2H2, MADS, NF-YB, NF-YA, and NF-YC) in the 1-kb pro- confirm that the plants are transgenic. Based on the expres- moter region, which indicated that it may well be induced sion of SlWD40, two independent lines of RNAi (RNAi-1 and by the ripening and ethylene burst (Supplemental Figure -2 lines) and OE (OE-1 and -2 lines) were chosen for subse- S2). Moreover, the evaluation of publicly available expression quent experiments (Supplemental Figure S3). data with tissue-specific expression analysis of SlWD40 con- To analyze fruit phenotype at the identical stage, fruits of firmed the hypothesis that SlWD40 was only slightly each genotype were labeled upon anthesis and harvested for expressed in the leaf, bud, flower, root, and young fruit but phenotyping, transcriptome, and metabolite profiling at ma- that its expression increases exponentially following mature ture green (MG, 34 DPA), breaker (Br, 37DPA), and pink green stage (Tomato Genome Consortium, 2012). (Pink, 45 DPA) stages of the wild type (WT). As seen in Intriguingly, its expression in different cell types of the to- Figure 4, the development and ripening process were sub- mato fruit revealed that it is highly similar to that of the stantially hindered in the RNAi fruit while the ripening pro- other known ripening regulators, such as RIN and NOR cess was significantly accelerated in comparison to the OE (Shinozaki et al., 2018). lines. Moreover, especially at the Br stage of WT fruits, the Moreover, given that assembly of co-expression networks size of RNAi fruits was significantly smaller than that of the is an efficient method to identify the important interactions OE and WT lines and the RNAi fruits were still at the ma- and relationship among different genes (Mutwil et al., 2011), ture green stage while the OE fruit were almost at the pink available transcriptome data of different organ and fruit de- stage. The contents of chlorophylls and carotenoids, some velopment stages were used to construct tomato co- of the most important parameters of fruit ripening, also expression network (Figure 3 and Supplemental Table S1; confirmed the positive function of SlWD40 on tomato ripen- Tomato Genome Consortium, 2012). The co-expression sub- ing process: The degradation of chlorophylls and synthesis network containing SlWD40 included 171 structure genes/ of the predominant carotenoid, lycopene, were significantly regulators, which are involved in chlorophyll and carotenoid hindered in RNAi fruits but accelerated in the OE fruits metabolism as well as tomato fruit ripening and cell wall (Figure 4B). metabolic pathways. Among the 171 genes, a total of 62 In order to estimate the effect of SlWD40 on the global genes exhibited high co-expression phenotype (jCo-expres- difference of gene expression during the different fruit devel- sion Coefficientsj 40.6, P5 0.05) with SlWD40 opmental stages, we additionally analyzed the differentially (Supplemental Table S1). Consistent with the results of expressed genes (DEGs) among the RNAi, OE lines, and WT cis-regulatory element analysis, three MADS TFs, including fruit at MG, Br, and Pink stages. For this purpose, we used RIN, two AP2s TFs, and one ARF TF were significantly posi- FPKM (fragments per kilobase per million mapped frag- tively co-expressed with SlWD40. Moreover, another vital ments) and identified genes with jlog (fold change) j5 1 ripening-related TF, NOR, exhibited a co-expression coeffi- and false discovery rate (FDR) (corrected P value)5 0.05 cient of 0.86 with SlWD40 (Supplemental Table S1). Besides (Supplemental Table S2). Firstly, we checked the SlWD40 Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 255 Figure 3 Co-expression network of SlWD40 with tomato ripening pathway-specific genes. Well-characterized key regulators such as RIN (ripening inhibitor), NOR (nonripening), and FUL1 (FRUITFUL1) (labeled in blue) strongly co-expressed with SlWD40. ELIP, early light induced protein; NF-Y, nuclear factor Y; MADS, MADS domain protein. Figure 4 Photographs and pigments content of WT and T1 generation RNAi (lines 1 and 2) and OE (lines 1 and 2) lines at 28, 34 (MG), 37 (Br), and 45 (Pink) DPA fruits. A, Photographs of WT and transgenic SlWD40 fruits. Images were digitally extracted for comparison. B, Chlorophylls and carotenoids of WT and transgenic SlWD40 fruits. The values in each column are the mean of at least three biological replicates. Error bars indicate SD. The asterisks indicate statistically significant differences determined by the Student’s t test (two-tail): *P5 0.05; **P5 0.01. ND, not detected. expression among the different genotypes at the MG stage. the endogenous SlWD40 expression, the OE effect of B33 Given the low expression level of SlWD40 of WT fruit at promoter was concealed and SlWD40 expression was not MG stage, RNAi fruit did not exhibit significantly different significantly different between the OE and WT fruit at the expression from WT fruit. However, as the ripening process Br and Pink stages (Supplemental Table S2). These results was initiated, the expression of SlWD40 was significantly in- were further confirmed by the principal component analysis duced and its expression was remarkably lower in the RNAi (PCA) and cluster analysis based on the transcriptome data fruit compared with that of WT fruit at Br and Pink stages of different samples. RNAi samples were closely grouped without affecting the expression of other WD40 family genes with WT at MG stage but substantially separated samples at (Supplemental Table S2). In the OE fruits, SlWD40 expres- Br and Pink stages. Conversely, OE samples were clustered sion was 5.64- and 5.71-fold higher than that of WT fruit at with WT sample especially at the Pink stage and subse- the MG stage. That said owing to the massive induction of quently separated from the WT sample at the MG stage 256 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. (Figure 5, A and B). In order to further mine the important SlWRKY6 (Solyc02g080890.3.1), SlWRKY17 (Solyc07g051840. DEGs under the effect of SlWD40, we further analyzed the 4.1), SlWRKY31 (Solyc06g066370.4.1), and SlWRKY79 overlapping DEGs of OE-WT fruit at MG stage and RNAi- (Solyc02g072190.4.1) was increased by 5.7-, 3.4-, 2.6-, 2.6-, WT fruits at Br and Pink stage. The results indicate that 244 2.4-, 2.2-, 2.2-, and 1.8-fold, respectively. Moreover, since fruit genes were stably downregulated in the OE-MG and upregu- size of SlWD40-RNAi fruits was smaller and IAA content di- lated in the RNAi Br and Pink stages, while 60 genes were rectly affects organ size, we found that the expression of stably upregulated in the OE-MG and downregulated in the SlGH3-2 (gene regulating auxin homeostasis) was increased RNAi Br and Pink stages (Supplemental Table S3). To further by three-fold in OE lines while the level of the same gene mine the functional categorization of DEGs, AgriGO v2.0 was decreased in RNAi lines by three- to six-fold. analysis tools (http://bioinfo.cau.edu.cn/agriGO/) by singular Additionally, given that SlWD40 was strongly co-expressed enrichment analysis has been used based on the conserved with RIN and NOR, the conserved DEGs of rin, nor mutants, DEGs (Wang et al., 2017; Supplemental Table S4). Among and SlWD40 transgenic fruits were analyzed and 31 genes were found as conserved DEGs across the genotypes the GO terms included in the “Molecular Function” category of the DEGs upregulated in the OE-MG and downregulated (Figure 5C and Supplemental Table S5; Fujisawa et al., 2012; in the RNAi Br and Pink stages, the pathways that affected Gao et al., 2020). Among the 31 genes, several important the lysis and enzyme activity, such as lyase activity ripening-related genes, such as ACS4 for ethylene biosynthe- (FDR = 0.0014), oxidoreductase activity (FDR = 0.011), and sis, PSY and Z-ISO for carotenoid biosynthesis, PL and PMEI monooxygenase activity (FDR = 0.013), were enriched. for cell wall modification, INV for sugar metabolism, and branched chain amino transferase 2 (BCAT2)and THA1 for Moreover, in the “Biological Process” category of DEGs downregulated in the OE-MG and upregulated in the RNAi amino acid metabolism, were significantly downregulated in Br and Pink stages, several cell-wall-related pathways were the SlWD40-RNAi, rin,and nor fruits and upregulated in the significantly enriched (Supplemental Table S4). Interestingly, SlWD40-OE fruits, which indicates that SlWD40 may partici- in the SlWD40-OE lines, the expression of SlWRKY75 pate in some or all of the primary regulatory functions of (Solyc05g015850.4.1), SlWRKY37 (Solyc02g021680.3.1), SlWR these important regulators, such as RIN and NOR.As the transcriptome data indicated that SlWD40 exhibited KY23 (Solyc01g079260.4.1), SlWRKY30 (Solyc07g056280.3.1), Figure 5 The effect of SlWD40 on fruit transcriptome. A, PCA of transcriptome of different genotypes in the three fruit stages. PC, principal com- ponent. B, Heat map of transcriptome profiles of pericarp tissues of T1 SlWD40 RNAi, OE lines, and WT. X-axis indicates different transgenic lines along with different fruit ripening stages. The hierarchical clustering based on transcriptome indicated that the development of RNAi samples is delayed especially on Br stage which are cluster with MG stage samples. C, Venn diagram showing the overlap DEGs between SlWD40 transgenic fruits, rin and nor mutant fruits. Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 257 substantially changed expression in the MG stage of two OE lines and Br and Pink stage of two RNAi lines compared with that of in the WT, to further confirm the expression difference, we checked the expression of SlWD40 and other 17 ripening-related genes such as RIN, NOR, PSY,and PL in these samples. The results indicate that all of these genes were significantly upregulated in the two OE lines at the MG stage but significantly downregulated in the two RNAi lines at Br and Pink stages, which further confirm the posi- tive function of SlWD40 on tomato ripening with the strongly co-expressed TFs (Figure 6). Metabolome analysis of SlWD40 transgenic fruits Given that metabolites are important parameters for esti- mating progression of the fruit ripening process (Carrari et al., 2006), we analyzed their levels in the different geno- types at a time concurrent to the MG, Br, and Pink stages of WTs. To obtain the global metabolome variation of differ- ent samples, a PCA analysis was carried out based on the primary metabolites, lipids, and secondary metabolites. The results indicate substantial differences among RNAi, OE, and WT fruits, which were particularly prominent at the Pink stage. RNAi pink fruits were closest to Br fruits and sepa- rated clearly from OE and WT fruits (Figure 7, A and B). Given that SlWD40 was strongly co-expressed with RIN and NOR, we analyzed the overlapping differentially enriched pri- mary metabolites between SlWD40-RNAi lines, rin and nor mutants at the Pink stage (Osorio et al., 2011). The results indicated that as the principal free amino acid that most contributes to the “umami” flavor of ripe tomato fruits, glu- tamate exhibited a conserved lower accumulation in the SlWD40-RNAi lines, rin and nor pink fruits (Figure 7C). In detail, given that several amino acids are increased and or- ganic acid decreased during WT tomato ripening (Carrari Figure 6 The different expression of SlWD40 and other 17 ripening-re- et al., 2006), it is of interest that we observed that aspartic lated genes of T1 transformants of RNAi and OE lines compared with acid, glutamic acid, and tryptophan levels were significantly WT fruits. The values in each column are the mean of three biological replicates. Error bars indicate SD. lower in both RNAi lines compared with those exhibited by the OE and WT pink fruits (Figure 8 and Supplemental Table S6). Moreover, nicotinic acid and glyceric acid levels confirmed the positive influence of SlWD40 on the fruit rip- were significantly higher in at least one RNAi line compared ening process (Figure 9 and Supplemental Table S6). with those observed in OE and WT pink fruits (Figure 8 and Supplemental Table S6). Moreover, the former results indi- Discussion cated that the representative lipid components, triacylglycer- We adopted a computational approach to mine key candi- ols (TAGs) are significantly decreased in avocado fruit date genes involved in tomato fruit ripening by integrating ripening (Rodriguez-Lopez et al., 2017). The same trends also comparative transcriptomics and eQTL analysis. In doing so, have been found in the present study that the content of we identified 16 previously uncharacterized candidate genes TAG 48:0, TAG 52:5, TAG 52:6, TAG 54:6, TAG 54:7, and for involvement in ripening. Among them, we chose to fol- TAG 54:8 was remarkably higher in the RNAi fruits than low up on SlWD40 since it was identified in the ChIP seq that of OE and WT pink fruits (Figure 9 and Supplemental screen as a direct target gene of RIN and relatively little is Table S6). As the secondary metabolites (such as naringenin known concerning its regulatory role (Fujisawa et al., 2013). chalcone, naringenin hexose, and chlorogenic acid deriva- Intriguingly, cross-tissue co-expression networks for SlWD40 tives) dramatically accumulate during ripening and represent an important quality of fruit, the significantly higher and strongly suggest that it may act with the other important lower content of them in the OE MG fruits and RNAi pink ripening regulators, such as RIN and NOR in affecting to- fruits compared with that of WT fruits, respectively, further mato fruit ripening, and the detailed analysis of DEGs and 258 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. Figure 7 The effect of SlWD40 on fruit metabolism. A, PCA of metabolite levels of different genotypes in the three fruit stages. PC, principal com- ponent. B, Heat map of metabolite profiles of pericarp tissues of T1 SlWD40 RNAi, OE lines, and WT. X-axis indicates different transgenic lines along with different fruit ripening stages. Y-axis indicates metabolites. The hierarchical clustering based on all metabolites indicated that the devel- opment of RNAi samples is delayed especially in the Pink stage which are cluster with Br stage samples of OE and WT. C, Venn diagram showing the overlap different primary metabolite between SlWD40 RNAi fruits, rin and nor mutant fruits at the Pink stage. differentially abundant metabolites between SlWD40 trans- candidate gene filtration pipeline. Moreover, on the basis of genic fruits further illuminate the important function of the distance between QTL and the target transcript, eQTL SlWD40 on tomato ripening process. can either be classified as cis-eQTL (where the gene encod- ing the transcript resides within the QTL interval) or trans- Integrating comparative transcriptomics with the eQTL with both types being prominent in tomato eQTL approach to mine for ripening regulators (Rockman and Kruglyak, 2009; Zhu et al., 2018). Combining Several tomato genes with strong ripening phenotypes have the gene filtration pipeline and the comparative transcrip- been identified via mutagenesis-based breeding programs in- tomics, we filtered genes responsible for the red ripe pheno- cluding rin, nor, NR,and Cnr (Tigchelaar, 1973; Lanahan type of tomato fruits (i.e. genes highly expressing only in et al., 1994; Vrebalov et al., 2002; Manning et al., 2006). “red” ripe fruits). Since gene duplication and subsequent Moreover, the recombinant inbred line (RILs) population is functional diversification creates novel metabolic pathways also useful to identify key loci regulating quantitative traits and regulation, we focused on TF families that were evolved including fruit ripening as well as aroma, color, and disease by gene duplication and already reported to be regulating resistance (Kimbara et al., 2018). However, since RILs usually tomato fruit ripening (e.g. MADS box and basic helix loop harbor more than one introgression, the possibility of the helix (bHLH) TF families) (Hileman et al., 2006; Waseem introgressed loci having beneficial/inhibitory interactions et al., 2019). Out of the 16 candidate genes, 10 were found with the genes in the genetic background renders gene to be generated either through tandem or block duplication function elucidation more complex in RILs. In contrast, ILs (https://bioinformatics.psb.ugent.be/plaza). In the VIGS ex- are high resolution in that they normally carry only single periment, green and yellow phenotypes have been obtained introgressions and as such epistatic interactions masking sin- for the candidates Solyc11g010710 (AP2 like) and gle gene effects are largely minimized (Ofner et al., 2016). Solyc07g052700 (MADS TF, AGL66). Our previously pub- Here, we used available transcriptome data from red ripe lished work implicated AP2a in regulating tomato fruit rip- fruit of S. lycopersicum (M82) parent and a set of lines with ening via regulation of ethylene biosynthesis and signaling distinct introgressed S. pennellii segments (http://ted.bti.cor (Chung et al., 2010; Karlova et al., 2011). MADS box TFs nell.edu/cgi-bin/TFGD/array_data/home.cgi) and optimized a such as SlCMB1 (Solyc04g005320), TAGL1 (Solyc07g055920), Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 259 Figure 8 The scheme of major metabolic changes of the transgenic lines. The difference of sugar and amino acid-related metabolites between transgenic lines and WT fruit at MG (A), Br (B), and Pink (C) stage. Blue and red color depicts a decrease and increase in metabolic levels com- pared with the WT fruit samples, respectively. Figure 9 The difference of representative lipids and secondary metabolites of T1 transformants of RNAi and OE lines compared with WT fruits. The values in each column are the mean of at least three biological replicates. Error bars indicate SD. The asterisks indicate statistically significant differences determined by the Student’s t test (two-tail): *P5 0.05; **P5 0.01. 260 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. and the canonical TF RIN (Solyc05g012020) have previously enhances the regulation of the complex during anthocyanin been reported as positive regulators of tomato fruit ripening biosynthesis (Ramsay and Glover, 2005; Zhang et al., 2014). (Zhang et al., 2018). Therefore, VIGS for AP2 like and MADS In the present study, the green phenotype of fruits of box candidates (Solyc11g010710 and Solyc07g052700) acted SlWD40-RNAi lines clearly demonstrated a positive role of as positive control supporting the efficacy of our approach. SlWD40 in tomato fruit ripening (Figure 4). Moreover, given Moreover, among the 16 genes, VIGS for two bHLH TFs that the global gene co-expression analysis is a powerful ap- (Solyc03g044460 and Solyc12g098620) showed green and proach to identify the important interactions among differ- light red phenotypes (Supplemental Figure S1). Here, light ent genes during the development of certain organs, it is red phenotype for Solyc12g098620 is in line with the re- important to note that the co-expression network of cently published work by D’Amelia et al. (2019) in which SlWD40 revealed links with RIN and NOR, while transcrip- the authors reported that this bHLH TF regulates carotenoid tome analysis indicated they shared conserved regulated biosynthesis. Additionally, one of the candidates, genes associated with ethylene (ACSs) and carotenoid (PSY Solyc12g010950 (alcohol dehydrogenase), obtained a whitish and Z-ISO) biosynthesis as well as cell wall degradation (PL green fruit phenotype, and another candidate, SlWD40, and PMEI) and sugar (INV) and amino acid (BCAT2 and THA1)metabolism (Supplemental Table S5). Interestingly, obtained a yellowish fruit phenotype (Figure 2 and Supplemental Figure S1). While the link between the alcohol based on available ChIP and transcriptome data, the RIN dehydrogenase and ripening is currently unclear, that for protein can directly bind to the promoter of SlWD40 and SlWD40 is not without precedence since it is a transcrip- thereby increase its expression (Martel et al., 2011; Fujisawa tional regulator which has been linked to rin and nor in et al., 2013). Moreover, as an important TF family member involved in ChIP experiments but not characterized in detail in its own ethylene signal, AP2a belonging to the AP2/ERF superfamily right. The eQTL for SlWD40 can bedefined as acis-eQTLsince has been shown to inhibit ethylene biosynthesis as well as SlWD40 is located in the region which has been introgressed positively regulate chlorophyll degradation and carotenoid by S. pennellii chromosomal architecture in the IL4-1 and biosynthesis in tomato (Wang et al., 2019). And in Setaria italica, SiAP2 can bind to the SiWD40 promoter to mediate IL4-1-1 and SlWD40’s expression is only dramatically lowered for IL4-1 and 4-1-1 (Supplemental Table S7). At the onset of abiotic stress responses (Mishra et al., 2012). In the present fruit ripening process (MG stage), SlWD40 is moderately study, the promoter region of SlWD40 was found to harbor expressed in S. lycopersicum (RPKM mean value 31) but neg- eight different AP2 TF binding sites and AP2a was signifi- ligibly expressed in S. pennellii (RPKM mean value 0.17) and cantly induced in SlWD40-OE-MG fruits while repressed in the WD40-RNAi-Breaker fruits (Supplemental Table S2). consistent with the ripening stage, the expression of SlWD40 was substantially induced in concert with fruit ripening. Additionally, it has been well-documented that several Considering the dramatically lowered SlWD40 expression in SlWRKYs regulate tomato fruit ripening and lycopene accu- IL4-1 and 4-1-1 and substantial difference of the red/green mulation (Cheng et al., 2016; Wang et al., 2017). In our colored fruits of S. lycopersicum and S. pennellii,respectively, analysis, SlWD40 OE resulted in upregulation (by two- to five-fold) of a broad number of SlWRKYs (SlWRKY75, we hypothesized that SlWD40 may participate in the S. lyco- persicum ripening process. Furthermore, VIGS for SlWD40 in SlWRKY37, SlWRKY23, SlWRKY30, SlWRKY6, SlWRKY17, MicroTom inhibited normal red coloration. These results SlWRKY31,and SlWRKY79) in mature green fruits of OE demonstrate the utility of our candidate gene filtration pipe- lines. Moreover, SlWRKY17, which was shown to interact line integrating comparative transcriptomics with eQTL in with RIN, SlERF2b, and SlERF7, is strongly co-expressed with ELIP2 and RIN in our analysis (Wang et al., 2017; identifying candidate ripening genes. Supplemental Table S1). All of these results indicate that SlWD40 is an important regulator of tomato fruit SlWD40 might act as a junction point and facilitate binding ripening of one or more above mentioned co-expressed TFs such as Among the eight genes which we validated by VIGS, the in- RIN, NOR, AP2a,and SlWRKYs to be involved in tomato rip- formation concerning the role of the WD40 family in regu- ening process. lating tomato ripening is the most limited. Ripening Besides the important ripening-related TFs, hormone sig- function of this gene was further confirmed by the stable naling also plays a vital role in the ripening process. As an OE and RNAi transformation (Figures 4 and 5). WD40 pro- important regulator of auxin-ethylene homoeostasis which teins contain a signature WD (Trp-Asp) dipeptide and 40 affects fruit ripening (Kumar et al., 2012; Sravankumar et al., amino acids in single repeats that then fold into four- 2018), the expression of SlGH3-2 (Solyc01g107390) increased stranded anti-parallel b-propeller sheets and are highly pro- in SlWD40-OE lines by three-fold while decreased in RNAi miscuous interactors, being both platforms for protein– lines by three- to six-fold (Figure 6 and Supplemental Table DNA and protein–protein interactions (Xu and Min, 2011; S2). Combined with the induction and repression effect of Mishra et al., 2012; Chen et al., 2022). In the canonical ACS4 in the SlWD40-OE and -RNAi fruit, respectively, these MYB–bHLH–WD40 protein complex, WD40 acts as a re- results collectively indicate that SlWD40 may also affect the cruiter and stabilizer of the MYB and bHLH protein which tomato ripening process through the regulation of the Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 261 ripening-related hormone homoeostasis. However, consider- (2) the presence of an eQTL for the gene, and (3) the pres- ably further experimentation will be required in order to ence of a functional annotation. Subsequently, these candi- test this hypothesis. dates were sorted into specific and nonspecific eQTL based on their expression in respective IL. Here, specific and non- Conclusion specific eQTL were defined based on the expression of a Our study aimed at better understanding the molecular particular S. pennellii candidate in a specific IL or several ILs, mechanisms underlying tomato fruit ripening. Comparative respectively. Specific eQTL candidates were then focused transcriptomics of small green fruited wild species with red and classified based on their function. For promoter analysis, fruited S. lycopersicum alongside eQTL mapping allowed the the 1-kb upstream sequence from the start codon of identification of key candidate genes involved in tomato SlWD40 was retrieved from the SGN tomato genome web fruit ripening. Utilizing co-expression networks alongside de- browser and the cis-regulatory elements were analyzed using tailed metabolome and transcriptome analysis indicated PlantCARE (http://bioinformatics.psb.ugentbe/webtools/ that SlWD40 has a positive impact on tomato ripening pro- plantcare/html/) and PlantPAN 2.0 (http://plantpan2.itps. cess and suggest that it may act in concert with strongly ncku.edu.tw/index.html) web tools. co-expressed TFs such as RIN, NOR, AP2a,and SlWRKYs (Figure 10). Beyond these insights into ripening, we believe SlWD40 co-expression network construction with our study also acts as a proof-of-concept study whereby the tomato ripening pathway genes transcriptome of phenotypically divergent wild relatives, We listed 171 target genes involved in carotenoid biosynthe- alongside eQTL mapping, can be used to identify causal sis, tomato fruit ripening, and cell wall metabolic pathways genes underlying trait variance. and their regulation. Some of these genes were well charac- terized. For all 171 genes, expression values were extracted Materials and methods from Tomato Genome Consortium (2012). The R script Narrowing down candidate genes involved in fruit written by Contreras-Lopez et al. (2018) was used to calcu- late correlation values and P values, both positive and nega- ripening tive correlation values were calculated and cytoscape was For M82 and S. pennellii (Penn) fruit, RNA-seq data are used to visualize network (Shannon et al., 2003). available (Bolger et al., 2014). Moreover, RNA-seq data for S. pennellii ILs fruit were also available (http://ted.bti.cornell. Virus-induced gene silencing edu/cgi-bin/TFGD/array_data/home.cgi). Starting with the Vector construction, infiltration, and fruit harvesting proce- RNA-seq data for M82 and S. pennellii fruit (Bolger et al., dures were performed as previously described (Orzaez et al., 2014) all 34,727 genes in the transcriptome were sorted in 2006, 2009). Briefly, an approximately 300-bp fragment of two ways. Firstly, the ratio of their expression value in M82 the candidate gene was amplified from tomato M82 fruit relative to that in S. pennellii and secondly, by ratio of their expression value of S. pennellii to that of M82 in order to cDNA using gateway compatible primers and recombined get genes that are highly expressed in M82 and S. pennellii, into the GATEWAY vector pDONR207 (Invitrogen, http:// www.invitrogen.com/) by the BP reaction following the respectively. Candidates from these two lists were further fil- manufacturer’s protocol to generate an entry clone. An er- tered by using three different criteria namely (1) that the relative fold change (FC) in the expression was at least 45, ror-free entry vector was confirmed by sequencing and then Figure 10 Proposed schematic overview of network of regulatory factors controlling tomato fruit ripening. 262 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. recombined with the pTRV2-Ros/Del/GW destination vector available in Zenodo (https://zenodo.org/) (doi:10.5281/zen- using an LR reaction to produce the expression clones odo.5525948 and 10.5281/zenodo.5525946). The RNA seq pTRV2-Ros/Del/GW-Respective Gene ID. Agrobacterium data were analyzed using LSTrAP (Proost et al., 2017). The tumefaciens strain GV3101:pMP90 was then transformed clean reads of each sample were aligned to the Tomato with sequenced expression vectors by electroporation. In or- Genome version SL4.0 and Annotation ITAG4.0 (ftp://ftp.sol der to infiltrate fruit for VIGS, purple MicroTom tomato genomics.net/tomato_genome/annotation/ITAG4.0_release/). was used and agroinfiltration was performed as previously The DEGs between transgenic fruit and WT fruit were identi- described (Alseekh et al., 2015). fied under the parameter of FC 5 2and FDR 5 0.05. Development of OE and RNAi lines Metabolic profiling The sequence encoding Solyc04g005020 was amplified from Fruit pericarp samples were harvested, immediately frozen in S. lycopersicum cv. Moneymaker (MM) cDNA by using gene- liquid nitrogen, and stored at –80 C until further analysis. -1 specific primers and inserted into the pDONR207 by attB re- Samples were then powdered by using retsch mill at 30 Ls , combination to generate entry clone. Primer sequences are for 30 s. Extraction of pigments, primary metabolites, lipid, provided in Supplemental Table S8.An error-free entry and secondary metabolites was performed as described previ- clone was confirmed by sequence analysis before recombina- ously (Salem et al., 2016). In brief, 500mL of the upper lipid tion into destination vector B33BinAR for fruit-specific OE and pigments containing phase was dried in a SpeedVac and named as B33BinAR_SlWD40. Additionally, artificial concentrator and resuspended in 250mL acetonitrile: 2-prop- miRNA (amiRNA) cassette was designed for Solyc04g005020. anol (7:3, v/v) solution. Two microliters of the solution were For this, Solyc04g005020 cDNA sequence was used as target analyzed by the Waters Acquity ultra-performance LC system sequence, employing the WMD3 program (http://wmd3.wei coupled with Fourier transform MS in positive ionization gelworld.org/cgi-bin/webapp.cgi) to design corresponding mode. Moreover, 150 and 300 lL of the polar phase were amiRs. An overlapping PCR (polymerase chain reaction) dried in a centrifugal vacuum concentrator for primary and strategy was employed with in-hand precursor DNA, follow- secondary metabolite profiling. The primary metabolite ing the WMD3 protocol (http://wmd3.weigelworld.or/down pellet was resuspended in 40 lL of methoxyaminhydrochlor- loads/CloningofartificialmicroRNAs.pdf). The pre-amiRs -1 ide (20 mgmL in pyridine) and derivatized for 2 h at obtained from overlapping PCR (using the athmir-319a 37 C. Afterward, 70 lLof N-methyl-N-[trimethylsilyl] trifluor- backbone) were cloned into the pENTR/D-TOPO vector and -1 oacetamide was added containing 20 lLmL fatty acid the clones were confirmed by DNA sequencing. methyl esters mixture as retention time standards. The mix- Subsequently, these sequences were cloned into B33BinAR ture was incubated for 30 min at 37 C at 400 rpm. A vol- via Asp718 and BamHI digestion and cohesive end ligation. ume of 1 lL of this solution was used for injection. The gas Primer sequences are provided in Supplemental Table S8. chromatography–mass spectroscopy system comprised a This and other final LR plasmids were then introduced into CTC CombiPAL autosampler, an Agilent 6890N gas chro- A. tumefaciens strain GV2260 by electroporation and subse- matograph, and a LECO Pegasus III time of flight mass spec- quently submitted for transformation into MM plants using trometry (TOF-MS) running in EI + mode. The secondary the leaf disc transformation method (McCormick et al., metabolite pellet was resuspended in 200-mL50% (v/v) 1986). methanol in water and 2 mL was injected on RP high strength silica T3 C column using a Waters Acquity UPLC Plant material and growth conditions system. The analysis workflow included peak detection, re- Transgenic plants for each genotype were selected on kana- -1 tention time alignment, and removal of chemical noise fol- mycin containing MS medium (50 mgL ). SNN and MM lowing the method of Salem et al. (2016). For metabolites (WT) were germinated on MS medium without kanamycin. and transcriptome data processing, the PCA and heat map Both transgenic lines and WT were selected and transferred analysis were performed by MetaboAnalyst 5.0 (https://www. to soil pot for cultivation under long-day conditions (16-h/ metaboanalyst.ca/). 8-h day/night cycle) at 22 C and 50% humidity, as described previously in the literature (Carrari et al., 2003). Upon anthe- RT-qPCR analysis sis, flowers were labeled with that particular date. Total RNA was extracted from fruit using TRIzol reagent (Invitrogen, Waltham, MA, USA). And the first-strand cDNA Transcriptome analysis synthesis was carried out as the manufacturer’s instructions Two biological replicate samples from two independent of PrimeScript RT Reagent Kit with gDNA Eraser (Takara, plants of each genotype of MG, Br, and Pink stages have Shiga, Japan). RT-qPCR was analyzed on an ABI Prism 7900 been harvested. Total RNA was extracted using the HT real-time PCR system (Applied Biosystems/Life NucleoSpin RNA Plant kit (Macherey-Nagel) and sent to the Novogene Company (Beijing, China) for Illumina HiSeq Technologies, Darmstadt, Germany) in 384-well PCR plates. –DDCt PE150 sequencing. The cDNA library was constructed follow- The RT-qPCR data were analyzed using the 2 analysis ing the manufacturer’s recommendations and then purified method according to Bustin et al. (2009) and all primers are to remove the low-quality sequences. The clean data are listed in Supplemental Table S8. Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 263 Statistical analysis FPA No. 664620). S.S.J. acknowledges funding by ICAR, India, Student’s paired t test was performed to assess whether the in the form of ICAR-International Fellowship. differences between different genotypes were statistically sig- Conflict of interest statement. None declared. nificant. The asterisks indicate statistically significant differ- ences determined by the Student’s t test (two-tail): *P5 0.05; **P5 0.01. References Accession numbers Alseekh S, Ofner I, Pleban T, Tripodi P, Di Dato F, Cammareri M, Mohammad A, Grandillo S, Fernie AR, Zamir D (2013) Sequence data from this article can be found in the Resolution by recombination: breaking up Solanum pennellii intro- GenBank/EMBL data libraries under accession numbers gressions. Trends Plant Sci 18: 536–538 SlWD40, Solyc04g005020. Alseekh S, Tohge T, Wendenberg R, Scossa F, Omranian N, Li J, Kleessen S, Giavalisco P, Pleban T, Mueller-Roeber B, et al. Supplemental data (2015) Identification and mode of inheritance of quantitative trait loci for secondary metabolite abundance in tomato. Plant Cell 27: The following materials are available in the online version of 485–512 this article. Ballester AR, Molthoff J, de Vos R, Hekkert B, Orzaez D, Supplemental Figure S1. VIGS phenotype of candidate Fernandez-Moreno JP, Tripodi P, Grandillo S, Martin C, Heldens J, et al. (2010) Biochemical and molecular analysis of pink genes. tomatoes: deregulated expression of the gene encoding transcrip- Supplemental Figure S2. Promoter analysis of SlWD40 for tion factor SlMYB12 leads to pink tomato fruit color. Plant Physiol thepresenceof ethylene (C2H2, AP2, EIN), auxin, and 152: 71–84 MADS-box binding-related cis-regulatory elements. Baranwal VK, Negi N, Khurana P (2021) Comparative transcriptom- Supplemental Figure S3. Genotyping of the SlWD40 of ics of leaves of five mulberry accessions and cataloguing structural and expression variants for future prospects. PLoS ONE 16: T0 transformants. e0252246 Supplemental Table S1. Co-expression network of Bartley GE, Scolnik PA (1993) cDNA cloning, expression during de- SlWD40 with tomato ripening pathway specific genes. velopment, and genome mapping of PSY2, a second tomato gene Supplemental Table S2. Transcriptome profiling of encoding phytoene synthase. J Biol Chem 268: 25718–25721 SlWD40 transgenic fruits. Batyrshina ZS, Yaakov B, Shavit R, Singh A, Tzin V (2020) Comparative transcriptomic and metabolic analysis of wild and Supplemental Table S3. The overlapped DEGs of SlWD40 domesticated wheat genotypes reveals differences in chemical and OE and RNAi fruit. physical defense responses against aphids. BMC Plant Biol 20:19 Supplemental Table S4. Functional categorization of Baxter CJ, Carrari F, Bauke A, Overy S, Hill SA, Quick PW, Fernie DEGs of SlWD40. AR, Sweetlove LJ (2005) Fruit carbohydrate metabolism in an in- Supplemental Table S5. The overlap DEGs of SlWD40, trogression line of tomato with increased fruit soluble solids. Plant Cell Physiol 46: 425–437 rin,and nor mutants. Bird CR, Ray JA, Fletcher JD, Boniwell JM, Bird AS, Teulieres C, Supplemental Table S6. Metabolite profiling of SlWD40 Blain I, Bramley PM, Schuch W (1991) Using antisense RNA to transgenic fruits. study gene-function—inhibition of carotenoid biosynthesis in Supplemental Table S7. Expression of SlWD40 in ILs. transgenic tomatoes. Bio-Technology 9: 635–639 Supplemental Table S8. Primer sequences used in this Bolger A, Scossa F, Bolger ME, Lanz C, Maumus F, Tohge T, Quesneville H, Alseekh S, Sorensen I, Lichtenstein G, et al. study. (2014) The genome of the stress-tolerant wild tomato species Supplemental Data Set S1. Finalized potential candidates Solanum pennellii. Nat Genet 46: 1034–1038 from both eQTL and TF approaches and their VIGS Breschi A, Gingeras TR, Guigo R (2017) Comparative transcriptom- phenotypes. ics in human and mouse. Nat Rev Genet 18: 425–440 Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista Acknowledgments M, Mueller R, Nolan T, Pfaffl MW, Shipley GL, et al. (2009) The MIQE guidelines: minimum information for publication of quanti- We thank Dr Youjun Zhang and Regina Wendenburg from tative real-time PCR experiments. Clin Chem 55: 611–622 Max-Planck-Institut fu ¨r Molekulare Pflanzenphysiologie for Carrari F, Baxter C, Usadel B, Urbanczyk-Wochniak E, Zanor MI, useful discussion and experiment assistance. Nunes-Nesi A, Nikiforova V, Centero D, Ratzka A, Pauly M, et al (2006) Integrated analysis of metabolite and transcript levels Funding reveals the metabolic shifts that underlie tomato fruit develop- ment and highlight regulatory aspects of metabolic network be- F.Z. acknowledges funding of The Key R&D Program havior. Plant Physiol 142: 1380–1396 of Hubei Province (2021BBA095) and the National Carrari F, Fernie AR (2006) Metabolic regulation underlying tomato Natural Science Foundation of China (32002102) and fruit development. J Exp Bot 57: 1883–1897 the work in Fernie Lab was supported by the Deutsche Carrari F, Nunes-Nesi A, Gibon Y, Lytovchenko A, Loureiro ME, Fernie AR (2003) Reduced expression of aconitase results in an en- Forschungsgemeinschaft in the framework of Deutsche hanced rate of photosynthesis and marked shifts in carbon parti- Israeli Project FE 552/12-1. In addition, S.A. and A.R.F. ac- tioning in illuminated leaves of wild species tomato. Plant Physiol knowledge funding of the PlantaSYST project by the 133: 1322–1335 European Union’s Horizon 2020 Research and Innovation Cazzonelli CI, Pogson BJ (2010) Source to sink: regulation of carot- Programme (SGA-CSA No. 664621 and No. 739582 under enoid biosynthesis in plants. Trends Plant Sci 15: 266–274 264 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. Centeno DC, Osorio S, Nunes-Nesi A, Bertolo ALF, Carneiro RT, Giovannoni J, Nguyen C, Ampofo B, Zhong S, Fei Z (2017) The epi- Arau´jo WL, Steinhauser M-C, Michalska J, Rohrmann J, genome and transcriptional dynamics of fruit ripening. Annu Rev Geigenberger P, et al. (2011) Malate plays a crucial role in starch Plant Biol 68: 61–84 metabolism, ripening, and soluble solid content of tomato fruit Hileman LC, Sundstrom JF, Litt A, Chen M, Shumba T, Irish VF (2006) Molecular and phylogenetic analyses of the MADS-box and affects postharvest softening. Plant Cell 23: 162–184 gene family in tomato. Mol Biol Evol 23: 2245–2258 Chang YM, Lin HH, Liu WY, Yu CP, Chen HJ, Wartini PP, Kao YY, Irfan M, Ghosh S, Meli VS, Kumar A, Kumar V, Chakraborty N, Wu YH, Lin JJ, Lu MJ, et al. (2019) Comparative transcriptomics Chakraborty S, Datta A (2016) Fruit ripening regulation of method to infer gene coexpression networks and its applications alpha-mannosidase expression by the MADS box transcription fac- to maize and rice leaf transcriptomes. Proc Natl Acad Sci USA tor RIPENING INHIBITOR and ethylene. Front Plant Sci 7:10 116: 3091–3099 Ito Y, Nishizawa-Yokoi A, Endo M, Mikami M, Shima Y, Chen L, Li W, Li Y, Feng X, Du K, Wang G, Zhao L (2019) Nakamura N, Kotake-Nara E, Kawasaki S, Toki S (2017) Identified trans-splicing of YELLOW-FRUITED TOMATO 2 encoding Re-evaluation of the rin mutation and the role of RIN in the in- the PHYTOENE SYNTHASE 1 protein alters fruit color by duction of tomato ripening. Nat Plants 3: 866–874 map-based cloning, functional complementation and RACE. Plant Karlova R, Rosin FM, Busscher-Lange J, Parapunova V, Do PT, Mol Biol 100: 647–658 Fernie AR, Fraser PD, Baxter C, Angenent GC, de Maagd RA Chen W, Chen L, Zhang X, Yang N, Guo J, Wang M, Ji S, Zhao X, (2011) Transcriptome and metabolite profiling show that Yin P, Cai L, et al. (2022) Convergent selection of a WD40 protein APETALA2a is a major regulator of tomato fruit ripening. Plant that enhances grain yield in maize and rice. Science 375: eabg7985 Cell 23: 923–941 Cheng Y, Ahammed GJ, Yu J, Yao Z, Ruan M, Ye Q, Li Z, Wang R, Kimbara J, Ohyama A, Chikano H, Ito H, Hosoi K, Negoro S, Feng K, Zhou G, et al. (2016) Putative WRKYs associated with Miyatake K, Yamaguchi H, Nunome T, Fukuoka H, et al. (2018) regulation of fruit ripening revealed by detailed expression analysis QTL mapping of fruit nutritional and flavor components in to- of the WRKY gene family in pepper. Sci Rep 6: 39000 mato (Solanum lycopersicum) using genome-wide SSR markers and Chitwood DH, Kumar R, Headland LR, Ranjan A, Covington MF, recombinant inbred lines (RILs) from an intra-specific cross. Ichihashi Y, Fulop D, Jimenez-Go´mez JM, Peng J, Maloof JN, Euphytica 214: 210 et al (2013) A quantitative genetic basis for leaf morphology in a Kumar R, Agarwal P, Tyagi AK, Sharma AK (2012) Genome-wide set of precisely defined tomato introgression lines. Plant Cell 25: investigation and expression analysis suggest diverse roles of 2465–2481 auxin-responsive GH3 genes during development and response to Chung MY, Vrebalov J, Alba R, Lee J, McQuinn R, Chung JD, Klein different stimuli in tomato (Solanum lycopersicum). Mol Genet P, Giovannoni J (2010) A tomato (Solanum lycopersicum) Genomics 287: 221–235 APETALA2/ERF gene, SlAP2a, is a negative regulator of fruit ripen- Lanahan MB, Yen HC, Giovannoni JJ, Klee HJ (1994) The never ing. Plant J 64: 936–947 ripe mutation blocks ethylene perception in tomato. Plant Cell 6: Contreras-Lopez O, Moyano TC, Soto DC, Gutierrez RA (2018) 521–530 Step-by-step construction of gene co-expression networks from Li S, Chen K, Grierson D (2021) Molecular and hormonal mecha- high-throughput Arabidopsis RNA sequencing data. Methods Mol nisms regulating fleshy fruit ripening. Cells 10: 1136 Biol 1761: 275–301 Li Y, Chen Y, Zhou L, You S, Deng H, Chen Y, Alseekh S, Yuan Y, D’Amelia V, Raiola A, Carputo D, Filippone E, Barone A, Rigano Fu R, Zhang Z, et al. (2020) MicroTom metabolic network: rewir- MM (2019) A basic helix-loop-helix (SlARANCIO), identified from ing tomato metabolic regulatory network throughout the growth a Solanum pennellii introgression line, affects carotenoid accumula- cycle. Mol Plant 13: 1203–1218 tion in tomato fruits. Sci Rep 9: 3699 Liu M, Gomes BL, Mila I, Purgatto E, Peres LE, Frasse P, Maza E, Eshed Y, Zamir D (1995) An introgression line population of Zouine M, Roustan JP, Bouzayen M, et al. (2016) Comprehensive Lycopersicon pennellii in the cultivated tomato enables the identifi- profiling of ethylene response factor expression identifies cation and fine mapping of yield-associated QTL. Genetics 141: ripening-associated ERF genes and their link to key regulators of 1147–1162 fruit ripening in tomato. Plant Physiol 170: 1732–1744 Fernandez-Moreno JP, Tzfadia O, Forment J, Presa S, Rogachev I, Lu P, Yu S, Zhu N, Chen YR, Zhou B, Pan Y, Tzeng D, Fabi JP, Meir S, Orzaez D, Aharoni A, Granell A (2016) Characterization Argyris J, Garcia-Mas J, et al. (2018) Genome encode analyses re- of a new pink-fruited tomato mutant results in the identification veal the basis of convergent evolution of fleshy fruit ripening. Nat of a null allele of the SlMYB12 transcription factor. Plant Physiol Plants 4: 784–791 171: 1821–1836 Manning K, Tor M, Poole M, Hong Y, Thompson AJ, King GJ, Fridman E, Carrari F, Liu YS, Fernie AR, Zamir D (2004) Zooming Giovannoni JJ, Seymour GB (2006) A naturally occurring epige- in on a quantitative trait for tomato yield using interspecific intro- netic mutation in a gene encoding an SBP-box transcription factor gressions. Science 305: 1786–1789 inhibits tomato fruit ripening. Nat Genet 38: 948–952 Fujisawa M, Nakano T, Shima Y, Ito Y (2013) A large-scale identifi- Martel C, Vrebalov J, Tafelmeyer P, Giovannoni JJ (2011) The to- cation of direct targets of the tomato MADS box transcription fac- mato MADS-box transcription factor RIPENING INHIBITOR inter- tor RIPENING INHIBITOR reveals the regulation of fruit ripening. acts with promoters involved in numerous ripening processes in a Plant Cell 25: 371–386 COLORLESS NONRIPENING-dependent manner. Plant Physiol 157: Fujisawa M, Shima Y, Higuchi N, Nakano T, Koyama Y, Kasumi T, 1568–1579 Ito Y (2012) Direct targets of the tomato-ripening regulator RIN McCormick S, Niedermeyer J, Fry J, Barnason A, Horsch R, Fraley identified by transcriptome and chromatin immunoprecipitation R (1986) Leaf disc transformation of cultivated tomato (L. esculen- analyses. Planta 235: 1107–1122 tum) using Agrobacterium tumefaciens. Plant Cell Rep 5: 81–84 Gao L, Gonda I, Sun H, Ma Q, Bao K, Tieman DM, Burzynski- Mishra AK, Puranik S, Bahadur RP, Prasad M (2012) The Chang EA, Fish TL, Stromberg KA, Sacks GL, et al. (2019) The DNA-binding activity of an AP2 protein is involved in transcrip- tomato pan-genome uncovers new genes and a rare allele regulat- tional regulation of a stress-responsive gene, SiWD40, in foxtail mil- ing fruit flavor. Nat Genet 51: 1044–1051 let. Genomics 100: 252–263 Gao Y, Wei W, Fan Z, Zhao X, Zhang Y, Jing Y, Zhu B, Zhu H, Mutwil M, Klie S, Tohge T, Giorgi FM, Wilkins O, Campbell MM, Shan W, Chen JJ (2020) Re-evaluation of the nor mutation and Fernie AR, Usadel B, Nikoloski Z, Persson S (2011) PlaNet: com- the role of the NAC-NOR transcription factor in tomato fruit rip- bined sequence and expression comparisons across plant networks ening. J Exp Bot 71: 3560–3574 derived from seven species. Plant Cell 23: 895–910 Integrated analysis of a tomato ripening regulator PLANT PHYSIOLOGY 2022: 190; 250–266 | 265 Ofner I, Lashbrooke J, Pleban T, Aharoni A, Zamir D (2016) Shi Y, Vrebalov J, Zheng H, Xu Y, Yin X, Liu W, Liu Z, Sorensen I, Solanum pennellii backcross inbred lines (BILs) link small genomic Su G, Ma Q, et al. (2021) A tomato LATERAL ORGAN bins with tomato traits. Plant J 87: 151–160 BOUNDARIES transcription factor, SlLOB1, predominantly regu- Orzaez D, Medina A, Torre S, Fernandez-Moreno JP, Rambla JL, lates cell wall and softening components of ripening. Proc Natl Fernandez-Del-Carmen A, Butelli E, Martin C, Granell A (2009) Acad Sci USA 118: e2102486118 A visual reporter system for virus-induced gene silencing in tomato Shinozaki Y, Nicolas P, Fernandez-Pozo N, Ma Q, Evanich DJ, Shi fruit based on anthocyanin accumulation. Plant Physiol 150: Y, Xu Y, Zheng Y, Snyder SI, Martin LBB, et al. (2018) 1122–1134 High-resolution spatiotemporal transcriptome mapping of tomato Orzaez D, Mirabel S, Wieland WH, Granell A (2006) Agroinjection fruit development and ripening. Nat Commun 9: 364 of tomato fruits. A tool for rapid functional analysis of transgenes Sønderby IE, Hansen BG, Bjarnholt N, Ticconi C, Halkier BA, directly in fruit. Plant Physiol 140: 3–11 Kliebenstein DJ (2007) A systems biology approach identifies a Osorio S, Alba R, Damasceno CMB, Lopez-Casado G, Lohse M, R2R3 MYB gene subfamily with distinct and overlapping functions Zanor MI, Tohge T, Usadel B, Rose JKC, Fei Z, et al. (2011) in regulation of aliphatic glucosinolates. PLoS ONE 2: e1322 Systems biology of tomato fruit development: combined transcript, Sravankumar T, Patel A, Naik N, Kumar R (2018) A protein, and metabolite analysis of tomato transcription factor ripening-induced SlGH3-2 gene regulates fruit ripening via adjust- (nor, rin) and ethylene receptor (Nr) mutants reveals novel regula- ing auxin-ethylene levels in tomato (Solanum lycopersicum L.). tory interactions. Plant Physiol 157: 405–425 Plant Mol Biol 98: 455–469 Proost S, Krawczyk A, Mutwil M (2017) LSTrAP: efficiently combin- Steinhauser M-C, Steinhauser D, Koehl K, Carrari F, Gibon Y, ing RNA sequencing data into co-expression networks. BMC Fernie AR, Stitt M (2010) Enzyme activity profiles during fruit de- Bioinformatics 18: 444 velopment in tomato cultivars and Solanum pennellii. Plant Physiol Ramsay NA, Glover BJ (2005) MYB–bHLH–WD40 protein complex 153: 80–98 and the evolution of cellular diversity. Trends Plant Sci 10: 63–70 Szymanski J, Bocobza S, Panda S, Sonawane P, Cardenas PD, Ranjan A, Budke JM, Rowland SD, Chitwood DH, Kumar R, Lashbrooke J, Kamble A, Shahaf N, Meir S, Bovy A, et al. (2020) Carriedo L, Ichihashi Y, Zumstein K, Maloof JN, Sinha NR Analysis of wild tomato introgression lines elucidates the genetic (2016) eQTL regulating transcript levels associated with diverse bi- basis of transcriptome and metabolome variation underlying fruit ological processes in tomato. Plant Physiol 172: 328–340 traits and pathogen response. Nat Genet 52: 1111–1121 Robinson R (1968) Ripening inhibitor: a gene with multiple effects Tieman D, Zhu G, Resende MF Jr, Lin T, Nguyen C, Bies D, Rambla on ripening. Rep Tomato Genet Coop 18: 36–37 JL, Beltran KS, Taylor M, Zhang B, et al. (2017) A chemical genetic Rocha-Sosa M, Sonnewald U, Frommer W, Stratmann M, Schell J, roadmap to improved tomato flavor. Science 355: 391–394 Willmitzer L (1989) Both developmental and metabolic signals ac- Tieman DM, Zeigler M, Schmelz EA, Taylor MG, Bliss P, Kirst M, tivate the promoter of a class I patatin gene. EMBO J 8: 23–29 Klee HJ (2006) Identification of loci affecting flavour volatile emis- Rockman MV, Kruglyak L (2009) Recombinational landscape and sions in tomato fruits. J Exp Bot 57: 887–896 population genomics of Caenorhabditis elegans. PLoS Genet 5: Tigchelaar E (1973) A new ripening mutant, non-ripening (nor). Rep e1000419 Tomato Genet Coop 35:20 Rodriguez-Lopez CE, Hernandez-Brenes C, Trevino V, Diaz de la Tomato Genome Consortium (2012) The tomato genome sequence Garza RI (2017) Avocado fruit maturation and ripening: dynamics provides insights into fleshy fruit evolution. Nature 485: 635–641 of aliphatic acetogenins and lipidomic profiles from mesocarp, idi- Vallarino JG, Kubiszewski-Jakubiak S, Ruf S, Rossner M, Timm S, oblasts and seed. BMC Plant Biol 17: 159 Bauwe H, Carrari F, Rentsch D, Bock R, Sweetlove LJ, et al. Rohrmann J, McQuinn R, Giovannoni JJ, Fernie AR, Tohge T (2020) Multi-gene metabolic engineering of tomato plants results (2012) Tissue specificity and differential expression of transcription in increased fruit yield up to 23%. Sci Rep 10: 17219 factors in tomato provide hints of unique regulatory networks dur- Vrebalov J, Ruezinsky D, Padmanabhan V, White R, Medrano D, ing fruit ripening. Plant Signal Behav 7: 1639–1647 Drake R, Schuch W, Giovannoni J (2002) A MADS-box gene nec- Rohrmann J, Tohge T, Alba R, Osorio S, Caldana C, McQuinn R, essary for fruit ripening at the tomato ripening-inhibitor (rin) locus. Arvidsson S, van der Merwe MJ, Riano-Pachon DM, Mueller- Science 296: 343–346 Roeber B, et al. (2011) Combined transcription factor profiling, Wang L, Zhang XL, Wang L, Tian Y, Jia N, Chen S, Shi NB, Huang microarray analysis and metabolite profiling reveals the transcrip- X, Zhou C, Yu Y, et al. (2017) Regulation of ethylene-responsive tional control of metabolic shifts occurring during tomato fruit de- SlWRKYs involved in color change during tomato fruit ripening. Sci velopment. Plant J 68: 999–1013 Rep 7: 16674 Salem MA, Juppner J, Bajdzienko K, Giavalisco P (2016) Protocol: Wang R, Tavano E, Lammers M, Martinelli AP, Angenent GC, de a fast, comprehensive and reproducible one-step extraction Maagd RA (2019) Re-evaluation of transcription factor function in to- method for the rapid preparation of polar and semi-polar metabo- mato fruit development and ripening with CRISPR/Cas9-mutagenesis. lites, lipids, proteins, starch and cell wall polymers from a single Sci Rep 9: 1696 sample. Plant Methods 12:45 Wang S, Lu G, Hou Z, Luo Z, Wang T, Li H, Zhang J, Ye Z (2014) Sauvage C, Segura V, Bauchet G, Stevens R, Do PT, Nikoloski Z, Members of the tomato FRUITFULL MADS-box family regulate Fernie AR, Causse M (2014) Genome-wide association in tomato style abscission and fruit ripening. J Exp Bot 65: 3005–3014 reveals 44 candidate loci for fruit metabolic traits. Plant Physiol Waseem M, Li N, Su D, Chen J, Li Z (2019) Overexpression of a ba- 165: 1120–1132 sic helix–loop–helix transcription factor gene, SlbHLH22, promotes Schauer N, Semel Y, Balbo I, Steinfath M, Repsilber D, Selbig J, early flowering and accelerates fruit ripening in tomato (Solanum Pleban T, Zamir D, Fernie AR (2008) Mode of inheritance of pri- lycopersicum L.). Planta 250: 173–185 mary metabolic traits in tomato. Plant Cell 20: 509–523 Xu C, Min J (2011) Structure and function of WD40 domain pro- Semel Y, Nissenbaum J, Menda N, Zinder M, Krieger U, Issman N, teins. Protein Cell 2: 202–214 Pleban T, Lippman Z, Gur A, Zamir D (2006) Overdominant Yang L, Huang W, Xiong F, Xian Z, Su D, Ren M, Li Z (2017) quantitative trait loci for yield and fitness in tomato. Proc Natl Silencing of SlPL, which encodes a pectate lyase in tomato, confers Acad Sci USA 103: 12981–12986 enhanced fruit firmness, prolonged shelf-life and reduced suscepti- Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, bility to grey mould. Plant Biotechnol J 15: 1544–1555 Amin N, Schwikowski B, Ideker T (2003) Cytoscape: a software Ye J, Wang X, Hu T, Zhang F, Wang B, Li C, Yang T, Li H, Lu Y, environment for integrated models of biomolecular interaction Giovannoni JJ, et al. (2017) An InDel in the promoter of networks. Genome Res 13: 2498–2504 Al-ACTIVATED MALATE TRANSPORTER9 selected during tomato 266 | PLANT PHYSIOLOGY 2022: 190; 250–266 Zhu et al. domestication determines fruit malate contents and aluminum Zhang Y, Butelli E, Martin C (2014) Engineering anthocyanin bio- tolerance. Plant Cell 29: 2249–2268 synthesis in plants. Curr Opin Plant Biol 19: 81–90 Zhang J, Hu Z, Yao Q, Guo X, Nguyen V, Li F, Chen G (2018) A to- Zhu G, Wang S, Huang Z, Zhang S, Liao Q, Zhang C, Lin T, Qin mato MADS-box protein, SlCMB1, regulates ethylene biosynthesis M, Peng M, Yang C, et al. (2018) Rewiring of the fruit metabo- and carotenoid accumulation during fruit ripening. Sci Rep 8: 3413 lome in tomato breeding. Cell 172: 249–261.e12

Journal

PLANT PHYSIOLOGYOxford University Press

Published: May 4, 2022

There are no references for this article.